Genetic Diversity and Population structure
analysis of bambara groundnut [Vigna
subterranea (L.) Verdc.] Landraces using
Morpho-agronomic
Characters and
SSR Markers
ODIRELENG OZIE MOLOSIWA
Thesis submitted to the
University of Nottingham for the degree of
Doctor of Philosophy
March 2012
School of Biosciences
Division of Plant and Crop Sciences
The University of Nottingham Sutton Bonington Campus
Loughborough, Leicestershire
LE12 5RD
UK
Dedication
In memory of my grandmother, Twaambo Molosiwa and my mother Mmaletsatsi
Molosiwa who had to go, during the course of my studies.
‘We only part to meet again’
ii
Acknowledgements
I would like to express my sincere gratitude to my supervisors, Dr Sean Mayes
and Professor Sayed Azam-Ali for their guidance, constructive advice, and
assistance given to me during the course of my studies. Many thanks for giving
me this opportunity to study at the University of Nottingham. I would also like to
thank Dr Erik Murchie for the guidance during my studies.
I highly appreciate the supervision I got from Professor Abu Sesay from
Botswana College of Agriculture (BCA) and Dr Sununguko Wata Mpoloka of
Biological Science Department (University of Botswana) during field work
studies in Botswana. I am also grateful to all the technical assistance and
encouragement from Dr Geleta Legesse Fite and Mr Chiyapo Gwafila both from
the Department of Agricultural Research (Ministry of Agriculture).
I am very thankful to the technical advice and assistance for laboratory work
experiments from Dr Katie Mayes, Dr Shravani Basu, Fiona Wilkinson, Dr
GraciaRibas-Vargas, for their sincere help and appreciation. A special thanks to
Jim Craigon for his statistical input and interest in my project. Further, EUBAMLINK project team assistance is acknowledged.
Many thanks from Mark Meacham, John Alcock, Matthew Tovey, and David
Hodson for their support in glasshouse and growth room experiments. In addition
the help from Emma Hooley, Sonoko Mitsui-Angwin, Diane Jones, Sue Flint, Sue
Golds and Linda Staniford is highly appreciated.
I am grateful and thankful to my colleagues and friends we share a journey in our
quest for knowledge, Stanley Noah, Endah Sri Redjeki, Nariman Ahmad, Presidor
Kendabie, Chai Hui Hui, Kidist Bogale Kibret, Ghaliya Al-Mamari, Upendra
Kumar Devisetty, Chanate Malumpong and Steve Hayes
I am grateful to the Commonwealth Scholarship and Fellowship Plan (CSFP) for
awarding me this scholarship, and the Botswana Government (Ministry of
Agriculture) to award me a study leave.
Finally, I thank my ‘big family’ of the Nazareth Church, my friends and my
family especially, Makgopho Nthebolang, Boemo Molosiwa and Moagi Molosiwa
for their kindness and support.
iii
Abstract
Bambara groundnut is an indigenous African legume grown mainly in subSaharan Africa; it is an important source of protein to the rural majority. There
are no established varieties and subsistence farmers grow locally adapted
landraces which are generally low yielding. Bambara groundnut is a
predominantly self-pollinating crop and is expected to exist as non-identical
inbred lines, although the previous lack of co-dominant markers has prevented a
formal assessment of heterozygosity within bambara groundnut genotypes.
A total set of 75 microsatellites that were characterised in this study were used to
investigate the genetic diversity of a set of 24 bambara groundnut landraces, to
provide an evaluation of the markers for polymorphism and provide a link with
DArT marker data that were previously analysed.
Sixty eight microsatellites were identified that were found to be consistent and
reproducible, from which a set of markers were selected and used for genetic
variability studies of bambara groundnut, to compare the use of molecular
markers with morphological markers, and to investigate using SSR markers in
pure line selection.
The genetic diversity of bambara groundnut was assessed based on morphological
characters for two seasons; in a glasshouse experiment at the University of
Nottingham, Sutton Bonington Campus, UK and in a field experiment that was
conducted at the Botswana College of Agriculture (Notwane farm), Gaborone in a
randomised block design with three replicates. The landraces were characterised
for 24 quantitative and 13 qualitative characters. The results indicated
considerable variation for quantitative characters, while significant morphological
differences were also recorded for most characters. Multivariate data analysis was
conducted using principal component analysis, cluster analysis and heritability
estimates were developed. The low cost, simplicity and agricultural relevance of
morphological characterisation makes it an important tool in germplasm genetic
variation studies.
Thirty four lines from field experiments were investigated for genetic diversity
based on 20 microsatellites. The expected heterozygosity (He) had an average of
iv
1in agreement with the fact that bambara groundnut is predominantly selfpollinating.
Both cluster analysis and principle component analysis (PCoA)
grouped landraces based mainly on their areas of origin.
A thorough molecular analysis of genetic and morphological variation in bambara
groundnut was conducted to investigate the relationship between the two
assessment techniques. This comparison will assist in breeders making informed
decisions as to which approach is best to use in germplasm characterisation and
plant breeding and how best to apply such knowledge in practical situations. DNA
markers could then aid with the selection of germplasm for breeding, quality
control within breeding programmes and, potentially, direct selection via Marker
Assisted Selection (MAS). Euclidean distance estimates for morphological data
and (Nei’s 1972) genetic distance estimates for SSR data were strongly correlated
(r = 0.7; P < 0.001) in the agronomy bay and (r = 0.6; P< 0.001) in the controlled
growth room. These results suggest the two approaches are generating the same
pattern of genetic diversity, and as such can be used as a surrogate for each other.
v
Table of Contents
Dedication............................................................................................................... ii
Abstract ................................................................................................................. iv
List of Figures ..................................................................................................... xvi
List of Tables ........................................................................................................ xx
Abbreviations .................................................................................................... xxiv
CHAPTER ONE: Introduction ............................................................................ 1
1.1
Leguminous crops ........................................................................................ 1
1.2
Bambara groundnut taxonomy ..................................................................... 4
1.3
The bambara groundnut crop ....................................................................... 6
1.4
Potential of bambara groundnut ................................................................... 7
1.4.1
Some interesting agronomic characters of the crop. ...................................... 8
1.4.2
Bambara groundnut yield potential ................................................................. 8
1.4.3.
Some uses of bambara groundnut ................................................................... 9
1.4.4
Genetic diversity resources ............................................................................. 9
1.4.5.
Potential areas of expansion............................................................................ 9
1.5
Constraint to bambara groundnut production ............................................. 10
1.5.1
Influence of sowing date/photoperiod........................................................... 10
1.5.2
Low moisture, pests and diseases ................................................................. 11
1.5.3
Anti-nutritional factors in bambara groundnut ............................................. 11
1.5.4
Genetic resources .......................................................................................... 12
1.5.5
Mating systems ............................................................................................. 12
1.6 Diversity evaluation using both morphological and molecular markers ......... 12
1.6.1
Genetic markers ............................................................................................ 12
vi
1.6.2
Morpho-agronomic markers ..................................................................... 13
1.6.3
Biochemical markers ................................................................................ 15
1.6.4
Molecular Markers .................................................................................... 16
1.6.4.1
Hybridisation (Sequence dependent) ........................................................ 16
1.6.4.1.1 Random Amplified Fragment Polymorphism (RFLP).............................. 16
1.6.4.1.2 Diversity Array Technology (DArT) ........................................................ 17
1.6.5
PCR-based molecular markers ...................................................................... 19
1.6.5.1
Random Amplified Polymorphic DNA (RAPD) ...................................... 19
1.6.5.2
Amplified Fragment Length Polymorphism (AFLP)................................ 20
1.6.5.3
Microsatellites: Simple Sequence Repeats (SSR) ..................................... 21
1.7
Microsatellites development and application ............................................. 22
1.8
Potential application of microsatellites in bambara groundnut .................. 23
1.9
1.8.1
Conservation of genetic resources ............................................................ 23
1.8.2
Molecular mappings.................................................................................. 25
1.8.3
Marker Assisted Selection and QTL ......................................................... 25
Comparison of genetic diversity estimates methods .................................. 28
1.9.1
The objectives of the study ....................................................................... 30
CHAPTER TWO: Materials and Methods....................................................... 31
2.1
Introduction ................................................................................................ 31
2.1.1.
Standard solutions ......................................................................................... 31
2.1.2
List of plant materials .................................................................................. 32
2.1.3
Overview of experiments. ............................................................................. 35
2.2
Methodology for Marker and DNA techniques ......................................... 37
2.2.1
Introduction ................................................................................................... 37
vii
2.2.2
Plant materials............................................................................................... 37
2.2.3
DNA extraction ............................................................................................. 38
2.2.3.1
Sigma DNA extraction Kit........................................................................ 38
2.2.4
DNA quantitation ......................................................................................... 39
2.2.5
Microsatellite development .......................................................................... 39
2.2.6
PCR gradient optimisation for primer annealing temperature ..................... 40
2.2.7
Gel electrophoresis of PCR products ........................................................... 41
2.2.8
Three primer systems ................................................................................... 41
2.2.9
PCR amplification of microsatellites ........................................................... 42
2.3.0
Gel electrophoresis of PCR products of Tagged primers............................. 43
2.3.1
Capillary electrophoresis .............................................................................. 43
2.3.2
Analysis of microsatellites ............................................................................ 44
2.4
Potential genotyping errors and some mitigating strategies in microsatellite
analysis .................................................................................................................. 44
2.4.1
DNA degradation ......................................................................................... 44
2.4.2
PCR based sources of error .......................................................................... 44
2.4.3
Interpretation of capillary electrophoresis .................................................. 45
2.4.4
Spectral overlap ........................................................................................... 45
2.4.5
Stutter and A-addition .................................................................................. 47
2.4.6
Short allele dominance................................................................................. 49
2.4.7
Allele size binning (Automated binning) ..................................................... 49
2.4.8
Deviation from Hardy Weinberg equilibrium .............................................. 54
2.5
Data analysis............................................................................................... 55
2.5.1
Data analysis for microsatellites, development and characterisation ............ 55
viii
2.5.1.1
Microsatellites marker analysis................................................................. 55
2.5.1.2
Principal component analysis (PCO) ........................................................ 55
2.5.1.3
Cluster analysis ......................................................................................... 55
2.5.1.4
Comparison of DArT and SSR genetic estimates ..................................... 56
2.5.2
Population structure and genetic diversity of bambara groundnut................ 56
2.5.2.1
Estimation of genetic diversity in the population ..................................... 56
2.5.2.2
Estimation of genetic diversity within and among bambara
groundnutpopulations ............................................................................................... 56
2.5.2.3
2.5.3
Estimation of population structure ............................................................ 57
Genetic diversity of bambara groundnut based on SSR markers and the
comparison with morpho-agronomic characters ........................................................... 58
2.5.3.1
Polymorphism of microsatellites in bambara groundnut .......................... 58
2.5.3.2
Principal component (PCO) and cluster analysis ...................................... 58
2.5.3.3
Analysis of Molecular Variance (AMOVA) ............................................. 58
2.5.3.4
Morphological data analysis ........................................................................ 58
2.5.3.5
Comparison of SSR marker and morphological marker data ....................... 59
2.6
Morpho-agronomic characterisation and evaluation of bambara
groundnut............................................................................................................... 59
2.6.1 Introduction: ............................................................................................... 59
2.6.2 Glasshouse experiment ............................................................................... 59
2.6.3
Plant materials.............................................................................................. 60
2.6.4
Experimental design..................................................................................... 60
2.6.5
Crop management ........................................................................................ 60
2.6.6 Morpho-agronomic traits measurements collected in the greenhouse ....... 63
2.6.6.1
Quantitative traits measurements in the green house ................................... 64
ix
2.6.6.2
2.7
Qualitative traits measurements in the glasshouse ....................................... 64
Field work experiment in Botswana ........................................................... 66
2.7.1
Introduction ................................................................................................... 66
2.7.2
Field site and experimental preparation ........................................................ 66
2.7.3
Plant material ................................................................................................ 66
2.7.4
Experimental design...................................................................................... 66
2.7.5
Crop management ......................................................................................... 67
2.7.6 Agro-morphological traits measurements in the field experiment ............. 70
2.7.7
Statistical analysis of agronomic traits......................................................... 70
2.7.8
Data analysis of agronomic traits ................................................................. 70
2.7.8.1
Descriptive characteristics ........................................................................ 70
2.7.8.2
Principal component analysis ................................................................... 71
2.7.8.3
Cluster analysis ......................................................................................... 71
2.7.8.4
Shannon-Weaver diversity ........................................................................ 72
2.7.8.5
Correlation coefficient .............................................................................. 72
2.7.8.6
Quantitative variances............................................................................... 72
2.7.8.7
Selection index (SI) and Duncan Multiple Range Test (DMRT) ............. 74
CHAPTER THREE: Microsatellites, development and characterisation ..... 76
3.1
Introduction .............................................................................................. 76
3.1.1
Breeding systems in bambara groundnut ...................................................... 76
3.1.2
Floral biology of bambara groundnut ........................................................... 77
3.1.3
Seed dissemination systems .......................................................................... 78
3.1.4
Analysis of breeding systems in bambara groundnut ................................... 78
3.1.5
Breeding system studies in other leguminous species .................................. 79
x
3.1.6
3.2
Materials and Methods ............................................................................ 81
3.2.1
3.3
Applications of microsatellites in this study ................................................. 80
DArT marker screening ................................................................................ 81
Results........................................................................................................ 83
3.3.1
Microsatellites marker analysis..................................................................... 83
3.3.1.1
Hardy Weinberg Equilibrium (HWE) ....................................................... 85
3.3.1.2
Estimation of Null alleles......................................................................... 87
3.3.2
Principal Component Analysis (PCO) .......................................................... 90
3.3.3
Cluster analysis ............................................................................................. 92
3.4
Discussions: ............................................................................................... 97
3.5
Conclusions ............................................................................................... 99
CHAPTER FOUR: Phenotypic diversity for morphological and agronomic
characters of bambara groundnut ................................................................... 101
4.1
Introduction ............................................................................................ 101
4.1.1
Correlation analysis studies ........................................................................ 102
4.1.2.
Selection of lines for breeding .................................................................... 106
4.1.3
The objectives of this study ........................................................................ 107
4.2
Results...................................................................................................... 108
4.2.1
Qualitative analysis of the genotypes ..................................................... 108
4.3.2
Shannon Weaver (H’) diversity analysis ................................................ 111
4.3.3
Descriptive analysis of the genotypes ..................................................... 113
4.3.4
Principal component analysis: ................................................................ 117
4.3.5
Cluster analysis ....................................................................................... 122
4.3.6
Correlation coefficients among traits ...................................................... 125
xi
4.3.7
Quantitative variance analysis ................................................................ 131
4.3.8
Comparison of agronomy bay and field experiment ............................... 135
4.3.8
Selection for breeding bambara groundnut ............................................. 137
4.4
Discussions .............................................................................................. 140
4.5
Conclusions ............................................................................................. 145
CHAPTER FIVE: Population structure and genetic diversity of bambara
groundnut ........................................................................................................... 146
5.1
Introduction ............................................................................................ 146
5.1.1
Genetic diversity in bambara groundnut ..................................................... 147
5.1.2
Genetic diversity and population structure of other legumes...................... 147
5.1.3The objectives of the study ..................................................................................... 149
5.2
Materials and Methods .......................................................................... 149
5.2.1
5.3
Phenotypic data analysis ............................................................................. 149
Results...................................................................................................... 150
5.3.1
Genetic diversity analysis ........................................................................... 150
5.3.2
Genetic diversity within and among regions............................................... 151
5.3.3
Principal coordinates analysis (PCoA) ....................................................... 152
5.3.4
Cluster analysis ........................................................................................... 153
5.3.5
Genetic differentiation based on FST ........................................................... 155
5.3.5.1
5.3.6
Pairwise comparison ............................................................................... 155
Analysis of Molecular Variance Analysis (AMOVA) ................................ 156
5.3.7 Comparison of molecular markers with pod and seed characters ...................... 157
5.4
Discussion ................................................................................................ 162
5.5
Conclusions ............................................................................................ 165
xii
CHAPTER SIX: Genetic diversity of bambara groundnut based on SSR
markers and the comparison with morpho-agronomic characters .............. 166
6.1
Introduction ............................................................................................ 166
6.1.1
Genetic diversity of bambara groundnut ..................................................... 167
6.1.2
Genetic diversity in other leguminous crops............................................... 167
6.1.3
Efficiency of molecular and morphological markers in genetic diversity
estimates ......................................................................................................................169
6.2
Materials and methods ........................................................................... 171
6.2.1
6.3
Plant Materials used .................................................................................... 171
Results...................................................................................................... 171
6.3.1
Polymorphism of microsatellites in bambara groundnut ............................ 171
6.3.2
Principal Component Analysis (PCO) ........................................................ 176
6.3.3
Comparison of SSR and morphological markers ........................................ 178
6.3.3.1
Principal Component Analysis ............................................................... 178
6.3.4
Genetic distance estimates between landraces ............................................ 183
6.3.5
Correlation between molecular and morphological distance estimates ...... 183
6.3.6
Molecular variance among bambara groundnut landraces.......................... 185
6.3.7
Breeding strategy ........................................................................................ 186
6.4
Discussion ................................................................................................ 189
6.5
Conclusions ............................................................................................ 193
CHAPTER SEVEN: General discussions ....................................................... 194
7.1
Introduction .............................................................................................. 194
7.2
Recap of the study .................................................................................... 196
7.3
Microsatellites development and characterisation ................................... 197
xiii
7.4
Morphological characterisation ................................................................ 199
7.5
Genetic diversity based on SSR markers and a comparison
withmorphological characters ............................................................................. 203
7.6
Population structure analysis .................................................................... 208
7.7
Impact of the findings and future work .................................................... 210
7.8
Future work .............................................................................................. 211
REFERENCE .................................................................................................... 212
APPENDICES ................................................................................................... 242
Appendix 1: Preparations of standard solutions ........................................................ 242
Appendix 2: List of characterised 75 primers used in bambara groundnut diversity,
and list of primer combinations used in multiple experiments. .................................. 243
Appendix 3: Estimated repeat length of alleles and adjustment for the characterisation
of 75 markers used in the analysis of 24 landraces ..................................................... 245
Appendix 4: A comparison of Nei and Li, (1979) similarity estimates for DArT
marker (upper) and SSR markers (bottom) matrices calculated using MVSP version
3.1 for the 24 bambara groundnut landraces. .............................................................. 247
Appendix 5: Mean values for the characters of the 35 landraces grown in the
agronomy bay experiment (UK) ................................................................................. 248
Appendix 6: Mean values for the characters of the 34 lines grown in the field
experiment (Botswana). .............................................................................................. 250
Appendix 7: Range of classes for the quantitative traits used for both the glasshouse
and the field experiment ............................................................................................. 252
Appendix 8: Hardy Weinberg Equilibrium (HWE) and the exact p-values estimated
using PowerMarker (Version 3.25)............................................................................. 253
Appendix 9: Cluster analysis, genetic similarity among the 105 bambara groundnut
genotypes, analysis using 141 variables and 105samples/cases ................................. 254
Appendix 10: Scatter plots for morpho-agronomic markers on (Euclidean distance
estimates) and molecular markers on (Nei’s 1972) conducted using Mantel’s test on
xiv
NTSYS, Pearson correlation and Spearman’s rank correlations on SPSS in the
Agronomy bay and controlled growth room experiment: Appendix 10.1 and 10.2. .. 257
Appendix 11: Mean for the phenotypic measures of 5 lines from the controlled growth
room experiment ......................................................................................................... 259
xv
List of Figures
Figure 1.1: Phylogenetic trees showing the relationship between Vigna species
from various Vigna subgenus and section. Adapted from (Wang et al., 2008).......3
Figure 1.2: A typical bambara groundnut crop in the field, unshelled pods,
flowering and pod initiation, and bambara groundnut seeds...................................6
Figure 2.0: Diagrammatic representation of the setup of the project....................36
Figure 2.4.4: A pair of capillary electrophoresis traces of PCR products for blue
labelled. The genuinely labelled blue PCR products, has bled through into
spectrum of the green labelled PCR product and false peaks are shown for sample
M18 and M19 (green)............................................................................................46
Figure 2.4.5: Capillary electrophoresis showing a potential scoring error due to
the effects of stutter band and overlap on sample PR 45-H. H11 and
44.E05....................................................................................................................48
Figure 2.4.6: Capillary electrophoresis showing limited short allele dominance
for marker PR 15 top and marker PR 42 bottom, since both are clearly visible and
complete drop out did not occur, correct calling of the peaks could be done…....49
Figure 2.4.7: Capillary electrophoresis showing potential sources of mistyping
errors due to rounding off alleles during binning...................................................51
Figure 2.4.8: Capillary electrophoresis showing some potential miscalling errors,
therefore the use of allele shapes, their height and size ranges are set as standard
way to identify genuine peaks for correct allele calling.......................................52
Figure 2.4.9: A graphical output of the cumulative allele length for marker 16, it
illustrates an example of an accurately binned marker with clearly defined colours
for different alleles as red and blue. The analysis was conducted with FLEXIBIN
(Automated binning) using a one unit repeat.........................................................53
Figure 2.6.5.1: The maximum and minimum temperature in the agronomy bay
(Glasshouse) experiment for the 119 bambara groundnut landraces grown in the
2008 season days from sowing..............................................................................62
Figure 2.6.5.2: The maximum and minimum relative humidity in the agronomy
bay (Glasshouse) experiment for the 119 bambara groundnut landraces grown in
the 2008 season......................................................................................................63
Figure 2.7.5.1: The amount and distribution of rainfall in bambara groundnut
field experiment at (Notwane) Sebele, in 2008- 2009
season.....................................................................................................................68
Figure 2.7.5.2: Maximum and minimum temperature in the field experiment for
the 34 bambara groundnut landraces grown at (Notwane) Sebele in the 2008 2009 season...........................................................................................................69
xvi
Figure 2.7.5.3: Maximum and minimum relative humidity in the field experiment
for the 34 bambara groundnut landraces grown at (Notwane) Sebele in the 2008 2009 season............................................................................................................69
Figure 3.1.2 Bambara groundnut flower, showing the floral morphology...........77
Figure 3.2: The first two axes of the PCO case scores, generated from the 24
landraces using MVSP for figure 3 (a) DArT Axis 1 represents 28.45% and Axis
2 represents 8.84 % of the molecular variation, figure 3 (b) SSR markers; Axis 1
represents 10.91 % and Axis 2 represents 8.61% of the molecular variation in the
24 selected bambara groundnut landraces..............................................................91
Figure 3.3:(a) Cluster analysis of the 24 bambara groundnut landraces. The
UPGMA dendrogram is based on the similarity matrix obtained from 201 DArT
markers using the Nei and Li, (1979). The number at the nodes of branches
represents the percentage bootstrap support of individual nodes at resampling at
1000........................................................................................................................93
Figure 3.3:(b) Cluster analysis based on the 24 bambara groundnut landraces, the
dendrogram was obtained based on 68 SSR markers, the UPGMA tree is based
on the Nei and Li, (1979) similarity coefficient. The number at the nodes of
branches represents the percentage bootstrap support of individual nodes at
resampling at 1000.................................................................................................93
Figure 3.4: (a) A scatter plot produced based on the matrix for DArT and SSR
markers genetic distance estimates from Nei and Li, 1979 (Appendix 4) using
Pearson product-moment coefficient correlation based on SPSS version 16 (b) A
scatter plot based on the matrix for DArT and SSR produced from the same
genetic distance estimates in Appendix 4 using Mantel-matrix correspondence test
on NTSYS pc version 2.1 program (MXCOMP module) based on 1000
permutation............................................................................................................96
Figure 4.2.2: Dendrogram of 35 bambara groundnut landraces showing a
(UPGMA) Euclidean cluster analysis based on 34 agro-morphological markers in
glasshouse experiment. The colour code for West Africa = Green, Southern Africa
=Red, East Africa =Yellow, Central Africa = Blue, Indonesia = Purple. The
number at the nodes of branches represents the percentage bootstrap support of
individual nodes resampling at 1000…………………………………………....123
Figure 4.2.3: Dendrogram of 34 bambara groundnut lines showing genetic
similarities based on 37 morpho-agronomic traits 24 quantitative traits and 13
qualitative traits, using the UPGMA cluster analysis (field experiment Botswana).
The number at the nodes of branches represents the percentage bootstrap support
of individual nodes resampling at 1000………………………………………...124
Figure 4.2.4: A regression analysis plot of mean over all the genotypes for the 24
variables recorded from agronomy bay (UK) and field experiment in
Botswana..............................................................................................................136
Figure 5.1.0 A PCO scatter plot for the 123 bambara groundnut genotypes from
Africa and Indonesia generated from 12 microsatellites with MVSP program
xvii
with a molecular variation of 16.15 %, with axis 1 contributing (9.87%) and
while Axis 2 explained (6.28 %). The two cluster groups were hand drawn on
Microsoft Word………………………………………........................................153
Figure 5.2.1: Cluster analysis of bambara groundnut landraces from five regions,
from Africa and Indonesia (Asia). The dendrogram is based on 12 SSR markers.
The Unweighted pair group method with arithmetic averages (UPGMA) tree was
based on Nei and Li’s coefficient of genetic similarity generated from the
presence/absence binary matrix on 123 bambara groundnut landrace
accessions.............................................................................................................154
Figure 5.3.0: A PCO scatter plot for the 87 bambara groundnut that produced
pods and seeds among 119 bambara groundnut planted. The data is based on 7
pod and 8 seed characters analysed using the MVSP program. The percentage
variation for Axis 1 represents 33.59% and the Axis 2 represent 16.8 % with a
cumulative percentage of 49.87% for the first two Axes.....................................158
Figure 5.4.0: A PCO scatter plot on case scores for the 87 bambara groundnut
that produced pods and seed based on 12 microsatellites, generated on MVSP
program. The cumulative percentage of variation explained for the first two Axes
is 16.08 %, Axis 1 contributes 9.45% and Axis 2 contributes 6.63%.................159
Figure 5.5.0:Scatter plot of correlation for morphological marker genetic
distances estimate based on standard Euclidean and SSR marker genetic distance
based on Nei’s 1972, the analysis was conducted on (a) Mantel test
correspondence test on NTSYS and (b) on Pearson correlation on SPSS version
16………………………………………………………………………………..160
Figure 6.1.1: UPGMA dendrogram of 105 bambara groundnut genotypes
revealed by UPGMA cluster analysis of 20 SSR markers based on Nei and Li,
1979 similarity estimate. Bootstrap values of 1000 replications more than 50% are
shown on corresponding nodes............................................................................175
Figure 6.2.1: The first two axes of the PCO case scores, generated from the 105
bambara groundnut genotypes based on 20 SSR markers generated on MVSP,
the first Axis accounts for 8.69 % while Axis 2 represent 6.27 % and together
explain a cumulative 14.95 % of the molecular variation. Figure a: shows a PCO
plot demarcated on 5 symbols to identify three individuals from one landrace
while Figure b: shows the grouping of the five regions into two major groups. The
two cluster groups were hand drawn on Microsoft Word……………..........…..177
Figure 6.3.1: The first two axes of the PCO case scores, generated from the 34
bambara groundnut landraces using MVSP for figure 6.3.1 (a) SSR marker Axis
1 represent 9.90 % and Axis 2 represent 8.35 %, figure 6.3.1 (b) Morphology
marker; Axis 1 represent 13.56 % and Axis 2 represent 8.80 % molecular
variation with a cumulative % of 18.25 % and 22.36 % respectively.................179
Figure 6.4.1: Cluster analysis of 34 bambara groundnut analysis with Unweighted
pair group method with arithmetic method (UPGMA) were generated using
NTSYS version 2.1, Figure 6.4.1 (a) is SSR marker dendrogram generated from
20 microsatellites markers based on Nei’s 1972 distance estimates, Figure (b) is a
xviii
morphology dendrogram generated on 37 morpho-agronomic traits generated on
Euclidean distance estimates................................................................................182
Figure 6.5.1:A scatter plot of correlation for morpho-agronomic and molecular
marker based on Pearson, Spearman (rank) and Mantel test, analysis conducted
on (A) NTSYS pcversion 2.1 and (B) on SPSS version 16, the morphological
markers were based on Euclidean distances estimates while the molecular marker
were on Nei’s 1972 coefficient………………….……………………………...184
Figure 6.5.2: Schematic diagram showing the selection strategy for the three
round of selection of bambara groundnut............................................................187
xix
List of Tables
Table 1.1: Summary of Vigna classifications based on 6 sub-genera, and some
examples from each section, adapted from African Vigna......................................2
Table 2.1.2.1: List of selected landraces used for the characterisation of SSR
markers and DArT analysis, their areas of origin and the clusters where the
landraces were selected. The selection was on the basis of a study conducted by
Singrün and Schenkel (2003), where a total of 223 bambara groundnut landraces
were analysed for genetic diversity using enzyme system EcoRi/MseI amplified
fragment length polymorphism (AFLP).................................................................32
Table 2.1.2.2:A list of 123 bambara groundnut accessions, source and their areas
of origin used in the experiment; 105 bambara groundnut accessions selected in
the greenhouse (35 x 3) samples, and 34 accessions that were selected and planted
in the field experiment in (Botswana)....................................................................33
Table 2.3.1: A summary of Flexibin analysis for marker 16, showing repeat
length, standard deviation and count of each repeat length...................................54
Table 2.6.1: Amount of irrigation water (mm) applied in the bambara groundnut
experiment in the agronomy bay (Glasshouse) expressed in days after sowing
(DAS) for the duration of the experiment in 2008 season.....................................61
Table 2.6.6: Quantitative and qualitative traits recorded and brief description as
listed from (IPGRI, 2000)......................................................................................65
Table 2.7.5: Amount of irrigation water (mm) applied in the bambara groundnut
experiment in the field experiment in Botswana, expressed in days after sowing
(DAS) for the duration of the experiment in the 2008/ 2009 season....................67
Table 3.1: A summary of PowerMarker data analysis of 24 bambara groundnut
landraces, based on 68 microsatellites...................................................................83
Table 3.2: The 68 markers used in the 24 bambara groundnut analysis were
subjected to Chi square and HWE exact test using MVSP version 3.25, with the
exception of nine non-polymorphic markers.........................................................86
Table 3.3: Estimation of null allele frequencies for each locus, using the
population inbreeding model (PIM) and the individual inbreeding model (IIM)
using INEst (Chybicki and Burczyk, 2009).........................................................88
Table 3.4: PCO case scores for the population structure of the selected 24 bambra
groundnut landraces, determined based on 201 DArT markers.............................90
Table 3.5: PCO case scores for the population structure of the selected 24 bambra
groundnut landraces, determined based on 65 SSR markers.................................90
Table 3.6: A comparison of the distribution of the 24 bambara groundut
landraces based on the UPGMA clustering analysis done using a set of 201 DArT
markers and 65 SSR markers.................................................................................94
xx
Table 3.7 Pearson, Spearman and Mantel test correlations between the genetic
similarity matrices based on the two markers systems (DArT vs SSR)...............95
Table 4.1.1: A comparison of correlations between yield components; seed yield
per plant, number of pods per plant, seed yield per hectare and 100 seed weight
and a number of characters, sourced from Karikari and Tabona, (2004); Misangu
et al., (2007); Ouedraogo et al., (2008); Goli et al., (1995); Jonah et al., (2010);
Karikari, (2000), and (Oyiga and Uguru, 2011)...................................................103
Table 4.1.2: Descriptor, classes and frequency distribution among the 35
landraces planted in the agronomy bay and 34bambara groundnut lines selected
and planted in the field in Botswana....................................................................109
Table 4.1.3: Shannon-Weaver index on the phenotypic diversity of 24
quantitative characters in the agronomy bay experiment and the field experiment
(Botswana)...........................................................................................................111
Table 4.1.4: Shannon weaver index on phenotypic diversity of qualitative
characters for the studied landraces in agronomy bay and field experiment.......112
Table 4.1.5: Descriptive characteristics for the 35 bambara groundnut planted UK
(Agronomy bay, 2008) from three replications....................................................114
Table 4.1.6: Descriptive characteristics for the 34 bambara groundnut planted in
field (Notwane, Botswana, 2008/2009 season) with three replications, derived
from seed from single plant from the agronomy bay experiment. These exclude 1
landrace which had few seeds..............................................................................115
Table 4.1.7: Principal components, matrix of eigenvalues and vectors for 24
quantitative characters of bambara groundnut landraces planted in the agronomy
bay (UK)...............................................................................................................119
Table 4.1.8: Principal component, matrix of eigenvalues and vectors for 24
quantitative characters of bambara groundnut lines planted in Botswana...........121
Table 4.1.9: Correlation coefficients among the 24 traits based on the 35 bambara
groundnut planted in the Agronomy bay (UK), traits were measured 10 weeks
after planting........................................................................................................127
Table 4.2.1: Correlation coefficient for 24 quantitative traits of the 34 bambara
groundnut planted in the field experiment in (Botswana) traits were measured 10
weeks after planting.............................................................................................129
Table 4.2.2: Quantitative variances based on phenotypic coefficient of variability
(PCV), genotypic coefficient of variability (GCV), broad sense heritability (h2B)
and genetic advance (GA) in the 35 landraces in the agronomy bay (UK).........132
Table 4.2.3: Quantitative variances based on phenotypic coefficient of variability
(PCV), genotypic coefficient of variability (GCV), broad sense heritability (h2B)
and genetic advance (GA) in the 34 lines (Field experiment).............................133
xxi
Table 4.2.4: A summary of analysis for the relationship between the agronomy
(UK) experiment and the field experiment in (Botswana), computed on Genstat
version 13.0..........................................................................................................136
Table 4.2.5: The Duncan multiple range test and the selection index of bambara
groundnut based on the vegetative and yield characters (Agronomy bay, UK)..138
Table 4.2.6: The Duncan multiple range test and the selection index of bambara
groundnut based on the vegetative and yield (field experiment in
Botswana).............................................................................................................139
Table 5.1: PowerMaker summary data analysis for the 12 microsatellites used
amplified from 123 bambara groundnut landraces (118 from Africa and 5 from
Asia/Indonesia)....................................................................................................150
Table 5.2: A comparison of the genetic diversity estimates for the among the five
regions of Africa and Asia (Indonesia) analysis conducted using FSTAT 2.9.3 for
all the 123 bambara groundnut landraces.............................................................151
Table 5.3: Principal Coordinate analysis (PCoA) from the investigation of
population structure of 118 bambara groundnut landraces collected from Africa
and 5 from Indonesia based on MVSP program..................................................152
Table 5.4: Genetic differentiation of the 123 bambara groundnut landraces from 4
regions of Africa and also Asia (Indonesia), estimated using Weir and Cockerham
(1984) on Genepop version 4.0............................................................................155
Table 5.5: Pairwise genetic distance based on FST values between populations,
calculated on 12 microsatellites based on five regions of Africa including Asia
(Indonesia)............................................................................................................156
Table 5.6: Analysis of Molecular Variance for the 123 bambara groundnut
landraces based on 12 SSR markers using Arlequin version 3.1.........................157
Table 5.7: Principal Coordinate analysis (PCoA) for 15 characters of pods and
seeds for 87 landraces that set reasonable seed numbers among the 119 landraces
planted in the agronomy bay for bambara groundnut germplasm
characterisation....................................................................................................157
Table 5.8: Principal Coordinate Analysis (PCoA, Euclidean) for 87 bambara
groundnut landraces that set seed, based on 12 microsatellites...........................158
Table 5.9: Correlation of molecular marker distance matrices, based on Pearson
correlation, Spearman rank correlation and Mantel test for the 12 qualitative
character and 12 molecular markers....................................................................161
Table 6.1: Summary of PowerMarker data analysis for the 35 bambara groundnut
landraces using 20 microsatellites analysis conducted on each of the 105
individual genotype..............................................................................................172
Table 6.2: Intra-landrace diversity among the 35 genotypes conducted on each of
three genotypes per landrace using 20 SSR markers based on Arlequin version
3.1.........................................................................................................................173
xxii
Table 6.3: PCO case scores for the population structure of the 105 genotypes
determined from each of the three samples of the 35 bambara groundnut landraces
based on 20 SSR markers.....................................................................................176
Table 6.4: PCO case scores for the population structure of the 34 bambara
groundnut selected for field studies in Botswana, analyses based on 20 SSR
markers.................................................................................................................178
Table 6.5: PCO case scores for the population structure of the 34 bambara
groundnut based on 37 morpho-agronomic characters, from the field experiment
conducted in Botswana........................................................................................178
Table 6.6: Correlation between morpho-agronomic markers and molecular
markers for the 35 and 34 bambara groundnut genotypes based on 20
microsatellites and 37 morph-agronomic characters, and for 5 lines based on 12
markers and 22 morpho-agronomic characters....................................................185
Table 6.7:Analysis of molecular variance (AMOVA) for the 105 bambara
groundnut genotypes for the comparison based on the five selected regions,
analysis using Arlequin version 3.5.....................................................................186
Table 6.8: Mean and range of the genetic distances values for three different
selection cycles of bambara groundnut from single seed descent estimated based
on 12 microsatellites markers using Popgene version1.31 (Yeh and Boyle,
1997)....................................................................................................................188
xxiii
Abbreviations
ANOVA
Analysis of Variance
AMOVA
Analysis of Molecular Variation
DNA
Deoxyribonucleic Acid
dNTP
Deoxynucleotide Triphosphates (usually mix of
dATP/dTTP/dCTP/dGTP)
dATP
deoxyadenosine Triphosphate
dTTP
deoxythymidine Triphosphate
dCTP
deoxycytidine Triphosphate
dGTP
deoxyguanine Triphosphate
DAR
Department of Agricultural Research
cM
centiMorgan
CV
Coefficient of variation
DAS
Days after sowing
EDTA
ethylene diamine Tetracetic Acid
f
Inbreeding coefficient
FIS
Inbreeding coefficient
FIT
Overall fixation index
FST
Fixation index
H`
Gene diversity
GA
Genotypic advance
GVC
Genotypic coefficient of variation
GST
Nei’s total fixation index
He
Expected heterozygosity
Ho
Observed heterozygosity
HWE
Hardy-Weinberg equilibrium
IBPGR
International Board for Plant Genetic Resources
xxiv
IITA
International Institute of Tropical Agriculture
ISSR
Inter Simple Sequence Repeat
MAS
Marker Assisted Selection
MVSP
Multivariate Statistical Package
NTSYS
Numerical Taxonomy and Multivariate Analysis System
PCoA
Principle coordinates analysis
PCV
Phenotypic coefficient of variation
PCO
Principal component analysis
PIC
Polymorphic information content
PCR
Polymerase chain reaction
RAPD
Randomly Amplified Polymorphic DNA
RFLP
Restriction Fragment Length Polymorphism
SS
Sizestandard
SLS
Sample loading solutions
SSR
Simple Sequence Repeat
UPGMA
Unweighted pair group method with arithmetic means
xxv
CHAPTER ONE: Introduction
1.1
Leguminous crops
The genus Vigna is a member of the family Leguminosae (= Fabaceae), subfamily
Papilionoideae, tribe Phaseoleae. Leguminosae are morphologically diverse and
include a number of trees and some aquatic plants such as in the genus Neptuniain
the subfamily Mimosoideae (Polhill and Raven, 1981) which consist of a number
of species that are aquatic. It is the third largest family of flowering plants behind
orchids (Orchidaceae) and asters (Asteraceae) and consists of approximately 650
genera and 18,000 species (Polhill et al., 1981; Doyle and Luckow, 2003). In
terms of agricultural importance Leguminosae comes second to cereals. The
Leguminosae have been divided into three major groups mainly on the basis of
their morphological and floral differences, that is the Caesalpinioideae,
Mimosoideae and Papilionoidaeae (Doyle and Luckow, 2003).Papilionoideae
with approximately 70% of the Leguminosae species is the largest subfamily, it
includes most of the crops and major model legume species (Doyle and Luckow,
2003; Cannon et al., 2009) it is subdivided into four large groups the galegoid,
millettioids, dalbergioids and genistoids (Doyle and Luckow, 2003).
The galegoid contains the robinioid clade with several forages and trees (Sesbania
and Robinia); it also consists of inverted-repeat-loss clade (IRLC) which includes
clovers (Trifolium spp.), vetch (vicia spp.), pea(Pisum sativum), chickpea (Cicer
arietinum), lentil (Lens culinaris) and alfalfa(Medicago sativum) (Doyle and
Luckow, 2003).The milletioid clade consists of common bean (Phaseolus
vulgaris), soybean (Glycine max), cowpeas (Vigna unguiculata), pigeonpea
(Cajanus cajan), mungbean (Vigna radiata), adzuki bean (Vigna angularis),
tepary bean (Phaseolus acutifolius), lima bean (Phaseolus lunatus), hyacinth bean
(Lablab purpureus), bambara groundnut (Vigna subterranea) and hausa
groundnut (Macrotyloma geocarpum). The dalbergioid clade consists of a number
of tropical trees (eg Dalbergis spp) and peanut (Arachis hypogaea).The genistoid
contains many tropical and temperate genera for example the lupins (Lupinus spp)
(Cannon et al., 2009).
Vigna consists of approximately 80 species that are grouped into six subgenera:
Vigna, Ceratotropis, Plectotropis,Sigmoidotropis,Lasiosporon and Haydonia.
1
Subgenus Vigna comprises 39 species, and it includes some important agricultural
species such as, Cowpea (Vigna unguiculata L. Walp), bambara groundnut (Vigna
subterranea L. Verdc) and mungbean (Vigna radiata) (Goel et al., 2002;
Vijaykumar et al, 2009). These species are of considerable importance in many
developing countries with cowpea and bambara groundnut from Africa, while
mungbean is from Asia (Smartt, 1985; Doi et al., 2002).
Table 1.1: Summary of Vigna classifications based on 6 sub-genera, and some
examples
from
each
sectionadaptedfrom
(African
Vigna,
bioversityinternational.org)
Subgenus
Vigna
Section
Vigna
Comosae
Macrodontae
Reticulatae
Liebrechtsia
Catiang
Specie
V. subterranea
V. comosa , V. haumaniana
V. somaliensis
V. reticulata
V. frutescens
V. anguiculata
Haydonia
Haydonia
Microspermae
Glossostylus
V.monophylla
V.microsperma
V. nigritia
Plectotropis
Plectotropis
Pseudoliebrechtsia
V. vexillata
V. nuda
Ceratotropis
Ceratotropis
Aconitifoliae
Angulares
V. mungo ,V. radiata
V. aconitifolia
V. angularis
Lasiospron
Lasiospron
V. longifolia
Sigmoidotropis
Sigmoidotropis
Pedunculares
Caracallae
Condylostylis
Leptospron
V. elegans
V. peduncularis
V. caracalla
V. venusta
V. adenantha
2
V. adenantha
V. caracalla
V. longifolia
V. acontifolia
V. mungo
V. angularis
V. umbellata
V. radiata
V. luteola
V. oblongifolia
V. vexillata
V. subterranea
V. anguilata
Figure 1.1: Phylogenetic trees showing the relationship between Vigna species from
various Vigna subgenus and section. Adapted from (Wang et al., 2008)
Legumes are an important part of subsistence agriculture as they provide proteinrich food, and ameliorate the soil by improving the structure (Sato et al., 2010).
They are a source of oil production for human consumption and production of
animal feed. Thanks to their symbiosis with nitrogen-fixing bacteria, legumes can
be grown without the addition of nitrogen fertilizer and most can be grown on
poor soils (Sandal et al., 2002). Leguminous plants have two model species, Lotus
japonicas and Medicago truncatula, together with significant research in soybean
(Glycine max), and these have been chosen to represent the diverse legume
family. A number of studies have been conducted on model species to develop
genome resources, such as the production of cDNA libraries, DNA marker
3
production and analysis, high density genetic linkage map production and genome
sequencing. The experience gathered from model species can facilitate basic
genetics in crop legumes and accelerate crop breeding (Sato et al., 2010).
1.2
Bambara groundnut taxonomy
Bambara groundnut (Vigna subterranea (L.)Verdc) synonym [Voandzeia
subterranea (L.) Thouars] is a herbaceous, self-pollinating plant with an
indeterminate growth habit. The domesticated bambara groundnut landraces have
quite a distinct tap root and numerous short lateral stems on which the trifoliate
leaves are borne, while the wild forms have a limited number of elongated lateral
stems with no clear tap root. The petiole is long, stiff and grooved with a base of a
wide range of colours such as green, purple or brown (Swanevelder, 1998). The
species subterranea is further divided into two groups: var. spontanea, comprising
the wild forms, found in a small area around northern Cameroon and Nigeria, and
var. subterranea comprising the cultivated forms in parts of the tropics, mostly in
sub-Saharan Africa (Pasquet et al., 1999; Basu et al., 2007). The chromosome
number in both wild and cultivated plants is 2n = 2x = 22 (Forni-Martins, 1986).
The wild bambara groundnut landraces usually have a spreading growth habit,
compared to the compact type of domesticated landraces (Swanevelder, 1998).
The other major difference between the two types is that of pod size, with
domesticated landraces having bigger seeds which do not wrinkle upon drying,
compared to the wild type (Pasquet, 2003, Basu et al., 2007). The germination of
cultivated forms is rapid and uniform while in the wild forms it is erratic and takes
longer, from approximately 15 to 30 days to germinate (Basu et al.,
2007).Generally, the domestication of crops involves a number of major steps,
with the development of altered plant architecture and also of harvest ability traits,
so that a wild form plant can be domesticated and made more amenable to
intensive agriculture (Basu et al., 2007). Both morphological and isozyme data
were used to demonstrate that wild bambara groundnut (spontanea) is the true
progenitor of domesticated bambara groundnut (subterranea) by Pasquet et al.,
(1999).
Domestication of bambara groundnut is believed to have occurred within the area
where the wild forms are found, which is the Jos plateau and Yola regions of
4
northern Nigeria through to Garuoa in Cameroon and, possibly, as far as the
Central African Republic (Hepper, 1963), with some authors including the areas
from Nigeria to Sudan, which includes Cameroon, Chad, and the Central African
Republic (Pasquet et al., 1999; Hanelt, 2001). Recently, Olukolu et al., 2011,
using both DArT molecular markers and phenotypic descriptors, provided
evidence that pointed out Cameroon/Nigeria as the putative area of origin of
bambara groundnut. The region showed a higher phenotypic diversity for both
quantitative and qualitative characters compared to regions of East Africa, Central
Africa and a combination of other countries in West Africa. The crop is believed
to have been brought first to East Africa and Madagascar, then later to South and
South East Asia, with the slave trade to Suriname, Brazil and later to the New
World (Hanelt, 2001).
It is reported to be cultivated in South and Central
America, India, Indonesia, Malaysia, the Philippines, Sri Lanka and parts of
northern Australia (Linnemann and Azam-Ali, 1993; Suwanprasert et al., 2006).
Bambara groundnut is related to cowpea and has a podding habit similar to that of
peanut (Arachis hypogaea) in that the pale yellow flower stalk bends downward
after fertilization, pushing the young pod into the soil, where it develops and
matures (Doku and Karikari, 1970; Uguru and Ezere, 1997), however, it is not
believed to require complete coverage with soil for the pods to develop.
5
20mm
Figure 1.2: A typical bambara groundnut crop in the field, unshelled pods, flowering and
pod initiation, and bambara groundnut seeds. Scale bar = 20 mm
1.3
The bambara groundnut crop
Bambara groundnut is an important food legume crop, cultivated mainly in subSaharan Africa. Through many years of successive cultivation, farmers have
selected for desirable traits such as growth habit and seed colour (Linnemann and
Azam-Ali, 1993). Farmers prefer the stable, reliable and low yield of bambara
groundnut to high yields of groundnut, which has been associated with more yield
volatility (Linnemann, 1994).
Bambara groundnut is adapted to wide climatic zones, it can be cultivated from
sea-level to up to 1600 m altitude, and an average temperature of 20-28oC is
considered ideal for the crop. A growth period of 110 to 150 days is required for
the crop to develop, although some records of reduced growth cycle landraces of
approximately 90 days have been recorded in Ghana (Berchie et al., 2010). It is
usually grown in mixed intercropping systems with no addition of fertilizers
(Karikari et al., 1995). The crop does well on poor soils which are low in
6
nutrients; however the application of phosphorus results in better nitrogen fixation
and an increase in stover and kernel yield (Ellah and Singh, 2008). It grows well
on well-drained soils, but sandy loams with a pH of 5.0 to 6.5 are most suitable
(Swanevelder, 1998).
The seed makes a complete food as it contains sufficient protein, carbohydrate, fat
and micronutrients (Poulter and Caygill, 1980). The seeds are consumed in a
variety of ways, as fresh pods or boiled with salt and pepper, or eaten as a snack
or mixed with maize seeds or with maize flour as a relish. Nutritional composition
undertaken by several researchers revealed that on average the seeds contain 63 %
carbohydrates, 19% protein, and 6.5 % oil (Ijarotimi and Esho, 2009; Nwokolo,
1987; Borough and Azam-Ali, 1992). The protein is of high quality having a
good balance of the essential amino acids and a relatively high lysine (6.8%) and
methionine (1.3%) content (Ellah and Singh, 2008). The gross energy has been
reported to be higher than that of other pulses including cowpea, lentils and
pigeonpea (Poulter, 1980). The high nutritional value of bambara groundnut
provides a cheap source of protein to poorly-resourced farmers in semi-arid areas
(Doku et al, 1978; Borough and Azam-Ali, 1992; Amarteifio et al., 2006) making
it a good supplement to a cereal-based diet.
The production records of bambara groundnut in some countries are scanty since
it is recorded together with other pulses and sometimes records are not easy to get
because it has not entered the formal market (Mbewe et al., 1995). According to
the Food and Agriculture Organisation of the United Nations: FAOSTAT (2009)
most of bambara groundnut production is taking place in West African countries
with Burkina Faso producing (44712) metric tonnes (MT), Mali (25165) MT,
Cameroon (24000) MT, and Democratic Republic of Congo (1000) MT.
1.4
Potential of bambara groundnut
Most African countries rely on rainfed agriculture, but such agriculture is
particularly vulnerable to climate change. In addition, there are usually other
concerns such as poverty, soil degradation and recurring drought (Mendelsohn,
2000).
In most countries in sub-Saharan Africa that are prone to drought,
unreliable rainfall, poor soils and poor crop productivity, the production of more
7
drought tolerant, indigenous crops, such as bambara groundnut are encouraged.
There is evidence that demonstrates that the crop is more resilient to adverse
environmental conditions as it tolerates low soil fertility soils and low rainfall, and
it is one of the most favoured crops by indigenous people (Azam-Ali et al., 2001).
1.4.1
Some interesting agronomic characters of the crop.
Bambara groundnut landraces have been shown to be able to tolerate drought as
they can sustain leaf turgor pressure by employing a combination of osmotic
adjustment, leaf area reduction and effective stomatal regulation of water loss
(Collinson et al., 1997). Some changes in the leaf orientation, which assist the
crop to reduce incident radiation on the leaf surface, are reported in droughted
landraces such as DipC from Botswana and DodR from Tanzania, reducing water
loss through transpiration (Collinson et al., 1999). Recently, drought response
mechanisms of bambara groundnut were revealed in two landraces, one from a
drought-prone environment (Namibia), S19-3, and from a high rainfall area
(Swaziland), UniswaRed. UniswaRed had a relatively higher transpiration rate
under drought conditions compared to S19-3 which showed a delay in reduction
in transpiration. This mechanism allowed S19-3 to maximise its water use and
escape drought better than UniswaRed (JØrgensen et al., 2010). The crop is
endowed with the advantages of being relatively resistant to pests and diseases,
and has substantial morphological diversity, with good adaptation to marginal
areas and poor conditions (Azam-Aliet al., 2001; Sesay et al., 1996). It also
contributes to the soil fertility through biological nitrogen fixation making it
beneficial in crop rotations and intercropping (Mukumbira, 1985, Karikari et al,
1995), hence farmers do not normally apply chemical fertilizers to bambara
groundnut (Mkandawire, 2007).
1.4.2
Bambara groundnut yield potential
The crop managed to outperform groundnut in controlled environment experiment
to survive and produced some pods when groundnut failed, which is a clear
indication of the crop potential (Azam-Ali et al, 2001). Bambara groundnut
landraces produced good yield in controlled environment and field experiments,
such as the University of Nottingham’s Tropical Crops Research Unit (TCRU)
where pods yields as high as 4 tha-1 were obtained (Collinson et al., 1999). In the
8
fields in Swaziland, Sesay et al., (2008) recorded seed yield of 2.6tha-1 while in
CÔte d’ Ivoire (Kouassi and Zoro, 2009) recorded seed yield as high as 4 tha-1. If
these landraces are developed further to produce cultivars and varieties they could
possibly produce even greater yields. The fresh seed of bambara groundnut often
have a high market price, with demand outweighing supply in many areas
(Coudert, 1984). In Botswana it is more expensive than cowpea and groundnut
(Botswana Agricultural Marketing Board, 2008), making it a good source of
income.
1.4.3. Some uses of bambara groundnut
In Botswana, soybean (Glycine max) is the ingredient for most weaning foods,
although bambara groundnut has been found to be promising in initial results as a
replacement, but has not yet been fully explored (Wambete and Mpotokwane,
2003; Ohiokpea, 2003). In Kenya, it is slowly replacing peanut as a substitute for
weaning food (Mkandawire, 2007). In recent years, bambara groundnut’s
importance as a cash crop has increased, as it is now being canned at commercial
levels in Zimbabwe (Makanda et al., 2009). The haulm for bambara groundnut is
also a valuable source of animal feed (Tibe et al, 2007).
1.4.4
Genetic diversity resources
There are substantial bambara groundnut genetic resources for the future
improvement of the crop since there are approximately 2000 seed accessions in
gene banks held by International Institute of Tropical Agriculture (IITA), and
about 972 accessions in the Southern Africa Development Community (SADC)
countries (Massawe et al., 2005). This provides a good opportunity for bambara
groundnut variety development and improvement of yields, which are still
relatively low.
1.4.5. Potential areas of expansion
Bambara groundnut has wide adaptability, since it is able to grow in ecological
zones of varying climates, ranging from areas with annual rainfall as low as 300
mm annually in Botswana to high annual rainfall of 1250 mm in Swaziland
(Azam-Ali et al., 2001). By scrutinising the world for potential bambara
groundnut production using Geographic Information Systems (GIS) technology,
9
Azam-Ali et al., (2001) identified some areas in America, Australia, Asia, as well
as in Africa, where it could produce significant pod yields, and some areas in the
Mediterranean were it is predicted to have the potential of producing yields as
high as 8.5tha-1
1.5
Constraint to bambara groundnut production
The introduction of peanut- groundnut (Arachis hypogaea) in many developing
countries has replaced bambara groundnut as a major crop (Azam-Ali et al., 2001;
Pasquet, 2003). Since bambara groundnut is grown by smallholders, especially
women, in drier regions (Linneman and Azam-Ali 1993) with limited resources
there is more likely to be poor management of the crop and thus yields are usually
low. In addition there are no established bambara groundnut varieties as yet, and
farmers are using landraces which are a mixture of genotypes (Zeven, 1998).
Bambara groundnut, as an underutilized species, until recently has been largely
ignored by research. To some extend this is due to lack of funds for research in
developing countries where the crop is grown (Azam-Ali et al, 2001).
1.5.1
Influence of sowing date/Photoperiod
Sowing date has been reported to influence the yield and yield variability, through
the effects of temperature and day length on plant development (Collinson et al.,
1996; Sesay et al., 2008). It is a short day species; in most bambara groundnut
genotypes the onset of flowering is not affected by photoperiod while the onset of
podding is adversely affected by photoperiod (Brink, 1997).There is variation
among genotypes in regard to response to photoperiod both at onset of flowering
and onset of podding with landraces Ankap 4, Yola and Ankap 2 from Nigeria
appearing to show different responses to photoperiod sensitivity to onset of
podding (Linnemann and Craufurd, 1994). This suggests that the crop generally
has a facultative response to photoperiod (Jackson, 2008). No genetic studies as
yet have been undertaken on the photoperiod response of bambara groundnut to
identify genomic regions affecting the response of the crop. This is despite
photoperiod being an important characteristic, in attempts to further adapt the
crop, particularly in countries away from the equator.
10
1.5.2
Low moisture, pests and diseases
Low yields of approximately 700 kgha-1 and as low as 200kgha-1 have been
recorded in bambara groundnut. Despite the crop being tolerant to drought (Balole
et al., 2003), dry matter production and yield of bambara groundnut are adversely
affected by soil moisture stress (Collinson et al., 1996; Mwale et al., 2007) as is
the case with all crops. Even though it is a hardy crop and susceptible to few pests
and diseases some had been observed to cause damage to the crop. Some diseases
such as leaf spot and blight (Phoma exigua var. exigua), root rot (Pythium
parocandrum) and wilt (Fusarium solani and F. oxysporium) and root knot
nematode (Meloidogyne javanica) have been observed in Swaziland and
Botswana (Magagula et al., 2003; Karikari et al., 1995). Aphids, which in turn
spread rosette viral diseases and groundnut plant hopper (Hilda patruelis) which
feeds on pegs and pods have been reported on bambara groundnut (Mkandawire,
2007). In storage, shelled bambara groundnut seeds are susceptible to attack by
bruchids (Callosobruchus maculatas), the shelled ones being less susceptible
(Munthali and Ramoranthudi, 2003). Observations have shown that the crop is
vulnerable to fungal disease attacks caused by Colletotrichum capsici with
adverse effects on grain yields (Obagwu, 2003).
1.5.3
Anti-nutritional factors in bambara groundnut
Even though bambara groundnut is an important source of protein in developing
countries, research has revealed the presence of condensed tannins, especially
among the brown, tan and red coloured landraces. Tibe et al., (2007) in their study
found 13 out of 27 landraces from Namibia, Botswana and Swaziland to contain
tannin content below the allowed limit of 0.1% in weaning food in Botswana. The
condensed tannin content ranged from 0.02 % to 0.49%, the cream coloured
landraces recorded levels well below the allowed limit and are recommended to
be used as weaning formula. However, Akaninwor and Okechukwu (2004), in
Nigeria found that tannin content in bambara groundnut can be reduced by
approximately 50% through processing techniques, such as soaking, dehulling,
drying and autoclaving. Farmers also claim that the cream coloured seed requires
shorter cooking time and taste better compared to red and dark coloured seed
(Ramolemana et al., 2003).
11
1.5.4
Genetic resources
There are substantial amount of genetic resources held by the IITA and various
gene banks in SADC countries. Despite these abundant genetic resources, at the
moment there is no Consultative Group on International Agricultural Research
Institution (CGIAR) that has a mandate to do research on bambara groundnut
(Mayes et al., 2009). IITA list their legume crops as cowpea and soybean
(http://www.iita.org) while International Crops Research Institute for the SemiArid Tropics (ICRISAT) list their legume crops as chickpea, pigeonpea and
groundnut (http://www.icrisat.org). The genetic potential of bambara groundnut is
not yet fully exploited, but with the introduction of biotechnology, new techniques
such as molecular markers will assist researchers to better understand the genetics
of bambara groundnut.
1.5.5
Mating systems
Bambara groundnut produces perfect flowers, it is self-pollinating and the
fertilization of the ovule occurs at the day of anthesis (Linnemann, 1994). It is
difficult to undertake hybridisation, with several attempts at artificial
hybridisation reported as unsuccessful (Suwanprasert et al., 2006) and a few
reported cases achieved (Massawe et al., 2003). Therefore, relatively few studies
have been undertaken on the inheritance of yield and related traits in bambara
groundnut (Basu et al., 2007), hence no breeding programme aimed at improving
bambara groundnut has so far been initiated to develop cultivars or varieties
(Oyiga et al., 2010).
1.6 Diversity evaluation using both morphological and molecular markers
1.6.1
Genetic markers
Crop genetic diversity is important for crop adaptation to withstand pests and
diseases and it is an important precondition for plant breeders to enhance the
progress of traits of economic value such as yield. Various methods are available
for use in estimating the genetic diversity of crops, such as morphological,
biochemical and molecular markers. Measurements of genetic diversity can be
generated using conserved accessions in gene banks (Gilbert et al., 1999; Parzies
12
et al., 2000). DNA-based molecular markers have several advantages over the
conventional phenotypic markers since their presence is not dependent on the
growth stage of the crop and can be found in all tissues (Mondini et al., 2009).
1.6.2
Morpho-agronomic markers
The morphological method is the oldest and considered the first step in
description and classification of germplasm (Hedrick, 2005). Evaluation of
genetic diversity through morphological traits is direct, inexpensive and easy.
However, morphological estimations are more dependent on environment and are
more subjective than other measurements (Li et al., 2009). Morphological
variability depends on a limited number of genes, and may not access much of the
potential variability for the agronomic traits present in a crop (Mayes et al., 2009).
The use of morphological and agronomic traits is a standard way of assessing
genetic variation for many species, especially under-researched crops such as
bambara groundnut (Azam-Ali et al., 2001).
Since bambara groundnut is an underutilised crop, studies of its genetic diversity
are scarce. However, Goli et al., (1995), characterized and evaluated
approximately 1400 bambara groundnut accessions at the International Institute of
Tropical Agriculture (IITA) in Nigeria based on 38 characters, which included
both quantitative and qualitative traits. Substantial agro-morphological diversity
was revealed, which they recommended to be confirmed using molecular markers.
Ntundu et al., (2006) identified some vegetative traits that had prominent loadings
in principal components analysis, and these are useful in distinguishing bambara
groundnut landraces. Similar traits, like seed weight, internode length, petiole
length, leaflet length, leaflet width, were identified as important traits in
distinguishing between wild and domesticated bambara groundnuts when
analysed with isozyme markers (Pasquet et al., 1999). In addition morphological
characters which can be highly correlated to grain yield give breeders the choice
to make decisions as to which traits to select for in bambara groundnut landraces
(Karikari, 2000).
Morphological markers have been used for phenotypic diversity studies in a
number of crops. Several numerical taxonomic techniques have been successfully
employed to classify and measure the patterns of genetic diversity in the
13
germplasm collection by other researchers working on crops such as black gram
(Vigna mungo) and Mungbean (Vigna radiata) (Ghafoor et al., 2001), soybean
(Glycine max) (Cater et al., 2001) and wheat (Triticum aestivum) (Bechere et al.,
1996). The comparison of phenotypic and genotypic variation within and between
several other crops has been examined to provide accurate taxonomic and genetic
differentiation in Musa spp, (Crouch et al., 2000), cowpeas (Vigna unguiculata)
(Omiogui et al., 2006) and sorghum (Sorghum bicolor)(Can and Yoshida, 1999).
Agronomic and morphological characters have been used to identify traits
contributing to important traits such as yield in crops like bambara groundnut
(Makanda et al., 2009) and soybean (Malik et al., 2007).
In a strategy to develop what they termed phenotypic similarity index (PS), Cui et
al., (2001) used morphological and agronomic traits to study the phenotypic
diversity of Chinese and North American soybean. A total of 47 Chinese and 25
North American cultivars were assessed for 25 characters. Their results showed
more phenotypic diversity among the Chinese cultivars, than the North American
cultivars, they also found clear differences between the two groups. From the use
of morphological markers they managed to come up with a strategic plan to
broaden the North American germplasm by the introgression of Chinese cultivars,
especially those from different clusters.
Swamy et al., (2003) used 20 agronomic characters to study the phenotypic
diversity and identify traits with higher loadings in principal component analysis
(PCA) for use as best descriptors in the core collection of Asian groundnut
(Arachis hypogaea). A total of 504 accessions which consist of 274 accessions of
subs. fastigiata (var. fastigiata and vulgaris) and 230 subs. hypogaea (var.
hypogaea) were evaluated. A significant difference between fastigiata and
hypogaea groups was found, and the principal component analysis showed that all
the traits contribute significantly to variation for both groups except pod yield per
plant, which did not appear in the first five principal components for both groups.
Low variation in the pod yield per plant indicated that it was not significantly
contributing as a descriptor in these accessions.
In studies to determine the selection criteria for cowpea (Vignaunguiculata)
breeding, Omoigui et al., (2006) analysed the genetic variability and heritability
14
of reproductive traits of cowpeas. They found a substantial amount of genotypic
coefficient of variation (GCV) and broad sense of heritability (h2) among the
selected cultivars on a number of traits. Higher heritability for 100-seed weight
(0.98), plant height (0.94), days to flowering (0.83) and days to maturity (0.77)
were recorded which was an indication that progress could be achieved in
selecting these traits for cowpea improvement
1.6.3
Biochemical markers
Isozyme analysis was the first technique used in the estimate of genetic variance
developed by Lewinton and Hubby in (1966). Isozymes are protein molecules
with different charges, and can be separated by gel electrophoresis based on their
molecular sizes, weight and electrical charges (Hedrick, 2005). The use of
isozyme is simple and cheap, since no DNA or sequence information, primers and
expensive PCR machines are need as in other marker types. Isozyme markers
have the advantage of being co-dominant, giving them an advantage over other
markers such as RAPDs, which are dominant markers and they are reproducible
(Spooner et al., 2005). The main disadvantage is that there are few isozyme assays
per species, and the enzymatic loci account for a small and non-random part of the
entire genome. Isozyme analysis is also affected by plant tissue and plant
developmental stage (Mondini et al, 2009). Different tissues in the same plant can
reveal different isozyme variation.
Koenig and Gepts, (1989) employed nine polymorphic isozyme loci to study the
genetic diversity of 83 wild common beans (Pharsalus vulgaris)from both the
Mesoamerican and Andean regions. The study was able to confirm the existence
of the two gene pools, Mesoamerican and Andean accessions. Genetic diversity of
Dst, Hs, and Ht were estimated. Dst estimates the total gene diversity distributed
among populations, Hs estimates mean heterozygosity with the population, while
Ht measures the mean heterozygosity in the entire population. The level of genetic
diversity within the wild species was Ht =0.13, and non-significant within
accession of Hs =0.006, and between accessions a moderate between Dst = 0.126
was recorded. Pasquet et al., (1999) used isozymes to investigate the population
structure of bambara groundnut and partition the genetic diversity between
domesticated and wild forms. A high genetic Nei genetic identity
of 0.948
15
between the wild and domesticated bambara groundnut landraces lead to a
conclusion that the wild bambara groundnut is the progenitor of domesticated
landraces.
To augment the initial description based on morphological markers, biochemical
markers were introduced and later replaced by DNA molecular markers which are
more robust as compared to both morphological and biochemical markers.
1.6.4
Molecular Markers
Molecular markers are fixed marks in the genome found at specific locations of
the genome, there are used to identify specific genetic differences. In order to
precisely identify traits of interest, the marker must be close to the gene of interest
so that the allele of both the marker and the gene could be inherited together.
DNA markers are passed on from one generation to another through the laws of
inheritance (Semagn et al., 2006). Several markers are available to choose for
genetic diversity studies. The selection criteria could be based on cost, technical
labour, level of polymorphism, reproducibility, locus specificity and genomic
abundance (Garcia et al., 2004). Molecular markers are useful in the development
of genetic and physical maps, and have increased the efficiency of indirect
selection of marker linked traits, generally markers are classified into
hybridisation based DNA markers and PCR-based DNA markers (Gupta et al.,
1999).
1.6.4.1
Hybridisation (Sequence dependent)
1.6.4.1.1
Random Amplified Fragment Polymorphism (RFLP)
RFLP was the first DNA marker system which was widely used and is based on
sequence differences which affect restriction enzyme recognition sequences.
Anumbers of steps are required in RFLP analysis. Restriction enzyme digested
genomic DNA is size fractionated by gel electrophoresis then transferred to a
hybridisation membrane. A ‘DNA probe’, a short fragment of labelled DNA, is
hybridised to the filter (Saiki et al., 1985; Kumar et al., 2009).The differences are
caused by evolutionary processes, spontaneous mutations and unequal crossing
over(González-Chavira et al., 2006). RFLP can also result from differences in
16
DNA sequences (additions or deletions, or gross chromosomal changes such as
inversions or translocations) and these changes the fragment sizes detectable as
restriction fragment length polymorphisms (Michelmore and Hubert, 1987).
Velasquez and Gepts, (1994) employed RFLP for diversity analysis of 85
common bean accessions in their center of origin. The accessions were classified
into two major groups the Middle America and the Andes. The genetic diversity
they recorded (Ht =0.38) was twice that they found when using isozyme markers.
Overall their analysis of both RFLP and Isozyme showed that RFLP revealed
more polymorphism.
However, RFLP has a number of disadvantages. It is time consuming, often uses
radioactive reagents, and requires large quantities of high quality genomic DNA
(Mondini et al, 2009). The RFLP technique has a problem of detecting low
polymorphism and few loci per assay; it is also not amenable to automation
(Semagn et al., 2006). The limitations in terms of routine use of RFLP lead to the
development of other markers such as RAPDs (Roy, 2000). Garcia et al., (2004),
compared the efficiencies of random amplification of polymorphic DNA (RAPD),
restriction fragment length polymorphism (RFLP), amplified fragments length
polymorphism (AFLP) and simple sequence repeat (SSR) to assess the genetic
diversity of 18 tropical maize inbred lines. They employed a total number of 774
(AFLPs), 262 (RAPDs), 185 (RFLP) and 68 SSR markers for genetic diversity
studies. The estimates of genetic distance correlation was higher for AFLP and
RFLP (r =0.87), followed by AFLP and SSR (r =0.78), and RAPDs and SSR (r
=0.33). The higher similarity between AFLP and RFLP markers are attributed to
the fact that the two techniques are based on restriction site changes and both
produced relatively higher polymorphism among the selected maize inbred lines.
1.6.4.1.2
Diversity Array Technology (DArT)
DArT is a micro-array hybridisation based technique that enables whole genome,
high throughput and screening (Jaccourd et al., 2001). In DArT, DNA samples
from a representative sampling of the germplasm are assembled to make up a
diversity panel. A complexity reduction method is carried out for the genomic
DNA of the representative germplasm. The genomic representation derived is
then cloned and individual inserts arrayed onto a microarray to form a discovery
17
array. The labelled DNA representations from individual test samples are
hybridised to the discovery array. The polymorphic DArT markers can be
identified as present or absent (Wenzel et al., 2004; Semagn et al, 2006) in a
single genotype. Various complexity reduction methods can be applied. A number
of DNA based molecular markers available are hampered by their dependence on
gel electrophoresis, therefore resulting in lower throughput. DArT is a genetic
marker system which requires low quantities of DNA and can provide
comprehensive genome coverage in organisms without prior DNA sequence
information (Jaccoud et al., 2001).
DArT markers revealed low levels of genetic diversity between cultivated and
wild pigeonpea (Cajanus cajan) (Yang et al., 2006), in bambara groundnut DArT
markers revealed a higher genetic diversity among a subset of 40 accessions
selected from a representative of the 124 landraces (Olukolu et al., 2011).Genetic
diversity and mapping have also been carried out in crops such as barley (Wenzl
et al., 2004; Zhang et al., 2009), wheat (Akbari et al., 2006), and sorghum (Mace
et al., 2009). They have also been used in QTL analysis of root-lesion nematode
resistance in barley (Sharma et al., 2011), and mapping kernel characteristics in
hard red spring whear lines (Tsilo et al., 2010). Recently, Briñez et al., (2011),
used DArT markers to assess the genetic diveristy of 89 common bean
accesssions. The Neighbour-Joining distance matrices was employed to
distinguish two major gene pools of common beans, the Mesoamerican and the
Andean, which was in agreement with previous studies conducted, based on
morphological markers, biochemical, and other molecular markers such as AFLP.
DArT are dominant markers thus are unable to differentiate heterozygous loci
from homozygous, but have the advantage of high locus specificity, due to their
detection by hybridisation (Jaccoud et al., 2001). While SSR markers have an
advantage over DArT markers because they are co-dominant, highly polymorphic
and widely distributed in the genome (Yang et al., 2006) they have the
disadvantage that they require substantial sequence information to generate.
18
1.6.5
PCR-based molecular markers
The ‘Southern transfer process’ has been almost replaced by the polymerase chain
reaction (PCR) (Mullis, 1990).
PCR is useful in studying DNA sequence
variation as it provides amplification of the DNA between two specific priming
sites in the genome. Polymerase chain reaction based markers require less DNA
per assay than RFLP and are higher throughput.
1.6.5.1
Random Amplified Polymorphic DNA (RAPD)
RAPD markers offered an opportunity to reduce the time and expense taken in
RFLP for genetic diversity and molecular mapping. It is based on PCR
amplification of random DNA segments with short, arbitrary primers (William et
al., 1990). An oligonucleotide used for RAPDs is usually ten base pairs long and
amplifies many loci simultaneously and therefore a number of multiple markers
can be assayed within a single PCR reaction. The amplified DNA is visualised
after ethidium staining and there is no need for hybridisation with labelled probes
as in RFLPs (Kumar et al., 2009).
The technique has been used for identification and mapping QTLs conferring
resistance to Aschochyta blight in chickpea (Santra et al., 2000) and identification
of the Uvf-1 gene which confers resistance against rust in Vicia faba (Avila et al.,
2003). RAPDS have been used in bambara groundnut for some landraces in a
genetic diversity assessment (Amadou et al., 2001; Massawe et al., 2003; Mine et
al., 2003). High levels of polymorphism were reported among landraces using
RAPDs markers in contrast to isozyme markers used by Pasquet et al., (1999).
Twenty-one RAPDs and 29 SSR markers were used to assess the genetic variation
and relationships between subspecies and botanical varieties of cultivated peanut
(Arachis hypogaea) and their relationships with the wild peanut species of the
genus Arachis, Heteanthae, Rhizomatae and Procumbentes. A high polymorphism
of 42.7% for RAPDs and 54.4 % for SSR was recorded for the 13 Arachis
selected, which was relatively high genetic variation for peanut as it is considered
to generally have a lower genetic variation (Raina et al., 2001).
The RAPDs technique is simple and inexpensive and can be used in laboratories
with limited resources. Some short comings of RAPDs include its poor
19
reproducibility and when used in linkage map production, the same loci may not
be detectable in different populations. The false positives observed in RAPDs
emanates from the rearrangement of fragments produced by primer binding sites
and intrastrand annealing and interactions during PCR reactions (Semagn et al.,
2006). RAPDs are dominant markers and do not differentiate between
homozygous and heterozygous markers. The inherent problems of reproducibility
make RAPDs unsuitable markers for transferability of results.
1.6.5.2
Amplified Fragment length polymorphism (AFLP)
AFLP was developed to overcome some of the shortcomings of reproducibility of
RAPDs as the technique combines the digestion of DNA with some specific
restriction endonucleases with a PCR-based technique (Sandal et al., 2002). AFLP
analysis involves the restriction digestion of genomic DNA with a combination of
rare cutting (EcoRI or PstI) and frequent cutting (MseI or TaqI) restriction
enzymes (Vos et al, 1995). Only DNA fragments with nucleotides that flank the
restriction sites that match the selective nucleotides of the primer are amplified
during PCR (Loh et al., 1999). The technique is amenable to high-throughput
analysis which is an added advantage. It is also more efficient and reproducible as
compared to the RAPD (Semagn et al., 2006). AFLPs are highly effective
markers and could be useful in genetic resource exploitation and identification of
novel traits (Crouch and Ortiz, 2004).
AFLP has been used in genetic diversity analysis studies such as in common bean
(Phaseolus vulgaris) (Maciel et al., 2003) in cowpeas (Vigna unguiculata)
(Coulbaly et al., 2002), and in bambara groundnut by Massawe et al., (2002) and
Ntundu et al., (2004). AFLP has also been used in the mapping of the nodulation
loci sym9 and sym10 in pea (Pisum sativum) (Schneider et al., 2002). The first
outline linkage map of bambara groundnut was developed mostly from AFLP
markers; 67 AFLP and one SSR (Basu et al., 2007).
20
1.6.5.3
Microsatellites: Simple Sequence Repeats (SSR)
Microsatellites, or simple sequence repeats (SSR), are tandem di- to tetranucleotides sequence motifs flanked by sequences and are present in most
eukaryotes genomes (McCouch et al., 1997). They arise due to slippage-like
events occurring randomly in stretches of repetitive sequence (Tautz, 1989). This
makes microsatellite a more powerful genetic maker and because of their high
reproducibility and co-dominance they are the marker of choice (Gupta and
Varshney, 2000; Reusch, 2001). Microsatellites are mostly useful in comparative
and association studies, genetic diversity, marker-assisted selection, population
and evolutionary studies (Nunome et al., 2006; Shi et al., 2011). Because of their
high variability they are especially good at distinguishing closely related
individuals (Kumar et al., 2009). A number of microsatellites are now available
for a wide range of crops, such as groundnut (Arachis hypogaea) (He et al., 2003;
Cuc et al., 2008), pigeonpea (Cajanus cajan) (Odeny et al., 2007; Saxena et al.,
2010), bambara groundnut (Basu et al., 2007), chickpea (Cicer arietinum) (Sethy
et al., 2003) and common bean (Phaseolus vulgaris) (Blair et al., 2011).
The technical simplicity, small amount of DNA required and high power of
genetic resolution had led to SSR markers slowly replacing other markers. The
microsatellite amplification protocol is easy, once primers have been designed for
a specific locus. After amplification of microsatellites by PCR, the products are
separated by capillary gel electrophoresis and detection of amplified allele can be
achieved by a laser induced fluorescence detection system. The use of
fluorescence labelled primer and laser detection (automated genotyping),
improves throughput, accuracy of call. The cost of fluorescent label attached to
each primer, which could be prohibitive, could be reduced by the three primer
procedure (Schueke, 2000). Multiple loci can be analysed simultaneously through
multiplexing. The major problem with microsatellites is that they need to be
isolated de novo from each species (Zane et al., 2002). In addition, there is poor
transferability of markers developed for one taxon to another (Ellis and Burke,
2007).
21
1.7
Microsatellites development and application
Bambara groundnut has genetic resources that offer potential for food security,
but the lack of molecular marker systems for their diversity assessment poses a
challenge for its genetic improvement and promotion as a crop (Yu et al., 2009).
Microsatellites have proven to be the marker of choice for genetic studies.
Despite their usefulness for many applications, the difficulty, expense and time in
obtaining microsatellite markers is a major hindrance in their use (Zane et al.,
2002). It is important that more microsatellites markers are developed. Basically
there are two strategies used for microsatellite development: microsatellites
markers can be sourced based on DNA sequence information deposited in the
databases (mining in public libraries/databases) or through screening of
genomicDNA libraries specifically constructed for discovery of repeated
sequences in the genome (Ritschel et al., 2004).
1.7.1
Microsatellites markers sourced from databases
The development of SSR markers had been reported through searching expressed
sequence tags (ESTs) databases. An EST is a DNA segment representing the
sequence from a cDNA clone that is derived by reverse transcription from an
mRNA molecule, or a part of it (Gupta et al., 1999). In silico mining of
microsatellites for the plant of interest can be done in the available DNA sequence
databases at the National Centre for Biotechnology Information (NCBI) and
European Molecular Biology Laboratory (EMBL) (Gupta and Varshney, 2000).
The sequences, after having been downloaded and aligned, can be used to identify
unique flanking sequence for microsatellite marker development. The markers
developed have been found to have the same utility as those derived from an
enriched genomic library (Sharma et al., 2007). This marker has been developed
in crops such as peas (Pisumsativum) (Moreno and Polans, 2006), mungbean
(Vigna radiata) (Seehalak et al., 2009), lima bean (Phaseolus lunatus) (GaitánSolís et al., 2002), common bean (Phaseolus vulgaris) Garcia et al., (2011) and
chickpea (Cicer arietinum) (Qadir et al., 2007; Varshney et al., 2007). The
transferability of EST-SSRs has been found to be relatively better compared to
non-ESTs SSR markers (Ellis and Burke, 2007). EST-SSRs markers derived from
Medicago truncatula revealed a significant transferability among other three
22
pulses, peas (Pisum sativum), faba bean (Vicia faba) and chickpea (Cicer
arietinum) (Gutierrez et al., 2005) and SSR markers have been used between
cultivated peanut and wild peanut (Liang et al., 2009).
1.7.2
Construction of genomic library
The simple approach or ‘traditional method’ of obtaining microsatellites has been
to create small insert in a plasmid library then to screen the clones by repeated
rounds of filter hybridisation using oligonucleotides (Akkaya et al., 1992; Strus
and Plieske, 1998). This method was found to be laborious, time consuming and
had low efficiency. The numbers of microsatellites discovered are low, and range
from approximately 0.04 to 12%, especially in those species with low levels of
microsatellite repeat (Nunome et al., 2006).
The technique has since been improved through selective hybridisation. There are
several approaches used to enrich the genomic library for microsatellites, detailed
in Zane et al., (2002). The approach of using enrichment for genomic library with
microsatellites has been modified by Edwards et al., (1996) and has proved to be
popular and applied by many researchers e.g. pigeonpea (Cajaus cajan) Burns et
al., (2001), bambara groundnut (Vigna subterranea) Basu et al., (2007) and
pigeonpea (Cajanus cajan) Odeny et al., (2009).
Factors such as the cloning efficiency, the need to increase the throughput by
sequencing large clones, and hybridisation limit the scope of microsatellites. The
advent of next generation sequencing is most likely to resolve these problems
(Santana et al., 2009). Microsatellites were chosen as the preferred method for
studying the genetic diversity of bambara groundnut. Even though there have been
great advances in genomic technology in several crops species, the availability of
molecular tools such as microsatellites have been limited in bambara groundnut.
1.8
Potential application of microsatellites in bambara groundnut
1.8.1 Conservation of genetic resources
The management and characterisation of germplasm is a starting point for crop
improvement. The germplasm collection is usually too large to be easily accessed
23
by plant breeders; hence the concept of core collection was developed so that a
few representative accessions are selected for use (Glaszmann et al., 2010).
There is a substantial number ofbambara groundnut accessions held by respective
countries in sub-Saharan Africa, some of which have undergone some
morphological characterisation such as those from Burkina Faso (Ouedraogo et
al., 2008) and Tanzania (Ntundu et al., 2006).Some genotypes from IITA have
undergone field characterisation and evaluation using phenotypic markers (Goli et
al., 1995). The use of microsatellites in particular should be able to identify
clusters among closely related materials and identify genotypes distantly related
for selection for breeding purposes. The use of SSR markers can also be helpful in
adding more data to the IITA current passport data (Mayes et al, 2009). In
germplasm some redundancies can occur and microsatellites can be used to
identify the redundant or closely related accessions.
The International Crops Research Institute for Semi-Arid Tropics (ICRISAT)
Upadhyaya et al., (2008) employed 48 microsatellites to analyse a core collection
of 3000 accessions of chickpea (Cicer arietum).They managed to divide the
accessions into four manageable subsets of Desi, kabuli, peas shaped and wild
accessions of Cicer among these accessions.
Genetic improvement has been successfully achieved in other leguminous crops
such as in common bean and soybean.
Similar approaches could be applied on
bambara groundnut, which is lagging behind in terms of genomic research. The
availability of microsatellites in bambara groundnut would enable breeders to
target genes of interest. The application of marker assisted selection (MAS)
techniques in bambara groundnut could be helpful in tagging those traits that are
economically important, for example early maturity, photoperiod insensitive and
high yield, understand drought tolerance in bambara groundnut and identify traits
necessary to enhance water use efficiency. The genomics research already done
(especially ‘omic’) in model legumes can be useful for studying ‘orphan’ crop
such as bambara groundnut.
24
1.8.2
Molecular mappings
Genetic mapping assist in identifying simply inherited markers which are close to
genetic factors affecting quantitative traits (QTLs). Molecular markers allow high
density DNA marker maps to be made for a number of crops, and this provides
the structure needed for the application of MAS. The traits could be genetically
simple or complex quantitative traits, which involve many genes the quantitative
traits loci (QTL) (Doerge, 2003).Sato et al., (2010), observed that MAS and
genomics have not yet been practically deployed significantly for underutilised
crops, even though there is a lot of potential to have a significant impact on these
crops.
The limited availability of microsatellites developed in some leguminous crops
has been attributed to a number of factors, such as low variability in the crop, to
lack of resources for marker development such as in chickpea (Millan et al.,
2006), groundnut (Varshney et al., 2009), pigeonpea (Yang et al., 2006) and
bambara groundnut (Basu et al., 2007).
For cultivated groundnut, the construction of linkage maps for the crop was only
recently reported by Varshney et al., (2009) using 318 recombinant inbred lines
(RIL) population derived from a cross between two cultivated genotypes (RIL-1
:TAG 24 x ICGV 86031). The map consists of 191 marker loci on 22 linkage
groups covering a total of 1785.40 cM with an average distance of 9.24cM. The
mapping population segregates for drought tolerance traits like transpiration
efficiency, specific leaf area and SPAD chlorophyll meter reading (SCMR). Even
though several QTLs were identified, none revealed a high phenotypic variation
that could be used in marker-assisted selection (Varshney et al., 2009).
1.8.3
Marker Assisted Selection and QTL
QTL mapping assists in identifying most heritable variation attributed to the
interaction between two or more genes and their environment. The knowledge
acquired is useful in designing crosses that may lead to improvements in crop
breeding (Collard et al., 2005). Genetic markers have made it possible to identify
regions of the genome (QTL) that contribute to the variation of traits of economic
importance in crops. Such markers can be useful in introgression and to facilitate
backcrossing which would otherwise take several years using just morphological
25
markers (Charcosset and Moreau, 2004). Some major achievements have been
recorded in chickpea and pigeonpea, with tremendous crop improvement (Kumar
et al., 2011).
The present map for bambara groundnut developed by Basu et al., (2007) is based
on F2 population derived by crossing V. subterranea var. subterranea (cultivated)
x V. subterranea var.spontanea (wild). Sixty-seven AFLP and one SSR markers
were mapped on 20 linkage groups spanning a total length of 516cM.Four major
QTLs have been located on the map for seed weight, specific leaf area (SLA),
carbon isotope discrimination (CID; an indicator of water use efficiency in other
species) and number of stems per plant (Basu et al., 2007).
More traits of
economic importance have to be studied.
A study in groundnut was undertaken by Khedikar et al., (2010) for quantitative
trait locus (QTL) analysis for late leaf spot (LLS) and rust, which are two major
foliar diseases in groundnut which cause yield losses of approximately 50-70% in
the crop. Parental genotypes TAG 24 and GPBD 24 were screened with 67 SSR
markers which were found to be polymorphic out of a set of 1,039 SSR markers.
56 of these markers produced 14 linkage groups, spanning 462.4 cM with an
average of 8.25 cM. The 268 recombinant inbred lines of TAG 24 and GPBD 24
were used in the QTL analysis, 11 QTL were produced for late leaf spot with 1.7
to 6.5 % phenotypic variation, 12 QTLs were produced for rust with 1.7 to 55.2 %
phenotypic variation. In this study, they identified a candidate SSR marker
(IPAHM 103) which is linked with a major QTL (rust 01), 55.2%. The marker
was validated for use in marker assisted selection in rust disease in a large number
of germplasm lines (Khedikar et al., 2010).
Some important abiotic and biotic stress, pests and diseases that cause damage
and losses to bambara groundnut have already been intensively studied in other
leguminous crops and their QTLs mapped successfully. Detection and mapping of
major locus resistance for fusarium wilt in common bean (Fall et al., 2001),
resistance to bacterial blight (Singh and Muñoz, 1999), white mold resistance
(Ender and Kelly, 2005), and phosphorus acquisition ability (Beebe et al., 2006)
have been reported. In groundnut, QTLs linked to drought resistance had been
identified (Ravi et al., 2011),
26
High resolution maps and ability to determine marker order is largely dependent
on population size. The smaller populations sizes often results in detection of few
QTLs which could have large phenotypic effects (Semagn et al., 2010).The latest
trend has been to combine QTL mapping with methods in functional genomics.
More saturated maps that include SNPs, ESTs derived markers, and STSs
provides a good opportunity for QTL mapping of highly saturated maps and could
be useful in MAS and comparative mapping (Collard et al., 2005). Expressed
sequence tag collections provide a platform for microarray technology that gives
and provides a potential source of candidate genes.
In a study to identify transcribed portion of the pigeonpea (Cajanuscajan) genome
for genes associated with Fusarium wilt (FW) and Sterility Mosaic disease
(SMD), 16 cDNA libraries were generated from Fusarium infected root tissues
from four genotypes ICPL 20162 and ICP2376 for FW and ICP7035 and TTB7
for SMD. A total of 5,860 expressed (ESTs) for FW and 3,788 for SMD tissues
were also discovered and deposited in the NCBI. This is a good opportunity for
marker development, gene discovery and functional studies for other orphan crops
(Varshney et al., 2009).
The study of rice as a model species for cereal crops has indicated that individual
rice chromosomes were largely collinear with those of other crops species such as
maize, rye, sorghum, barley and wheat and other important agricultural grasses, at
least at a gross level. Researchers identified QTL controlling important agronomic
traits, such as shattering and plant height that had been mapped to collinear
regions among grass species (Xu et al., 2005).
A significant collinearity in gene order had also been reported in a number of
legumes such as common bean and soybean. Yang et al., (2010) undertook a
study to evaluate the efficacy of using soybean gene chip for transcript profiling
in common bean. They hybrised cRNAs from nodule, leaf and roots for soybean
and common bean in triplicate on a soybean Gene Chip, their results revealed that
genes for basic cellular functions and metabolism were highly conserved between
the two species. Their result is an indication of a functional orthologs between this
species, and the study could be extended to other legumes for crop improvement.
27
An example is reported by Zhu et al., (2005) when information from a model
specie Medicago truncatula has been used to map the nodulation receptor kinase
(NORK) gene which is responsible for both bacterial and fungal symbiosis in
other legumes.
1.9
Comparison of genetic diversity estimates methods
A limit to plant breeding has been due to the lack of robust markers such as
molecular markers, previous work was based on pedigree data, morphological,
physiological and cytological measurements (Garcia et al., 2004). The advent of
molecular markers has meant that plant breeders could estimate genetic diversity
faster and easier. Since different marker types differ in their properties, it is
possible they give different estimates of genetic diversity (Rauf et al., 2010). The
comparison of molecular markers for estimating genetic diversity could show how
useful a marker is for a plant breeding purpose(Franco et al., 2001). The estimates
of genetic diversity of makers can be done based on correlation, regression, scatter
plots and cluster analysis (Weir, 1996).
The efficiency and utility of six primer combinations for AFLP and RAPD, 100
RFLP and 36 SSR markers were investigated in12 soybean germplasm by Powell
et al., (1996.). The study consisted of a total of 12 genotypes of Glycine max of
which 2 are wild type Glycine soja, the similarity matrices for the markers were
compared, it revealed that the average similarity matrix was lower for SSR(0.341)
while the other markers were similar AFLP (0.655), RFLP (0.639), RAPD
(0.664). The Mantel test was used to determine the correlation between the
markers and found significant correlation between all maker types (P<0.001). The
highest correlation was between SSR and AFLP (0.855) while the lowest was
recorded between RAPD and RFLP (0.744). Both markers proved to be useful in
the assessment of the selected genotypes.
Lu et al., (1996) compared PCR based methods (RAPDs, AFLP, microsatellitesAFLP, and inter-SSR) with DNA based RFLP to determine the most informative,
and useful in genetic diversity studies based on ten pea genotypes. Their results
revealed that the PCR based method were more informative than RFLP, and trees
28
derived from PCR based markers were significantly correlated with the exception
of inter-SSR derived tree.
Other studies for the comparison of markers were conducted in other crops as
well, Pejic et al.,(1998) investigated the efficiency of RAPD, SSR and AFLP in
the analysis of maize, inbred lines. Garcia et al., (2004), compared the utility of
RAPDs, RFLP, AFLP and SSR markers to find the best marker suitable for maize
inbred lines selection. In wheat, Stodart et al., (2005) compared AFLP and SSR
markers to determine their utility in genetic diversity measurements among the 44
bread wheat landraces from different regions.
The geneticdistanceestimates compared in leguminous crops, include the one from
Maras et al., (2008) who evaluated the ability of AFLP and SSR to detect genetic
diversity among 29 common bean (Phaseolus vulgaris) accessions. Ten primer
combinations of AFLP produced 112 polymorphic bands, while 14 SSR markers
produced 100 polymorphic bands and both markers were able to separate the two
gene pools of Andean and Mesoamerican origin. Jaccard coefficient of similarity
was employed to generate similarity matrix in both markers, the two genetic
distances GSAFLP and GSSSR were evaluated for correlation using the Mantel
correspondence test (Mantel, 1967), and a significant correlation r =0.67 was
found, which shows a good similarity between the two markers.
In comparison of the morphological and RAPDs markers in estimating the
differences among 15 common beans (Phaseolus vulgaris), Dursun et al., (2010),
employed 8 RAPDs and 16 morpho-agronomic markers. The difference between
the two markers was revealed in the displaying of clusters as they differed in
topology. The Euclidean matrix produced by the morphological marker and the
Dice similarity matrix from the RAPDs markers were compared using Mantel
matrix correspondence tests, the results showed no correlation between the two
markers. This lack of correlation was thought to be possibly incorrect
measurements for morphological traits and few samples sizes for RAPDs used in
the study (Dursun et al., 2010). However, in most of the studies conducted to
reveal the genetic distances estimates the relationship between molecular and
morphological markers had been observed to show non-significant correlations
(Burstin and Charcosset, 1997).
29
No study has been conducted to compare the genetic distance estimates of
markers in the germplasm of bambara groundnut. Therefore this study aims to
determine the genetic diversity among the selected bambara groundnut germplasm
employing both morpho-agronomic (qualitative and quantitative) markers, and
molecular markers, and determine the relationship between the two techniques.
1.9.1
The objectives of the study
To develop and characterise microsatellites markers; the development of
markers will have a major impact on the genetic analysis and breeding of
bambara groundnut, particularly in genetic diversity, population structure
analysis implementation of pure line selection.
To characterise selected landraces based on morpho-agronomic characters
and to determine the agro-morphological diversity among landrace and
consequently produce a genetic distances estimate to correlate with the
genetic distance estimates based on SSR.
To conduct a genetic diversity estimate based on SSR markers, which will
consequently produce a genetic distance estimates to correlate with the
morphological marker distance estimates.
To compare morpho-agronomic markers with the SSR markers and
identify any significant correlations and evaluate which is more
informative and whether the costs associated with molecular analysis are
justified.
To establish the genetic similarity among bambara groundnut landraces
sampled across a vast area of sub-Saharan Africa using microsatellites
(SSR) markers since there is little information about this germplasm.
There is constant movement of bambara groundnut germplasm between
various neighbouring countries, and among farmers within the same
country.
The existence of landraces in bambara groundnut means that there are
likely to be multiple genotypes planted in any trial for a landrace. This will
add genetic variability to the already existing environmental variability
and interaction (i.e. VP = VG + EG + VGXE). Co-dominant microsatellite
markers will allow us to determine whether this is more of a problem in
some landraces than others
30
CHAPTER TWO: Materials and Methods
2.1
Introduction
This chapter is divided into two sections: molecular biology (DNA and marker
techniques) and phenotypic (morpho-agronomic) assessment of the germplasm.
Materials and methods that are common in each section are described. Those
procedures described that are specific to some experiments are described under
appropriate chapters. The procedures described were used to carry out
experiments at the University of Nottingham, Sutton Bonington, Campus, (UK)
and Botswana College of Agriculture, Sebele, and Department of Biological
Science, University of Botswana (Botswana).
2.1.1. Standard solutions
A list of standard solutions that were used in the molecular biology experiments
are in appendix 1, while section 2.1.2 is a list of plant materials used in the
phenotypic assessment and molecular biology experiments.
31
2.1.2
List of plant material
Table 2.1.2.1: List of selected landraces used for the characterisation of SSR
markers and DArT analysis, their areas of origin and the cluster where the
landraces were selected. The selection was on the basis of a study conducted by
Singrün and Schenkel (2003), where a total of 223 bambara groundnut landraces
were analysed for genetic diversity using enzyme system EcoRi/MseI amplified
fragment length polymorphism (AFLP).
Landraces
Origin
Region
Cluster
DodR
Tanzania(TZA)
East Africa
DodC
Tanzania (TZA)
East Africa
AS17
South Africa (RSA)
Southern Africa
DipC
Botswana (BWA)
Southern Africa
SwaziRed
Swaziland (SWA)
Southern Africa
TicaNicuru
Mali (MLI)
West Africa
Ramayana
Indonesia(IND)
Asia
LunT
Sierra Leone (SLA)
West Africa
Vssp6
Cameroon (CMR)
West Africa
Nav 4
Ghana (GHA)
West Africa
Nav red
Ghana (GHA)
West Africa
Mahenene black
Namibia (NAM)
Southern Africa
S19/3
Namibia (NAM)
Southern Africa
S19-3
Namibia (NAM)
Southern Africa
UniswaRed
Swaziland (SWA)
Southern Africa
SB16 5A
Namibia (NAM)
Southern Africa
AHM968
Namibia (NAM)
Southern Africa
NAM 1761/3
Namibia (NAM)
Southern Africa
Malawi 3
Malawi (MW)
Southern Africa
Tvsu 569
Cameroon (CMR)
West Africa
Tvsu 610
Nigeria (NGA)
West Africa
Tvsu 747
Zambia (ZMB)
Southern Africa
GabC
Botswana (BWA)
Southern Africa
Tvsu 999
Zimbabwe (ZWE)
Southern Africa
6
1
15
12
14
Core
1
Core
Core
Core
Core
Core
Core
8
Core
1
15
8
Core
4
6
7
16
17
Nine core accessions, have been previously used in BAMLINK experiments. (BAMLINKMolecular, Environmental and Nutritional Evaluation of Bambara Groundnut (Vigna
subterraneaL. Verdc.) for Food Production in Semi-Arid Africa and India.
32
Table 2.1.2.2:A list of 123 bambara groundnut accessions, source and their areas
of origin used in the experiment; 105 bambara groundnut accessions selected in
the greenhouse (35 x 3) samples, and 34 accessions that were selected and planted
in the field experiment in (Botswana).
Tag No.
Accession
Origin
Source
Regions
Grown in greenhouse
Field
Sample 1
Sample 2
Sample 3
Selected
1
(Wild type) 1
Nigeria
IITA
West Africa
2
(Wild type) 13
Nigeria
IITA
West Africa
3
9
Nigeria
IITA
West Africa
A
B
C
C
4
(Wild type) 144
Ghana
IITA
West Africa
A
B
C
C
5
191
Benin
IITA
West Africa
A
B
C
B
A
B
C
B
A
B
C
A
A
B
C
C
A
B
C
C
A
B
C
C
A
B
C
C
6
289
Benin
IITA
West Africa
7
85
Burkina Faso
IITA
West Africa
8
292
Burkina Faso
IITA
West Africa
9
308
Burkina Faso
IITA
West Africa
10
1276
Ivory Coast
IITA
West Africa
11
1284
Central A.Republic
IITA
Central Africa
12
1288
Central A.Republic
IITA
Central Africa
13
1307
Central A.Republic
IITA
Central Africa
14
1315
Central A.Republic
IITA
Central Africa
15
1324
Central A.Republic
IITA
Central Africa
16
1329
Central A.Republic
IITA
Central Africa
17
1337
Central A.Republic
IITA
Central Africa
18
(Wild type) 1206
19
1352
Burkina Faso
Central A.Republic
IITA
West Africa
IITA
Central Africa
20
118
Ivory Coast
IITA
West Africa
21
438
Cameroon
IITA
West Africa
22
440
Cameroon
IITA
West Africa
23
447
Cameroon
IITA
West Africa
24
448
Cameroon
IITA
West Africa
25
(Wild type) 1164
Burkina Faso
IITA
West Africa
26
460
Cameroon
IITA
West Africa
IITA
West Africa
27
467
Cameroon
28
472
Cameroon
IITA
West Africa
IITA
West Africa
29
473
Cameroon
30
476
Cameroon
IITA
West Africa
IITA
West Africa
31
480
Cameroon
32
483
Cameroon
IITA
West Africa
33
484
Cameroon
IITA
West Africa
34
492
Cameroon
IITA
West Africa
35
501
Cameroon
IITA
West Africa
IITA
West Africa
36
502
Cameroon
37
503
Cameroon
IITA
West Africa
IITA
West Africa
38
506
Cameroon
39
529
Cameroon
IITA
West Africa
40
536
Cameroon
IITA
West Africa
41
210
Ghana
IITA
West Africa
42
214
Ghana
IITA
West Africa
43
216
Ghana
IITA
West Africa
44
229
Ghana
IITA
West Africa
45
231
Gambia
IITA
West Africa
46
243
Gambia
IITA
West Africa
47
246
IITA
West Africa
48
790
Zambia
IITA
Southern Africa
A
B
C
C
49
793
Kenya
IITA
East Africa
A
B
C
N/A
IITA
East Africa
A
B
C
A
IITA
Southern Africa
A
B
C
A
A
B
C
C
50
792
51
799
Gambia
Kenya
Madagascar
52
806
Madagascar
IITA
Southern Africa
53
808
Madagascar
IITA
Southern Africa
54
810
Madagascar
IITA
Southern Africa
55
88
Mali
IITA
West Africa
56
89
Mali
IITA
West Africa
57
91
Mali
IITA
West Africa
58
23
Nigeria
IITA
West Africa
59
25
Nigeria
IITA
West Africa
60
32
Nigeria
IITA
West Africa
61
33
Nigeria
IITA
West Africa
62
119
Nigeria
IITA
West Africa
33
Table 2.1.2.2: continued
Tag No.
Accession
Origin
Source
Regions
Grown in greenhouse
Field
Sample 1
Sample 2
Sample 3
Selected
63
120
Nigeria
IITA
West Africa
64
172
Nigeria
IITA
West Africa
65
395
Cameroon
IITA
West Africa
66
275
Nigeria
IITA
West Africa
67
278
Nigeria
IITA
West Africa
68
283
Nigeria
IITA
West Africa
69
286
Nigeria
IITA
West Africa
A
B
C
B
70
329
Nigeria
IITA
West Africa
A
B
C
C
71
330
Nigeria
IITA
West Africa
72
331
Nigeria
IITA
West Africa
73
334
Nigeria
IITA
West Africa
74
335
Nigeria
IITA
West Africa
A
B
C
A
75
348
Nigeria
IITA
West Africa
76
390
Sudan
IITA
Central Africa
A
B
C
B
77
391
Sudan
IITA
Central Africa
78
369
Tanzania
IITA
East Africa
79
371
Tanzania
IITA
East Africa
80
379
Tanzania
IITA
East Africa
81
385
Tanzania
IITA
East Africa
A
B
C
C
82
682
Zambia
IITA
Southern Africa
83
683
Zambia
IITA
Southern Africa
84
696
Zambia
IITA
Southern Africa
A
B
C
B
85
754
Zambia
IITA
Southern Africa
A
B
C
B
86
757
Zambia
IITA
Southern Africa
87
1033
Zimbabwe
IITA
Southern Africa
88
AHM753
Namibia
UoN
Southern Africa
A
B
C
A
89
DipC
Botswana
UoN
Southern Africa
90
S19-3
Namibia
UoN
Southern Africa
A
B
C
C
91
UNIS R
Swaziland
UoN
Southern Africa
A
B
C
C
92
AHM968
Namibia
UoN
Southern Africa
A
B
C
B
93
AS17
South Africa
UoN
Southern Africa
94
Dod C
Tanzania
UoN
East Africa
95
Dod R
Tanzania
UoN
East Africa
96
GAB C
Botswana
UoN
Southern Africa
97
JAC B
Botswana
UoN
Southern Africa
98
KABCA 4
Sierra Leone
UoN
West Africa
99
SB4-2
Namibia
UoN
Southern Africa
SB16 5A
Namibia
UoN
Southern Africa
100
101
UNIS C
102
V5 6O A
Swaziland
Botswana
UoN
Southern Africa
UoN
Southern Africa
103
TICANICARU
Mali
UoN
West Africa
104
S-1913
Namibia
UoN
Southern Africa
A
B
C
A
105
MaheneneBlack
Namibia
UoN
Southern Africa
A
B
C
A
A
B
C
A
A
B
C
A
106
YOLA
Nigeria
UoN
West Africa
107
NAV-RED
Ghana
UoN
West Africa
108
NAV-4
Ghana
UoN
West Africa
109
BOTS 1
Botswana
DAR
Southern Africa
110
BOTS 2
Botswana
DAR
Southern Africa
111
BOTS 3
Botswana
DAR
Southern Africa
112
BOTS 4
Botswana
DAR
Southern Africa
113
BOTS 5
Botswana
DAR
Southern Africa
114
CS37(RP)
Kenya
UoN
East Africa
115
CS129 (RP)
Kenya
UoN
East Africa
116
VSSP 11
Cameroon
UoN
West Africa
117
VSSP 6
Cameroon
UoN
West Africa
A
B
C
C
118
RAMAYANA
Indonesia
UoN
Asia
A
B
C
B
Southern Africa
A
B
C
B
Asia
only DNA samples sourced
119
Hybrid
120
BC
UoN
Indonesia
UoN
Bogor
121
BH
Indonesia
Bogor
Asia
only DNA samples sourced
122
GC
Indonesia
Gresik
Asia
only DNA samples sourced
123
GH
Indonesia
Gresik
Asia
only DNA samples sourced
N/A: Not applicable for accession 49-Acc793KEN which was not planted in the field due to
shortage of seeds. A, B, C: represent plant/genotype 1, 2, 3 respectively, for the same landrace
34
2.1.3
Overview of experiments.
Laboratory experiments were conducted to investigate the genetic diversity of
bambara groundnut, using a core of 24 landraces. These landraces were selected
based on a study conducted by Singrün and Schenkel (2003) (Table 2.1.2.1).
These formed an initial test set to evaluate marker polymorphism and provided a
link to previous DArT analysis. A total of 75 pairs of microsatellites were
characterised (Appendix 2).
The second aspect of the project involved the characterisation and evaluation of
bambara groundnut landraces for quantitative and qualitative characters. This
was initiated in the agronomy bay glasshouse at the University of Nottingham
using 119 bambara groundnut landraces (87 sourced from International Institute
of Tropical Agriculture, 27 from The University of Nottingham, and 5 from
Department of Agricultural Research, Ministry of Agriculture, Botswana) the
experiment was conducted at the School of Biosciences, UK (27May planting to
4November 2008 harvest). DNA was extracted from the 119 accessions which
were planted in the glasshouse and sent for DArT analysis (Mayes et al., 2009) as
well as a subset of samples being analysed with microsatellite markers
characterised in this project.
Field work was conducted on 34 lines derived from seed from single plants
selected from the previous year’s experiment (among the 119 bambara groundnut
landraces) (Table 2.1.2.2)and planted at the Botswana College of Agriculture,
Notwane (Sebele) field (11December 2008 to 11May, 2009). Extraction of DNA
was carried out at the University of Botswana, Biological Science Department
from (20May to 5June, 2009).
A selection of the five best lines from the field experiment for use as potential
varieties to use in (Botswana), and a growth room experiment was conducted for
characterisation, evaluation and genetic analysis of these set of lines
35
Development of
microsatellites
24 genotypes were
selected based on
Singrü n and
Schen kel, (2003 )
landraces Table
2.1. 2.1
201 DArT marker
analyses co nducted
on 24 genotypes
75 mic rosatellites
characterised
119 landraces planted in
the agronomy bay (Green
house) Table 2.1. 2.2
Microsatellites
analyses conducted
using 12 markers
3 individuals of 35 lines
were selected (among the
119) and characterized (3 x
35 = 105) individuals Table
2.1.2. 2
34 lines from a single plant
were selected for field
experiment (amon g the 105
genotypes) Table 2.1.2. 2
Microsatellites
analyse s
conducted using
20 markers
Microsatellites
analyses
conduc ted using
similar set of 20
markers
5 individuals of genotypes
from the best lines were
Characterized
Microsatellites
analyses conducted
using similar set of
12 markers
Figure 2: Diagrammatic representation of the setup of the project
36
2.2
Methodology for Marker and DNA techniques
2.2.1
Introduction
Characterisation of molecular markers is dependent on the amplification of DNA
extracted from samples based on available markers. This section describes the
procedures used to extract DNA from the plant materials listed in section 2.1.2.2,
the
quantitation
and
amplification
employed.
The
development
and
characterisation of markers listed in Appendix 2, and their subsequent analysis are
described. The use of capillary electrophoresis to size the amplified fragments and
the potential genotyping errors and some mitigating strategies in microsatellites
analysis are described.
2.2.2
Plant materials
2.2.2.1 Plant materials for microsatellites characterisation and DArT analysis
Isolation of bambara groundnut DNA was undertaken for use in the
characterisation of microsatellites. A core set of 24bambara groundnut landraces
listed in (Table 2.1.2.1) formed an initial test to evaluate the marker
polymorphism and provide a linkage to DArT analysis.
2.2.2.2 Plant materials used for population structure analysis
A total of 119 bambara groundnut accessions were planted in the agronomy bay,
118 are from Africa while 5 are originally from Indonesia (4 were directly
sourced from Indonesia, and 1 is from The University of Nottingham Stock)
(Table 2.1.2.2.).African accessions were sourced as follows; eighty-seven
accessions were sourced from the International Institute of Tropical Agriculture
(IITA; Nigeria) while 27 accessions were sourced from the University of
Nottingham stocks and five were supplied from
Department of Agricultural
Research, Ministry of Agriculture, Botswana. The origin of the complete
accession list was derived from five major regions; 5 genotypes are from Asia
(Indonesia), 11 from Central Africa region (Central African Republic and Sudan),
11 from East Africa (Kenya and Tanzania), 29 from Southern Africa (Botswana,
Namibia, Madagascar, Swaziland, Zambia and Zimbabwe) and 67 from West
Africa (Nigeria, Ghana, Benin, Burkina Faso, Ivory Coast, Cameroon and Sierra
Leone).
37
2.2.2.3 Plant materials used for genetic diversity analysis
A total of 35 bambara groundnut landraces in (Table 2.1.2.2) were used in the
study. Twenty one accessions were sourced from IITA (Nigeria), while 12 were
from the University of Nottingham stocks and two were sourced from Botswana.
Three plants per accession which makes 105 genotypes were used for the genetic
diversity study analysis to estimate the genetic diversity within landraces.
2.2.3
DNA extraction
In this experiment the intention was to get good quality genomic DNA to use in
PCR for the optimisation of the 75 available primer pairs that has been developed
and characterised in this experiment section 2.2.5. The 119 landraces were used in
population structure analysis, while the 105 genotypes were used for genetic
diversity analysis. The GenElute Plant Genomic DNA kit (Sigma Aldrich) was
used in the DNA extraction, since it has been shown to produce high quality DNA
for PCR (Basu et al., 2007).
2.2.3.1
Sigma DNA extraction Kit
The GenElute Plant Genomic DNA kit (Sigma Aldrich) protocol was followed for
DNA extraction. Fresh young growing leaves were picked and collected in 50 mL
Falcon tubes (Sarstedt) and placed on liquid nitrogen. About200mg of leaf tissue
was ground with a pestle in a mortar under liquid nitrogen until sample became
fine powder and transferred into a pre-chilled Eppendorf tubes. Then 350µL of
lysis A solution and 50µL of lysis B solutions were added and mixed thoroughly
by vortexing. The mixture was then incubated at 65oC for 10 minutes with
occasional inversion. 130µL of precipitation solution was added to the mixture
and mixed by inversion before incubating on ice for 5 minutes. The tube was
centrifuged at 13,000 rpm for 5 minutes to separate debris, proteins and
polysaccharides. The supernatant was transferred to a GenElute filtration column,
and centrifuged at 13,000 rpm for 1 minute. 700µL of binding solution was added
to the flow through in the collection tube and mixed by inversion. 500µ L of
column preparation solution was added to the binding column to activate the
retention of DNA and then centrifuged at 13000 rpm for 1 minute and the flow
through discarded. 700µL of supernatant was added into the prepared column and
centrifuged for 13,000 rpm for 1 minute, and the flow through was discarded.
38
The binding column was placed into another tube for the first wash with 500µL of
wash solution and centrifuged at 13000 rpm for 1 minute. The washing with
500µL of wash solution was repeated but now centrifuged at 13,000 rpm for 3
minutes. The flow through was discarded while the binding column was
transferred to a new collection tube, and 100 µL of pre-heated elution solution at
65oC is added to each column and centrifuged at 13,000 rpm for 1 minute. The
genomic DNA was quantitated (as below) and stored at -20oC for later use.
2.2.4
DNA quantitation
Agarose gel visualisation under (UV light) was used to estimate the quantity and
approximate size (quantity) of the DNA. 5 µL of each sample of isolated DNA
was loaded onto a 1% Agarose Molecular Grade (Bioline) gel in 0.5 x TBE buffer
alongside a range of uncut lambda DNA standards containing 500 to 25 ng DNA.
The DNA was stained by adding (2.5 µL of a 10mg/mL) ethidium bromide before
pouring and quantification of the DNA was achieved by comparing the intensity
of the DNA bands from the DNA extraction with the intensity of the bands from
the lambda DNA standards. Approximate DNA loading (± 20ng) can be obtained
by comparing band intensities by the eye. After quantitation DNA samples were
diluted to 10ng/µL for PCR.
2.2.5
Microsatellite development
Development of microsatellites libraries was undertaken based on the method of
Edwards et al., (1996). The technique involves the hybridisation of restriction
digested and PCR amplified genomic DNA to small filters carrying simple
sequence repeat oligonucleotides (SSRs), followed by the elution and
amplification. Rather than cloning, a mixture of enriched libraries was
pyrosequenced (Roche 454).
The basic approach is given in Basu et al., (2007). Sequences containing
microsatellites repeat motifs were identified using the MISA.pl Perl script.
Primers were then designed flanking the motifs with the aid of the Primer3 web
interface program
(http://fokker.wi.mit.edu/primer3/input.htm) (Roven and
Skaletsky, 2000). A total of 75 primer sets (Appendix 2) were designed, PCR
amplification and optimum annealing temperatures were determined.
The
following criteria were used for primer design: primer length of 18-27, GC
39
content 20-80, Tm 57-63oC, product size 70-300 bases. Primers were designed
and synthesized by MWG Eurofins. Microsatellites were not directly labelled with
WellRed dyes from Beckman Coulter, but they were labelled using a three-primer
‘tagged’ reaction (Schuelke, 2000).
2.2.6
PCR gradient optimisation for primer annealing temperature
The polymerase chain reaction (PCR) involves in vitro amplification of DNA
through a series of three polymerization cycles, the DNA denaturation, primer
templates annealing and DNA synthesis by thermostable DNA polymerase.
Optimization of PCR involves testing a number of factors, such as annealing
temperature (Ta), poor results showing multiple bands on agarose are reflected
when the Tais too low, even when Tais too high the desired products quality is also
reduced due to the poor annealing of primers (Rychlik et al., 1990). Gradient PCR
helps to identify the optimal annealing temperature for pairs of primers. The range
of annealing temperatures over which amplification occurs also gives an
indication of how reliable the primer pairs are in PCR.Seventy five primer pairs
were screened and optimised for annealing temperature using the genomic DNA
extracted from the 24 genotypes in (Table 2.1.2.1) this was done to ensure optimal
primer performance and to identify the best primers for tagging.
The PCR reaction mixtures (20µL final volumes) that contained approximately
10ng (2µL) of template DNA were constructed as given below in 96-well plates
(Thermo Scientific): PCR Buffer (New England BioLabs; includes MgCl 2 to
1.5mM final (2µL). 20 µM Forward primer (0.5µL), 20 µM Reverse primer
(0.5µL). 10 mM dNTPs (0.4µL) (Promega corporation), Thermus aquaticus
polymerase (Taq) (New England BioLabs) (0.2µL) and 14.4 µL of sterilized
distilled water. The plate was briefly centrifuged at 3,700 rpm in an Eppendorf
refrigerated centrifuge (5180) to bring down the contents and sealed with
Thermowell® Sealing mat (Fisher Scientific). Amplification was carried out in a
Thermo HybaidPCR gradient machine (Thermo Hybaid Express) programmed
with the following cycling regime: 94oC for 3 minutes, 35 cycles of 94oC for 1
minute, 12 temperatures ranging between 45-60oC for 1 minute, 72oC for 1 minute
and final extension at 72oC for 10 minutes. The optimised annealing temperatures
for each primer are shown in Appendix 2.
40
2.2.7
Gel electrophoresis of PCR products
5µL of 6x loading buffer (standard reagent) was added to each sample and gently
mixed before being given few seconds spin at 3,700 rpm in an Eppendorf
refrigerated centrifuge (5180). After amplification, the reaction products were
analysed by gel electrophoresis alongside a 2-log ladder (New England Biolabs)
on a 2 % agarose gel in 0.5 x TBE, with ethidium bromide (Promega corporation)
(2.5 µL of a 10mg/stock added before pouring, using 26-well combs (Biorad
Maxi gel, model). After running the gel at 90 Voltage for 45 min, it was visualised
by illumination with UV light and images taken using a Biorad (Gel DOC 2000),
and hard copies of images were printed on a thermal printer (Mitsubishi P91) for
the analysis of bands. The optimal annealing temperatures were determined based
on the strongest band intensity temperature, that the product was approximately
the expected size and also that the amplification was reasonably clean with little
track background.
2.2.8
Three primer systems
To reduce the costs and facilitate screening of large numbers of potential
microsatellites, a three primer system was used. One of these primers carries a
fluorescent label and these fluorescent labels are relatively expensive. To
overcome this especially when using a large number of microsatellites for
genotyping, Schuelke (2000) devised a three primer system method, whereby, a
sequence-specific (M13) tail is added at the 5’ of the forward primer to give the
‘Tagged-Forward’ primer. A sequence–specific reverse primer is used in the
reaction, together with a labelled M13 sequence primer. The amount of the tagged
forward primer should be roughly 1/10th of the reverse primer. The remaining
9/10th of the forward reaction primer is made from the fluorescently-labelled M13
primer. The PCR conditions are set in such a way that during the early PCR
cycles, the specific forward primer with its M13 (-21) sequence is incorporated
into the accumulating PCR products. As the tagged forward primer is exhausted in
the PCR reaction, the universal M13 takes over as the forward primer due to the
PCR products now having this priming sequence and this incorporates the
fluorescent dye into the final PCR product. The M13 (-21) primer genotyping
protocol provides a cheaper way to use commercially available fluorescent
labelled dye primers.
41
The ‘Tag’ CAC GAC GTT GTA AAA CGA C sequence was fused to each
forward primer at the 5’ end that had optimised well in the first round of untagged
priming. In this study 69 primers were tagged out of the 75. Example of Tagged
Forward primer D1: 5’ CAC GAC GTT GTA AAA CGA CTG CTT CTT CAA
GGA GGA AGT AAG T 3’ where the underlined sequence represents the M13
Universal Tag, while the rest of the sequence is the bambara groundnut specific
microsatellite primer sequence. Tagged forward primers were ordered directly
from MWG Eurofins. The main disadvantage of this approach is that extending
the 5’ end of the forward primer with a non-specific sequence of 18bp can
destabilize the reaction and make successful amplifications less likely (Basu et al.,
2007). Primers that showed multiple bands and or for which no clear
amplification product occurred were not selected for further use.
2.2.9
PCR amplification of microsatellites
A total of 68 microsatellite markers in (Appendix 2) were assayed against 24
diverse bambara groundnut genotypes listed (Table 2.1.2.1), which were expected
to be a reasonable representation of the diversity present in bambara and had
differences in morphological traits and collection sites. The 24 accessions were
selected from 17 clusters identified from the analysis of 223 bambara groundnut
accessions, selective amplification was conducted using EcoRI and MseI primer,
10 AFLP primer combinations and one heterologous SSR primer were used
(Singrun and Schenkel, 2003).The distribution of the 24 accessions among the 17
clusters is given on (Table 2.1.2.1).
The M13 labelling primer was chosen to label the PCR products with blue, green
or black fluorescent dye (WellRED primers; Sigma). For the preparation of a 10x
working stock, 10µL from the x1000 (200pmol/µL) stock was mixed with 990µL
of dH2O. This gave all working primer stocks for the three primer reactions as x10
(including M13). The following components were used in the polymerase
reaction: dH20 (11.4µL), PCR Buffer (2µL), 20 µM Forward primer (0.2µL),
20µM Reverse primer (2µL), M13 Tag (1.8 µL), 10 mM dNTPs (0.4µL) (NEB)
Taq (0.2 µL), 10ng/µL bambara genomic DNA (2µL). After sample mixing
primer reactions were dispensed into a 96 well plate and spun briefly in an
Eppendorf 5180 refrigerated centrifuge for a few seconds at 3700 rpm. The plate
42
was then placed in the PCR System 9700 (Applied Biosystems), programmed
with the following cycling regimes: 94oC for 3 minutes, 35 cycles of (94oC 1 min,
selected primer annealing temperature for 1 min, 72oC for 2 min) and final
extension at 72oC for 10 minutes. The PCR amplification products were checked
using a 2% agarose gel before size analysis using a CEQ 8000 fragment analyser
(Beckman coulter inc, USA).
2.3.0
Gel electrophoresis of PCR products of Tagged primers
After amplification, the reaction products were analysed by gel electrophoresis
alongside a 2-log ladder (NEB) on a 2 % agarose gel as described in (section
2.2.7). After running the gel at 90 Voltage for 45 minutes, the gel was visualised
by illumination with UV light and images taken for analysis of bands. The bands
intensities were used to determine the amount of PCR product to pool for
multiplexed fragment analysis on the Beckmann CEQ 8000.
An individual
sample which gave a strong amplification product is reduced in a pool, while one
which gave weaker amplification is increased to give a better overall balance.
PCR amplifications which did not work well were repeated.
2.3.1
Capillary electrophoresis
A single PCR product or a set of two to four PCR products were pooled together,
PCR products used usually ranged from (2 – 6 µL) per primer depending on the
intensity of bands recorded on agarose gel. The sample loading solution (SLS)
(Beckman coulter Inc, Fullerton, USA) was mixed with the size standard (SS)
(Beckman Coulter Inc, Fullerton, USA) in the ratio of 1:100 (v/v) and 25µL of the
mix was loaded into each well in a new PCR plate. 4µl of the PCR product from
the different genotypes was added to the SLS:SS mix and covered with a drop of
mineral oil (Beckman Coulter, Inc Fullerton, USA). All PCR products (SSR
fragments) were sized on CEQTM Genetic Analysis System with a 400 bp size
standard. Fragments were analysed with the CEQM 8000 Fragments Analysis
Software Version 8 (Beckman Coulter Inc., Fullerton, USA) the sizes were
manually scored.
43
2.3.2
Analysis of microsatellites
Microsatellite’s hypervariability, abundance and co-dominance has led to them
being employed in many research fields such as population genetics, linkage map
construction, quantitative trait locus (QTL) mapping and molecular markerassisted selection (MAS). When scoring microsatellites a number of errors can
occur, the most common being those due to stuttering, large-allele dropout and
null alleles (Bonin et al., 2004). The occurrence of genotypic errors in the data
can be limited by undertaking some mitigating measures.
In a protocol for
estimating error rates it is recommended that these measures be systematically
reported to attest the reliability of published genotyping studies (Pompanon et al.,
2005). All types of molecular markers are prone to genotyping error and they
occur when the observed genotype of an individual does not correspond to the
true genotype (Bonin et al., 2004).
2.4
Potential genotyping errors and some mitigating strategies in
microsatellite analysis
2.4.1
DNA degradation
The quality of DNA can deteriorate during sampling, extraction and during
storage. A low number of DNA molecules in an extract due to extreme dilution or
degradation leads to low numbers of intact molecules of DNA template for
amplification and this favours allelic drop out and false alleles (Pompano et al,
2005). To minimise DNA degradation DNA stocks were stored in TE buffer and
stored at -20oC. Poor quality DNA produces poor amplification of PCR products,
which could lead to some missing data. In this study the quality of DNA was
tested by analysis on a 1% agarose gel electrophoresis with ethidium bromide, and
stocks were diluted to 10 ng for working stock for all samples.
2.4.2
PCR based sources of error
PCR inhibitors in DNA preparations can contribute to genotyping errors. Low
quality reagents, high temperature and high concentrations of PCR products have
been reported to cause allelic dropout (Pompano et al., 2005). Marker assays
were conducted using standard protocols for bambara groundnut of (Basu et al.,
2007). Microsatellites were optimised through the use of an annealing temperature
44
gradient from (45 oC to 60oC; Hybaid PCR Express) and the optimal temperature
was determined and used in the amplification experiments. Markers were
screened individually using the blue M13 WellRed Dye to provide information on
peak patterns and size ranges of alleles. This helps to avoid unidentified largeallele dropout in multiplexing, identify primers with difficult to interpret patterns
and helps to spot SSRs which are amplifying from more than one locus.
Multiplexing combinations of PCR products were as described in (section 2.3.0).
The amount of PCR product used in a pooled sample was adjusted based on how
strong the signal was after capillary analysis for the single samples. Generally,
twice as much D3 (green) labelled product was added as D4 (blue) labelled
product. It is during the process of identification isolation and amplification by
PCR that some errors, such as null alleles, stuttering and large allele drop out
occur (Van Oosterhout et al., 2004).
2.4.3
Interpretation of capillary electrophoresis
Agarose and slab gel polyacrylamide gel electrophoresis which were widely used
for microsatellites have limitations, particularly in terms of accurate sizing of
alleles (Wang et al., 2009).
The CEQTM 8000 (CEQ 8000: Genetic Analysis
System, Beckman Coulter, USA) provides automated and accurate estimates of
allele sizes, together with the use of a combination of three fluorescently labelled
primers (and a fourth Dye for ladder) (Hirst and Illand, 2001). While the
CEQTM8000 contains automated binning wizard software which can be used to
determine the sizes of alleles of markers, the allele size determination was done
manually to avoid errors that can be brought about by the automated sizing, which
may not differentiate between stutters and true peaks. Visual inspection of
electrophoretic patterns is highly recommended during screening of markers in
order to solve problems of stutter patterns, low height large alleles, which may not
be detected by automatic automation. Dewoody et al., (2006) recommend the use
of both automated calling software and human inspection.
2.4.4
Spectral overlap
Primers were fluorescently labelled and PCR products were multiplexed with a
general guide that D4 (blue) dye labelled products are diluted more than the D3
(green) or D2 (black) labelled PCR products. Despite that, spectral overlaps
45
sometimes create false peaks as shown by figure 2.4.4. The genuinely labelled D4
(blue) labelled PCR product, has bled through into the spectrum of the D3 (green)
labelled PCR product leading to the formation of false peak. This problem was
resolved by multiplexing PCR products with expected size differences, and
scoring multiplexed PCR products simultaneously so that false peaks can be
identified. Generally, the false signal is at a far lower intensity than the genuine
signal, so can be distinguished without major problems.
Figure 2.4.4: A pair of capillary electrophoresis traces of PCR products for blue labelled. The
genuinely labelled blue PCR products, has bled through into spectrum of the green labelled PCR
product and false peaks are shown for sample M18 and M19 (green).
46
2.4.5
Stutter and A-addition
Most loci tend to produce ‘stutter’ bands due to slipping of the polymerase with
respect to template during the Taq polymerase extension step. In addition, Taq
polymerase also has a tendency to add a non-template adenine to the 3’ end of the
newly synthesised strand (Pompano et al., 2005; Bonin et al., 2007). Interpreting
stutter loci can be difficult since heterozygotes can be scored as homozygote for
difficult to interpret alleles which are shallow and contain many stutter alleles.
This consistent mistyping will bias allele frequencies (Dewoody et al., 2006).
Figure 2.4.5; demonstrates one of the potential errors that can occur due to stutter
bands when two alleles with different sizes overlap. Sample at PR 45-H. H11 in
the upper pane and sample E44.E 05 in the lower pane are not easy to interpret on
their own. The alleles appear to be heterozygotes with complex stutter bands, but
it becomes clearer when compared with other samples from the same locus with
similar pattern and shapes. For sample PR 45-H. H11 it becomes clear that it is a
heterozygote when analysing it together with sample PR45-G-G11, while sample
E-44.E05 is clearly resolved to be a heterozygote when compared with E-47.G05
and E51.H05.
47
Figure 2.4.5: Capillary electrophoresis showing a potential scoring error due to the effects of
stutter band and overlap on sample PR 45-H. H11 and 44.E05
48
2.4.6
Short allele dominance
Short allele dominance (large allele dropout) when not detected can lead to a
decrease in sample heterozygosity in microsatellites analysis. Large allele dropout
occurs when during amplification smaller alleles amplify better than the larger
alleles and the larger allele occasionally fails to appear altogether. This
phenomenon can be prevalent in loci with large differences in allele sizes
(Dewoody et al, 2006). Two examples of short allele dominance are shown in
figure 2.4.6 for Primer 15 and Primer 42.
Figure 2.4.6: Capillary electrophoresis showing limited short allele dominance for marker PR 15
top and marker PR 42 bottom, since both are clearly visible and complete drop out did not occur,
correct calling of the peaks could be done.
2.4.7
Allele size binning (Automated binning)
Binning is sorting allele lengths into discrete classes and there could be errors
associated with this process. Several strategies are in place to undertake allele
calling, such as comparing raw data to the database of the expected length and
assign it to the closest data in automated binning and this has been found not to be
a suitable approach in non-model species with no reference data set, such as
bambara groundnut (Amos et al., 2007).
49
Automated binning has been resolved with the software FLEXIBIN which uses
least square minimization and allows allelic drift (Amos et al., 2007). The
common practice of rounding to the nearest whole number usually result in
miscalls and most likely under estimation of allelic richness. In addition the issue
of ‘allelic drift’ also makes it difficult to undertake automatic binning of alleles
(Matschiner and Salzburger, 2009). Allelic drift is a source of some errors because
it is tendency of true allele bins to display a slightly different value from the
known repeat length (Idury and Cardon, 1997). A fixed repeat length is set so that
only allele bins with a specific difference are allowed. Each dinucleotide repeat
unit contributes an effective repeat unit of length in the range of 1.7 to 2.3bp
(Amos et al., 2007; Idury and Cardon, 1997).
Called alleles can be binned using Flexibin (Amos et al, 2007) to produce a
graphical output that allows any potential problems to be identified and rectified.
Figure 2.4.9 shows marker 16, which has a clear discrete allele distribution. For
using microsatellite genotyping, allele sizes should be whole numbers but the
genotype software (CEQ 8000) creates output to two decimals, which is partly a
reflection of the effect of different base composition having slightly different
molecular weights. Their conversion to integer values or ‘alleles’ poses a potential
danger of mis-typing. The usual way of rounding off integers (i.e. <X.4999 = X.0
and Y.5000 = Y+1) was not followed since it could introduce errors. In practice,
the calling needs to reflect the movement of the microsatellite ‘shape’ up and
down the size range.
An example of a potential source of error in rounding off is illustrated in figure
2.4.7. Sample D92 would be called differently at 194 from the rest of the samples
which would be recorded as 195 for D95 and D99 on the basis of rounding, but in
this case all the alleles were recorded as 195. Size calling was done using
FLEXIBIN (Amos et al., 2007), which forces all alleles into a one base pair
periodicity.
50
Figure 2.4.7: Capillary electrophoresis showing potential sources of mistyping errors due to
rounding off alleles during binning. Analysis based on CEQ 8000 software.
Another potential source of genotyping error is illustrated in figure 2.4.8. The
peak shapes that make up the microsatellite were carefully observed to set up a
standard way to identify genuine peaks and to correct allele calling, based on the
shape of peak, their height and size ranges. Sample I3 B09 has a higher peak with
a smaller recording of 211.99 it has a similar pattern to sample I20E09 which has
lower peak with a recording of 212.03. This difference could lead to the samples
called differently, but looking at sample I40G09 it gives a characteristic shape and
height of these alleles and all are called at 213. A similar potential genotyping
error is observed in sample L21 and L23 with a call size of 257 different from
L24, but their similar shape and height led to them being called at 257.
51
Figure 2.4.8: Capillary electrophoresis showing some potential miscalling errors, therefore the
use of allele shapes, their height and size ranges are set as standard way to identify genuine peaks
for correct allele calling.
52
Marker 16
192
Allele sizes ( bp)
190
188
186
184
182
180
0
10
20
30
40
Number of samples in each bin
50
60
Figure 2.4.9: A graphical output of the cumulative allele length for marker 16, which illustrates an
example of an accurately binned marker with clearly defined colours for different alleles as red
and blue. The analysis was conducted with FLEXIBIN (Automated binning) using a one unit
repeat.
The automated binning for marker 16 shows measures of allele size ranging from
183.65 to 190.51 grouped into 6 group repeats (Figure 2.4.9). A Summary of
Flexibin analysis for marker 16 estimated repeat length, standard deviation and
counts of repeats of each length are given in (Table 2.3.1). The estimated repeat
length summary for each allele and the adjustment factors are listed in appendix 3.
53
Table 2.3.1: Summary of Flexibin analysis for marker 16, showing repeat length,
standard deviation and count of each repeat length
Repeats
Length
Mean bp
s.d.
3
183.65
183.65
0.026
5
185.61
185.60
0.020
6
186.59
186.60
0.039
7
187.57
187.56
0.053
9
189.53
189.52
0.048
10
190.51
190.55
0.029
bp= base pairs
sd = standard deviation
2.4.8
Count
6
8
10
10
10
4
Deviation from Hardy Weinberg equilibrium
Testing for Hardy-Weinberg equilibrium has become an important quality control
in genetic data under the assumption that a high error rate will generate some
disequilibrium. However other causes lead to disequilibrium, including selection,
inbreeding and population admixtures through migration or fusion. Genotyping
errors are another primary suspect in any observed deviations from HWE, and if
genotypic error can be ruled out, other possibilities such as admixture should be
investigated (Chen et al, 2005). According to (Pompano et al., 2005) errors can
cause disequilibrium, such as null alleles and allelic dropout.
54
2.5
Data analysis
2.5.1
Data analysis for microsatellites, development and characterisation
2.5.1.1
Microsatellites marker analysis
A total of 68 microsatellite markers in (Appendix 2) were assayed against 24
diverse bambara groundnut genotypes listed (Table 2.1.2.1). A summary of SSR
statistics such as number of alleles, observed heterozygosity (Ho), expected
heterozygosity (He), polymorphic information content and inbreeding coefficient
(f) for each locus were computed using the program PowerMarker version 3.25
(Lui and Muse, 2005).All alleles were binary coded as 1 or 0 for their presence or
absence in each genotype and used for data analysis.
The proportion of alleles shared between two genotypes averaged over loci was
used as a measure of similarity for both markers(DArT and SSR) based on Nei
and Li, (1979)similarity coefficient. The estimation was based on the formula:
GSij = 2Nij/(Ni + Nj), where Nij represents the number of fragments shared by
accession i and j, Ni represents amplified fragments in sample i and Nj represents
amplified fragments in sample j (Nei and Li, 1979).
2.5.1.2
Principal component analysis (PCO)
To examine the genetic relationship between and within all individual landraces
of bambara groundnut, principal coordinate analysis (PCoA) was used on the data
set using multivariate statistical package (MVSP) (Kovach, 2006). The ordination
does not make any assumptions about the distribution of variates or the population
genetics of the population (Kloda et al., 2007).
2.5.1.3
Cluster analysis
A similarity matrix produced was used to generate dendrograms based on the
unweighted pair group method with arithmetic averages (UPGMA) cluster
analysis was performed using multivariate statistical analysis (MVSP) (Kovach,
2006), and dendrogram were produced to show the similarities and differences
between bambara groundnut genotypes. The cophenetic correlation between the
genetic similarity and dendrogram generated was estimated to validate the relation
55
of the original similarity estimates and the binary data matrix analysed using
NTSYS pc version 2.1 (Rohlf, 2000). Cluster analysis was also conducted using
Winboot (Yap and Nelson, 1996) bootstrap analysis with 1000 replications.
2.5.1.4
Comparison of DArT and SSR genetic estimates
The genetic similarity obtained from DArT and SSR were compared by
measuring the degree of correlation between them using the Matrix correlation
correspondence test Mantel Z statistics based on 1000 permutations. The
computations were conducted on NTSYSversion 2.1 (Rohlf, 2000).The
correlations between the similarity matrices was also analysed based on Pearson
product-moment correlation coefficient and Spearman’s rank correlation
coefficient using SPSS version 16.0. These comparisons were conducted to
investigate whether there are any similarities between the genetic distance
estimates generated by these markers.
2.5.2
2.5.2.1
Population structure and genetic diversity of bambara groundnut
Estimation of genetic diversity in the population
For population structure analysis of 123 bambara groundnut accessions, analysis
was based on 12 markers in (Appendix 2). The standard parameters of genetic
diversityas described in (section 2.5.1.1) were analysed based on PowerMarker
Version 3.25 (Lui and Muse, 2005).
2.5.2.2
Estimation of genetic diversity within and among bambara
groundnut populations
To evaluate the relationships between the 123 bambara groundnut accessions,
Principal Coordinate Analysis (PCoA) and cluster analysis were employed based
on the binary matrices that were generated for the presence or absence of alleles at
each locus. The PCoA reveals the major components of molecular differentiation.
The accessions were also colour and shape coded according to origin to help
reveal any relations between the geographical location and genetic differentiation
present in each dataset for PCoA.
Another measure of genetic variation in a population is gene diversity, sometimes
referred to as average heterozygosity; however the two genetic measures are not
56
identical. Gene diversity measures the frequency of alleles at a gene locus, while
average heterozygosity estimates the mean proportion of heterozygosity over all
loci studied (Bergmann and Ruetz, 1991).Inbred populations show few
heterozygotes, but mostly different homozygotes, thus the use of gene diversity
estimates is more appropriate (Weir, 1990).
Genetic diversity estimate gene
diversity per locus and the data is calculated from the sample andestimated using
unbiased estimator (Nei, 1987) on FSTAT version 2.9.3.
Different numbers of genotypes/samples were assayed from various countries.
The estimate of observed number of alleles in a sample is dependent on the
sample size. This problem was resolved through the use of FSTAT software to
calculate the allelic richness in each population based on smallest number of
individual samples (Leberg, 2002). The program estimates allelic richness (Rs)
independent of the sample sizes, and this allows a comparison of genetic diversity
between populations with different sample sizes. It estimates the expected number
of alleles in a sub-sample of 2n genes, given that 2N have been sampled (N ≥ n).
In FSTAT, n is fixed as the smallest number of individuals typed in a sample
RS =
n
i 1
2 N Ni
2N
2n
where Ni is the number of alleles of type i among the 2N gene (Goudet, 2001)
2.5.2.3
Estimation of population structure
To quantify the structure of the populations F-statistics, FST (Wright, 1978) was
calculated using Arlequin version 3.1 based on Weir and Cockerham, (1984) and
pairwise genetic distance among populations (Excoffier et al, 2005) were
generated. The significance threshold of FST was generated by 1000 permutation
testing to get an unbiased P-value for the test data.
To investigate the genetic structure of bambara groundnut landraces, analysis of
molecular variance (AMOVA) was conducted using Arlequin version 3.1. The
accessions were grouped into a three level hierarchy according to the
classification based on PCoA structure analysis in figure 5.1.0.
57
2.5.3
Genetic diversity of bambara groundnut based on SSR markers and
the comparison with morpho-agronomic characters
2.5.3.1
Polymorphism of microsatellites in bambara groundnut
To determine the genetic relationships within and between populations of
bambara groundnut samples, three plants per landrace were used (Table 2.1.2.2).
All three samples of the 35 accessions were counted as individual cases for the
construction of a binary matrix, scored as presence (1) or (0) for absence for each
possible allele to make a total of 105 samples. The 0/1 matrix was used for the
calculation of genetic distances and the generation of cluster data to determine
how the selected bambara groundnut were related.
2.5.3.2
Principal component (PCO) and cluster analysis
The matrix generated with genetic similarity estimates was used to cluster
genotypes and the generation of principal components and principal coordinate
analysis as described in section 2.5.1.2 and 2.5.1.3, and to examine the genetic
relationship between and among all individual genotypes based on MVSP
program (Kovach, 2006).
2.5.3.3
Analysis of Molecular Variance (AMOVA)
Since three genotypes per landrace were studied, analysis of genetic diversity
within each landrace was conducted using an AMOVA analysis with Arlequin 3.5
(Excoffier and Lishcher, 2010).
The total variance among genotypes was
partitioned into variance among populations, among individuals within
populations and within populations; the populations were defined based on the
two groups in figure 6.2.1. The significance of the partitioning of the genetic
variance components was tested using 1000 permutations.
2.5.3.4
Morphological data analysis
Thirty five genotypes and 34 bambaragroundnut lines selected (Table 2.1.2.2)
were studied for variation of morphological and agronomic traits following the
IPGRI descriptors (IITA, BAMNET, 2000) (section 2.6.6; Table 2.6.6).The
description of morpho-agronomic data analysis and the generation of cluster
analysis are described in (section 2.6.6 and 2.7.8).
58
2.5.3.5
Comparison of SSR marker and morphological marker data
For the comparison of morpho-agronomic and molecular (SSR)markers, PCO
analysis, cluster analysis and correlation matrix was conducted on both data set
based on 20 SSR markers and 34 and 37 morpho-agronomic traits recorded in the
agronomy bay and field experiment respectively.
Nei’s 1972 genetic distance was estimated for SSR markers while the Euclidean
distances were estimated for morphological marker. The estimated means were
tested by means of Matrix correspondence test (Mantel, 1967), which uses 1,000
permutations to estimate the correlations significance between distance matrices
and this was calculated using the NTSYS pc software. Simple Pearson productmoment coefficient correlation and Spearman rank’s coefficient correlation were
used to test the correlations based on SPSS version 16. In addition the results for
cluster and PCO analysis were compared to identify any similarities between the
two marker types.
2.6
Morpho-agronomic characterisation and evaluation of bambara
groundnut
2.6.1
Introduction:
Among the 119 accessions planted in the agronomy bay (greenhouse), morphoagronomic assessment on the germplasm was conducted on three individuals of
the 35 bambara groundnut genotypes listed on (Table 2.1.2.2).A field work
experiment was conducted on 34 lines derived from seed from single plants
selected from the greenhouse (Table 2.1.2.2.). The details of the experiment
procedures are described below.
2.6.2
Glasshouse experiment
The experiment was set up in an unheated agronomy bay glasshouse at the
University of Nottingham, School of Biosciences, Sutton Bonington, in the United
Kingdom. The dimension of the glasshouse is 10.1 m x 4.7 metres wide and 2.3
metres high. The glasshouse is made up of conventional aluminium and glass and
it had vent for manual regulation of heat inside the glasshouse.
59
2.6.3
Plant materials
Eighty seven bambara groundnut accessions from the 119 accessions were
sourced from the International Institute of Tropical Agriculture (IITA; Nigeria),
while 27 were from the University of Nottingham and five were brought from
Botswana.
2.6.4
Experimental design
Seed bed preparation was done by digging and raking the soil and applying 290
kg/ha of Ammonium Nitrate fertilizer. Soil was raked to level, and the seedbed
was covered with black plastic to suppress weed growth before planting. The
glasshouse was fitted with a Tiny Tag (Gemini Data Loggers, UK) to measure
temperature and relative humidity every 10 minutes for the entire duration of the
experiment.
Accessions were planted in a randomised complete block design replicated three
times. Two seeds were sown per hole at a depth of 5 cm and spacing of 30 cm x
30 cm (inter- and intra-row) giving 15 plants per row. Seeds were surface
sterilised with 15% by volume NaClO (Sodium hypochlorite) for 15 minutes and
rinsed 3 times in sterile water before sowing which was done on the 27May 2008
and later thinned to 1 plant 21 days after sowing. Each landrace was represented
once per replication, thus each replication was used in emergence counts.
2.6.5
Crop management
2.6.5.1
Photoperiod
Bambara groundnut is a short day length crop, and the glasshouse was receiving
natural long day light, that could adversely affect the pod formation of the crop.
The crop received natural daylight with no supplementary lighting. Day length
was controlled at 12hrs per day by covering with a black polythene screen fitted
over a metal frame above the crop starting at 2000hrs and uncover at 0800hrs to
maintain a 12 hrs photoperiod, from 20June 2008.
60
2.6.5.2
Crop protection
Phytoseilus persimilis was used as a biological pest control against red spider mite
(Tetranychus urticae) every two weeks.
2.6.5.3
Irrigation
The trickle irrigation system was used which consists of PVC micro-porous
tubing placed at each row. Crops received non limiting moisture on a weekly basis
starting from day 0 to 112 days after sowing and approximately a total of 330 mm
of water was supplied (Table 2.6.1)
Table 2.6.1: Amount of irrigation water (mm) applied in the bambara groundnut
experiment in the agronomy bay (Glasshouse) expressed in days after sowing
(DAS) for the duration of the experiment in 2008 season.
DAS
0
8
10
17
22
26
35
42
47
56
61
68
75
84
90
97
102
107
112
Total
Amount (mm)
20
20
10
10
20
20
10
20
20
20
20
20
10
10
20
20
20
20
20
330
61
2.6.5.4
Climatic factors
Air temperature and relative humidity were recorded every 10 minutes
automatically on the tiny tag, the average minimum and maximum temperatures
of 10.9oC and 29.5oC respectively were recorded with an average of 17.4oC
(Figure 2.6.5.1). The average relative humidity recording was 78.7 %, with a
maximum and minimum of 95.6 % and 78.7 % respectively (Figure 2.6.5.2).
60.0 °C
50.0 °C
Temperature o C
40.0 °C
30.0 °C
Average
min
20.0 °C
max
10.0 °C
0.0 °C
0
-10.0 °C
20
40
60
80
100
120
Days after sowing
Figure 2.6.5.1: The maximum and minimum temperature in the agronomy bay (Glasshouse)
experiment for the 119 bambara groundnut landraces grown in the 2008 season
62
120
Relative humidity (%)
100
80
60
Average
min
max
40
20
120
100
80
60
40
20
0
0
Days after sowing
Figure 2.6.5.2 The maximum and minimum relative humidity in the agronomy bay (Glasshouse)
experiment for the 119 bambara groundnut landraces grown in the 2008 season.
2.6.6
Morpho-agronomic traits measurements collected in the greenhouse
In the agronomy bay (green house) experiment 35 accessions (Table 2.1.2.2) that
emerged from all three replications were followed through for data collection. The
accessions were evaluated for 24 quantitative and 10 qualitative characters,
according to the bambara groundnut descriptor list (IPGRI, IITA, BAMNET
2000)and measured both during vegetative growth and after harvesting (Table
2.6.6).
Additional characters measured were days toseedling emergence and leaf area. A
seedling was considered to have emerged when the first true leaf become
visible.Non-destructive leaf area assessment was determined in the glasshouse
using the measurement of the middle leaflet width and length using the equation:
Leaf Area = 0.74 x 3 x leaf number x (leaflet Length x Width x π/4)] developed
by Deswarte (2001). The equation was confirmed by Cornelissen (2004) using the
leaf area meter (LI-COR 3000), and was applied in field experiments in
Swaziland by Edje and Sesay, (2003).
63
Days to maturity was observed with the yellowing and browning of leaves. The
date of final harvest was based on the observation of leaf senescence, then shoot
dry weight was measured by oven drying the above ground of selected plants.The
finalpod yield was determined from each single plant. The pods were oven dried
at 37oC for one week while the plant biomass was oven dried for 48 hrs at 72oC.
2.6.6.1
Quantitative traits measurements in the green house
Data for the 24 quantitative traits consists of; days to emergence, days to 50%
flowering, number of leaves, plant spread(Canopy) (cm), leaflet length (mm),
leaflet width (mm), plant height (mm), internode length (mm), petiole length
(mm), petiole-internode ratio, petiolule length (mm), peduncle length (mm) and
number of stems per plant, were recorded at 10 weeks after sowing. Yield
characters scored after harvest include; number of pods per plant, pod length
(mm), pod width (mm), pod dry weight (g), number of seeds per plant, seed
weight per plant (g), seed length (mm) and seed width (mm) (Table 2.6.6).
2.6.6.2
Qualitative traits measurements in the glasshouse
For qualitative data, individual plants were recorded per plot to represent each
genotype. Thirteen qualitative characters recorded were for testa colour, eye
pattern, testa pattern, pod colour, pod texture, pod shape, seed shape, terminal
leaflet colour, stress susceptibility, leaf shape and stem hairiness and leaf colour
at germination and growth habit.
In the agronomy bay experiment only 10 qualitative data were recorded with
exception of stem hairiness, leaf colour at germinationand growth habit, while in
the field experiment all the 13 characters were recorded.
64
Table 2.6.6: Quantitative and qualitative traits recorded and brief description as
listed from (IPGRI, 2000).
Characters
Days to emergence
(DAE)
Days to 50% flowering
(DAF)
Number leaves per plant
(LNO)
Characters and description
Number of days from sowing to when the first fully expanded leave appears in a
plot
Plant spread (SPRD)
Widest point between two opposite points recorded at 10 WPA
Leaflet length (LL)
Length of median leaflet at the fourth node recorded at 10 WPA
leaflet width (LW)
Width of median leaflet at the fourth node recorded at 10 WPA
Multiply leaflet width X length and number of leaves and use a formula at
10WPA
Measured from the ground level to the tip of the highest point recorded at
10WPA
Leaf Area(LA)
Plant height (PHT)
Internode length (ITN)
Petiole length (PTL)
Petiole-Internode ratio
(PITN)
Petiolule length (PTLL)
Number of days from sowing to first flower opening on 50% of plants per plot
Total number of leaves per plant at 10 weeks after planting (WAP)
Length of fourth internode of the longest stem, recorded at 10WPA
Measured from the stem node to the junction of the three leaflets at the longest
stem at the fourth node at 10WPA
The ratio of the measurement of the petiole and internode
Recorded on the base of the leaflet of the longest petiolule at the fourth node at
10WPA
Peduncle length (PNL)
Number of stems
(STEM)
Recorded on the fourth internode of the longest stem recorded at 10WPA
Days to maturity (DAM)
Number of days from planting to maturity
Shoot dry weight (SDW)
Number of pods per
plant (POD)
Weight of above ground biomass of harvested plants
Number of stems recorded from selected plants at 10WPA
Pod dry weight (PDW)
Average of 5 plants recorded per plot recorded within two months of harvest
Average weight of pods taken from 5 plants per plot within two months of
harvest
Pod length (PODL)
Average length of pod taken from 5 plants per plot within two months of harvest
Pod width (PODW)
Seeds per plant (SNO)
Average width of pod taken from 5 pods per plant within two months of harvest
Average number of seeds taken from 5 plants per plot within two months of
harvest
Seed length (SL)
Average of seed length from 5 seeds taken per plot
Seed width (SW)
Average seed width from 5 seeds taken per plot
Average weight of seed from 5 plants taken with a plot after drying, within two
months of harvest
Seed weight (SWE)
65
2.7
Field work experiment in Botswana
2.7.1
Introduction
A field experiment was conducted at Botswana College of Agriculture (Notwane
farm) Sebele, Botswana from 11 December 2008 to 11 May, 2009 (2008/2009).
Detailed descriptions of the study site, experimental design, and crop management
are given below.
2.7.2
Field site and experimental preparation
The field experiment was undertaken at Botswana College of Agriculture
(Notwane farm) Sebele in Botswana approximately at latitude 24o33’S and
longitude 25o54’E, 994 metres above sea level. The analysis of soils in Sebele
have been recorded as; shallow, ferruginous tropical soils, medium to coarse grain
sands and sandy loams with a low water holding capacity and subject to crusting
after heavy rains (Baker, 1987). After tractor ploughing and harrowing each plot
was hand harrowed to make a fine seedbed and a basal application of single
superphosphate at a rate of 25 P kg ha-1 of fertiliser was applied just before
planting.
2.7.3
Plant material
Thirty four bambara groundnut lines, with seed selected from a single plant of 35
accessions planted in the glasshouse. One individual among the three plants selfed
in the greenhouse was selected for field experiment (Table 2.1.2.2) except for
landrace 49-Acc 793 from Kenya, which produced low number of seeds for the
field experiment.
2.7.4
Experimental design
The design of the experiment was a randomised complete block design with three
replicates, and each bambara groundnut line was assigned randomly to the plots
with sowing done on the 11December 2008. Individual plot sizes were 3.2 m x 0.4
m, with 0.5 m guard row surrounding each experimental plot. Seeds were sown
with 10 cm space between plants and later thinned to 30 cm between plants at 21
days after sowing to remain with 10 plants per row.
66
2.7.5
Crop management
2.7.5.1
Crop protection
During sowing a nematicide (Nemacur 10 GR, Bayer AG) was applied to each
row at a rate of 1.5 gm-1 to prevent the infestation of root-knot nematodes. Plants
were
sprayed
with
insecticide
Malathion
50%
EC
(S-1,2–bis
(ethoxycarbonyl)ethylO,O-dimethyl phosphorodithioate) and fungicide Eria
(triazole, binzimidazole ) (Republic of South Africa) using knapsack sprayer to
control aphids and diseases as needed.
2.7.5.2
Irrigation
The watering regime for the experiment was synchronised with the BAMLINK
project, which had the same planting date and connected through same source of
water supply. Trickle irrigation system was used and PVC micro-porous tubing
placed in each row. Irrigation was applied up to 79 DAS and a total amount of
156 mm of water was applied (Table 2.7.5). In addition the crops received a total
amount of366.87 mm of rainwater, which was recorded through the Hobo
Weather Station Data logger (Weather Tempcon L.t.d.) installed at the site, the
rainfall distribution is shown on (Figure 2.7.5.1).
Table 2.7.5: Amount of irrigation water (mm) applied in the bambara groundnut
experiment in the field experiment in Botswana, expressed in days after sowing
(DAS) for the duration of the experiment in the 2008/ 2009 season
DAS
2
8
9
14
15
21
28
79
Total
Amount (mm)
18
18
18
36
18
18
15
15
156
67
60
50
Rainfall (mm)
40
30
20
10
0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100103106109112115118121124127130133136139142145148
Days after sowing
Figure 2.7.5.1: The amount and distribution of rainfall in bambara groundnut field experiment at
(Notwane) Sebele, in 2008- 2009 season.
2.7.5.3
Climatic factors
A total amount of approximately 522 mm of moisture was received by the crop in
the field including a combination of both applied moisture and precipitation
received. The average temperature was 21.7oC, while the minimum and maximum
were 15.6oC and 28.9 oC respectively as shown in (Figure 2.7.5.2). The relative
humidity recorded an average of 76%, and a minimum and maximum of 41.6%
and 98% respectively (Figure 2.7.5.3), and these were fairly similar to the
recordings found in the agronomy bay experiment in UK.
68
Figure 2.7.5.2: Maximum and minimum temperature in the field experiment for the 34 bambara
groundnut landraces grown at (Notwane) Sebele in the 2008 - 2009 season.
Figure 2.7.5.3: Maximum and minimum relative humidity in the field experiment for the 34
bambara groundnut landraces grown at (Notwane) Sebele in the 2008 - 2009 season.
69
2.7.6
Agro-morphological traits measurements in the field experiment
Data was collected from 15 plants per line, 5 plants from each replication. The
number of pods and seeds per plant were recorded from each harvested plant. The
date of final harvest was based on the observations of leaf senescence and the
final pod yield was determined from five plants per replication. Leaf area (cm2),
measurements were determined in the field based on the middle leaflet width and
length as described in (section 2.6.6).
2.7.7
Statistical analysis of agronomic traits
Similar statistical analysis was conducted for both agronomy house experiment
and the field experiment. To identify any structures on the bambara groundnut
landraces based on the phenotypic diversity, cluster and principle coordinate
analysis (PCoA) were conducted. In order to identify characters that are
contributing more to morphological diversity, principal component analysis
(PCA) was undertaken and Eigenvalues were examined, and for selection of the
best landraces heritability estimates were done, to try to identify characters which
are expected to respond more to selection (Roy, 2000).
2.7.8
Data analysis of agronomic traits
2.7.8.1
Descriptive characteristics
Data analysis for all quantitative characters were subjected to analysis of variance
(ANOVA) using the Genstat version 13.0 (Lawes Agricultural Trust, Rothamsted
Experimental Station, UK) General linearized models (GLM) package to
determine the statistical differences on the traits for the given genotypes. The
mean values, ranges standard error of means and coefficient of variation were
calculated on the 24 agro-morphological data.
The measurements for each trait for all the genotypes were standardized by
subtracting the mean from respective traits and dividing by the standard deviation
using Genstat version 13.0 in order to reduce the influence of the scale differences
(Upadhyaya, 2003). The transformed values for each character had an average of
zero and standard deviation of one and these values were used for principal
component analysis and cluster analysis. The standardized data was used to
70
estimate the matrix of distances between all pairwise combinations of genotypes.
The dendrograms were constructed from measurements of a combination of both
the qualitative and quantitative characters (Hill et al., 1998).
2.7.8.2
Principal component analysis
Principal component analysis (PCA) is a technique that summarises patterns of
correlations among observed variables and reduces a large number of observed
variables to a smaller number of components with several linear combinations
called principal components (Tabachnick and Fidel, 2007) with each principle
component or Eigenvalues being independent of other components. The
importance of PCA is to extract maximum variance from the data set with each
component. The new sets of transformed uncorrelated variables are close to the
original variables but arranged in decreasing order of variance. PCA further
enables plotting data in two dimensions to look at outliers, groups or clusters
based on biological data (Chatfield and Collins, 1980). The quantitative and
qualitative data were analysed using Principal Component Analysis in Genstat
version 13, based on the correlation matrix, which basically gives traits equal
weightings.
Principal Component analysis (PCA) was used to reveal traits that account for
most variation between lines. The Eigenvalues ≥1 were selected and used to
define the agro-morphological diversity. Principal component analysis was
constructed using MVSP (Kovach, 2006).
2.7.8.3
Cluster analysis
Cluster analysis, basically aims to find groupings in a set of individuals, objects or
units such that individuals within a group are similar to each other but individuals
in a different group are dissimilar to others. The quantitative and the qualitative
data were converted to binary data. For qualitative characters the absence of a trait
was recorded as 0 and the presence of a trait as 1. For quantitative data genotypes
that were significantly different were scored 1 and those not significantly different
were scored 0. Then the unweighted pair group method with arithmetic averages
(UPGMA) cluster analysis was performed using NTSYC version 2.1 (Rholf,
2000) using average linkage based on Euclidean distance and dendrograms
71
produced to show the similarities and differences between bambara groundnut
genotypes.
2.7.8.4
Shannon-Weaver diversity
The Shannon-Weaver diversity index (H’) of Shannon and Weaver (1949) was
estimated to measure the phenotypic diversity between traits. In the calculations
of Shannon Weaver index (H’), in order to avoid bias of the contribution of the
individual trait the diversity of the mean, phenotypic range are arbitrarily divided
into similar number of classes for both the glasshouse and field experiment
(Engels, 1994) (Appendix 7)
The Shannon-Weaver diversity index (H’) was calculated using the phenotypic
frequencies to assess the phenotypic diversity for each trait. Where (H’= ∑
(Pilnp)where pi is the proportion of accessions in the ith class of an n- class
character and n is the number of phenotypic classes for a character. Pi calculates
the abundance of the given categories for each character, which is then multiplied
by the natural log. The traits measurements were divided by their maximum
values, log n, and normalised to keep the values between 0 and 1, on Genstat
version 13 based on ECDIVERSITY procedure. The analysis was conducted on
both the quantitative and qualitative traits.
2.7.8.5
Correlation coefficient
Correlation coefficient (r) was used in the study to determine inter-relations
between all quantitative characters (Table 2.6.6).Correlation ranges between -1
and 1 and measures the extent to which two variable scores increase at the same
time in positive correlation while in negative correlation the other goes down
while the other increases. A correlation of 0 implies that there is no linear
relationship between the variables.
Pearson correlation was conducted using
SPSS version 16.0 and the significance test for correlation coefficient was tested
on a two-tailed test on the same program.
2.7.8.6
Quantitative variances
Data for each trait was subjected to analyses of variance to estimate the genetic
variability of the selected genotypes and to partition the phenotypic variability
into components due to genetic and environmental factors. Measures of variability
72
such as genotypic coefficient of variability (GCV), phenotypic coefficient of
variability (PCV), broad sense heritability (h2), and genetic advance (GA) based
on percentage of the mean were estimated.
There are a number of methods available for estimating heritability (h2), which
includes using the resemblance among relatives, from generations derived from a
cross between two pure-breeding lines, experimental mating designs and from
components calculated from replication experiments (Hill et al., 1998). The latter
method was selected for use and each trait that was subjected to analyses of
variance was used to estimate the genetic variability of the selected genotypes and
to partition the phenotypic variability into components due to genetic and
environmental factors. The genetic parameters were estimated using formulas
adapted from Allard, (1960), Singh and Chadhary, (1985) as follows:
Vg = [ Mean Square Genotype-Mean Square Error/r]
Vp = [Mean Square Genotype/r]
Ve = [Mean Square Error/r]
r is the number of replications
The Mean Square Genotype (MSG) and Mean Square Error (MSE) are variance
components estimated as functions of the mean square estimates from ANOVA
table. Mean square genotype (MSG): estimates genotypic variance, this value is
observed variance among the line means, while mean square error (MSE)
measures variance from plot residuals.
Phenotypic (PCV) and genotypic (GCV) coefficient are estimated using the
following formulas
PCV = (√Vp/X) x 100
GCV = (√Vg/X) x 100
Vp represents the phenotypic variance: Vg represents the phenotypic variance,
while X represents the mean.
73
Heritability (h2B) expressed as the percentage of the ratio of the genotypic
variance (Vg) to the phenotypic variance (Vp) was estimated based on the
genotypic mean
Expected genetic advance (GA) was estimated using a formulaof Allard, (1960) as
GA = K (Sp) h2B, GA (as % of mean) = (GA/X) x100
where h2B and Sp is the heritability ratio and the phenotypic standard deviation
(√Vp) and K is a selection differential that varies depending on the selection
intensity. In the present analysis 2.06 was considered for K, which is 5% selection
intensity. The phenotypic standard deviations among traits were calculated using
Genstat version 13.
2.7.8.7
Selection index (SI) and Duncan Multiple Range Test (DMRT)
For a development of new varieties it is important that selection of the best
genotypes is conducted. Usually, it is those traits of interest or economic value
that breeder select for. Plant breeders could decide to select for one or more traits
at a time and this is referred to as multiple trait selection. And appropriate weight
is given to each character, for example its heritability (h2), genetic and phenotypic
correlations between different characters of interest could be used.
The
component characters are then combined into a score, or a selection index
(Falconer and Mackay, 1996). For multi-trait selection the classical selection
index (I) proposed by Smith (1936) and Hazel (1943) which is a linear
combination of traits of interest could be used with the formula:
I =b1x1 +b2x2+c3x3+ ...+bnxn
where x1, x2, x3...xn are the phenotypic performance of different traits, while b1,
b2, b3, are relative weight attached to each traits.
The weights attached depend on the economic importance attached to traits
depending on their heritability and correlation between various traits (Hill et al.,
1998).
In this study, a similar index was used, based on four characters of importance,
shoot dry weight, leaf area, seeds number plant and pod number per plant.
74
Economic weight attached to these characters was based on the genetic advance
(GA) (5% of the mean), which is described in section 2.7.8.6.
SI= (X1xW1) + (X2xW2) + (X3xW3) + (XnxWn)
where W1, W2, W3...Wn are the respective weights for each variable. Since the
variables are measured in different units with large differences in magnitude, the
variables were standardized with the following formula;
Xi = (Xi -µ)/ st. Dev
X1 = shoot dry weight,
X2 = leaf area
X3 = seed number plant
X4 = pod
number per plant.
To identify lines which have a potential to produce higher yields in a Botswana
environment based on four selected characters, the selection index was used with
a weighting of the genetic advance found in the field experiment in Chapter 4,
table 4.2.7. SI = (X1 x 0.378) + (X2 x 0.424) + (X3+0.828) + (X4 x 0.881). X1 =
Leaf area, X2 =Shoot dry weight, X3 =Seed number per plant, X4 = Pod number
plant. Selection index ranks values were obtained using Microsoft Excel.
Duncan Multiple Range Test (DMRT) was used to identify genotypes that were
significantly different from each other on selected traits and calculated based on
Genstat version 13.0. Based on Duncan Multiple Range Test and selection index
the best performing genotypes were identified and ranked.
75
CHAPTER THREE: Microsatellites, development and
characterisation
3.1
Introduction
Bambara groundnut (Vigna subterranea (L.) Verdc) is an important indigenous
leguminous crop that is cultivated especially by women in most regions of subSaharan Africa (Azam-Ali et al., 2001). In some countries, for example in
Botswana, bambara groundnut is usually grown both for home consumption and
for sale and has considerable importance for subsistence farmers for their local
market and commercialisation on a small scale. Despite the fact that bambara
groundnut has a potential to contribute to food security in Africa, it has no
established varieties. Resource poor farmersgrow crops that are adapted to local
environmental conditions (some commercial) landraces, which are genetically
diverse populations selected under low-input agriculture (Zeven, 1998).
3.1.1
Breeding systems in bambara groundnut
Knowledge of the mating systems of species or plant populations is important for
establishing controlled breeding programs, as breeding methods for self-pollinated
crops can be different from the cross-pollinated ones, and those with mixed
mating systems (Ferriara et al., 2000). The mechanisms by which plants produce
their offspring has a far reaching impact on how the diversity is partitioned and
spread within and between populations. Outcrossing species are genetically
variable with lower genetic differentiation, while inbreeding plants are less
variable, with higher local structure and diversity between populations (Rymer et
al., 2002). Breeding systems can significantly affect population ecology and
evolution in several ways, since it determines the homozygosity/ heterozygosity
of individuals. Inbreeding in plants can be measured by using genetic markers, by
estimating the frequencies of homozygotes and heterozygotes (Charlesworth,
2006). Inbreeding also leads to a reduction in effective population size and lowers
the genetic recombination that occurs within the population (Lui et al., 1999).
Inbreeding leads to an increased homozygosis within a population and random
changes in gene frequencies from subsequent generations (Robertson, 1961). The
mating systems can be investigated using molecular markers such as isozymes,
RAPDs, AFLP and microsatellites.
76
3.1.2
Floral biology of bambara groundnut
The crop has perfect flowers, with stamens and pistil borne in the same flower.
Flowers are borne on a raceme on long, hairy peduncles which arise from the
nodes (Doku and Karikari, 1970). The flower has a pair of hairy epicalyces.The
calyx consists of five hairy sepals, four on the upper side and the lower sepal is
free. The standard petal encloses the wing and keel petal until the flower opens. It
is usually bright golden yellow and wraps around keel. Generally the wing petals
are yellow and enclose the stigma, style and stigma. The stamens are diadelphous
which means that there are nine partly fused filaments. In young flower buds, the
stigma is slightly above the anthers, while in mature flowers the filaments
elongate to place the anthers at a level with the stigma (Massawe et al., 2003).
Node
Peduncle
Petals
P
Sepal
Figure 3.1.2 Bambara groundnut flower, showing the floral morphology. Scale bar = 1 cm
Bambara groundnut is believed to be mainly self-pollinated and anthers dehisce as
the stigma becomes receptive even before the flowers open and sometimes
fertilisation takes place on the same day as anthesis (Linnemann, 1994). A similar
observation was made by Doku and Karikari, (1971), who noticed that pollen
maturity and stigma receptivity occurs just before or immediately after the flower
opens. The flower structure of bambara groundnut discourages outcrossing since
the staminate and pistilate parts are covered by a bract of the cap-like operculum.
In addition,the rapid loss of pollen viability reduces the transfer of viable pollen
grains (Chijioke et al., 2010). Although there is a lack of detailed studies on the
breeding system of bambara groundnut, it appears to be preferably cleistogamous
77
and would be expected to be inbreeding. However, Doku and Karikari, (1970)
reported that ants can also facilitate self- and cross- pollinate bambara groundnut.
Mkandawire (2007) reported that self polination in bambara groundnut is mainly
found in bunched plants while cross pollination occurs in spreading types.
3.1.3
Seed dissemination systems
Seed dissemination is important when investigating the potential for migration
and geneflow between populations. The higher demand for bambara groundnut
seed in countries in southern Africa cannot be met within country, thus farmers
are sourcing seeds from Zimbabwe, which exports some of its seeds to countries
such as Botswana, South Africa and Swaziland (Azam-Ali et al., 2001). A survey
carried out in Botswana (Brink et al., 1996) showed that while most of the farmers
prefered to use the previous season’s harvest as their seed stock, they do also
exchange seeds with friends and family members. This movement of seeds across
regions is likely to have a impact on the genetic diversity and population structure
of bambara groundnut.
3.1.4
Analysis of breeding systems in bambara groundnut
Since bambara groundnut is a self-pollinating crop, intra-landrace variation might
be expected to be low. Pasquet et al., (1999) investigated the genetic diversity and
population structure of bambara groundunt using 79 domesticated and 21 wild
accessions. They employed a total of 41 isozyme markers representing 23 enzyme
systems and reported a higher genetic diversity for wild accessions (Ht = 0.087)
and a lower genetic diversity for the domesticted type (Ht = 0.052) with 14 and 7
polymophic loci each, respectively. However, their results revealed a relatively
higher intrapopulation diversity among the domesticated accessions (Hs=0.033)
and lower levels in the wild type (Hs =0.025) and both accessions showed very
low levels of heterozygosity which was attributed to the self-pollination nature
for both wild and domesticated bambara groundnut.
Massawe, et al., (2002) employed AFLP while Massawe et al., (2003) used
RAPDs to determine the heterogeneity within bambara groundnut landraces. They
found significant variation among landraces and also among individuals within
each landrace. This observation was attributed to the autogamous breeding system
of bambara groundnut.
78
Investigations of intra-landrace genetic diversity were conducted in 10 landraces
and 15 individual genotypes of the bambara groundnut landrace by Singrun and
Schenkel (2003) who used EcoRI/MseI amplified fragment length polymorphism
(AFLP) and the heterologous primer pair of AG81 from soybean. Their results
demonstated that none of the landraces consisted of a single genotype. Although
these studies have shown high levels of variability in bambara groundnut and shed
some light on the mating system of the crop, because of the relatively limited
polymorphism of isozymes and the dominant marker nature of RAPD and AFLP
analysis, there is still more work to be done, particularly in relation to the levels
of heterozygosity present within individuals. Initial microsatellite work has
confirmed the presence of multiple genotypes within bambara groundnut
landraces which has implications for the breeding of the crop. However,
determining the level of heterozygosity present within individual genotypes is
important for coming up with possible breeding routes available, especially in the
production of pure line seed (Basu et al., 2007; Mayes et al., 2009).
3.1.5
Breeding system studies in other leguminous species
Plant mating systems have been generally divided into three main sections, that is
predominantly outcrossing, mixed self fertilizing and out crossing and
predominantly self fertilizing (Hedrick, 2005). Some leguminous species like
Medicago trunculata (Kamphius et al., 2007), common bean (Phaseolus vulgaris)
Tosti and Negri, (2005) and pigeonpea (Cajanus cajan) Songok et al., (2010) are
predominantly self pollinating, but with a low level of cross-pollination. Although
it had been previously believed that wild soybean Glycine soja was autogamous,
as is cultivated soybean (Glycine max), a mean multilocus outcrossing rate
estimate of 13% showed that it is also cross-pollinated (Ohara and Shimamoto,
2002). Chickpea (Cicer arietum) also commonly known to be a self pollinating
crop was shown to have the capability to cross-pollinate with other wild Cicer
species such as, Cicer echinospermum and Cicer reticulatum (Upadhayaya et al.,
2008).
Maquet et al., (1997) used isozymes markers to confirm the self-pollinating
mating sytem of lima bean (Phaseolus lunatus). They studied a collection of 235
lima bean accessions originating from Latin America and the Carib zone using 10
79
allozyme markers. The study revealed a high inbreeding coefficient (f = 0.891), a
low intrapopulation gene diversity (Hs = 0.032) as compared to a higher
interpopulation gene diversity (DST =0.235).
When using 48 SSR markers to analyse the genetic diversity of 39 parental lines
of mung bean (Vigna radiata), Somta et al., (2009) observed lower observed
heterozygosity ofHo = 0.04 compared to a higher expected heterozygosity of He
=0.39 which was an indication of the inbreeding nature of mung bean.
3.1.6
Applications of microsatellites in this study
The aim of this part of the study was to develop a comprehensive set of
microsatellites for bambara groundnut and select the best markers for
fingerprinting and other breeding applications. Microsatellites have desirable
features that makes them well suited for this application, compared to other
markers. Microsatellites or simple sequence repeats (SSRs), are tandem arrays of
nucleotide repeats (one to six bases motifs) with SSR loci spread all over the
genome. They are a marker of choice due to their higher information content and
other features such as high reproducibility and their co-dominant nature (Gupta
and Varshney, 2000). They are multi-allelic, highly abundant, analysis is simple
and methodologies are easily transferable. Therefore, they are more useful than
RAPD or AFLPs and can yield twice as much information per locus as the AFLP
and three times as much as RAPDs, according to Gallego et al., (2005).
In this study the characterisation of microsatellites has been undertaken using 24
bambara groundnut landrace accessions. The 24 bambara groundnut landraces
were selected on the basis of a study conducted by Singruin and Schenkel (2003),
where a total 223 bambara groundnuts were analysed for genetic diversity, 46
accessions were originaly from West Africa (Benin, Ghana and Nigeria), 6 from
East Africa (Kenya and Tanzania), 7 from Madagascar, 4 from Indonesia, while
the rest (160) were from Southern Africa (Botswana, Namibia, Swaziland,
Zambia and Zimbabwe). Analysis was undertaken using EcoRI/MseI amplified
fragment length polymophisms (AFLP) and one heterologous SSR primer pair
AG81 derived from soybean. Their results produced 17 clusters and the 24
landraces were used selected from these clusters and are listed on (Table 2.1.2.1).
80
The same 24 landraces were analysed with Diversity Arrays Technology (DArT)
markers for comparison of the efficiency of the two techniques.
Diversity Arrays Technology was developed as a hybridisation-based technology
and is valued for the high level of data production due to its microarray platform.
It can type thousands of loci in a single assay, and generates whole genome
fingerprints of genomic representations, generated from sub-samples of genomic
DNA (Jaccourd et al., 2001). DArT has the advantage of low cost, high
throughput and it does not require sequencing, this makes it more suited for use in
‘orphan’ crops such as bambara groundnut, as compared to SSR markers which
requires prior sequence information (Yang et al., 2006). A number of marker
types have been employed in crop breeding studies with different efficacy and
ease of use to quickly develop or assay large number of markers (Akbari et al.,
2006). The importance of comparing different marker systems is to assist in
making informed decisions as to which marker is best to use in germplasm
characterisation and plant breeding. The aim of this part of the study is to compare
the use of DArT and SSR in assessing the genetic diversity of 24 bambara
groundnut landraces and genetic diversity analysis of bambara groundnut.
3.2
Materials and Methods
3.2.1
DArT marker screening
DArT marker screening and genotyping were undertaken by Diversity Array Pty,
Ltd, Yarralumla, Australia as described by Jaccoud et al., 2001. This basically
consists of three major steps; array development, genotyping, and data analysis.
The number of markers that can be obtained does not only depend on the levels of
genetic diversity in the germplasm but also on the combination of restriction
enzymes used to generate the representation used to produce the clones.Therefore
a suitable complexity reduction method has to be identifed. Two restriction
endonucleases combinations were tested by DArT Pty Ltd. for the treatment of
combined DNA samples, with PstI used as the rare cutter (restriction 6 bp), while
enzymes AluI, BanII, BsoBI, BstNI, MseI, RasI, TaqI and Tsp5091) sourced from
(New England Biolabs Ltd., Pickering Canada) with a 4 bp were tested as
frequent cutter. Gel electrophoresis suggested that AluI was a suitable 4 bp cutter
81
as it produced a homogenous smear without repetitive bands after visualisation
with Ethidium bromide stained agarose gel, thus it was selected to develop intial
Discovery Array. Further details on the development of the DArT array for
bambara groundnut is described in Stadler, (2009).
A full genotyping array containing 7,680 clones was generated from two
complexity reduction methods, using PstI/AluI and PstI/TaqI. The restriction
endonuclease PstI/AluI produced 157 polymorphic clones, while in the second
complexity reduction method PstI/TaqI produced 168 polymorphic clones. When
data sets were combined, and after removing all repeated discrimation patterns a
final remainder of 296 polymorphic clones were used for DArT genetic diversity
analysis, based on the initial 94 genotypes. However, when these polymorphic
markers were used in the large scale analysis of 342 bambara groundnut a total of
201 robust markers polymorphic across all samples analysed remained and these
were used in the analysis of the full bambara groundut germplasm. It is this
dataset that the 24 bambara groundut accessions were drawn from, the same
accessions having also been analysed using 68 SSR markers by the author.
82
3.3
Results
3.3.1
Microsatellites marker analysis
From the initial set of 75 markers listed in appendix 2, seven markers had poor
amplification; producing smeared/complex bands or no PCR product at all. As a
result they were discarded. Therefore 68 microsatellites were used to characterise
and evaluate the genetic diversity of 24 bambara groundnut landraces (Table 3.1).
Table 3.1: Summary of PowerMarkers data analysis of 24 bambara groundnut landraces,
based on 68 microsatellites.
Marker
Primer 1
Primer 2
Primer 3
Primer 4
Primer 5
Primer 6
Primer 7
Primer 8
Primer 9
Primer 10
Primer 11
Primer 12
Primer 13
Primer 14
Primer 15
Primer 16
Primer 17
Primer 18
Primer 19
Primer 20
Primer 21
Primer 22
Primer 23
Primer 24
Primer 25
Primer 26
Primer 27
Primer 28
Primer 29
Primer 30
Primer 31
Primer 32
MAF
0.71
0.69
0.88
0.54
0.50
0.50
0.29
0.71
0.92
0.46
0.92
0.54
0.92
1.00
0.17
0.21
0.67
0.46
0.29
1.00
0.83
1.00
0.81
0.88
1.00
0.40
0.75
0.50
1.00
0.29
0.88
0.25
GN
3
4
3
4
4
4
6
2
2
4
2
3
2
1
15
6
5
5
9
1
3
1
4
4
1
12
5
5
1
6
3
11
SS
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
No.
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
AN Avail.
3
1
3
1
3
1
4
1
4
1
4
1
6
1
2
1
2
1
4
1
2
1
3
1
2
1
1
1
14
1
6
1
5
1
5
1
9
1
1
1
3
1
1
1
3
1
4
1
1
1
12
1
5
1
5
1
1
1
6
1
3
1
10
1
GD
0.45
0.45
0.23
0.62
0.57
0.64
0.74
0.41
0.15
0.61
0.15
0.56
0.15
0.00
0.90
0.82
0.52
0.71
0.82
0.00
0.29
0.00
0.32
0.23
0.00
0.80
0.42
0.63
0.00
0.80
0.23
0.86
Het.
0.00
0.04
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.04
0.00
0.00
0.04
0.04
0.00
0.00
0.00
0.00
0.04
PIC
0.40
0.38
0.21
0.57
0.48
0.58
0.69
0.33
0.14
0.53
0.14
0.47
0.14
0.00
0.89
0.79
0.48
0.67
0.80
0.00
0.26
0.00
0.30
0.22
0.00
0.79
0.39
0.57
0.00
0.77
0.21
0.84
f
1.00
0.91
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
NaN
0.87
1.00
1.00
1.00
1.00
NaN
1.00
NaN
0.88
1.00
NaN
0.95
0.90
1.00
NaN
1.00
1.00
0.95
83
Table 3.1 (Continued)
Marker
Primer 33
Primer 34
Primer 35
Primer 36
Primer 37
Primer 38
Primer 40
Primer 41
Primer 42
Primer 43
Primer 44
Primer 45
Primer 48
mBam2co80
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
D12
D13
D14
D15
E1
E3
E5
E7
E9
E10
E11
Mean
MAF
0.29
0.96
1.00
1.00
0.33
0.58
0.63
0.83
0.63
0.42
0.42
0.73
0.29
0.17
0.21
0.67
0.81
0.79
0.42
0.83
0.96
1.00
0.46
0.71
0.42
0.42
0.96
0.23
0.27
0.79
0.96
0.94
0.52
0.96
1.00
0.88
0.65
GN
9
2
1
1
7
5
2
3
4
5
5
3
13
12
11
5
3
4
7
3
2
1
7
6
8
9
2
12
11
3
2
3
5
2
1
3
5
SS
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
No.
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
AN Avail.
9
1
2
1
1
1
1
1
7
1
5
1
2
1
3
1
4
1
5
1
5
1
2
1
13
1
12
1
11
1
5
1
3
1
4
1
7
1
3
1
2
1
1
1
6
1
6
1
7
1
9
1
2
1
11
1
10
1
3
1
2
1
3
1
4
1
2
1
1
1
3
1
5
1
GD
0.82
0.08
0.00
0.00
0.75
0.61
0.47
0.29
0.54
0.70
0.74
0.39
0.83
0.90
0.89
0.52
0.31
0.36
0.74
0.29
0.08
0.00
0.71
0.48
0.76
0.76
0.08
0.86
0.83
0.34
0.08
0.12
0.61
0.08
0.00
0.23
0.45
Het.
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.04
0.00
0.00
0.04
0.00
0.00
0.00
0.00
0.04
0.00
0.00
0.00
0.00
0.00
0.04
0.00
0.04
0.00
0.00
0.08
0.04
0.00
0.00
0.04
0.08
0.00
0.00
0.17
0.01
PIC
0.79
0.08
0.00
0.00
0.71
0.58
0.36
0.26
0.49
0.64
0.70
0.32
0.82
0.89
0.87
0.48
0.27
0.34
0.71
0.26
0.08
0.00
0.68
0.45
0.73
0.74
0.08
0.85
0.81
0.31
0.08
0.12
0.54
0.08
0.00
0.21
0.42
f
1.00
1.00
NaN
NaN
1.00
1.00
1.00
1.00
0.93
1.00
1.00
0.90
1.00
1.00
1.00
1.00
0.87
1.00
1.00
1.00
1.00
NaN
0.94
1.00
0.95
1.00
1.00
0.91
0.95
1.00
1.00
0.66
0.87
1.00
NaN
0.28
0.97
MAF-Major Allele Frequency
GN- Genotype number
SS- Sample Size
No.- Number of observations for a marker
AN- Allele number
Avail. Availability
GD- Gene diversity or expected heterozygosity, the probability that two randomly chosen alleles from population are
different
Het. The proportion of heterozygous individuals in the population
PIC-Polymorphic Information Content
f-inbreeding coefficient
84
A total of 313 alleles were identified, with nine markers non-polymorphic (marker
14, marker 20, marker 22, marker 25, marker 29, marker 35, marker 36, marker
D8 and marker E10). The number of alleles per marker ranged from 1 for nonpolymorphic markers to 14 in marker 15, with a mean of 5 alleles per marker
(Table 3.1). The polymorphic information content (PIC) values ranged from 0.08
to 0.89 from primer D7, primer D13, primer E9, primer E3, primer 34 to marker
15 and marker mBam2co80with an average of 0.42, with nine markers been
monomorphic. 39.7 % of the markers were highly polymorphic with PIC values
ranging from 0.5 to 0.89, while 33.8 % of the markers were just informative with
PIC values ranging from 0.21 to 0.49. The remaining 26.5% include nine markers
that were monomorphic, and eight markers with low polymorphic information
content and a range of 0.08 to 0.15.A one sample t-test conducted on the data, for
the average He(0.45) and PIC (0.42) revealed no significant difference to those
obtained by Basu et al., (2007) for He(0.50) and PIC (0.47).
Bambara groundnut is an inbreeding crop, so as expected all the markers showed
a lower observed heterozygosity (Ho) compared to the expected heterozygosity
(He). However, 23.5% of the markers showed some heterozygosity with a range
of 0.04 to 0.17. Markers E5 and E11showed a low inbreeding coefficient at 0.66
and 0.28 respectively, while the rest of the markers had an inbreeding coefficient
of 1. As both markers E5 and E11 appear to be outliers, it seems possible that
these primers detected more than a single locus. Interestingly, both are derived
from Roche 454 sequence (average read length 92bp) derived from RNA.
3.3.1.1
Hardy Weinberg Equilibrium (HWE)
HWE was tested using PowerMarker, which uses three different methods to test
for Hardy-Weinberg equilibrium, the Chi square statistics and the permutation
version of the exact test given (Table 3.2). All markers are highly significant at
(P<0.05), an indication that the population is not in HW, as would be expected for
structured accessions of a germplasm collection derived from an inbreeding
species.
85
Table 3.2: The 68 markers used in the 24 bambara groundnut analysis were
subjected to Chi square and HWE exact test using MVSP version 3.25, with the
exception of nine non-polymorphic markers.
Marker
Primer 1
Primer 2
Primer 3
Primer 4
Primer 5
Primer 6
Primer 7
Primer 8
Primer 9
Primer 10
Primer 11
Primer 12
Primer 13
Primer 15
Primer 16
Primer 17
Primer 18
Primer 19
Primer 21
Primer 23
Primer 24
Primer 26
Primer 27
Primer 28
Primer 30
Primer 30
Primer 32
Primer 33
Primer 34
Primer 37
Primer 38
Primer 40
Primer 41
Primer 42
Primer 43
Primer 44
Primer 45
2
X value
48.00
43.13
48.00
72.00
72.00
72.00
120.00
24.00
24.00
72.00
24.00
48.00
24.00
254.43
120.00
96.00
96.00
192.00
48.00
38.39
72.00
240.07
72.96
96.00
120.00
48.00
202.36
192.00
24.00
144.00
96.00
24.00
48.00
48.20
96.00
96.00
19.20
2
X d.f.
3
3
3
6
6
6
15
1
1
6
1
3
1
91
15
10
10
36
3
3
6
66
10
10
15
3
45
36
1
21
10
1
3
6
10
10
1
Exact p-value
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
86
Table 3.2 (Continued)
Marker
Primer 48
mBam2co80
D1
D2
D3
D4
D5
D6
D7
D9
D10
D11
D12
D13
D14
D15
E1
E3
E5
E7
E9
E11
3.3.1.2
X2 value
288.00
264.00
240.00
96.00
24.02
72.00
144.00
48.00
24.00
102.59
120.00
125.23
192.00
24.00
208.76
204.55
48.00
24.00
24.01
40.13
24.00
24.22
X2 d.f.
78
66
55
10
3
6
21
3
1
15
15
21
36
1
55
45
3
1
3
6
1
3
Exact p-value
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.01
0.02
0.00
0.02
0.03
Estimation of Null alleles.
Deviation from HWE in a population can be caused by selection, non-random
mating, inbreeding/self-fertilisation, migration and the presence of null alleles.
The presence of null alleles was investigated in the data set (Table 3.3).
Estimation of null alleles was conducted using the INEst (Inbreeding /Null allele
Estimation) software, which takes into account the possibility of inbreeding
within a population during estimation of null frequencies (Chybicki and Burczyk,
2009). The population inbreeding model (PIM) and the individual inbreeding
model (IIM) are calculated. The PIM estimate uses the jacknife algorithm, while
the IIM estimates uses the Gibbs sample command which uses a number of run-in
steps of approximately 10,000. The estimates of PIM and IIM are given in table
3.3.
87
Table 3.3: Estimation of null allele frequencies for each locus, using the
population inbreeding model (PIM) and the individual inbreeding model (IIM)
using INEst (Chybicki and Burczyk, 2009).
Marker
Primer 1
Primer 2
Primer 3
Primer 4
Primer 5
Primer 6
Primer 7
Primer 8
Primer 9
Primer 10
Primer 11
Primer 12
Primer 13
Primer 14
Primer 15
Primer 16
Primer 17
Primer 18
Primer 19
Primer 20
Primer 21
Primer 22
Primer 23
Primer 24
Primer 25
Primer 26
Primer 27
Primer 28
Primer 29
Primer 30
Primer 31
Primer 32
Primer 33
Primer 34
Primer 35
Primer 36
Primer 37
Primer 38
Primer 40
PIM
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
IMM
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
Ho
0.000
0.000
0.000
0.000
0.000
0.083
0.000
0.042
0.000
0.083
0.000
0.000
0.042
0.000
0.000
0.000
0.042
0.083
0.042
0.000
0.042
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.042
0.000
He
0.050
0.621
0.082
0.642
0.369
0.909
0.000
0.839
0.553
0.852
0.684
0.858
0.329
0.723
0.684
0.656
0.776
0.883
0.488
0.911
0.042
0.082
0.284
0.223
0.383
0.000
0.000
0.000
0.156
0.156
0.000
0.528
0.755
0.000
0.000
0.230
0.000
0.692
0.000
F(Wright index)
1.000
1.000
1.000
1.000
1.000
0.908
1.000
0.950
1.000
0.902
1.000
1.000
0.873
1.000
1.000
1.000
0.946
0.906
0.915
1.000
0.901
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
PIM: Population Inbreeding Model: (IIM) Individual Inbreeding Model
88
Table 3.3 (Continued)
Marker
Primer 41
Primer 42
Primer 43
Primer 44
Primer 45
Primer 48
mBam2co80
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
D12
D13
D14
D15
E1
E3
E5
E7
E9
E10
E11
PIM
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
IMM
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
Ho
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.083
0.000
0.042
0.000
0.000
0.042
0.042
0.000
0.000
0.000
0.000
0.042
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.167
He
0.574
0.000
0.000
0.000
0.000
0.610
0.000
0.294
0.536
0.730
0.403
0.691
0.883
0.042
0.042
0.230
0.741
0.082
0.000
0.728
0.305
0.723
0.000
0.284
0.000
0.000
0.000
0.000
0.156
F(Wright index)
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.845
1.000
0.897
1.000
1.000
0.000
0.000
1.000
1.000
1.000
1.000
0.943
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
-0.068
PIM: Population Inbreeding Model: (IIM) Individual Inbreeding Model
Both the population inbreeding model (PIM) and individual inbreeding model
(IIM) showed that none of the markers have null alleles which usually leads to
erroneous interpretation of the data (Wagner et al., 2006). However, marker E5
with unusual low inbreeding coefficient (0.66) was excluded from the data
analysis. Other two markers which were excluded are markers E11 and 15 with
high heterozygous values of 0.17 and 0.13 which maybe more indicative of the
SSR amplifying from two loci, rather than true heterozygosity at the locus.
89
Therefore only 65 SSR markers were used in the finale analysis of the 24 bambara
groundnut landraces.
3.3.2
Principal Component Analysis (PCO)
Eigenvalues and the cumulative percentage of the principal component case
scores were used for the analysis of the population structure for the selected 24
bambara groundnut for the comparison of DArT and SSR , the results are given
in (Table 3.4 and Table 3.5). Data for the two markers looking at the first two
axes suggests that DArT marker is revealing more variation at (37.3%) compared
to SSR markers at (19.5 %) in the first two axes.
Table 3.4: PCO case scores for the population structure of the selected 24 bambra
groundnut landraces, determined based on 201 DArT markers
Axis
1
Axis
2
Axis
3
Axis
4
Axis
5
Axis
6
Axis
7
Axis
8
Axis
9
Axis
10
Eigenvalues
Percentage
9.58
28.45
2.98
8.84
2.15
6.39
2.02
6.00
1.79
5.32
1.71
5.07
1.53
4.54
1.39
4.13
1.35
4.01
1.22
3.61
% Accumulation
28.45
37.30
43.69
49.68
55.00
60.07
64.61
68.74
72.75
76.36
Table 3.5: PCO case scores for the population structure of the selected 24 bambra
groundnut landraces, determined based on 65 SSR markers.
Axis 1 Axis 2
Axis 3
Axis 4 Axis 5
Axis 6 Axis 7
Axis 8 Axis 9 Axis 10
Eigenvalues
2.58
2.04
1.87
1.80
1.57
1.47
1.27
1.21
1.13
1.05
Percentage
10.91
8.61
7.90
7.60
6.63
6.19
5.37
5.13
4.77
4.44
% Accumulation
10.91
19.51
27.42
35.02
41.64
47.83
53.20
58.33
63.10
67.53
The PCO analysis with each marker type indicates that both DArT and SSR
analysis show differentiation between the selected 24 bambara groundnut
landraces. However, it was DArT which clearly gave a separation of landraces
according to their areas of origin, West African landraces were separated from the
East African landraces and the Southern Africa landraces, but Ramayana from
Indonesia formed a group with the Southern African landraces (Figure 3.2 a). The
SSR marker analysis did not display such a clear separation of landraces based on
place of origin (Figure 3.2 b). The landraces from Southern Africa are scattered
across both axes while East African landraces grouped together. Ramayana from
Indonesia-Asia shows a close proximity to the Southern African landraces.
90
a) DArT marker
PCA case scores
0.9
0.7
East Africa
0.4
Axis 2
0.2
Southern Africa
0.0
-0.2
West Africa
-0.4
-0.7
-0.9
Indonesia
-1.1
-1.1
-0.9
-0.7
-0.4
-0.2
Axis 1
0.0
0.2
0.4
0.7
0.9
b) SSR marker
PCA case scores
0.7
0.6
East Africa
0.4
Axis 2
0.3
Southern Africa
0.1
0.0
-0.1
West Africa
-0.3
-0.4
-0.6
-0.7
-0.7
Indonesia
-0.6
-0.4
-0.3
-0.1
0.0
Axis 1
0.1
0.3
0.4
0.6
0.7
Figure 3.2: The first two axes of the PCO case scores, generated from the 24 landraces using
MVSP for figure 3.2 (a) DArT Axis 1 represents 28.45% and Axis 2 represents 8.84 % of the
molecular variation, figure 3.2 (b) SSR markers; Axis 1 represents 10.91 % and Axis 2 represents
8.61% of the molecular variation in the 24 selected bambara groundnut landraces.
91
3.3.3
Cluster analysis
The UPGMA dendrogram (Figure 3.3a and Figure 3.3b) show a moderate genetic
variability for both markers. DArT markers showed three clear and distinct
clusters (Figure 3.3a) and grouped landraces based on their areas of origin. Cluster
1 consists of landraces mainly from West Africa, while cluster 2 consists of a
mixture of three Southern Africa landraces, this group also includes Ramayana
originally from Indonesia-Asia, and this could be a reflection of the origin of this
landrace from Africa, this cluster also consists of wild landrace VSSP6 from
Cameroon and DodR from East Africa. Cluster three is mainly landraces from
Southern Africa with the exception of DodC from Tanzania located on the edge of
the cluster.
The SSR marker dendrogram figure 3.3 (b) grouped the landraces into three
clusters with, two major groups and one which consist of Tiganicuru from Mali
and DodR from Tanzania only. The landraces are largely grouped based on their
areas of origin like the DArT markers since the majority of landraces within a
cluster originate from the same region. Cluster one, consists of a mixture of four
landraces from West Africa and four from Southern Africa. Cluster two consist
mostlylandraces from Southern Africa with the exception of Nav Red and Tvsu
569 from West Africa, DodC from East Africa and Ramayana from Indonesia.
The comparison for the two markers is tabulated in Table 3.6.
DArT markers had more bootstrap values more than 50% as compared to the SSR
marker DArT which indicates the reliability of the marker. Landraces S19-3 and
S19/3, both from Namibia, were identified by DArT at 97% similarity while SSR
marker showed minor differences at 99% similarity. It is interesting to note that
DArT markers found GabC from Botswana to be identical to AHM968 from
Namibia. This could probably be caused by the seeds exchange between these two
neighbouring countries or potentially a mistake has been made during seed
handling.
92
UPGMA
LunTSLA
Tvsu 569CMR
I
Nav RedGHA
Nav 4GHA
Tvsu 610NGA
TiganicuruMLI
VSSP6CMR
DodRTZA
II
RamayanaIND
NAM 1761/3NAM
MHNblackNAM
DipCBWA
Tvsu 999ZWE
Tvsu 747ZMB
Malawi 3MLW
SB16 5ANAM
GabCBWA
III
AHM968NAM
S19-3 2007NAM
S19/3NAM
Uniswa 2007 NAM
SwaziRedSWA
AS17RSA
DodCTZA
75
70
98
84
100
97
52
76
0.52
0.6
0.68
0.76
0.84
0.92
1
Nei & Li's Coefficient
Figure 3.3 (a) Cluster analysis of the 24 bambara groundnut landraces. The UPGMA dendrogram
is based on the similarity matrix obtained from 201 DArT markers using the Nei and Li, (1979).
The number at the nodes of branches represents the percentage bootstrap support of individual
nodes at resampling at 1000.
UPGMA
Nav 4GHA
Tvsu 610NGA
NAM 1761/3NAM
Uniswa 2007 NAM
SwaziRedSWA
SB16 5ANAM
VSSP6CMR
LunTSLA
S19-3 2007NAM
S19/3NAM
MHNblackNAM
Nav RedGHA
Tvsu 747ZMB
Tvsu 999ZWE
GabCBWA
DipCBWA
Tvsu 569CMR
AHM968NAM
AS17RSA
Malawi 3MLW
RamayanaIND
DodCTZA
TiganicuruMLI
DodRTZA
54
99
99
81
0.52
0.6
0.68
0.76
0.84
0.92
I
II
III
1
Nei & Li's Coefficient
Figure 3.3 (b) Cluster analysis based on the 24 bambara groundnut landraces, the dendrogram was
obtained based on 65 SSR markers, the UPGMA tree is based on the Nei and Li, 1979 similarity
coefficient. The number at the nodes of branches represents the percentage bootstrap support of
individual nodes at resampling at 1000.
93
Table 3.6: A comparison of the distribution of the 24 bambara groundnut
landraces based on the UPGMA clustering analysis done using a set of 201 DArT
markers and 65 SSR markers.
a) DArT Marker
Cluster I
LunT SLA
Tvsu 569 CMR
Nav 4 GHA
Nav Red GHA
Tvsu 610 NGA
Tiganicuru MLI
b) SSR marker
Cluster I
Nav 4 GHA
Tvsu 610 NGA
Uniswa 2007 NAM
NAM 1761/3 NAM
SwaziRedSWA
SB16 5A NAM
VSSP6 CMR
LunT SLA
Cluster II
VSSP6 CMR
DodR TZA
Ramayana IND
NAM 1761/3 NAM
Mahenene black NAM
Cluster III
DodC TZA
S19/3 NAM
S19-3 2007 NAM
Uniswa 2007 NAM
SB16 5A NAM
AHM968 NAM
Malawi 3 MLW
Tvsu 747 ZMB
Gabc BWA
Tvsu 999 ZWE
AS17 RSA
DipC BWA
SwaziRedSWA
Cluster II
S19/3 NAM
S19-3 2007 NAM
Mahenene black NAM
Nav Red GHA
Tvsu 747 ZMB
Tvsu 999 ZWE
Gabc BWA
DipC BWA
Tvsu 569 CMR
AHM968 NAM
AS17 RSA
Malawi 3 MLW
Ramayana IND
DodC TZA
Cluster III
Tiganicuru MLI
DodR TZA
DArT markers revealed higher similarities between the landraces with an average
of 0.71 and a wider range of 0.48 to 0.99 between the landraces compared to SSR
markers which had a genetic similarity estimate mean of 0.65 and a range of 0.43
to 0.87 (Appendix 4). The lowest difference between the landraces according to
DArT markers is between AHM968 from Namibia and GabC from Botswana
94
(0.99), and the largest difference is between landraces DodR from Tanzania and
Tvsu 610 from Nigeria (0.48). For SSR markers the least difference was between
S19-3 and S19/3 at 0.87, and the largest genetic distance is between Nav 4 and
Tvsu 999 at 0.43. The differences in genetic distance/similarity estimates by
markers has been attributed to the extent of distribution of genome coverage by
markers and their evolutionary different properties and the individual loci used for
analysis (Geleta et al., 2005). SSR markers are also likely to show more intralandrace variability, as they are likely to evolve at higher rates than DArT
markers.
The cophenetic estimates measures how the dendrogram produced clearly reveals
pairwise distances in the original data. Cophenetic values for the two marker types
displayed significant values with r = 0.97 for DArT and r = 0.83 for SSR which is
an indication of a very good fit and a good fit, respectively (Figure 3.3a and
Figure 3.3b). The correlation between genetic similarity estimates for the two
markers were highly significant, using the correlations estimates produced using
Pearson product-moment coefficient correlation,
Spearman rank’s coefficient
correlation and Mantel tests correlation (Table 3.7) and (Figure 3.4 a and Figure
3.4 b).
Table 3.7: Pearson, Spearman and Mantel test correlations between the genetic
similarity matrices based on the two markers systems (DArT vs SSR).
Marker
DArT
SSR
N
P value
Pearson productmoment coefficient
Correlation
DArT
1
0.346
276
0.01
Spearman(rank)
coefficient
correlation
DArT
1
0.336
276
0.01
Mantel test
DArT
1
0.354
276
0.005
N = Number of observations in matrix
95
a)
b)
0.99
DArT
0.86
0.73
0.61
0.48
0.43
0.54
0.65
0.76
0.87
SSR
Figure 3.4 (a) A scatter plot produced based on the matrix for DArT and SSR markers genetic
distance estimates from Nei and Li, 1979 (Appendix 4) using Pearsonproduct-moment coefficient
correlation based on SPSS version 16 (b) A scatter plot based on the matrix for DArT and SSR
produced from the same genetic distance estimates in Appendix 4using Mantel-matrix
correspondence test on NTSYS pc version 2.1 program (MXCOMP module) based on 1000
permutation.
96
3.4
Discussions:
Diversity in bambara groundnut.
Tests for deviations from Hardy Weinberg proportions are usually used to check
for random mating in populations which will in turn be used to estimate the
inbreeding coefficient (Robertson and Hill, 1984). The results showed that all
markers did not conform to HWE. There was a deficit of heterozygosity observed
among the 24 bambara groundnut landraces as confirmed by the (f) inbreeding
coefficient average of0.97, which could be the most likely reason to account for
the deviations from Hardy Weinberg equilibrium. The number of alleles per
marker ranged from 1 to 14, with an average of 5 per marker and the observed
heterozygosity (Ho) was lower than the expected heterozygosity, which would be
consistent with the clear deviation observed from HWE. The heterozygosity of the
SSR markers is very low, reflecting the genetic composition and mating behaviour
of the tested landraces, as inbreeding, together with the lack of null alleles in these
markers suggesting that this marker type has usefully revealed very low levels of
out-crossing in bambara groundnut.
Even though there was a large inbreeding estimate in this study at an average of
0.97, the accessions showed far higher expected heterozygosity (He = 0.45) than
observed. The selected landraces are originally from 13 countries, with the
majority of landraces from Southern Africa (14), seven from West Africa, while
two are from (East Africa) Tanzania and one is from (Asia) Indonesia. A similar
observation was made in other highly self-fertilizing species. Siol et al., (2008) set
up a study on Medicagotruncatula, to find out the reason behind the high genetic
diversity among the selfing species. Seven microsatellite loci were used and
showed between two and five alleles per loci per locus, and an average observed
heterozygosity of 0.011 against an expected heterozygosity of 0.457.
Buso et al., (2006) carried out a similar exercise for common bean (Phaseolus
vulgaris) which is thought to have a similar breeding system to bambara
groundnut. They found that from 20 SSR markers evaluated using 85 accessions,
the number of alleles per locus ranged from 3 to 10, with a mean of 7. They also
recorded a lower observed heterozygosity (Ho) of 0.026 compared to the expected
heterozygosity (He) of 0.622, suggesting that it is also an inbreeding crop.
97
Principal Component Analysis (PCoA); in this study DArT markers were able to
clearly differentiate the 24 bambara groundnut landraces in a way that
corresponded to their areas of origin. Similar findings by Massawe et al., (2002)
using AFLP and Amadou et al., (2001) using RAPDs on studying bambara
groundnut landraces. However, Yang et al., (2006) when using DArT markers on
the analysis of pigeonpea, found that they could not be differentiated according to
their place of origin, but the markers were related to their morphological
characters. The DArT markers showed more molecular variation among the West
African bambara groundnut landraces compared to the Southern African
landraces, which may be a reflection of the domestication pattern of bambara
groundnut. The West African landraces as the putative area of origin has been
identified as more diverse using morphological markers (Pasquet et al., 1999). In
contrast, principle component analysis for the SSR markers did not clearly show
the differentiation of the 24 landraces based on their areas of origin, it showed that
there is greater genetic differentiation among the southern African landraces than
the West African materials. This shows that the two markers reveal different
levels of discrimination (Jaccoud et al., 2001), possibly due to their differing
mutation rates. However, the DArT markers explained the greater proportion of
the molecular variability in the first two axes among the 24 bambara groundnut
landraces at 37.3% as compared to SSR with 19.5 %.
Both DArT and SSR markers cluster analysis fits well with the dendrogram
produced. This was revealed by high cophenetic coefficients for each marker type
0.97 for DArT and 0.83 for SSR.
Other researchers have recorded similar
magnitudes of cophenetic correlation, Giancola et al., (2002) recorded cophenetic
coefficient r = 0.701 for SSR marker among 100 soybean cultivars when using 33
SSR markers, in 12 soybean accessions, Powell et al. (1996) recorded r = 0.958
for 36 SSR markers, Raman et al. (2008) recorded a similar cophenetic coefficient
of 0.97 for DArT markers, in a set of 94 genotypes of Lupinus albus L.
A highly significant Pearson product-moment correlation coefficient (r =0.35),
Spearman
rank
correlation
coefficient
(r
=0.34)
and
Mantel
matrix
correspondence test ( r = 0.35) which was low showed that both techniques, even
though there are targeting different parts of the genome, their results could still be
inferred to some extent from one to the other (Table 3.7). The comparison made
98
on wheat using Mantel test correspondence showed a relatively higher correlation
between DArT and SSR markers. Mantovani et al., (2008) found a correlation
between the genetic distance matrices of DArT and SSR of r = 0.68 among a set
of 31 accessions using 1,315 DArT markers and 103 SSR markers, which
indicated an agreement between the two markers. Stodart et al., (2007) observed a
strong positive correlation (r =0.84) between DArT and SSR markers when using
256 DArT markers and 63 SSR markers on 44 accessions of bread wheat
(Triticum aestivum L.) using Mantel test correspondence.
The DArT marker as in the PCO was able to clearly separate landraces based on
their areas of origin. The use of these relationships revealed by the PCO analysis
and UPGMA dendrogram could assist in formulating a breeding program for
bambara groundnut, for example, by selecting genetically far apart landraces for
cross breeding in this case which could combine the best attributes of landraces
from Southern Africa with those from West Africa, taking into account agroecological zones.
Both molecular techniques DArT and SSR, showed a relatively similar cluster
pattern. However it was the DArT marker which consistently showed more
efficiency, revealing higher PCA score values, bootstrap values and clearly
structured PCO and clusters consistent with known origins. Similar studies on the
comparison of DArT and SSR by Mantovani et al., (2008) revealed that cluster
classification for DArT was more robust as compared to the one obtained through
SSR markers or at least, is more functionally useful. Which they suggested could
be due to the relatively medium to high numbers of polymorphic markers for
DArT that can be identified and it is difficult to get a similar numbers for SSR
markers
3.5
Conclusions
Genetic analysis for the two techniques broadly showed a similar pattern of
clustering; grouping landraces was mainly based on their areas of origin. Both
clustering and PCoA, DArT markers consistently defined landrace on their areas
of origin. SSR marker revealed a higher differentiation among the landraces,
shown by lower average genetic similarities. Comparatively the initial costs for
the two markers are relatively similar (Hurtado et al., 2008), but the DArT
99
markers have the advantage of less cost per assay and this makes DArT markers
more attractive. In this study DArT markers proved to be relatively superior,
however fewer genotypes were used for the comparison of the two markers.
A summary of the achievements in this chapter is the development and
characterisation of bambara groundnut and the comparison of SSR markers and
DArT. A set of 68 markers were characterised and found to be suitable for later
use in genetic studies of bambara groundnut
100
CHAPTER FOUR: Phenotypic diversity for morphological and
agronomic characters of bambara groundnut
4.1
Introduction
Bambara groundnut grows well in low input cropping systems hence it is one of
the legume crops preferredby many subsistence farmers (Harris and Azam-Ali,
1993). Physiological variation in bambara groundnut has been recorded for
different responses to photoperiod (Linnemann et al., 1995), sowing date (Sesay
et al., 2008), moisture deficit (Collison et al., 1999; Mwale et al., 2007) and
growth rate (Massawe et al., 2003; Collinson et al., 1996). Confirming there could
mean single genotypes have different responses but this provides a good
opportunity for bambara groundnut variety development. However, at the moment
yields are unreliable and low; due to the lack of developed varieties and farmers
are still planting landraces (Zeven, 1998).
Crop genetic diversity is important for crops to withstand pest and diseases and is
useful for plant breeders to enhance the breeding progress of traits of economic
value such as yield. A wide range of phenotypes provide more insurance and
during harsh climatic changes crops with favourable characters survive better than
poor ones, therefore enough genetic diversity will ensure survival of the
crop/species.
Knowledge of phenotypic diversity has been employed in crop improvement in
developing breeding lines that have stable yields and across various environments;
for example in chickpea (Yaghatipoor and Farshdfar, 2007);and in breeding for
disease tolerant and drought tolerant genotypes in groundnut (Puttha et al., 2008;
Painawadee et al., 2009). Durán et al., (2005) used morphological characteristics
to estimate phylogenetic relationships among lines of Caribbean bean landraces
within the Andean and Mesoamerican gene pools. Morphological characters
managed to identify two clusters, one with Mesoamerican characteristics which
includes red mottled lines, while the Andean characteristics included all the lines
from Puerto Rico and the Dominican Republic. These examples demonstrate that
even though morphological markers can be influenced by the environment, their
application especially to underutilized crops is still very important.
101
Yield is usually a complex trait and controlled by a number of genes as well as
influenced by the environment. The application of correlation analysis association between two or more characters - is important to understand how an
improvement in one character could cause simultaneous changes to other
characters (Falconer and Mackay, 1996). In order to develop high yielding
varieties it is important to study the genetic variation for yield and yield
components which are in turn influenced by the genetic and environmental causes
(Maniee et al., 2009). There is limited amount of work on heritability and genetic
advance on the quantitative characters of bambara groundnut.
4.1.1
Correlation analysis studies
Yield is an important and complex trait difficult to manipulate for crop
improvement (Shi et al., 2009), however yield such as seed number per plant,
seed yield per hectare, pods number per plant and 100 seeds weight could be
correlated to other characters. This will then allow an indirect selection of yield
based on those characters.
Relatively low correlations were found among most of the traits among 1384
bambara groundnut accessions at the International Institute of Tropical
Agriculture (IITA) Nigeria, by Goli et al., (1995).
They found a strong
correlation of seed yield per plant to a number of characters which they identified
as potential characters to select for bambara groundnut improvement. They
recorded correlations of seed yield per plant to number of seeds per pod of(r =
0.88), and to pods per plant (r = 0.30). A positive correlation of r =0.13 was
observed between 100 seed weight andseed yield per plant which is a good
indication that these two characters could be used effectively in the selection of
bambara groundnut. Number of stems per plant was positively correlated with
days to maturity, which was an indication that plants with more stems matured
late, a negative correlation was found between days to maturity and cercospora
virus index, which could indicate that fast maturing plants have a lower
probability of infection; however no heritability studies were undertaken.
102
Table 4.1.1: A comparison of correlations between yield components; seed yield
per plant, number of pods per plant, seed yield per hectare and 100 seed weight
and a number of characters , sourced from Karikari and Tabona, (2004); Misangu
et al., (2007); Ouedraogo et al., (2008); Goli et al., (1995); Jonah et al., (2010);
Karikari, (2000), and Oyiga and Uguru, 2011.
No.
1
2
Characters
Seed yield per plant
Materials used in the study
12 landraces
9 landraces
310 accessions
1384 accessions
12 landraces
12 landraces
9 landraces
310 accessions
12 landraces
12 landraces
310 accessions
12 Landraces
9 landraces
310 accessions
1384 accessions
12 Landraces
9 landraces
1384 accessions
9 landraces
310 accessions
12 Landraces
9 landraces
310 accessions
12 Landraces
9 landraces
1384 accessions
9 landraces
1384 accessions
9 landraces
1384 accessions
310 accessions
1384 accessions
12 Landraces
310 accessions
1384 accessions
12 Landraces
310 accessions
1384 accessions
12 landraces
1384 acessions
12 landraces
12 landraces
Number of pods per plan
1384 accessions
12 landraces
12 landraces
310 accessions
9 landraces
1384 accessions
13 genotypes
12 landraces
310 accessions
12 landraces
310 accessions
12 landraces
9 landraces
1384 accessions
13 genotypes
9 landraces
1384 accessions
12 landraces
310 accessions
1384 accessions
12 landraces
310 accessions
1384 accessions
12 landraces
310 accessions
Correlation character Correlation Values Reference
Number of pods per plant
0.764
Karikari and Tabona, 2004
Misangu et al., 2007
0.83
0.852
Ouedraogo et al. , 2008
0.3
Goli et al., 1995
0.33
Jonah et al., 2010
100 seed weight per plant
0.415
Karikari and Tabona, 2004
Misangu et al., 2007
0.16
0.257
Ouedraogo et al., 2008
0.06
Jonah et al. , 2010
Shelling percentage
0.587
Karikari and Tabona, 2004
Ouedraogo et al., 2008
0.275
-0.1
Jonah et al., 2010
Plant height
0.38
Misangu et al., 2007
0.026
Ouedraogo et al., 2008
0.08
Ouedraogo et al., 2008
0.42
Jonah et al., 2010
Days to maturity
-0.41
Misangu et al., 2007
-0.01
Ouedraogo et al., 2008
Seed width
0.11
Misangu et al., 2007
0.224
Ouedraogo et al., 2008
-0.23
Jonah et al., 2010
Seed length
-0.08
Misangu et al., 2007
0.295
Ouedraogo et al., 2008
-0.05
Jonah et al., 2010
Number of leaves per plan
0.34
Misangu et al., 2007
0.12
Goli et al., 1995
Leaf width
0.11
Misangu et al., 2007
0.13
Goli et al., 1995
Leaf length
0.5
Misangu et al., 2007
0.14
Goli et al., 1995
Pod length
0.194
Ouedraogo et al., 2008
0.1
Goli et al., 1995
0.13
Jonah et al., 2010
Pod width
0.092
Ouedraogo et al., 2008
0.08
Goli et al., 1995
0.15
Jonah et al., 2010
Canopy spread
0.231
Ouedraogo et al., 2008
0.15
Goli et al., 1995
Number seeds per pod
0.202
Karikari and Tabona, 2004
Goli et al., 1995
0.04
Number of seeds per plant
0.882
Karikari and Tabona, 2004
Jonah et al. , 2010
Seed yield per hectare
0.8
100 seed weight
0.12
Goli et al., 1995
-0.71
Jonah et al., 2010
Shelling percentage
0.35
Jonah et al., 2010
-0.172
Ouedraogo et al., 2008
Plant height
0.55
Misangu et al., 2007
0.23
Goli et al., 1995
0.66
Oyiga and Uguru, 2011
0.25
Jonah et al., 2010
Seed width
0.023
Ouedraogo et al., 2008
-0.74
Jonah et al., 2010
Seed length
0.073
Ouedraogo et al., 2008
0.51
Jonah et al., 2010
Number of leaves per plan
0.31
Misangu et al., 2007
0.36
Goli et al., 1995
0.663
Oyiga and Uguru, 2011
Misangu et al., 2007
Leaf length
0.56
0.27
Goli et al., 1995
Pod length
0.67
Jonah et al., 2010
0.03
Ouedraogo et al., 2008
0.25
Goli et al., 1995
Pod width
-0.67
Jonah et al., 2010
-0.079
Ouedraogo et al., 2008
0.18
Goli et al., 1995
Seed yield per plant
0.33
Jonah et al., 2010
Ouedraogo et al., 2008
0.852
103
Table 4.1.2(Continued)
No.
3
4
Characters
Materials used in the study
Seed yield per hectare 9 landraces
12 landraces
9 landraces
310 accessions
12 landraces
100 seeds weight
310 accessions
1384 accessions
310 accessions
9 landraces
12 Landraces
310 accessions
1384 accessions
12 landraces
9 landraces
1384 accessions
310 accessions
9 landraces
310 accessions
12 landraces
310 accessions
1384 accessions
12 landraces
310 accessions
1384 accessions
12 landraces
310 accessions
9 landraces
1384 accessions
Correlation character Correlation Values Reference
100 seed weight
0.88
Karikari , 2000
Jonah et al., 2010
0.16
Shelling percentage
0.82
Karikari , 2000
Ouedraogo et al., 2008
-0.054
Jonah et al., 2010
-0.11
Number of pods per plant
-0.054
Ouedraogo et al., 2008
0.12
Goli et al., 1995
Shelling percentage
0.187
Ouedraogo et al., 2008
0.88
Karikari, 2000
Jonah et al., 2010
0.1
Plant height
0.096
Ouedraogo et al., 2008
0.11
Goli et al., 1995
0.27
Jonah et al., 2010
Days to maturity
-0.8
Karikari, 2000
Goli et al., 1995
-0.14
Seed width
0.524
Ouedraogo et al., 2008
0.85
Jonah et al., 2010
seed length
0.529
Ouedraogo et al., 2008
0.79
Jonah et al., 2010
Pod length
0.44
Ouedraogo et al., 2008
0.29
Goli et al., 1995
0.81
Jonah et al., 2010
Pod width
0.491
Ouedraogo et al., 2008
0.48
Goli et al., 1995
0.62
Jonah et al., 2010
Canopy spread
0.23
Ouedraogo et al., 2008
-0.86
Karikari, 2000
Goli et al., 1995
0.16
Karikari, (2000) when studying the variability of Botswana and Zimbabwean
landraces in a field experiment he found a significant correlation of days to
flowering (0.84), 100 seeds weight(r = 0.88), and shelling percentage (r = 0.82), to
grain yield kg ha-1 of bambara groundnut. A negative correlation was recorded
between grain yield kg ha-1to both canopy spread at (r = -0.85), and days to
maturity at (r = -0.63). He also carried out a heritability analysis to identify the
best traits for selection in the Botswanan environment. He recorded heritability
(h2) values of 0.72 for grain yield, 0.25 for 100 seed weight, 0.38 for shelling
percentage, and 0.36 for plant dry matter at harvest. Ouedraogo, et al., (2008)
characterised and evaluated 310 accessions from Burkina Faso, yield per plant
was correlated against a number of characters and revealed a positive correlation
todays to flowering (0.06),100seeds weight(r = 0.257) but low correlations. In
contrast canopy spread had a positive correlation(r = 0.231) while shelling
percentage had a negative correlation (r = -0.199) to yield per plant. The
characters were not subjected to heritability analysis.
Karikari and Tabona (2004) undertook a study on 12 bambara groundnut
landraces to identify characters associated most with adaptation to drought in the
Botswanan environment. Their results showed canopy spread, 100 seed weight,
104
root-shoot ratio, and number of seeds per pod as the most suitable characters. A
highly significant correlation was found between seed yield per plant with number
of pods at r = 0.76, and between seed yield per plant with number of seeds per
plant at r = 0.88. They emphasised the importance of root-shoot ratio in the semiarid environment of Botswana, and they found that the root-shoot ratio has
significant correlation with shelling percentage and seed yield per plant at r =
0.296 and r = 0.398, respectively.
Jonah et al., (2010), investigated the genetic correlations between yield and yield
related characters in 12 bambara groundnut landraces in Nigeria. Highly positive
correlations were found between seed yield per hectare and pod yield per plant (r
=0.87), and between seed yield per hectare and seed yield per plant (r = 0.91), and
between seed yield per plant and plant height at 8 weeks after sowing(r =0.77). A
high correlation was also identified between pod length and pod width at (r =0.89)
and seed length and seed width at (r =0.82), which is potentially useful for
selecting genotypes with bigger seeds. However they found a negative correlation
between pod number per plant with 100 seed weight at (r= -0.74) which implies
that selecting landraces for higher pods numbers could lead to, leaner pods
produced in turn.
Wigglesworth, (1996), undertook a field trial on six bambara groundnut landraces
in Botswana in order to study the genotypic variation and heritability of pod
numbers, 100 seed weight, seed weight per plant and to find some correlation
between the traits. The results recorded a significant phenotypic correlation
between pod numbers and seed weight per plant (r = 0.77), and between 100 seed
weight and seed weight per plant (r = 0.52). Heritability values recorded were
lower for seed weight per plant (0.25), pod number (0.39) and higher for 100 seed
weight at (0.94). Therefore 100 seed weight was singled out as an important trait
to select for, among the local landraces.
Thirteen bambara groundnut populations were evaluated for floral structure in
Nigeria in a field experiment byOyigaet al., (2010). They recorded an anther
diameter correlation with number of pods per plant (r =0.41), and to seed weight
per plant (r =0.51) which is an indication of the relationship between seed number
and the biomass synthesized during the growth stages of seed formation (Jeuffroy
105
and Chabanet, 1994). Selection for anthers with larger diameter was identified as
a strategy for yield improvement in bambara groundnut. They also found a
negative correlation between stigma anther separation with seed weight per plant
at (r = -0.59), which they thought the stigma-anther separation is an important
factor in the production of low seed weight in bambara groundnut.
In another study, Jonah et al., (2010), carried out a phenotypic diversity study on
12 bambara groundnut landraces. They undertook some broad sense heritability
and genetic advance estimates and recorded high heritability and genetic advance
in pod yield per plant (0.75; 16%), for seed width (0.85; 16%), and for 100 seed
weight (0.70; 12%) the high heritability and genetic advance suggest that these are
selectable traits. However, they reported moderate value for seed yield per hectare
at 0.54, which indicate a limitation to the improvement of this trait.
4.1.2. Selection of lines for breeding
For a robust plant breeding program the selection process should be effective
enough to capture those individuals with which are superior in a number of traits
(Strefeler and Wehner, 1986). It is usually genotypes with the superior characters
that are recommended over others for crop improvement and usually the selection
indices are used to identify those genotypes. In many breeding programs more
than one trait is been selected for at the same time, and thus multiple selection
indexes are used. Several simultaneous selection indices such as those by Smith
(1936) and Hazel (1943) are used (Tardin et al., 2007). The selection index
method is expected to be faster in generating benefits, the method assigns suitable
weights for each trait depending on its importance (Eshghi et al., 2011). The
concept of selection index as was developed by Smith (1936) was found to have
some difficulties like determining relative economic values thus several
modifications have been made (Baker, 1974).
The efficiency of selection indices does not depend only on the estimation ofthe
coefficient, but also on the crop and characters under study. Monirifar, (2010)
adds after evaluating and constructing some selection indices for use in alfalfa
(Medicago sativa). In this study, leaf area, shoot dry weight (biomass), yield such
as pod number per plant and seed number per plant were characters selected for
use in the selection index. Canopy development is reported as an important
106
determinant of crop radiation capture and is mainly influenced by temperature
(Massawe et al., 2003) while Collinson et al., (1999) reported the effect of soil
moisture deficit and its impact on leaf area development and yield reduction in
three bambara groundnut landraces.
The selection index (SI) adopted from Monirifar, (2010) comes up with one value
from several variables that had been selected as of economic value by the breeder.
This technique also requires the use of some weight for each variable to include in
the equation, but with different weight assigned depending on its importance as
deemed by the plant breeder (Baker, 1974). The selection index is often used in
the breeding program of cassava at the International Center for Tropical
Agriculture (CIAT) (Ceballos et al., 2007). The selection index (SI) has recently
been utilised in soybean germplasm evaluation for acid tidal swamp tolerance
(Kuswantoro et al., 2010). A greenhouse and a field experiment were conducted
based on assessing 17 genotypes for six characters. One genotype with the highest
ranking because of its root and shoot dry weight was identified. Ojulong et al.,
(2010) used a similar selection index to evaluate cassava seedlings developed for
yield characteristics, and identified traits which they suggested could be included
in selection of the crop such as root weight and root weight per tree.
4.1.3
The objectives of this study were
To evaluate and characterise bambara groundnut landraces based on agromorphological characters
To assess the genetic diversity of bambara groundnut landraces based on
Shannon weaver index, genotypic variability, phenotypic variation,
heritability and genetic advance
To classify bambara groundnut genotypes by means of cluster and
principal component analysis in order to select genotypes suitable for
further breeding
To identify better performing lines in a Botswana environment based on
seed yield and biomass production based on the selection index and
Duncan Multiple Range Test
107
4.2
Results
4.2.1
Qualitative analysis of the genotypes
The frequency distribution among the 35 bambara groundnut planted in the
glasshouse and 34 bambara groundnut lines planted in the field experiment, shows
that majority of the seed colour of the landraces were reddish in colour (77.2% of
the landraces planted in the agronomy bay, and 85.1% of the genotypes that were
planted in the field experiment) based on classes 3, 4, 5, and 6 for testa colour
(Table 4.1.2). Twolandraces had cream coloured seeds while three had black
seedsin both the glasshouse and field experiment. Most landraces had no eye
pattern (91.4%; 92.9%) and no testa pattern (82.9%; 82.1%) and only 2 seeds had
dotted spot. Even though farmers have been observed to plant a mixture of
colours for bambara groundnut, two surveys done in Swaziland and Botswana
revealed that the most preferred landraces are the cream coloured ones (Sesay et
al., 2003; Brink et al., 1996). In this study all pod colour classes were observed
with the exception of black ones, and most of them were those which are pointed
with a nook (60%; 64.2%)(Table 4.1.2).
The crops experienced little stress in both the experiments (88.6%; 85.3%) the
temperature recordings in the glasshouse had an average of 21.7oC while in the
field an average of 29.5oC was observed and these are ideal temperatures for
bambara groundnut (Swanevelder, 1998). Three types of plant growth habit were
observed 47.1% were bunch type, 38.2% were semi-spreading and 14.7% were
spreading types, which shows that farmers are mostly selecting for the bunch and
semi spreading types.
108
Table 4.1.2: Descriptor, classes and frequency distribution among the 35 landraces planted in the agronomy bay and 34 bambara groundnut lines
selected and planted in the field in Botswana
Descriptor and Classes
Frequency of class (%)
0
1
2
3
4
5
6
7
8
5.7
0
14.3
11.4
28.6
22.9
8.6
8.6
7.1
0.0
14.3
10.7
42.9
17.9
0.0
7.1
91.4
5.7
0.0
0.0
2.9
0.0
0.0
92.9
3.6
0.0
0.0
3.6
0.0
0.0
82.9
5.7
0.0
0.0
5.7
2.9
0.0
2.9
0.0
82.1
0.0
0.0
0.0
7.1
3.6
0.0
7.1
0.0
21.4
39.3
21.4
17.9
0.0
21.4
39.3
21.4
17.9
0.0
42.9
31.4
17.1
8.6
46.4
28.6
17.9
7.1
8.6
28.6
60.0
2.9
3.6
28.6
64.3
3.6
Testa colour
1= Cream
2= Grey
6= Dark
brown
5= Brownish red
3=Light red
4= Dark red
7=Dark purple
8= Black
Eye pattern
0=No eye
pattern
1= Butterfly
4= Thick dotted
lines
5=Circular
2=Triangular
3= Mottled
6=Thin lines
Testa pattern
0=no pattern
1= Entire
2=Striped
4=Dotted
5=Little rhomboid one side spotting
6=Little rhomboid two side spotting
3=Marbled
7=Much rhomboid
Pod colour
1=Yellowish brown
2= Brown
4 =Purple
5=Black
3=Reddish brown
Pod texture
1= Smooth
2= Little grooved
3=Much grooved
8= Holstein
4 = Much grooved
Pod shape
1=Without point
2=Pointed
3=Pointed and nooked
4=Pointed both sides
109
Table 4.1.2 continued
Seed shape
1 =Round
2= Oval
Terminal leaflet colour
1=Green
2=Red
3=Purple
Stress susceptibility
1= No visible sign
4= High
2=Low
3= high
85.7
21.4
78.6
74.3
0.0
25.7
73.5
0.0
26.5
88.6
2.9
5.7
2.9
0.0
85.3
2.9
5.9
5.9
0.0
0.0
65.7
31.4
2.9
0.0
58.8
38.2
2.9
47.1
2.9
41.2
58.8
47.1
38.2
5= very high
Leaf shape
1=Round
14.3
2=Oval
3=Lanceolate
4= Elliptic
50.0
Stem hairiness
0=absent
3=Sparse
5=Dense
Leaf colour at germination
1=Green
2=Purple
Growth habit
14.7
1=Bunch
2=Semi-bunch 3=Spreading
The ones in bold: Glasshouse experiment results
110
4.3.2
Shannon Weaver (H’) diversity analysis
Knowledge of variation of characters is important to plant breeders since they
should know which population is more varied for which characters. Shannon
Weaver (H’) within population index (Hennink and Zeven 1991), therefore (H’)
can be useful in identifying those traits that warrant the attention of breeders to
improve. The estimates of Shannon weaver (H’), was relatively high with a
similar mean diversity 0.70 in UK and 0.69 in Botswana (Table 4.1.3). The
diversity ranged from 0.19 (Leaflet length) to 0.97 (petiole-internode ratio) in UK
and in Botswana it ranged from 0.19 (leaflet width) to 0.99 (pod width).
Table 4.1.3: Shannon-Weaver index on the phenotypic diversity of 24
quantitative characters in the agronomy bay experiment and the field experiment
(Botswana).
UK
BOTSWANA
H'
H'
Days to emergence
0.86
0.32
Days to 50% flowering
0.96
0.43
Leaf number.
0.71
0.71
Spreading
0.77
0.74
Leaflet length
0.19
0.73
Leaf width
0.97
0.19
Leaf Area
0.88
0.84
Plant height
0.79
0.92
Internode
0.67
0.54
Petiole
0.79
0.92
Pet-Internode
0.97
0.86
Petiolule
0.24
0.73
Peduncle
0.19
0.73
Stem number.
0.96
0.94
Days to maturity
0.42
0.79
Shoot dry weight
0.93
0.88
Pod numbers per plant
0.93
0.61
Pod dry weight
0.94
0.56
Pod length
0.70
0.77
Pod width
0.32
0.99
Seed number.
0.90
0.56
Seed length
0.32
0.59
Seed width
0.51
0.86
Seed weight
0.95
0.28
Mean diversity (H’)
0.70
0.69
Characters
111
The characters that showed greatest variation between the two sites were the
leaflet length, leaflet width, seed weight, and pod width. This variance was
reflected in pods number per plant and seeds number per plant which was
drastically affected in the field experiment.
The Shannon-Weaver diversity index (H’) was calculated on the qualitative
characters to compare the genetic diversity among characters both in UK and
Botswana. The most diverse characters were pod colour (0.94) and testa colour
(0.93) in UK, leaf colour at emergence (0.98) and pod colour (0.96) in the
Botswana. In both UK and Botswana the least diverse characters was eye pattern
0.32 and 0.28 respectively.
Table 4.1.4: Shannon weaver index on phenotypic diversity of qualitative
characters for the studied landraces in agronomy bay and field experiment
Character
Pod texture
Pod colour
Pod shape
Seed shape
Testa colour
Testa pattern
Eye pattern
Leaflet colour
Leaf shape
Stress susceptibility
Leaf colour at emergence
Stem hairiness
Growth habit
Mean diversity (H’)
UK
H’
0.89
0.94
0.70
0.59
0.93
0.47
0.32
0.82
0.67
0.34
*
*
*
0.67
Botswana
H’
0.87
0.96
0.63
0.75
0.87
0.42
0.28
0.83
0.71
0.41
0.98
0.73
0.91
0.72
* Not recorded
These considerable variation, identified based on (H’) is important for bambara
groundnut improvement, however there is non-significant and low correlation(r =
0.168) between diversity analysis values in UK and Botswana field experiment for
quantitative characters, compared to a highly significant correlation (r = 0.953) for
112
the qualitative characters, based on Pearson correlation analysis. These also reflect
the lower effect of the environment on qualitative compared to quantitative
characters.
4.3.3
Descriptive analysis of the genotypes
The average genetic diversity was slightly higher in Botswana (0.71) compared to
UK (0.67) and substantial variability among the genotypes was revealed in most
of the 24 characters (Table 4.1.5, and 4.1.6) for both experiments in the UK
(agronomy bay) and Botswana (field experiment). For example, the minimum and
maximum shown in the UK agronomy for shoot dry weight are 6.9 g -113.9g, leaf
area 1341 cm2 – 4489 cm2,pod numbers per plant 7 - 182 and for seeds numbers
per plant7 – 155, while in the field experiment the ranges for shoot dry weight is
12.8 g - 113.7g, leaf area 52 cm2 - 3304 cm2, pod number per plant 2 - 138 and
for seed number per plant 1- 140. This show there is great potential for selection
in these traits for further crop improvement.
113
Table 4.1.5: Descriptive characteristics for the 35 bambara groundnut planted UK
(Agronomy bay, 2008) from an average of three plants, the vegetative characters
recorded at 10 weeks after planting, while the yield characters are recorded after
harvest.
Characters
Minimum
Maximum
Mean
Stdev
CV
F test
Days to emergence (d)
8
18
10.1
1.3
12.9
Days to 50% flowering (d)
38
54
42.4
2.4
5.8
Number leaves per plant
33
293
79.4
23.5
29.6
Canopy width (cm)
7
51
19.5
4.7
24.3
Leaflet length (cm)
6.8
10.9
8.6
0.5
5.2
Leaflet width (cm)
2.3
5.3
3.7
0.3
8.4
Leaf Area (cm )
1341
4489
17034
1795.7
10.5
Plant height (cm)
22
46
33.5
2.3
6.8
Internode length (mm)
0.8
6
2.3
0.4
15.7
Petiole length (cm)
8.5
25.5
16.8
1.5
8.7
Petiole-Internode ratio
3.3
17.6
8
1.5
18.1
Petiolule length (mm)
0.9
4.8
2.3
0.3
14.4
Peduncle length (mm)
1
4.9
2.3
0.5
19.8
Number of stem
4
22
9.9
2.5
25.4
*
***
***
***
***
***
***
***
***
***
***
***
***
**
Days to maturity (d)
109
161
155.3
7.5
4.9
ns
Shoot dry weight (g)
6.9
113.9
33.3
11.5
34.4
Number of pods plant
7
182
54.4
19.8
36.3
Pod dry weight (g)
2.1
97.5
32.7
12
36.7
Pod length (mm)
12.3
24
18.8
1.5
8
Pod width (mm)
7.8
15.8
12.1
0.9
7.6
Number of seed plant
7
155
59.1
18.4
31.2
Seed length (mm)
6.6
13.4
10.7
0.9
8.9
Seed width (mm)
5.3
10.3
8.4
0.7
8.2
Seed weight (g)
1
69.8
23.1
8.6
37.2
***
***
***
***
***
***
***
***
***
2
*, **, *** Significant at 5%, 1% and 0.1% respectively, ns = non-significant
114
Table 4.1.6: Descriptive characteristics for the 34 bambara groundnut planted in
field (Notwane, Botswana, 2008/2009 season) with three replications, the
vegetative characters recorded at 10 weeks after planting, while the yield
characters are recorded after harvest.
Characters
Minimum
Maximum
Mean
Stdev
CV
F test
Days to emergence (d)
10
19
15.1
1.7
11
Days to 50% flowering (d)
42
67
56.7
2.8
5
Number leaves per plant
34.5
233.8
112.7
20.4
18
Canopy width (cm)
98.4
577.8
240.6
31.8
13
2
Leaflet length (cm )
22.4
94
66.9
6.6
10
Leaflet width (cm)
14.8
40
24.9
2.2
9
Leaf Area (cm)
52
8369
3304
857
25.9
Plant height (cm)
183
361.8
278.1
22
8
Internode length (mm)
11.2
86.6
22.7
6.3
28
Petiole length (cm)
63
191.4
139.4
15.5
11.1
Petiole-Internode ratio
2.3
14.5
7.2
1.2
16
Petiolule length (mm)
6.5
38
17.9
3.2
18
Peduncle length (mm)
6.1
37.6
20.2
4.8
24
Number of stem
4
16.2
8.3
1.7
20
Days to maturity (d)
126.8
155
136.9
2.5
2
Shoot dry weight (g)
12.8
113.7
41.8
9.4
23
Number pods plant
1.5
137.5
16
12.7
79
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
Pod dry weight (g)
0.3
54.1
8.2
6.9
83
ns
Pod length (mm)
12
26
18.2
1.8
10
Pod width (mm)
7.3
14
10.5
1.1
10
Number of seed plant
1
140.5
15.6
12.4
79
Seed length (mm)
7
16
11
1.2
11
Seed width (mm)
4.7
10.4
8.1
0.9
11
Seed weight (g)
0.1
30.2
5.3
4.2
80
***
***
**
***
**
**
*, **, *** Significant at 5%, 1% and 0.1% respectively, ns = non-significant
The estimate of coefficient of variation (CV) was used to compare the variability
of different characters of the 35 genotypes grown in the UK (agronomy bay
experiment) and 34 genotypes in Botswana (field experiment), Table 4.1.5 and
4.1.6). The 34 bambara groundnut lines were derived from seeds from single
plants selected among the 35 accessions from the agronomy bay experiment
(Table 2.1.2.2). The characters that revealed higher variations are the yield and
yield related traits, with a relatively higher CV in the greenhouse experiment, such
as pod numbers (36%), pod weight (37%), seed numbers (31%) and seed weight
(22%). While in the field the coefficient of variation was higher for similar traits
115
pod numbers (79%), pod weight (83%), number of seed (79%) and seed weight
(80%).
Lower coefficients of variation were revealed in characters such as days to
emergence (4.9%), leaflet length (5.2%), days to flowering (5.8%), days to 50%
flowering (5.8%) and plant height (6.8%) in the agronomy bay experiment. In the
field experiment lower coefficients of variation were recorded in days to 50%
flowering (5.75%), leaflet length (5.23%), leaflet width (8.38%), plant height
(6.81%), petiole length (8.70%), days to maturity (4.85%), pod width (7.56%),
pod length (7.95%), seed length (8.88%) and seed width (8.24) which indicates
the small variation in both environments for these traits, it also suggests that these
are more likely to be highly heritable traits.
Highly significant differences (P<0.001) were detected among genotypes in most
of the 24 characters that were analysed in the Agronomy bay (glasshouse) and in
the field experiment in Botswana. With the exception of days to maturity in the
glasshouse, and pod dry weight in the field experiment (Table 4.15 and 4.1.6).
This is an indication that there is a substantial amount of genetic variability
among the traits analysed for this set of accessions.
The largest value for seed number and pod number per plant were shown by
genotype 88-AHM753 from Namibia, in UK (agronomy bay experiment) with
121 seeds per plant, and 121 pods per plant followed by genotype 91UNIS R from
Swaziland with 107 seeds per plant and genotype 85 Acc754 from Zambia with
106 pods numbers per plant. The largest shoot dry weight was shown by 118Ramayana from Indonesia with 81 g, while the largest leaf area was attained by
48Acc790 from Kenya at 10369cm2, both were followed by genotype 85 Acc 754
from Zambia with 76 g shoot dry weight and 9997cm2 leaf area. The two leaf
areas although different, there are not significantly different from each other
according to Duncan’s range multiple test (Table 4.2.5).
The genotypes with the lowest seed number and pod number per plant were
70Acc330 and 69Acc286 both from Nigeria with 15 and 12 seeds each, and only
10 and 11 seeds each, respectively. The lowest shoot dry weight and leaf area
were shown by genotypes 56Acc89 from Mali at 10g shoot dry weight and 1357
116
cm2 leaf area, followed by 6Acc289 from Benin for shoot dry weight at 10g, and
50Acc792 from Kenya with 1887 cm2 for leaf area.
For the field experiment in Botswana, line 90S19-3 from Namibia produced the
largest number of pods per plant and seeds per plant of 68 pods and 66 seeds,
respectively. The second highest yielding line was 84Acc696 from Zambia with
an average of 36 pods and 33 seeds per plant. The greatest shoot dry weight was
recorded from 76Acc390 from Sudan with an average 103 g per plant followed by
81Acc385 from Tanzania with an average of 83 g per plant, both lines also had
the largest leaf area, of 5986 cm2 for 76Acc390, with 81Acc385 coming second
with 5437 cm2. However, these two lines produced a relatively low number of
pods with an average of 10 pods per plant. This shows that they concentrated most
of the assimilates in leaf formation instead of pods and could indicate issues with
fertility/pod/seed set.
The lowest shoot dry weights were recorded from lines 69Acc286 and 70Acc329
from Nigeria, and 45Acc231 from Ghana, both lines failed to produce any pods or
seed. Fewer numbers of pods were recorded in lines 95DODRfrom Tanzania and
113Bots5 from Botswana. In a drought experiment conducted in the glasshouse by
Collinson et al., (1999) 95DODR compared favourablywell withone landrace
from Botswana (DIPC), in terms of pod number produced and final harvest
biomass. Its poor performance in Botswana could imply that it is not well adapted
to the Botswanan environment.
The line 113Bots5 was acquired from the Francistown market in the north eastern
part of Botswana in 2004/2005 season, in an effort to study and improve some
bambara groundnut landraces.
The landrace was not included in field trials
before; its poor performance shows that it may be originally from a different
environment from the Botswanan environment.
4.3.4
Principal component analysis:
The 24 characters assessed in the Agronomy bay experiment were subjected to
principal component analysis to identify characters that accounted for most of the
variance in the selected genotypes. Only the first six principal components with
eigenvalues more than one were selected, giving an accumulated total variation of
117
79.24% (Table 4.1.7). The first principal component (PC 1) contributed 35% of
variation and characters with higher loadings were canopy width, plant height,
shoot dry weight, petiole length, internode length, pod number and leaf area. The
second principal component accounted for 14% of the total variation and higher
loadings were observed from mostly the vegetative part of the crop, petioleinternode ratio, petiolule length, leaf area, petiolelength andshoot dry weight.
Leaflet length, days to 50% flowering, days to emergence, pod length and
peduncle length accounted for most of the 11.88% identified at the third principal
component. Characters contributing most variation among the genotypes in the
fourth principal component were days to maturity, leaf numbers, days to 50%
flowering, internode length, and days to emergence. Principal components (PC5)
and (PC6) contributed approximately 6% and 5 % respectively, with pod dry
weight, days to emergence, petiolule, petiole-internode ratio and days to maturity
accounting for most of the 6 % in principal component 5. Similarly leaflet length,
day to emergence, pod dry weight, number of stems and internode lengthare
contributing higher loadings in principal component 6. Generally when observing
PC1 and PC2 in the agronomy bay experiment it shows that characters that were
able to separate the genotypes are mainly vegetative, with shoot dry weight and
petiole length appearing in both PC1 and PC2, while pod number contributedonly
in the first principal component.
118
Table 4.1.7: Principal components, matrix of eigenvalues and vectors for 24
quantitative characters of bambara groundnut landraces planted in the agronomy
bay (UK).
Principal components (PC)
PC1
PC2
PC3
PC4
PC5
PC6
Eigenvalues Variance
8.459
3.359
2.851
1.74
1.384
1.223
% Total contribution
35.25
14.00
11.88
7.25
5.77
5.09
%Accumulated
35.25
49.25
61.13
68.38
74.15
79.24
Days to emergence
-0.108
0.095
0.264
0.115
0.282
0.357
Days to 50% flowering
-0.029
0.122
0.339
0.321
0.164
-0.326
Days to maturity
0.233
-0.181
-0.057
0.322
0.247
-0.242
Plant height
0.295
0.081
0.157
0.055
-0.168
0.104
Internode length
0.255
-0.102
0.045
0.165
-0.229
0.197
Leaflet length
0.102
0.018
0.343
0.101
0.126
0.384
Leaf number per plant
0.233
-0.181
-0.057
0.322
0.247
-0.242
Leaflet width
0.209
0.017
0.142
0.115
0.159
0.171
Leaf Area
0.249
0.265
0.052
-0.029
-0.120
-0.119
Pod dry weight
0.044
-0.237
-0.078
0.013
0.538
0.349
Peduncle length
0.033
-0.337
0.204
0.081
-0.265
0.119
Petiole length
0.275
0.168
0.144
0.106
-0.131
0.054
Petiolule length
0.162
0.291
0.154
-0.362
0.266
-0.136
Pet-Internode ratio
0.162
0.291
0.154
-0.362
0.266
-0.136
Pod length
0.033
-0.281
0.229
-0.387
0.114
-0.056
Pod width
0.152
-0.313
0.184
-0.070
0.002
-0.301
Pod number per plant
0.253
0.064
-0.359
0.071
0.053
0.009
Shoot dry weight
0.281
0.162
0.042
0.053
-0.084
-0.117
Seed Length
0.198
-0.355
0.121
-0.210
-0.055
0.034
Seed width
0.149
-0.318
0.057
-0.242
-0.061
-0.167
Seed number plant
0.243
0.058
-0.372
0.025
0.088
0.037
Canopy width
0.295
0.081
0.157
0.055
-0.168
0.104
Stem numbers
0.215
0.039
-0.142
-0.229
-0.164
0.271
Seed weight
0.239
-0.100
-0.324
-0.14
0.154
0.099
The principal component analysis for the 24 quantitative characters was also
conducted among the 34 lines planted in the field (Botswana) experiment. The
first six principal components with eigenvalues over one, which accounted for
74.15% of the total variation, were selected to analyse the characters for the
selected lines (Table 4.1.8). The first component (PC1) explained 23.64 % of the
overall variability among the lines with most of the variation coming from petiole
length, leaf area, petiolule length, shoot dry weight, and plant height. High
119
loadings for the second component (PC2) which described 17.08 % of the
variance was accounted for by pod weight, seed weight, seed number, pod number
and pod width.petiole-internode ratio, days to 50% flowering, seed length, leaf
number, and pod length contributed most of the 11.98% explained at principal
component three (PC3). Principal component 4 separates the lines based mainly
on the seed length, seed width, pod length, pod width and leaflet length with a
total variation of 8.68%. Principal component five and six contributed 6.82% and
5.95 % respectively. Days to 50% flowering, leaflet length, plant height, days to
emergence and pod length contributed approximately 7% of total variance to
(PC5), while days to maturity, days to 50% flowering, pod length, peduncle length
and leaf numbers accounted for most of the 5.95 % of total variance in (PC 6).
Observations made on the field experiment revealed that the vegetative characters
contributed most of the PC1 while seed characters significantly contributed in
PC2 separating the lines. However, it was the vegetative traits in both experiments
that appear as the main characters that can be used to clearly separate the selected
genotypes.
120
Table 4.1.8: Principal component, matrix of eigenvalues and vectors for 24
quantitative characters of bambara groundnut lines planted in Botswana.
Principal components (PC)
Eigenvalues Variance
PC1
PC2
PC3
PC4
PC5
PC6
5.67
4.1
2.88
2.08
1.64
1.43
% Total contribution
23.64
17.08
11.98
8.68
6.82
5.95
%Accumulated
23.64
40.72
52.7
61.38
68.2
74.15
Days to emergence
-0.245
0.156
-0.041
-0.047
0.275
0.123
Days to 50% flowering
0.005
-0.033
0.299
-0.126
0.348
0.430
Days maturity
0.109
-0.211
0.056
0.020
0.038
0.506
Plant height
0.292
-0.020
0.050
0.165
0.293
-0.286
Internode length
0.233
0.066
-0.433
0.017
0.034
0.055
Leaflet length
0.251
0.145
-0.019
0.181
0.329
-0.153
Leaf number
0.089
-0.191
0.233
-0.07
-0.485
0.201
Leaf width
0.281
0.032
-0.022
-0.178
0.201
0.147
Leaf Area
0.326
-0.056
0.176
-0.017
-0.185
0.135
Pod weight
0.040
0.443
0.078
-0.125
0.008
0.153
Peduncle length
0.219
0.135
-0.202
0.178
-0.135
0.215
Petiole length
0.341
-0.095
0.099
-0.065
0.061
-0.313
Petiolule length
0.308
-0.099
0.092
0.098
-0.022
-0.147
Pet-Inter ratio
-0.013
-0.120
0.472
-0.157
-0.03
-0.259
Pod length
-0.046
0.157
0.226
0.371
0.268
0.215
Pod width
-0.186
0.253
0.154
0.241
0.007
-0.013
Pod number per plant
0.111
0.351
0.18
-0.329
-0.089
-0.007
Shoot dry weight
0.306
-0.112
0.211
0.018
0.113
0.105
Seed length
0.082
0.190
0.239
0.442
-0.115
0.016
Seed width
-0.014
0.238
0.098
0.404
-0.336
-0.052
Seeds number per plant
0.120
0.377
0.174
-0.290
-0.074
-0.046
Canopy width
0.285
0.031
-0.244
0.067
-0.191
0.186
Stems numbers
0.111
-0.067
-0.083
0.083
0.096
0.032
Seed weight
0.120
0.389
-0.161
-0.201
-0.053
-0.025
The characters that could potential reveal greater diversity among germplasm and
possibly as identified by higher loadings in the PC1 in both agronomy bay and
field experiment are spreading length, shoot dry weight, petiole length, petiolule
length and leaf area.
121
4.3.5
Cluster analysis
Figure 4.2.2 shows a dendrogram for the Euclidean cluster analysis for the 35
bambara groundnut evaluated for a combined analysis for 24 quantitative and 10
qualitative characters in the glass house (UK). The landraces were grouped into
three clusters, mainly on their areas of origin since cluster I consist mostly
landraces from West Africa, while cluster II consists of landraces mainly from
Southern Africa. The landraces from Central Africa were grouped with those from
West Africa, while those from East Africa were grouped with those from
Southern African together with Ramayana from Indonesia. Three landraces from
Southern
Africa
90S19-3,
109BOTS1
and
119Hybrid
showed
some
morphological similarities with those in West African. The morphological
characters used could not distinguish between 3Acc9NGA and 45A231GHA from
Nigeria and Ghana respectively.
122
87
89
68
118-Ramayana-INDMW
89
94
5.93
5.06
4.19
Euclidean
3.32
3Acc9NGA
50Acc792KEN
56Acc89MLI
6Acc289BEN
60Acc32NGA
20Acc118BFA
117-VSSP6CMR
76Acc390SDN
90-S19-3NAM
I
33Acc484CMR
69Acc286NGA
45Acc231GHA
70Acc329NGA
10Acc1276CAF
40Acc563CMR
30Acc476CMR
109-BOTS1
119-Hybrid
4Acc144NGA
74Acc335NGA
49Acc793KEN
II
113-BOTS5
48Acc790KEN
85Acc754ZMB
99-SB4-2NAM
100-SB16ANAM
91-UNISRSWA
92-AHM968NAM
81Acc385TZA
88-AHM753NAM
III
104-S-1913NAM
95-DODRTZA
105-MHNBlackNAM
118-Ramayana-IND
84Acc696ZMB
2.45
Figure 4.2.2: Dendrogram of 35 bambara groundnut landraces showing a (UPGMA) Euclidean
cluster analysis based on 34 agro-morphological markers in glasshouse experiment. The colour
code for West Africa = Green, Southern Africa =Red, East Africa =Yellow, Central Africa = Blue,
Indonesia = Purple. The number at the nodes of branches represents the percentage bootstrap
support of individual nodes resampling at 1000
A dendrogram for cluster analysis performed in the field experiment (Botswana)
using 24 quantitative and 13 qualitative characters (37 agro-morphological) on 34
lines of bambara groundnut produced three clusters traits (figure 4.2.3). Cluster I
consists mostly of lines from West Africa, but with a mixture of some lines from
Southern Africa, like 92-AHM968NAM, 91-UNISRSWA, 99-SB4-2NAM and
104-S-193NAM which also include Ramayana from Indonesia. Cluster II consists
of mixture of four lines which are from West Africa, three from Southern Africa
and one from Central Africa. Lines 40Acc563CMR and 69Acc286NGA did not
produce any pods, but are clustered together with the Southern African lines
mainly due to their higher petiole length, and leaf area and these are some of the
traits that had higher loadings in PC 1 (Table 4.1.8).
123
57
118-Ramayana-INDMW
88
6.12
5.46
4.79
Euclidean
4.13
3Acc9NGA
6Acc289BEN
56Acc89MLI
20Acc118BFA
30Acc476CMR
74Acc335NGA
45Acc231GHA
70Acc329NGA
117-VSSP6CMR
50Acc79KEN
92-AHM968NAM
91-UNISRSWA
118-Ramayana-IND
99-SB4-2NAM
104-S-1913NAM
33Acc484CMR
10Acc1276CAF
119-Hybrid
85Acc754ZMB
40Acc563CMR
69Acc286NGA
113-BOTS5
60Acc32NGA
4Acc144NGA
48Acc790KEN
84Acc696ZMB
76Acc390SDN
81Acc385TZA
88-AHM753NAM
100-SB165ANAM
105-MHNBlackNAM
95-DODRTZA
109-BOTS1
90-S19-3NAM
I
II
III
3.46
Figure 4.2.3: Dendrogram of 34 bambara groundnut lines showing genetic similarities based on
37 morpho-agronomic traits 24 quantitative traits and 13 qualitative traits, using the UPGMA
cluster analysis (field experiment Botswana). The number at the nodes of branches represents the
percentage bootstrap support of individual nodes resampling at 1000
Three East African lines were clustered together in cluster III, which also contains
six lines from Southern Africa. Line 76Acc390 from Sudan was morphologically
similar to the Southern African lines and performed relatively well in terms of pod
number per plant, leaf area and had larger leaf area and plant height. The poor
adaptation of the West African lines was revealed by lower number of pods per
plant produced. East African lines that produced comparatively higher number of
pods per plant like those from Southern Africa were grouped together such as 81Acc 385 from Tanzania, 95DODR from Tanzania and 48-Acc790 from Kenya.
Those lines which originally are from Southern Africa but, produced low number
of pods per plant like 99-SB4-2 and 92-AHM968 both from Namibia were
clustered together with the West African lines in Cluster I. Even though the lines
124
were not clearly separated based on their areas of origin, individual clusters
consist of majority of lines from one region.
A comparison of the two clusters revealed that higher genetic distance estimates
were observed in the agronomy bay at 3.38 compared to 2.66 in the field
experiment. The tree from the agronomy bay experiment clearly defined landraces
according to their areas of origin much more than in the field experiment with 5
bootstrap values above 50% compared to only 2 for the field experiment.
4.3.6
Correlation coefficients among traits
To determine the relationship among the 24 measured characters, Pearson
correlations based on the mean of the genotypes were generated (Table 4.1.9 and
Table 4.2.1) for the agronomy bay experiment and field experiment, respectively.
In the agronomy bay experiment a number of the characters were positively
correlated to both pod number and seed number, while a number of traits such as
days to emergence, days to flowering, peduncle length, leaflet length, petioleinternode ratio and pod length showed a negative correlation. This suggest that
those landraces which, emerged late ended up having a lower number of leaves
and most of those landraces affected were the spreading ones, eventually
producing a lower number of seeds.
Seed weight appears to contribute
significantly to both pod number and seed number at (r = +0.86) and (r = +0.88)
respectively, and this implies that indirect selection for pod number and seed
number can be successfully achieved by selecting for the seed weight character.
There were moderate correlations of shoot dry weight and pod number and pod
number to leaf area at (r = +0.61 and r = +0.52), respectively, while leaf area and
shoot dry weight were highly correlated to each other at (r = +0.94). Pod number
and seed number were also highly correlated at (r = +0.94). The genotypes that
produced large shoot dry weight and leaf area managed to produce higher pod
number and seed number, as is reflected by the positive correlation between the
shoot dry weight, leaf area, pod number and seed number per plant. This could
be affected by the absence of plant to plant competition.
A similar trend was revealed in the field experiment, many of the characters were
positively correlated to pod number and seed number, except days to flowering,
number of stems, shoot dry weight, leaf numbers and days to maturity, which
125
suggests that the effect of days to flowering and days to maturity can be
detrimental on pod number and seed number produced. This is mainly because
both the leaf numbers, number of stems and shoot dry weight of the crop is
reduced. The late flowering landraces had less time to develop pod/ flowers. Seed
weight and pod dry weight showed a high correlation to pod number with (r =
+0.84) and (r = +0.83), respectively. A lower correlation was observed between
shoot dry weight and pod number at (r = +0.17) and leaf area and pod number at
(r = +0.25), but there was a moderate relatively high correlation between shoot
dry weight and leaf area at (r = +0.68), while the correlation between seed number
and pod number was a perfect correlation (r = +1.0) implying a fixed number of
seeds per pod.
126
Table 4.1.9: Correlation coefficients among the 24 traits based on the 35 bambara groundnut planted in the Agronomy bay (UK), traits were
measured 10 weeks after planting.
Characters
DAE
DAF
LEAFN
SPREADL
LEAFL
LEAFW
LEAFAREA
HEIGHT
INTERNODE
PETIOLE
PETINTERN
PETIOLOULE
DAE
1.00
DAF
0.39*
1.00
LEAFN
-0.25
0.04
1.00
SPREADL
-0.31
-0.17
0.44**
LEAFL
0.20
0.21
-0.16
0.17
1.00
LEAFW
0.05
0.07
0.22
0.46**
0.42*
1.00
LEAFAREA
-0.17
0.09
0.92**
0.56**
0.13
0.48**
1.00
HEIGHT
-0.17
0.04
0.49**
0.59**
0.44**
0.50**
0.67**
1.00
INTERNODE
-0.16
-0.08
0.27
0.80**
0.25
0.43*
0.42*
0.68**
1.00
PETIOLE
-0.12
0.08
0.53**
0.58**
0.36*
0.52**
0.71**
0.91**
0.63**
1.00
PETINTERN
0.19
0.09
-0.14
-0.63**
0.14
-0.19
-0.16
-0.11
-0.69**
-0.07
1.00
PETIOLOULE
0.00
0.13
0.48**
0.15
0.23
0.29
0.57**
0.43*
0.12
0.46**
0.20
1.00
1.00
PEDUNCLE
0.04
0.07
-0.22
0.30
0.13
0.05
-0.12
0.09
0.41*
-0.06
-0.48**
-0.26
STEMSNO
-0.23
-0.29
0.49**
0.37*
0.10
0.21
0.50**
0.50**
0.46**
0.37*
-0.20
0.31
DAM
-0.28
0.09
0.23
0.50**
0.14
0.36*
0.29
0.48**
0.53**
0.41*
-0.42*
0.08
SDW
-0.19
0.06
0.84**
0.67**
0.19
0.52**
0.94**
0.70**
0.54**
0.75**
-0.25
0.47**
PODNO
-0.38*
-0.25
0.57**
0.60**
-0.10
0.31
0.52**
0.47**
0.48**
0.46**
-0.31
0.21
PDW
0.07
-0.22
-0.22
0.02
0.13
0.16
-0.19
-0.04
0.07
-0.07
-0.13
-0.09
PODL
0.08
-0.03
-0.08
-0.17
0.04
-0.06
-0.06
0.07
0.07
-0.04
-0.01
0.11
PODW
-0.22
0.08
-0.01
0.25
0.16
0.35*
0.12
0.31
0.34*
0.22
-0.18
0.03
SEEDNO
-0.33
-0.29
0.51**
0.55**
-0.14
0.34*
0.47**
0.42*
0.46**
0.40*
-0.31
0.23
SEEDL
-0.26
-0.21
0.03
0.39*
0.24
0.39*
0.14
0.43**
0.54**
0.28
-0.26
0.08
SEEDW
-0.29
-0.17
-0.09
0.24
0.11
0.27
0.03
0.29
0.27
0.14
-0.09
0.06
-0.35* -0.43**
0.28
0.47**
-0.05
0.32
SEEDWEIGHT
* Correlation is significant at the 0.05 level; **Correlation is significant at the 0.01 level
0.29
0.39*
0.44**
0.31
-0.24
0.22
127
Table 4.1.9 (Continued)
Characters
PEDUNCLE
STEMSNO
DAM
SDW
PODNO
PDW
PODL
PEDUNCLE
1.00
STEMSNO
0.10
1.00
DAM
0.19
0.21
1.00
SDW
-0.02
0.48**
0.45**
1.00
PODNO
-0.19
0.59**
0.54**
0.61**
1.00
PDW
0.10
0.06
0.31
-0.06
0.13
1.00
PODL
0.33
0.08
0.07
-0.05
-0.25
0.26
1.00
PODW
0.36*
0.08
0.43*
0.31
0.08
0.15
0.46**
PODW
SEEDNO
SEEDL
SEEDW
1.00
-0.21
0.58**
0.51**
0.53**
0.98**
0.13
-0.21
0.05
1.00
SEEDL
0.49**
0.31
0.43*
0.28
0.21
0.26
0.61**
0.65**
0.21
1.00
SEEDW
0.32
0.24
0.34*
0.15
0.16
0.10
0.362*
0.64**
0.18
0.71**
1.00
0.32
0.03
0.27
0.88**
0.45**
0.421*
SEEDNO
-0.06
0.58**
0.50**
0.40*
0.86**
SEEDWEIGHT
* Correlation is significant at the 0.05 level; **Correlation is significant at the 0.01 level
DAE: days to emergence; DAF: days to 50% flowering;
LEAFN: Leaf number per plant;
PETINTERN =Petiole-Internode ratio, STEMSNO =Number of stems per plant
SEEDWEIGHT
SPREADL: Canopy size; LEAFW = Leaflet width
DAM=Days to maturity, SDW = shoot dry weight
1.00
HEIGHT =plant height
PODNO = pods number per plant
PDW =Pod dry weight SEEDNO = seeds number per plant SEEDW = seed width
128
Table 4.2.1: Correlation coefficient for 24 quantitative traits of the 34 bambara groundnut planted in the field experiment in (Botswana) traits
were measured 10 weeks after planting.
Characters
DAE
DAF
LEAFN
SPRDIN
LEAFW
LEAFL
LEAFAREA
HEIGHT
INTERNODE
PETIOLE
PETINTER
PETIOLULE
DAE
1.00
DAF
0.12
1.00
LEAFNO
-0.25
0.01
1.00
SPRDIN
-0.44*
-0.18
0.17
1.00
LEAFW
-0.44*
-0.05
-0.01
0.40*
1.00
LEAFL
-0.31
0.00
-0.18
0.40*
0.62**
1.00
LEAFAREA
-0.48**
-0.04
0.68**
0.46**
0.62**
0.52**
1.00
HEIGHT
-0.35*
-0.04
-0.04
0.31
0.60**
0.66**
0.48**
1.00
INTERNODE
-0.21
-0.34
0.04
0.69**
0.37*
0.37*
0.37*
0.38*
1.00
PETIOLE
-0.50*
-0.08
0.24
0.45**
0.58**
.054**
0.62**
0.82**
0.35*
1.00
PETINTER
-0.12
0.25
0.13
-0.35*
0.00
-0.10
0.02
0.08
-0.72**
0.30
1.00
PETIOLULE
-0.42*
0.02
0.25
0.34
0.24
0.35*
0.43*
0.55**
0.29
0.64**
0.17
1.00
PEDUNCLE
-0.05
-0.24
0.27
0.47**
0.23
0.16
0.424*
0.34*
0.53**
0.25
-0.43*
0.33
STEMSNO
-0.49**
0.00
-0.02
0.25
0.28
0.31
0.22
0.20
0.18
0.30
-0.03
0.15
DAM
-0.25
0.30
0.19
0.28
0.34
-0.11
0.24
0.08
0.02
0.13
0.09
0.19
SDW
-0.34*
0.10
0.52**
0.30
0.29
0.24
0.65**
0.43*
0.28
0.56**
0.10
0.53**
PODNO
0.03
0.16
-0.02
0.03
0.28
0.16
0.21
0.01
0.00
0.14
0.13
0.06
PDW
0.21
0.14
-0.23
0.11
0.16
0.21
0.01
-0.04
0.12
-0.12
-0.10
-0.10
PODL
0.19
0.32
-0.16
-0.12
-0.12
0.23
-0.08
0.00
-0.24
-0.18
0.04
-0.04
PODW
0.37
-0.05
-0.21
-0.30
-0.40*
-0.03
-0.34
-0.15
-.38*
-0.42*
0.08
-0.35
SEEDSNO
-0.03
0.10
-0.07
0.08
0.25
0.22
0.19
0.05
0.04
0.18
0.14
0.06
SEEDL
-0.05
0.03
0.10
0.04
0.04
0.31
0.30
0.23
-0.07
0.07
0.05
0.13
SEEDW
-0.06
-0.27
-0.09
0.07
-0.17
0.03
-0.05
0.01
-0.11
-0.16
-0.06
0.04
0.06
-0.17
-0.27
0.389*
0.28
0.33
SEEDWEIGHT
* Correlation is significant at the 0.05 level; **Correlation is significant at the 0.01 level
0.07
0.05
0.43*
0.08
-0.30
-0.02
129
Table 4.2.1: (Continued)
Characters
PEDUNCLE
STEMSNO
DAM
SDW
PODNO
PDW
PODL
PODW
SEEDSNO
SEEDL
SEEDW
PEDUNCLE
1.00
STEMSNO
-0.03
1.00
DAM
0.07
0.14
1.00
SDW
0.34
0.28
0.37
PODNO
0.15
-0.02
-0.25
0.13
1.00
PDW
0.20
-0.11
-0.18
-0.05
0.74**
1.00
PODL
0.01
0.05
-0.06
0.07
0.01
0.27
1.00
PODW
-0.05
-0.16
-0.20
-0.24
0.08
0.46*
0.41*
1.00
SEEDSNO
0.14
-0.02
-0.25
0.16
0.96**
0.80**
0.09
0.18
1.00
SEEDL
0.18
0.01
-0.11
0.14
0.19
0.25
0.52**
0.27
0.20
1.00
SEEDW
0.24
-0.16
-0.08
-0.13
0.20
0.33
0.26
0.44*
0.21
0.68**
1.00
-0.04
0.14
0.71**
0.07
0.14
SEEDWEIGHT
1.00
0.35
-0.01
-0.29
-0.14
0.63** 0.75**
SEEDWEIGHT
* Correlation is significant at the 0.05 level; **Correlation is significant at the 0.01 level
DAE: days to emergence; DAF: days to 50% flowering;
LEAFN: Leaf number per plant;
PETINTERN =Petiole-Internode ratio STEMSNO =Number of stems per plant
SPREDL: Canopy size; LEAFW = Leaflet width
DAM=Days to maturity, SDW = shoot dry weight
1.00
HEIGHT =plant height
PODNO = pods number per plant
PDW =pod dry weight SEEDNO = seeds number per plant SEEDW = seed width
130
4.3.7
Quantitative variance analysis
Bambara groundnut improvement is not only dependent on the magnitude of
phenotypic variation of the crop, but also on the extent of how the traits are
heritable. Therefore it is important to quantify the heritable and non-heritable
component from the phenotypic variation observed. Assessing the genotypic
coefficient of variation (GCV), heritability and genetic advance (as percentage of
the mean) at the same time gives a good estimation of the amount of advance
expected in selection (Baye, 2002). Since one objective of this study is to identify
the best performing lines, it is important the genotypic coefficient of variation
(GCV), phenotypic coefficient of variation (PCV), heritability, and genetic
advance (as percentage of the mean) are estimated in the selected genotypes.
The estimates of phenotypic coefficient of variability (PCV) and genotypic
coefficient of variability (GCV), heritability (in a broad sense), and genetic
advance as a percentage of the mean were analysed (Table 4.2.2) for the
glasshouse experiment (Table 4.2.3) for the field experiment in Botswana. The
highest phenotypic coefficient of variation (PCV) in the glasshouse experiment
was observed in the number of pods per plant at 51.73% and the lowest was on
the days to maturity at 1.13%. In the field experiment, the range of the phenotypic
coefficient of variation ranged from a high for number of pods per plant at
82.17% to a low for days to maturity at 5.6%. The phenotypic coefficient of
variation was relatively high in the agronomy bay experiment for shoot dry
weight at 55.1%, leaf area at 54.4% compared to 45.2% for shoot dry weight in
the field and 31.5% for leaf area in the field. But for the seed numbers phenotypic
coefficient of variation was much higher in the field at 79.0% compared with
51.5% in the agronomy bay experiment.
131
Table 4.2.2: Quantitative variances based on phenotypic coefficient of variability
(PCV), genotypic coefficient of variability (GCV), broad sense heritability (h2B)
and genetic advance (GA) in the 35 landraces in the agronomy bay (UK).
Traits
MSG
MSE
PCV%
GCV%
h2B
GA %of
mean
Days to emergence
4.18
2.57
7.23
11.65
0.39
10.3
Days to 50% flowering
33.79
8.93
6.78
7.91
0.74
8.7
Number leaves per plant
4758
826.2
45.59
50.16
0.83
50.3
Canopy spread
181.02
33.8
35.85
39.75
0.81
40.6
Leaflet length
1.28
0.311
6.61
7.60
0.76
8.2
leaflet width
0.77
0.14
12.56
13.88
0.82
14.1
Leaf area
17918884
4836578
46.52
54.44
0.73
60.2
Plant height
56.41
7.79
12.02
12.95
0.86
12.1
Internode length
1.77
0.193
31.52
33.40
0.89
28.8
Petiole length
31.66
3.23
18.33
19.35
0.90
16.1
Petiole-Internode ratio
13.98
3.16
23.71
26.95
0.77
28.9
Petiolule length
1.028
0.167
23.50
25.67
0.84
25.0
Peduncle length
0.742
0.321
16.08
21.34
0.57
23.2
Number of stems
17.8
9.45
16.90
24.68
0.47
24.6
Days to maturity
94.48
85.24
1.13
3.61
0.10
1.0
Shoot dry weight
1011.1
197
49.47
55.13
0.81
57.1
Number pod per plant
2960.5
584.9
51.73
57.75
0.80
60.0
Pod dry weight
778.7
215.4
41.90
49.27
0.72
54.6
Pod length
8.6
3.35
7.03
9.00
0.61
10.0
Pod width
4.39
1.24
8.49
10.02
0.72
11.2
Number of seeds per plant
2815.3
508.6
46.92
51.83
0.82
52.6
Seed weight
449.8
110.8
46.0
53.0
0.75
57.8
Seed length
2.94
1.34
6.84
9.27
0.54
10.0
Seed width
1.49
0.72
6.05
8.42
0.52
8.8
Mean square genotype (MSG): estimates genotypic variance, this value is observed variance
among the line means, while mean square error (MSE) measures variance from plot residuals.
Mean Square Error (MSE) are variance components estimated as functions of the means square
estimates from ANOVA table.
Genetic advance are estimates percentage based on the mean
132
Table 4.2.3: Quantitative variances based on phenotypic coefficient of variability
(PCV), genotypic coefficient of variability (GCV), broad sense heritability (h2B)
and genetic advance (GA) in the 34 lines (Field experiment).
Traits
MSG
MSE
PCV%
GCV%
h2B
GA %of
mean
Days to emergence
7.36
3.65
10.36
7.36
0.50
11.41
Days to 50% flowering
137.15
10.24
11.89
11.44
0.93
9.32
Number leaves per plant
3045.1
552.2
28.91
26.16
0.82
31.26
Canopy spread
19641
1335
33.74
32.57
0.93
25.42
Leaflet length
234.57
57.7
13.19
11.46
0.75
15.27
leaflet width
35.74
6.40
13.81
12.51
0.82
14.87
Leaf area
3240130
972383
26.31
31.45
0.70
37.84
Plant height
3133.3
652.5
11.62
10.34
0.79
13.10
Internode length
208.74
53.23
37.22
32.12
0.74
43.35
Petiole length
1504.4
316.4
16.09
14.30
0.79
18.06
Petiole-Internode ratio
9.8
1.83
25.08
22.62
0.81
27.34
Petiolule length
54.49
13.32
23.95
20.82
0.76
27.75
Peduncle length
86.19
30.83
27.08
21.70
0.64
32.25
Number of stems
11.8
3.66
23.58
19.58
0.69
28.21
Days to maturity
176.39
9.57
5.60
5.45
0.95
3.63
Shoot dry weight
1067.6
117.5
46.38
43.75
0.89
42.46
Number pod per plant
511.4
237.6
82.17
60.12
0.54
88.15
Pod dry weight
97.89
69.97
68.65
36.66
0.29
48.46
Pod length
25.98
4.57
16.06
14.58
0.82
16.32
Pod width
3.94
1.75
10.84
8.08
0.56
11.82
Number of seeds per plant
453.3
225
79.49
56.41
0.50
82.86
Seed weight
43.49
26.16
72.52
45.78
0.40
66.01
Seed length
4.62
2.07
11.36
8.44
0.55
12.34
Seed width
2.22
1.16
10.61
7.33
0.48 10.68
Mean square genotype (MSG): estimates genotypic variance, this value is observed variance
among the line means, while mean square error (MSE) measures variance from plot residuals.
Mean Square Error (MSE) are variance components estimated as functions of the means square
estimates from ANOVA table.
Genetic advance are estimates percentage based on the mean
The genotypic coefficient of variation in the agronomy bay experiment ranged
from number of pods per plant from 57.75% to 3.61% for days to maturity, while
in the field experiment the same traits revealed lowest genotypic coefficient of
variation for days to maturity (5.45%) andhighest genotypic coefficient of
variation60.12%. High genotypic coefficient of variation values as shown by pods
per plant, number of seeds per plant, seed weight in both sites (UK and in
133
Botswana) indicates that these traits could lead to good progress in crop
improvement. Lower genotypic coefficient of variation shown by, days to
flowering, days to maturity, seed width that are lower in both sites indicate that
these traits are less amenable to improvement through selection. Relatively low
values of both phenotypic coefficient of variation and genotypic coefficient of
variation were in agronomy bay experiment than in the field experiment.
To quantify the amount that is heritable for the 24 characters, broad sense
heritability was estimated in (Table 4.2.2 and Table 4.2.3). In the agronomy bay
experiment estimates of broad sense heritability ranged from 0.9 for petiole length
to 0.1 for days to maturity, most of the characters have a heritability of more than
0.7. Whereas, in the field experiment the estimates of broad sense heritability was
highest for days to maturity at 0.95 and lower for days to pod dry weight 0.29,
with most of the characters showing more than 0.7 heritability. There was higher
broad sense heritability in agronomy bay experiment for pod number per plant
and seed number per plant 0.8 and 0.82 compared to 0.54 and 0.50 respectively
for both characters in the field experiment. While the shoot dry weight increased
from 0.81 in the agronomy bay to 0.89 in the field experiment and the leaf area
was the same at 0.7 for both sites. The higher estimates of heritability in
agronomy bay experiment and field experiment for number of leaves, plant
spread, leaflet width, and plant height shows that these traits may not have been
affected by environment.
Genetic advance (GA) expected when selecting at 5% based on the percentage of
the mean reached a maximum of 60.2% for internode length and a minimum of
1% for days to maturity in the agronomy bay experiment, while for the field
experiment it ranged from 88.2% for number of pods per plant to 3.6 % for days
to emergence. Other characters that showed relatively high estimates of genetic
advancewere number of seeds per plant (54.6%), number of leaves per plant
(57.1%), petiolule (57.8 %) and plant height (60%) respectively, in the agronomy
bay. The characters that showed higher genetic advance in the field experiment
were internode length (43.3%), pod dry weight (48.5%), seed weight (66%) and
number of seeds per plant (82.9%). Lower genetic advance recorded in the field
were for days to 50% flowering (9.3%), seed width (10.7%), days to emergence
134
(11.4%) and pod width (11.8%), respectively, which suggest that these traits could
be difficult to select for in semi-arid Botswana environment.
A number of characters such as shoot dry weight, number of pods per plant,
number of leaves per plant, and plant spread have consistently shown higher
phenotypic coefficients of variation, genotypic coefficients of variation and
estimates of broad sense heritability in both experimental sites, which suggest
that these characters could be useful for selection in bambara groundnut. In this
study, pod number, seed number, leaf area and shoot dry weight (biomass) were
selected as the basis for multiple selection in bambara groundnut.
4.3.8
Comparison of agronomy bay and field experiment
Bambara groundnut evaluation and characterisation was conducted at two
experimental sites one in UK and one in Botswana, with different environmental
conditions. The performances of the genotypes in the two sites were compared
especially since the UK materials are used as the initial selection for breeding
purpose.
A regression analysis was conducted between the agronomy bay and field
experiment, a linear relation between the two experiments is shown on figure
4.2.4. The relationship between the two experiments is highly significant (P
<0.001), which suggest that, the agronomy bay experiment could be useful in the
selecting of materials to plant in the field (Table 4.2.4).
135
Table 4.2.4: A summary of analysis for the relationship between the agronomy
(UK) experiment and the field experiment in (Botswana), computed on Genstat
version 13.0
Source
Regression
Residual
Total
d.f.
1
22
23
s.s.
58646
6150
64796
m.s.
v.r.
58646.1 209.81
279.5
2817.2
F pr.
<.001
Percentage variance accounted for 90.1
Standard error of observations is estimated to be 16.7.
Fitted and observed relationship with 95% confidence limits
250
200
150
100
50
0
0
50
100
150
200
250
Field
Figure 4.2.4: shows a regression analysis plot of mean over all the genotypes for
the 24 variables recorded from agronomy bay (UK) and field experiment in
Botswana
The data from the two sites for the 24 characters, show that they are well
correlated (Figure 4.2.4).
136
4.3.8
Selection for breeding bambara groundnut
To identify lines which have a potential to produce higher yields in a Botswanan
environment based on four selected characters, the selection index was used with
a weighting of the genetic advance found in the field experiment: A linear
equation for SI = (X1 x 0.378) + (X2 x 0.424) + (X3+0.828) + (X4 x 0.881) was
derived using the genetic advance from the fields study (Table 4.2.6). This index
put more emphasis on yield in a multiple trait selection. X1 = Leaf area, X2
=Shoot dry weight, X3 =Seed number per plant, X4 = Pod number plant.
Selection index (SI) analysis produced single values for each genotype. These
values were then ranked for the agronomy bay and field experiment respectively.
The Duncan multiple range test (DMRT) was also performed to identify
genotypes with different characters. Table 4.2.8 and Table 4.2.9 show the ranking
of the genotypes using the selection index (SI) and (DMRT).
137
Table 4.2.5: The Duncan multiple range tests and the selection index of bambara
groundnut based on the vegetative and yield characters (Agronomy bay, UK).
Leaf Area
DMRT
SDW
DMRT
PODS
DMRT
SEEDS
DMRT
SI
RANK
3Acc9NGA
Landraces
2956
Abcd
17.2
Ab
36
Abcd
54
abcdefg
-1.33
25
4Acc144NGA
2953
Abcd
30.2
Abcd
46
Abcdef
50
abcdef
-0.8
18
6Acc289BEN
2587
Abcd
10.8
A
17
A
25
Abc
-2.47
34
10Acc1276CAF
4316
Abcd
35.8
Abcd
46
Abcdef
52
abcdef
-0.25
15
20Acc118BFA
2512
Abcd
17.2
Ab
34
Abcd
43
abcde
-1.66
26
30Acc476CMR
3391
Abcd
24.2
Abc
54
Abcdef
60
bcdefgh
-0.56
16
33Acc484CMR
3470
Abcd
23.4
Abc
50
Abcdef
56
abcdefg
-0.71
17
40Acc563CMR
4896
Abcd
33.5
Abcd
20
Ab
25
Abc
-1.13
23
45Acc231GHA
3318
Abcd
35.5
Abcd
40
Abcde
33
Abcd
-0.93
20
48Acc790KEN
10369
F
68.3
Fgh
106
Gh
105
Ij
4.3
1
49Acc793KEN
2281
Abc
20.1
Abc
52
Abcdef
55
abcdefg
-1.08
22
50Acc792ZWE
1887
Ab
14.6
Ab
33
Abcd
38
Abcd
-1.99
30
56Acc89MLI
1357
A
9.9
A
28
Ab
40
Abcd
-2.33
33
60Acc32NGA
2613
Abcd
20.2
Abc
22
Ab
30
Abc
-2
31
69Acc286NGA
2953
Abcd
13.2
A
10
A
15
Ab
-2.62
35
70Acc329NGA
4049
Abcd
28.9
Abcd
11
A
12
A
-1.89
28
74Acc335NGA
2944
Abcd
22.6
Abc
20
Ab
26
Abc
-1.93
29
76Acc390SDN
1955
Ab
11.2
A
30
Ab
39
Abcd
-2.13
32
81Acc385T ZA
6713
Cdef
42.7
Bcdef
86
Efgh
102
Hij
2.19
8
84Acc696ZMB
9939
Ef
54.1
Defg
35
Abcd
50
abcdef
1.45
10
85Acc754ZMB
9997
Ef
76.3
Gh
106
Gh
94
Fghij
4.27
2
88-AHM753NAM
4347
Abcd
33.7
Abcd
121
H
133
J
2.53
5
90-S19-3NAM
3364
Abcd
21.5
Abc
48
Abcdef
48
abcde
-0.95
21
91-UNISRSWA
6059
Bcde
53.2
Defg
105
Gh
107
Ij
2.79
3
92-AHM968NAM
4103
Abcd
35.7
Abcd
87
Efgh
96
Ghij
1.24
12
95-DODRT ZA
3822
Abcd
34.2
Abcd
79
Cdefgh
86
efghi
0.79
14
99-SB4-2NAM
6500
Cdef
46.8
Cdef
100
Gh
105
Ij
2.58
4
100-SB16ANAM
6429
Cdef
48.1
Cdef
67
Bcdefg
64
cdefghi
1.26
11
104-S-1913NAM
4240
Abcd
36.8
Abcde
92
Fgh
103
Ij
1.52
9
105-MaheneneBlackN
6794
Def
62.5
Efgh
81
Defgh
76
defghi
2.29
7
109-BOT S1
2393
Abcd
20.7
Abc
47
Abcdef
53
abcdef
-1.15
24
113-BOT S5
4557
Abcd
31.7
Abcd
84
Efgh
77
defghi
0.85
13
117-VSSP6CMR
2704
Abcd
20.5
Abc
31
Abc
30
Abc
-1.77
27
118-Ramayana-IND
9583
Ef
80.5
H
48
Abcdef
50
abcdef
2.45
6
119-Hybrid
4751
Abcd
29.7
Abcd
31
Abc
40
Abcd
-0.82
19
Means with similar letters in column are not significantly different at 5% Duncan Multiple Range test.
DMRT (Duncan Multiple Range Test)
138
Table 4.2.6: The Duncan multiple range tests and the selection index of bambara
groundnut based on the vegetative and yieldcharacters (field experiment in
Botswana).
Lines
Leaf Area
DMRT
3Acc9NGA
4329
Fghi
4Acc144NGA
2878
Bcdefg
Shoot
DMRT
Seeds No
DMRT
Pod No
DMRT
SX
RANK
35.7
Bcdefghi
7
A
7
Ab
-0.91
24
31
Abcdefghi
9
A
8
Ab
-1.31
29
6Acc289BEN
3391
Bcdefg
28.6
Abcdefgh
7
A
9
Ab
-1.24
27
10Acc1276CAF
2157
Abcde
44.8
Efghij
8
A
8
Ab
-1.32
30
20Acc118BFA
3988
Defgh
21.5
Abc
12
A
11
Ab
-0.72
20
17
30Acc476CMR
2861
Bcdefg
37.9
Cdefghi
14
A
13
Ab
-0.49
33Acc484CMR
2811
Bcdefg
27.1
Abcdefg
22
A
23
Ab
0.47
9
40Acc563CMR
3614
Defgh
45.4
Efghij
*
A
*
Ab
0.21
13
45Acc231GHA
1446
Ab
18.7
Abc
*
A
*
Ab
-1.08
26
48Acc790KEN
4586
Ghi
61.4
Jk
11
A
11
Ab
0.25
12
50Acc792ZWE
3059
Bcdefg
48.2
Ghij
*
A
*
Ab
0.09
14
56Acc89MLI
1614
Abc
21.9
Abc
6
A
6
Ab
-2.26
34
60Acc32NGA
2583
Bcdef
38.6
Cdefghi
13
A
12
Ab
-0.69
19
69Acc286NGA
4319
Fghi
16
Ab
*
A
*
Ab
-0.17
15
70Acc329NGA
387
A
13.7
A
*
A
*
Ab
-1.54
33
74Acc335NGA
3786
Defgh
23.4
Abcd
10
A
11
Ab
-0.9
23
76Acc390SDN
5986
I
82.9
L
12
A
10
Ab
1.19
6
81Acc385T ZA
5437
Hi
102.9
M
8
A
10
Ab
1.17
7
84Acc696ZMB
4343
Fghi
50.8
Ij
33
A
36
B
3.12
2
85Acc754ZMB
4098
Efgh
72.3
Kl
20
A
17
Ab
1.33
4
88-AHM753NAM
3879
Defgh
47.2
Fghij
31
A
35
Ab
2.68
3
90-S19-3NAM
4019
Defgh
48.3
Ghij
66
B
68
C
7.33
1
91-UNISRSWA
2442
Bcdef
36.6
Bcdefghi
9
A
9
Ab
-1.26
28
92-AHM968NAM
2062
Abcd
24.7
Abcde
18
A
21
Ab
-0.23
16
95-DODRT ZA
3152
Bcdefg
38
Cdefghi
5
A
5
A
-1.53
32
99-SB4-2NAM
3256
Bcdefg
29.4
Abcdefghi
21
A
21
Ab
0.46
10
100-SB16ANAM
3428
Cdefg
45.1
Efghij
18
A
19
Ab
0.53
8
104-S-1913NAM
2715
Bcdefg
21
Abc
22
A
23
Ab
0.31
11
22
105-MaheneneBlack
3306
Bcdefg
44.3
Defghij
9
A
9
Ab
-0.81
109-BOT S1
3982
Defgh
69.9
Kl
18
A
19
Ab
1.25
5
113-BOT S5
3865
Defgh
49.8
Hij
5
A
5
Ab
-1.04
25
117-VSSP6CMR
2252
Bcde
26.6
Abcdef
*
A
*
Ab
-0.63
18
118-Ramayana-IND
2609
Bcdefg
31.8
Abcdefghi
7
A
8
Ab
-1.51
31
119-Hybrid
2383
Bcdef
46.3
Fghij
12
A
11
Ab
-0.75
21
Means with similar letters in a column are not significantly different at (5%) Duncan Multiple Range test.
DMRT (Duncan Multiple Range Test)
The selection index (SI) as revealed by the ranking was able to identify the best
performing landraces in the agronomy bay experiment as, (1) 48-Acc790 from
Kenya, (2) 85-Acc754, (3) 91-UNISR from Swaziland, (4) 99-SB4-2 from
Namibia, (5) 88-AHM753 from Namibia, (6) 118-Ramayana from Indonesia, (7)
139
105-Mahenene black,
(8) 81-Acc385 from Tanzania,
(9) 104-S-1913 from
Namibia and (10) 84Acc696 from Zambia.
In the field experiment (Botswana), (1) 90-S19-3 from Namibia, (2) 84-Acc 696
from Zambia, (3) 88-AHM753, (4) 85-Acc754 from Zambia, (5) 109-BOTS1
from Botswana, (6) 76-Acc390 from Sudan, (7) 81-Acc385 from Tanzania, (8)
100SB16A from Namibia, (9) Acc3348 from Cameroon, and (10) 99-SB4-2 from
Namibia. Interestingly genotypes from Namibia, from a drier environment
performed well in the agronomy bay environment and similarly produced more
yield in the Botswanan environment.
At 5 % selection using Duncan multiple range test, in the agronomy bay
experiment no particular landrace stood outas revealed by the selection index, on
all traits. In the field experiment it was line 90-S19-3 from Namibia which
surpassed the rest on number of pods per plant and number of seeds per plant.
4.4
Discussion
Qualitative characters observed
Qualitative characters showed a substantial amount of variability in growth habit,
leaflet shape and in pod and seed characters. Pod texture, pod shape, pod colour,
testa colour and test pattern showed considerable variation while eye pattern and
stress susceptibility had low variation. The predominant testa colourwasthe
reddish colour, which implies that these are the ones farmers have selected for.
All the three plant growth habits were identified among the selected landraces,
which reflect the cropping system of bambara groundnut by farmers. The
spreading type landraces are useful in mixed cropping with other crops such as
cereals while the semi bunch and the bunch are good for monoculture.
Descriptive characters observed
The phenotypic ranges of various characters found in this study are in line with
findings by other researchers when evaluating bambara groundnut (Goli et al.,
(1995); Karikari and Tabona, (2004); Ntundu et al., (2006); Ouedraogo et al.,
140
(2008). Generally, characters which showed a greater range of difference were
observed for number of pods per plant, seed weight, and seed number. The broad
range of differences among these traits shows that there is a good possibility for
crop improvement.
Knowledge on the genetic variation, heritability and correlation between bambara
groundnut agro-morphological characters is important to initiate a feasible
breeding program. In this study considerable variation in all the characters was
observed, except for pod dry weight and days to maturity. Adeniji et al., (2008)
found significant differences for 11 characters among the 18 traits they measured
in their study for 10 bambara groundnut accessions to evaluate the interrelationship for pod and seed yield characters sourced from North-Eastern
Nigeria. Similarly, Ntundu et al., (2006) observed highly significant
morphological variation for 13 characters on 27 of the characters they measured
for the two seasons in Tanzania when they were studying the morphological
diversity among 100 Tanzania bambara groundnut accessions.
The application of principal component analysis
Principal component analysis was performed to further find out which characters
are important in explaining the variation among the selected genotypes. In the
agronomy bay experiment the first two principal components (PC 1 and PC 2)
separate mainly on the basis of vegetative traits such as the spreading length,
shoot dry weight, petiole length, petiolule length, leaf area and pod number per
plant. Shoot dry weight and petiole length had higher loadings in both (PC1 and
PC2). In the field experiment overall variability among the lines came primarily
from vegetative traits in PC 1 such as petiole length, leaf area, petiolule length,
shoot dry weight, plant height and spreading length, while PC 2 higher loadings
were mainly pod and seed characters pod dry weight, seed weight, seed number
per plant, pod number per plant, and pod width. This indicates the importance of
these characters in identifying bambara groundnut landraces. Ntundu et al., (2006)
observed similar patterns of loading in their study on 100 landraces in Tanzania,
whereby the high loading within principal component one was mainly due to
vegetative characters while the second was mainly seed characters.
141
Clusters analysis
Cluster analysis revealed that bambara groundnut is phenotypically distinct across
the regions of Africa and can easily be separated, based on the important
characters, in this case pod number per plant and shoot dry weight differences.
The degree of genetic differentiation in a crop is dependent on its breeding
systems, life history and geographical distribution (Roy, 2000). In bambara
groundnut the influence of rainfall and its distribution pattern on the productivity
was reported (Azam-Ali et al., 2001), while Collinson et al., (1999) identified
moisture as one important factor in the biomass production on bambara
groundnut. Sesay etal., (2008) observed that in the sub-tropical regions of Africa
sowing date had an impact on the final yield of bambara groundnut mainly due to
the effects of temperature and varying day-length.
The formation of clear clusters in bambara groundnut (figure 4.2.2 and figure
4.2.3) was expected, as bambara groundnut is an in-breeder it should reveal
greater inter-population diversity and less intra-population. Cui et al., (2001) used
phenotypic traits and clearly separated Chinese and North American soybean
cultivars, while Upadhyaya, (2003) used 16 morphological and 32 agronomic
traits in groundnut to differentiate subspecies, fastiga (var. fastiga,vulgaris,
aequatorian, peruviana) and subsp. Hypogaea (var. hypogaea, hirsuta).
However, there were some overlaps noticed from various regions, especially
between Southern Africa and East African lines, as revealed in both studies in the
agronomy bay and field experiment. Confirming the importance of seed sources in
bambara groundnut (Brink et al., 1996; Massawe et al., 2005)
Correlations coefficient
Correlations of greater than approximately (r =
+0.7) or less than (r = -0.7) are
the ones likely to be of biological importance (Hill et al., 1998). Such correlations
were found between, petiole length and plant height (r = +0.91), shoot dry weight
and plant height (r = +0.70), leaf area and petiole length (r = +0.71) for the
agronomy bay experiment. While for the field experiment correlation coefficients
of r = +0.7 or greater were recorded between; pod length and days to maturity (r =
0.93), leaf area and leaf number (r = +0.71) and plant height and petiole length (r
= +0.82).
142
Yield related traits recorded higher positive correlations; seed weight and pod dry
weight were highly correlated to pod number with r = +0.84 and r = +0.83
respectively, while seed number per plant was highly correlated with pod number
at r =+0.98 in the agronomy bay and r = +1.0 in the field experiment in Botswana.
This could be possibly due to sick plant producing few pods which are shrivelled,
and of low weight. Similar findings were found in bambara groundnut by Ofori
(1996), Karikari (2000), Karikari and Tabona (2004), Jonah et al., (2010) and
Onwubiko et al., (2011), which suggest that these characters can be selected for in
order to improve bambara groundnut yield.
The lower correlations of leaf area,shoot dry weight to both seed number and pod
number reveals that most of the lines which are not adapted to the Botswana
environment concentrated their growth into vegetative growth rather than in pod
and seed production, some high biomass producing lines are not necessarily those
that produced more seeds and pods. For example lines 76-Acc390 from Sudan,
81-Acc385 from Tanzania, 48-Acc790 from Kenya, 84-Acc696 from Zambia, 3Acc9, 69-Acc286 from Nigeria and 85-Acc754 from Zambia produced high leaf
area and shoot dry weight (biomass) more so than 90-S9-3 from Namibia which
produced the highest number of pods and seeds. However, this still reflects the
importance of leaf area and shoots biomass production in the final yield of
bambara groundnut and the different adaptation of landraces to their original
climatic environment (Mwaleet al., 2007).
Phenotypic coefficient of variation and genotypic coefficient of variation
Among the characters assessed both in the agronomy bay experiment and in the
field experiment in Botswana, the number of pods per plant showed a high
phenotypic coefficient of variation and genotypic coefficient of variation, while
broad sense heritability was relatively low. The phenotypic coefficient of variation
for pod number per plant ranged from 51.7% in the agronomy bay experiment to
82.1% in the field experiment, while the genotypic coefficient of variation ranged
from 57.8 % from the agronomy bay experiment to 60.1% in the field experiment.
Wigglesworth (1996) found a phenotypic coefficient of variation of 20.2 %, a
genotypic coefficient of variation of 12.6 % and a relatively lower heritability of
0.39 in pod number and lower phenotypic coefficient of variability and genotypic
143
coefficient of variability of 100 seed weight at 18 % and 17.7% , but higher
heritability of 0.94 when he assessed 10 landraces, six from Botswana, one from
Zimbabwe, and three as checks from West Africa (Mali, The Gambia , Niger) for
their
potential under irrigation using sewage water systems in Botswana
environment. In contrast, Karikari, (2000) reported heritability of bambara grain
yield at 0.71, 100 seed weight at 0.25, and shoot dry weightat0.36 when assessing
the adaptability of local and exotic landraces in a Botswana environment. Their
results suggest that selecting for 100 seed weight in Botswana would lead to a
yield increase as compared to selecting for lower heritable traits like seed weight,
shoot dry weight and pod number per plant.
In Nigeria, Jonah et al.,(2010) reported heritability for pod yield per plant as 0.69,
seed yield per plant 0.72, for 100 seed weight 0.92 and the genetic mean
advance was relatively low at 16.2% for pod number and 22.5% for seed number,
which suggests that these traits will be difficult to select for in the Nigerian
environment. Hill et al., (1998) argued that even though heritability is an
important tool in the selection of potential material, the estimates may also depend
on the environment upon which materials are under test, and can sometimes differ
within the same crops. In this study, number of pods per plant and number of seed
revealed moderate genetic advance of the mean at 60% and 52.6% in the
agronomy bay experiment and different values to those in the field experiment, at
88.2% for pod number per plant and 42.5% for seed number per plant (Table
4.2.2 and Table 4.2.3). This is an indication that these traits are under additive
genetic control, thus selection of these traits can lead to an improvement in
bambara groundnut.
Selections of the best lines
According to the selection index ranks obtained as shown in (Table 4.2.5 and
Table 4.2.6), genotype 90-S19-3 from Namibia was the highest ranked with an
index value of 7.33 followed by 84-Acc696 from Zambia with an index value of
3.12. Genotype 109-BWA1 from Botswana was used as a check as it was one of
the lines among the five selected lines for stable seed colour, uniform leaf
morphology, and yield stability in various regions of Botswana for more than five
seasons (Chui et al., 2003), but it recorded an index value of 1.25 at the fifth
144
position. This shows that among the selected lines, at least four more could be
potentially adopted for selection and breeding in Botswana environment. The
yield production of 18 g per plant for 90-S19-3, 13.5 g for 84-Acc696 compares
favourably well with those found by other researchers, Berchie et al., (2010) in
Ghana recorded yield per plant for landrace Zebra (23.6 g), and Burkina at (17.7
g).
4.5
Conclusions
There was a substantial amount of variation found in the selected material, and
this indicates that there is a great potential for crop improvement. This was also
shown by higher genotypic coefficient of variation, heritability and genetic
advance (5 % of the mean) in seed and pod characters. GCV, h2b, GA for number
ofpods per plant (82.9%; 60.12%; and 88%) is a good estimation that selection
could lead to crop improvement. The use of shoot dry weight, leaf area, seed
number per plant and pod number per plant in the selection index, successfully
managed to identify 5 genotypes that have the potential to use as varieties in
Botswana environment. The selection index also identified Namibian landraces as
the best performers since 4 were selected among the best in 10 in the greenhouse
experiment in UK and 5 were selected among the best 10 in the field experiment
in Botswana.
Contributions in this chapter
Multiple selection of characters was effectively employed in bambara
groundnut genotype selection
A combination of leaf area, shoot dry weight, seed number per plant and
seed weight, are important traits that can be useful in determining the
selection indices of landraces to identify best performing lines.
145
CHAPTER FIVE: Population structure and genetic diversity of
bambara groundnut
5.1
Introduction
Many different landraces of bambara groundnut have been cultivated for a long
time by farmers all over sub-Saharan Africa and have experienced a wide range of
environmental conditions (Sesay, 2009). Farmers have also kept a large number of
genetic resources on their farms.As a minor crop relatively little attention has
been given to its genetic structure, despite a large germplasm collection being
held at International Institute of Tropical Agriculture (IITA) and a number of
countries in sub-Saharan Africa. An example is a gene bank that was established
in 1988 by 15 member countries of Southern African Development Community
(SADC), the SADC Plant Genetic Resources Centre based in Zambia
(www.spgrc.org.zm). An appropriate application of genetic analysis requires a
detailed knowledge of the genetic and historical relationships among and within
landraces. This knowledge also assists in identifying inbred lines that have
maximal diversity for use in breeding programmes (Liu et al., 2003).
Genetic resources for crop breeding are comprised from populations of genotypes
collected from various places. It is important to maintain maximum genetic
variability as well as identify origins or genotypes which are particularly useful
for breeding. Plant genetic resources are the backbone of agriculture and play an
important role in development of new cultivars (Malik and Singh, 2006).
Knowledge of genetic and trait diversity within a population and among
populations is also important for conservation management and for identifying
rare traits or genetic origins within species and to determine which could merit
special attention (Zhuravlev et al., 2010). Genetic variability is important for
landraces to adapt to environmental changes for their future survival, and for
genetic and trait improvement in crop breeding (Upadhyayaet al., 2008; Mwale et
al., 2007).
Diversity and population estimations are useful measures for estimating different
aspects of genetic structure in population studies (Gregorius, 2010). The
population structure of a species can arise due to numerous factors such as the
breeding system of the crop, effects of cultivation, breeding history and usage
146
since they can have significant effects on the partitioning of genetic diversity
within and among populations (Hamrick and Godt, 1996).
Population
differentiation is affected by evolutionary processes, such as genetic drift,
population size, selection founder effects and migration (Hedrick, 2005; Roy,
2000).
5.1.1
Genetic diversity in bambara groundnut
Early work done on bambara groundnut using isozyme markers by Pasquet et al.,
(1999) to investigate the population structure and partitioning of the observed
genetic diversity between the wild and domesticated accessions lead to the
conclusion that genetic diversity is present in both wild and domesticated bambara
groundnut. The other major findings in their study are the almost complete
absence of heterozygotes in both wild and domesticated forms and the highgenetic
identities between the wild and domesticated forms. This led them to conclude
that wild bambara groundnut is a true progenitor of the domesticated type.
Further genetic diversity studies were conducted by Ntundu et al., (2004) they
used AFLP markers to assess the genetic diversity among 100 bambara groundnut
landraces from diverse geographical regions of Tanzania. They used 11
informative AFLP primer combinations and generated 49 scorable polymorphic
fragments across all the selected accessions. Genetic distances between accessions
ranged from 0.1 to 0.68 based on a Jaccard variability index, while the cluster
analysis revealed that bambara groundnut consist of two major groups based on
their geographic origins in Tanzania. These markers provided evidence that there
is substantial genetic diversity within bambara groundnut.
5.1.2
Genetic diversity and population structure of other legumes.
A comparison of genetic diversity and population structure was conducted in 88
pigeonpea (Cajanus cajan) accessions from India and East Africa, using 6
microsatellites (Songok et al., 2010). Since India is the putative centre of origin
while East Africa is the presumed secondary centre of diversity, as expected more
diversity were recorded in India as compared to East Africa. Higher number of
alleles (42) were observed in India and Nei’s unbiased estimates of gene diversity
(H) of (0.55) compared to East Africa with 31 alleles and (H) of 0.228.
147
Using 18 microsatellites Lazrek et al., (2009) investigated the genetic diversity of
136 lines of Medicago truncatula, of 10 populations from Tunisia. They detected
an average of 4.2 alleles per locus, and an average gene diversity of 0.35.
Population structure results based on analysis of molecular variance (AMOVA)
showed that the majority (53.6%) of the variation present is mainly from the
differences between populations. The significant variation was attributed to the
difference between thenorthern and southern part of the country mainly due to the
influences of eco-environmental factors.
Liu et al., (2008) in China undertook a study on a total of 440 lentil(Lens
culinaris)from the National Genebank (Chinese Academy of Agricultural
Sciences, Beijing) with 204 accessions originally from China while 132 were
introduced into China (‘foreign’) and the rest 104 were designated ‘alien’ with no
traceable records. Fourteen SSR markers were used to investigate the genetic
diversity and population structure of all three lentil accession groups. A total of 87
alleles were detected among the 440 accessions with an average of 6.2 alleles per
locus, a mean observed heterozygosity (Ho) of 0.08 and an expected
heterozygosity (HE) of 0.56 were recorded. The researchers employed Principle
Coordinate Analysis (PCoA) and cluster analysis for population structure analysis,
with both techniques in agreement with each other as they separated the
germplasm into three accession groups, with the ‘foreign’ materials proving to be
the most diverse.
A population genetic structure analysis was conducted in 604 common bean
accessions from the International Center for Tropical Agriculture (CIAT) using 36
SSR markers. A total of 679 alleles were detected with an average of 18.4 alleles
per locus. The use of PCoA divided the collection into two main gene pools
Mesoamerican and Andean (Blair et al., 2009). They conducted an analysis of
molecular variance (AMOVA) to determine variation in gene pools, races,
subgroups, and the difference between primary and secondary center of diversity
for the crop.
More variability was assigned to the genepools (36.77%) and
races(32.57%) as compared to subgroups (32.09) but most of the variation
remained within each subpopulation (Blair et al., 2009).
148
In the present study a set of 12 preselected SSR markers were employed to
investigate the genetic diversity and population structure among 123 bambara
groundnut landraces. One hundred and eighteen are African landraces originally
from most parts of sub-Saharan Africa which covers four regions, namely; Central
Africa, East Africa, Southern Africa and West Africa. Five landraces are from
Indonesia (Asia); four were sourced directly from Indonesia while one was
sourced from The University of Nottingham seed stock.
5.1.3
The objectives of the study:
To analyse the population structure and genetic diversity of bambara
groundnut based on pod and seed related characters and SSR markers
To determine the association of morpho-agronomic markers based on
(qualitative characters of seed and pods) with SSR marker.
5.2
Materials and Methods
5.2.1
Phenotypic data analysis
Eight-seven landraces from the 119 that were planted in the agronomy bay
experiment (Table 2.1.2.2) reached maturity and produced reasonable pod and
seed numbers. The pods and seed characterisation was based on IPGR descriptors
(IITA, 2000). The measurements were taken on 10 seeds per plant using a Vernier
calliper (Mitutoyo) for pod length, pod width, seed length, and seed width. Other
measures were taken for pod texture, pod colour, pod shape, seed testa colour,
testa pattern, and eye pattern. The pod weight and seed weight per plant were
measured on an electronic balance and values were standardized (normalised), to
remove scalar effects.
149
5.3
Results
5.3.1
Genetic diversity analysis
From an initial 75 microsatellites, 12 markers were selected for further use, due to
their good amplification and ability to detect high levels of polymorphism in the
initial analysis of SSR markers. These markers are used throughout for population
structure analysis of bambara groundnut. A detailed characterisation of the 12
microsatellites on 123 bambara groundnut landraces revealed that all markers
were polymorphic (Table 5.1).
Table 5.1: PowerMaker summary data analysis for the 12 microsatellites used in the
analysis of 123 bambara groundnut landraces (118 from Africa and 5 from
Asia/Indonesia).
Marker
Primer 7
Primer 15
Primer 16
Primer 19
Primer 23
Primer 33
Primer 37
Primer 44
mBam3co18
D11
D14
E7
Mean
MAJ
0.61
0.30
0.60
0.14
0.64
0.40
0.38
0.89
0.15
0.32
0.11
0.66
0.43
GN
6
21
8
24
6
16
13
5
16
19
32
5
14
SS
123
123
123
123
123
123
123
123
123
123
123
123
123
No.
123
123
123
123
123
123
123
123
123
123
123
123
123
AN
5
16
8
23
6
15
12
5
15
17
29
4
13
Avail.
1
1
1
1
1
1
1
1
1
1
1
1
1
He
0.55
0.80
0.59
0.93
0.51
0.75
0.78
0.21
0.89
0.83
0.95
0.47
0.69
Ho
0.01
0.08
0.00
0.01
0.00
0.02
0.01
0.00
0.01
0.02
0.03
0.03
0.02
PIC
0.49
0.78
0.55
0.93
0.45
0.72
0.76
0.20
0.87
0.81
0.94
0.39
0.66
f
0.99
0.90
1.00
0.99
1.00
0.97
0.99
1.00
0.99
0.97
0.97
0.93
0.97
MAJ: major allele frequency; GN: genotype number; SS: sample size;
No.: number of observation;
AN: number of alleles; Avail: availability;
He: exp. heterozygosity;Ho: observed heterozygosity; PIC: polymorphic information content
f: inbreeding coefficient
The 12 microsatellites had high average polymorphic information content (0.66)
and managed to detect a total of 155 alleles with an average of 14 alleles per
marker. Polymorphic information content ranged from 0.20 for marker 44to 0.94
for maker D14. The average observed heterozygosity was 0.02 leading to a
corresponding inbreeding coefficient of 0.97. The observed heterozygosity was
much lower than the expected heterozygosity with departures from Hardy
Weinberg equilibrium (Appendix 8) detected across all markers and estimates of
inbreeding coefficient (f) between roughly 0.90 and 1.0 for all the markers. This is
typical for a strongly inbreeding crop.
150
5.3.2
Genetic diversity within and among regions
A comparison of the genetic diversity of 123 bambara groundnut was compared
among the 5 regions, based on estimates of total number of alleles, number of
alleles per locus and two estimates of allelic richness and Nei unbiased estimates
of gene diversity (H`) (Table 5.2). Higher genetic diversity was observed in
Southern African populations at 0.70 and lower diversity in Asian populations at
0.18 probably due to fewer number of samples in this population.
Table 5.2: A comparison of the genetic diversity estimates among the five regions
of Africa and Asia (Indonesia) analysis conducted using FSTAT 2.9.3 for all the
123 bambara groundnut landraces.
Regions
Asia (Indonesia)
Central Africa
East Africa
Southern Africa
Western Africa
N
5
10
10
31
67
At
1.50
3.58
4.67
7.42
9.83
NA
18
43
56
9
118
Rs1
1.50
3.06
4.01
4.31
4.37
Rs2
N/A
3.58
4.67
5.73
6.06
Ho
0.00
0.02
0.01
0.04
0.01
H`
0.18
0.48
0.69
0.70
0.65
N: Number of samples/genotypes NA: Number of alleles per locus At = Total number of alleles
Ho = Observed heterozygosity N/a = Estimates not calculatedH` = Gene diversity
Rs1 = Allelic richness based on sample size of 5 individuals
Rs2= Allelic richness based on sample size of 10 individuals
To compare genetic diversity based on allelic richness, two estimates were
calculated. The first allelic richness (Rs1) was standardized based on the smallest
number of landraces from Asia (5), while the second allelic richness estimate
(Rs2) was based on small samples from Central Africa and East Africa of 10
samples each (Table 5.2). The first and second estimates of allelic richness show
West Africa to have a higher diversityat 4.37 and 6.06 respectively followed by
Southern Africa at (Rs1) of 4.31 and (Rs2) of 5.73.
151
5.3.3
Principal coordinates analysis (PCoA)
Principal coordinate analysis was used to investigate the population structure of
the 123 bambara groundnut genotypes. The results show a cumulative percentage
of 16.15 %, for the first two axes with a variation of 9.87 % for Axis 1 and 6.28 %
for Axis 2 as shown in Table 5.3.
Table 5.3: Principal Coordinate analysis (PCoA) from the investigation of
population structure of 118 bambara groundnut landraces collected from Africa
and 5 from Indonesia based on MVSP program.
Eigenvalues
Percentage
Cumulative %
Axis Axis Axis Axis Axis
1
2
3
4
5
0.78 0.49
0.47 0.42
0.36
9.87 6.28
6.01 5.31
4.51
9.87 16.15 22.16 27.48 31.98
Axis Axis Axis
6
7
8
0.30 0.28
0.27
3.79 3.53
3.42
35.77 39.30 42.73
Axis Axis
9
10
0.26
0.23
3.27
2.94
45.99 48.93
The PCoA for the 123 bambara groundnut landraces is shown in figure 5.1.0 for
the first two axes, the population structure of the landraces are clearly demarcated
based on their areas of origin. The most evident separation is the West AfricanCentral African separation from the Southern Africa-East Africa Indonesian
landraces which are clearly distinguished as two groups (Figure 5.1.0). There are
also four landraces from the Southern African origin that share some potential
characters/introgression with the West African landraces.
152
PCA case scores
Group II
Group I
0.19
0.15
0.11
0.08
Axis 2
0.04
0.00
-0.04
-0.08
-0.11
-0.15
-0.15
-0.11
-0.08
-0.04
0.00
0.04
0.08
0.11
0.15
0.19
Axis 1
West Africa
Central Africa
Southern Africa
East Africa
Indonesia
Figure 5.1.0 A PCO scatter plot for the 123 bambara groundnut genotypes from Africa and
Indonesia generated from 12 microsatellites with MVSP program with a molecular variation of
16.15 %, with axis 1 contributing (9.87%) and while Axis 2 explained (6.28 %). The two cluster
groups were hand drawn on Microsoft Word.
5.3.4
Cluster analysis
A dendrogram which shows a population analysis based on the four regions for
Africa and one for Indonesia is shown in Figure 5.2.1. The cluster analysis is
largely in agreement with the PCoA coordinates which clearly demarcated
landraces based on their areas of origin. Landraces from one region are mostly
grouped together, with some exception where some mixture are visible, as some
dark blue traces could be found among the green colour coded landraces. The
landraces from Southern Africa, East Africa and Indonesia are also clustered
together while the West African and the Central African landraces are grouped
together
153
UPGMA
70
80
69
100
18-1206BFA
93-AS17RSA
105-MHNBLACKNAM
79-371TZA
92AHM968NAM
54-810MDG
68-283NGA
63-120NGA
103-TIGANICURUMLI
57-91MLI
100-SB16 5ANAM
114-CS37(RP)KEN
110-BOTS2BWA
99-SB4-2NAM
104-S-1913NAM
90-S19-3NAM
123GHIND
122GCIND
121BHIND
120-BCIND
118-RAMAYANAIND
89-DipCBWA
94-DODCTZA
82-682ZMB
97-JACBBWA
80-379TZA
81-385TZA
78-369TZA
48-790ZMB
46-243GMB
83-683ZMB
115-CS129(RP)KEN
102-V560ABWA
96-GABCBWA
91-UNIS R SWA
113-BOTS5BWA
86-757ZMB
53-806MDG
51-799MDG
84-696ZMB
23-447CMR
21-438CMR
119-HYBRIDBWA
74-335NGA
69-286NGA
33-484CMR
30-476CMR
112-BOTS4BWA
49-793KEN
52-806MDG
39-529CMR
88-AHM753NAM
56-89MLI
85-754ZMB
6-289BEN
40-536CMR
71-330NGA
70-329NGA
64-172NGA
67-278NGA
38-506CMR
107-NAVREDGHA
20-118IVC
43-216GHA
41-210GHA
98-KABCA4SLA
47-246GMB
7-85BFA
42-214GHA
95-DODRTZA
72-331NGA
28-472CMR
45-231GMB
25-1164BFA
36-502CMR
37-503CMR
32-483CMR
29-473CMR
65-395CMR
27-467CMR
22-440CMR
76-390SDN
108-NAV4GHA
19-1352CAF
17-1337CAF
16-1329CAF
14-1315CAF
12-1288CAF
11-1284CAF
35-501CMR
15-1324CAF
13-1307CAF
2-13NGA
116-VSSP11CMR
10-1276IVC
9-308BFA
3-9NGA
101-UNISCSWA
60-32NGA
24-448CMR
66-275NGA
61-33NGA
58-23NGA
109-BOTS1BWA
8-292BFA
117VSSP6CMR
111-BOTS3BWA
26-460CMR
5-191BEN
73-334NGA
44-229GHA
59-25NGA
50-792KEN
34-492CMR
4-144GHA
87-1033ZWE
62-119NGA
77-391SDN
75-348NGA
55-88MLI
31-480CMR
106-YOLANGA
1-1NGA
100
58
52
60
72
0.04
0.2
0.36
0.52
0.68
0.84
1
Nei & Li's Coefficient
Figure 5.2.1. Cluster analysis of bambara groundnut landraces from five regions, from Africa and
Indonesia (Asia). The dendrogram is based on 12 SSR markers. The Unweighted pair group
method with arithmetic averages (UPGMA) tree was based on Nei and Li, (1979) coefficient of
genetic similarity generated from the presence/absence binary matrix on 123 bambara groundnut
landraces on MVSP. Colour codes :( Purple: Indonesia), (Yellow: East Africa), (Red: Southern
Africa), (Dark blue: Central Africa) and (Green: West Africa).
154
5.3.5
Genetic differentiation based on FST
Genetic differentiation was generally high ranging from 0.610 for West Africa to
0.645 for East Africa, while in Central Africa was moderate at 0.440. The
diversity was lower in the Indonesian landraces with 0.267 (Table 5.4).
Table 5.4: Genetic differentiation of the 123 bambara groundnut landraces from 4
regions of Africa and also Asia (Indonesia), estimated using Weir and Cockerham
(1984) on Genepop version 4.0
Differentiation level
Asia (Indonesia)
Central Africa
East Africa
Southern Africa
Western Africa
5.3.5.1
FST
0.267
0.440
0.645
0.630
0.610
Pairwise comparison
FST- based genetic differentiation revealed significant differentiation (P <0.05)
among the landraces in all the regions except between the East African and the
Southern African landraces and also between the East African and Asian
landraces (Table 5.5). The results are consistent with PCoA results and cluster
analysis, where the landraces from the two regions showed no clear separate
groups. There was low but significant genetic differentiation between the West
Africa landraces with the Central, East African and Southern African landraces
possibly due to some mixture of the landraces between regions which was also
revealed by PCoA and cluster analysis. The highest genetic distance was observed
between Asian landraces and Central African landraces, these populations also
have lower genetic differentiation
155
Table 5.5: Pairwise genetic distance based on FST values between populations,
calculated on 12 microsatellites based on five regions of Africa including Asia
(Indonesia).
Regions
Central
Africa
Asia (Indonesia)
Central Africa
0.510** 0.178ns
0.204**
East Africa
Southern Africa
East
Africa
Southern
Africa
West
Africa
0.166ns
0.202**
0.276**
0.095**
0.027ns
0.088**
0.103**
**Significant at (P <0.05) and (ns) are not significant
5.3.6
Analysis of molecular variance analysis
The analysis of molecular variance (AMOVA) for all 123 bambara groundnut
landraces was partitioned into a three-level of hierarchy which consists of
variation in between the 2 groups as revealed in Figure 5.1.0. Group 1 consist of
most of the West African landraces with the exception of 13, one from East
Africa, four from Southern Africa and all landraces from Central Africa. Group 2
is made up of all Southern African landraces except four, all East African
landraces except one, 13 landraces from West Africa and all the five landraces
from Indonesia (Asia). AMOVA identified a highly significant (P <0.000) and
meaningful variation at all the three hierarchy levels. The difference between the
two groups was significant at 12.45 %, but the majority of variation was among
individual landraces at 84.5% with little but significant variation within
individuals at 3.05% (Table 5.8).Similar observations were made in the PCoA,
cluster analysis and FST, where genetic differentiation was found between regions,
countries and within genotypes.
156
Table 5.6: Analysis of Molecular Variance for the 123 bambara groundnut
landraces based on 12 SSR markers using Arlequin version 3.1
Source of variation
df
Among populations
Sum of squares
Variance components
Percentage variation P value
1
69.133
0.514Va
12.45
<0.000
Among individuals within populations
121
859.578
3.489Vb
84.50
<0.000
Within individuals
Total
123
245
15.500
944.211
0.126Vc
4.129
3.05
<0.000
5.3.7 Comparison of molecular markers with pod and seed characters
For the morphological markers for the first two axes, the cumulative variation
explained was 49.9% (Table 5.7) which was higher (based on 87 bambara
groundnut accessions) compared to variation explained by the first two axes using
SSR markers (16.08 %; Table 5.8). The PCoA for the pod and seed characters was
able to group the West African landrace together, while the landraces from other
regions do not show a clear pattern of separation. Most of the landraces were
clustered together at the centre of the graph, which shows that they share common
characters (Figure 5.3.0).
The PCoA for the SSR marker data shows a greater
dispersion among landraces, with clear separation for landraces from different
regions (Figure 5.4.0).
Table 5.7: Principal Coordinate analysis (PCoA, Euclidean) for 15 characters of
pods and seeds for 87 landraces that set reasonable seed numbers among the 119
landraces planted in the agronomy bay for bambara groundnut germplasm
characterisation.
Eigenvalues
Percentage
Cumulative %
Axis
1
3.98
33.59
33.59
Axis
2
1.93
16.3
49.9
Axis
3
1.07
9.04
58.9
Axis
4
0.68
5.71
64.6
Axis
5
0.58
4.88
69.5
Axis
6
0.4
3.37
72.9
Axis
7
0.37
3.16
76
Axis
8
0.33
2.79
78.8
Axis
9
0.29
2.48
81.28
Axis
10
0.26
2.23
83.51
157
PCA case scores
0.9
0.7
0.6
Axis 2
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.6
-0.4
West Africa
-0.2
0.0
Central Africa
0.2
Axis 1
East Africa
0.4
0.6
Southern Africa
0.7
0.9
Indonesia
Figure 5.3.0: A PCO scatter plot for the 87 bambara groundnut that produced pods and seeds
among 119 bambara groundnut planted. The data is based on 7 pod and 8 seed characters analysed
using the MVSP program. The percentage variation for Axis 1 represents 33.59% and the Axis 2
represent 16.8 % with a cumulative percentage of 49.87% for the first two Axes.
Table 5.8: Principal Coordinate Analysis (PCoA) for 87 bambara groundnut
landraces that set seed, based on 12 microsatellites
Eigenvalues
Percentage
Cumulative %
Axis Axis Axis Axis Axis Axis Axis Axis Axis Axis
1
2
3
4
5
6
7
8
9
10
0.73 0.51 0.46 0.43 0.35 0.33 0.31 0.30 0.27 0.25
9.45 6.63 5.95 5.57 4.57 4.29 4.03 3.88 3.53 3.16
9.45 16.08 22.03 27.60 32.17 36.46 40.50 44.38 47.91 51.07
158
PCA case scores
0.17
0.14
0.10
Axis2
0.07
0.03
0.00
-0.03
-0.07
-0.10
-0.14
-0.17
-0.17
-0.14
-0.10
West Africa
-0.07
-0.03
0.00
Axis 1
Central Africa
East Africa
0.03
0.07
0.10
Southern Africa
0.14
0.17
Indonesia
Figure 5.4.0.A PCO scatter plot on case scores for the 87 bambara groundnut that produced pods
and seed based on 12 microsatellites, generated on MVSP program. The cumulative percentage of
variation explained for the first two Axes is 16.08%, Axis 1 contributes 9.45% and Axis 2
contributes 6.63%.
The Southern African landraces together with the East African and Indonesia
landrace are grouped together, while the West African and the Central African are
grouped together. These groups are clearly separate from each other, with the
exception of 4 landraces from Southern Africa and 7 from West Africa which may
reflect some exchange of material or other gene flow between these regions,
although simple errors in the accession records or samples order could potentially
give similar effects.
The relationship between the Euclidean distance matrix based on the 13 morphoagronomic characters of pod and seed and 12 SSR markers based on (Nei’s 1972)
genetic distance matrix were tested using a Mantel test correspondence test,
Spearman rank correlation coefficient and examined through Pearson productmoment correlation coefficient on NTSYS and SPSS 16 (Table 5.9 and Figure
5.5.0).
159
a)
0.06
Genetic
0.03
0.01
-0.02
- 0.04
- 0.03
0.00
0.03
Morphology
0.06
0.09
b)
Figure 5.5.0 Scatter plot of correlation for morphological marker genetic distances estimate based
on standard Euclidean and SSR marker genetic distance based on Nei’s 1972, the analysis was
conducted on (a) Mantel test correspondence test on NTSYS and (b) on Pearson correlation on
SPSS version 16
160
A low, negative and non-significant correlation between the genetic distance
matrices were recorded on Mantel correspondence test while Pearson correlation
and the Spearman rank correlation recorded a low but highly significant
correlations. This could be an indication that the two markers are explaining
different variation in the selected materials.
Table 5.9: Correlation of molecular marker distance matrices, based on Pearson
correlation, Spearman rank correlation and Mantel test for the 12 qualitative
character and 12 molecular markers.
Marker
SSR
Morphology
N
P value
Pearson
Morphology
-0.048**
1
3741
0.003
Spearman
Mantel test
Morphology Morphology
-0.038*
-0.0016
1
1
3741
3741
0.021
0.488
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
161
5.4
Discussion
Genetic diversity in the 123 landrace accessions
The 123 bambra groundnut accessions used in this study from 21 countries
covered all the regions of Africa, and also included 5 accessions from Indonesia.
The collection of landraces represented is expected to represent the breadth of the
diversity in bambara groundnut.
The genetic diversity and polymorphic information content recorded in this study
gave averages for expected heterozygosity of 0.69, polymorphic information
content of 0.66 and an average number of 13 alleles per locus (Table 5.1).
Relatively high genetic diversity was observed possibly due to the fact that the
material was from a wide geographic distribution from Africa and also Asia
(Indonesia).
Diouf and Hilu, (2005) identified an average of 5.3 alleles per locus in 11 cowpea
(Vigna unguiculata) varieties in Senegal, using 30 SSR markers. In common bean
(Phaseolus vulgaris), Masi et al., (2003) analysed 264 genotypes based on 30
SSR markers; they identified a total of 135 alleles and an average of 4.3 alleles
per locus. Their study revealed lower number of alleles and average number of
alleles per locus even though relatively higher number of SSR markers was used.
However, it should be borne in mind that the SSR markers used here were
deliberately selected after a pre-screen of an initial 75 showing polymorphism
within bambara groundnut landraces.
Genetic diversity measures for the landraces were estimated, based on the total
number of alleles and the number of alleles per locus. These measures are highly
dependent on sample size and large samples usually contain more alleles than
small samples (Kalinowski, 2004). The genetic analysis was conducted at a
regional level, where landraces from several countries were grouped together. The
problem of samples size was also resolved by the use of FSTAT software to
calculate the allelic richness in each population based on smallest number of
individual samples (Leberg, 2002). The program estimates allelic richness (Rs)
independent of the sample sizes, and this allows a comparison of genetic diversity
between populations with different sample sizes.
162
When allelic richness was measured based on a minimum of 5 samples (Rs1) and
10 samples for (Rs2), West African landraces revealed a slightly higher genetic
diversity at 4.37 and 6.06 followed by the Southern African landraces with (Rs1)
4.31 and (Rs2) at 5.73(Table 5.2). For the gene diversity H`, Southern African
and East African at 0.70 and 0.69 respectively revealed a slightly higher diversity
compared to West Africa at 0.65.The relatively low diversity estimates for the
West African landraces, which may be below what would be expected in the
major centre of diversity compared to the Southern African landraces could be
caused by the sampling strategy which biased the sampling to include more
landraces from 2 countries among West African samples. A total of 67 genotypes
that were sampled from West Africa, 45 came from Nigeria and Cameroon which
is approximately 67 % of biased towards these two countries. While the Southern
African countries contributing the majority of the landraces were Namibia and
Botswana with 16 out of 30 genotypes, which is approximately 55 %. Nei
unbiased estimates of gene diversity is mostly influenced by moderate allele
number and allelic frequency rather than alleles of high or low frequency (Shete,
2003; Songok et al., 2010).
Population structure and relationship among different landraces
Low FST measures were observed in the Indonesian landraces as compared to the
African landraces, and this could be caused by long period of time the Indonesian
landraces expansion to Indonesia took place. In addition the low differentiation
from Indonesian landraces could also be caused by few entries of bambara
groundnut from different places in Africa or the limited number of Indonesian
accessions.
The SSR-based population structure analysis using principal coordinate (PCoA)
analysis clearly defined landraces based on their areas of origin which is in
agreement with previous work on bambara groundnut by Amadou et al., (2001)
and Massawe et al., (2002). The landraces from West Africa tend to group with
those from Central Africa, while those from Southern Africa clustered with those
from East Africa and Indonesia. In fact there was no significant difference
between the Southern Africa and East African landraces based on FST pairwise
genetic differentiation between regions (Table 5.5). This is also a reflection of a
163
potentially extensive movement of seeds materials between these regions, by
farmers.
FST estimates correlation between genes of different individuals in the same
population and is used as a measure of genetic differentiation among populations
FST= 0, between subpopulations indicates that they are identical in all allele
frequencies, but when it is FST = 1 they are fixed for different alleles. In a pairwise
comparison between the five regions, the FST ranged from 0.027 between East
Africa and West Africa, to 0.510 between Indonesia and Central Africa. The East
African landraces also show a lot of common alleles with the West African
landraces at 0.088.
The cluster analysis, based on UPGMA (Nei and Li, 1979) displayed a minimum
similarity of 24% among the 123 accessions and clustering occurred based on
areas of origin, in a similar way to the PCoA results, although only two major
groups were clearly differentiated.
The study revealed a clear structure for these 123 bambara groundnut landraces.
The total analysis of molecular variance (AMOVA) results for the 123 landraces
revealed that most of the variation was among individual genotypes (84.5%)
followed by among populations (12.45%) and last is within individuals (3.05%).
Massawe et al., (2003) found high levels of polymorphism among landraces at
77.1% with 28.7% within genotypes using RAPDs markers. Wasike et al., (2005)
used AFLP to study the genetic diversity of 32 African and 9 Asian pigeonpea
(Cajanus cajan) varieties, the analysis of molecular variance estimates between
the two regions revealed a higher genetic variation of 92.16% within the
populations while only 7.84 was among the two regions. In cowpeas, Zannou et
al., (2008) also reported a higher percentage of variation within accessions (73%)
as compared to among groups (26%) which indicates a higher within population
diversity for these species, possibly associated with the accessions originating
from different ancestors.
164
Comparison of genetic distance estimates from morpho-agronomic and SSR
markers
Morphological character assessment is the first step in the characterisation of
germplasm, because breeding programmes rely on the magnitude of phenotypic
variability in crops. Qualitative characters examined are mostly influenced by the
consumer preference and the socio-economic conditions at a particular time
(Gafoor et al., 2002). In this study both morphological and SSR markers were
able to differentiate the landraces.
Correlation of the genetic distance estimates
from the two marker types showed non-significant and negative correlation
according to Mantel test, which indicates that, both marker typesdiscriminate
differently among the genotypes. The lower levels of negative but highly
significant correlations observed using Pearson and Spearman rank between
morphological and DNA markers could also indicate some agreement between the
phenotypic and molecular marker. However, the fewer number of markers used in
the study may also contribute to lower correlations because of sampling of the
genome is low (Vieira et al., 2007) and here the sampling of the genome is
relatively shallow (12 SSR). In addition morphological markers are less reliable,
and efficient in clearly discrimination genotypes as compared to molecular
markers in genetic relationships (Bayele et al., 2005).
5.5
Conclusions
The East African landraces had a slightly higher differentiation to both Southern
Africa and the West African landraces. AMOVA also revealed a low variance
among the two major groups which are mainly clustered based on regions origins
of (West African, Central Africa) and (Southern Africa, Indonesia, East Africa) as
more variation was observed between genotypes. The poor correspondence
between distance estimates based on morphological traits and SSR markers was
observed. The correlation estimates for Pearson product –moment coefficient and
spearman rank correlation coefficient gave negative and significant correlation
while Mantel test gave a negative and non-significant correlation.
165
CHAPTER SIX: Genetic diversity of bambara groundnut based on SSR
markers and the comparison with morpho-agronomic characters
6.1
Introduction
The application of molecular markers is widely accepted as a potentially powerful
tool in crop improvement of a number of crops. Characterisation of plant genetic
resources represents a good starting point to dissect allelic variation and identify
variation in crops (Upadhyaya et al., 2008). Genetic diversity can be measured at
various levels including within accessions (particularly for landraces), between
accessions and also among species, with the phylogenetic relationships revealing
how a group of species are related (Wang et al., 2009).
Molecular markers could be used to identify genetically different populations and
use them in selecting parents so that inbreeding can be avoided as has been done
in a highly inbreeding alfalfa (Medicago sativa)(Noeparvar et al., 2008). DNA
markers can be linked to agronomic characters and thus are useful in markerassisted selection (MAS) in plant breeding. MAS could then be used in the
selection of crops and make plant breeding more effective and efficient (Collard et
al., 2005).
A number of molecular marker systems have been employed in bambara
groundnut for genetic diversity assessment, including; RAPDs (Amadou et al.,
2001), AFLP (Massawe et al., 2002; Ntundu et al., 2004); SSR markers (Basu et
al., 2007) and isozymes (Pasquet et al., 1999). Recently DArT markers and
morphological markers have been used and compared (Olukolu et al., 2011). A
more robust approach to estimate genetic variation could be realised if both
morphological and molecular techniques are simultaneously used (Parsaeian et
al., 2010).
Botswana is a country with semi-arid climatic conditions and usually low levels of
cereal grain yield (mostly maize and sorghum) due to poor soils and low moisture.
Well adapted crop genotypes in that environment such as bambara groundnut
could be used to increase food production (Brink et al., 2000).
The first
expedition for a bambara groundnut survey in Botswana was reported in 1947,
followed by a second one in 1985 (Ministry of Agriculture, Botswana; Appa-Rao
et al., 1986), and these covered only the northern part of the country, which
166
covered only two, out of a potential ten, districts in the country (Karikari et al.,
1995). Few bambara groundnut landraces have been tested for adaptation and
undergone selection for high yields in the Botswana environment. It is therefore
important to identify genetic variation among bambara groundnut landraces, using
both the molecular markers and morphological markers in the target environment.
6.1.1
Genetic diversity of bambara groundnut
Massawe et al., (2002) reported substantial genetic diversity among 16 bambara
groundnut landrace single genotype samples when they used AFLP with a
combination of seven primer pairs. Pairwise similarities between landraces were
determined according to Jaccard coefficient and the matrices were used to
produce dendrograms on unweighted pair-group method with arithmetic mean
(UPGMA) cluster analysis. Landraces were grouped into three based on their
geographic areas of origin and Southern Africa landraces DipC1995 and Malawi5
were grouped together, with no landraces samples identical. Amadou et al.,
(2001) used Random Amplified Random Amplified DNA (RAPD) to assess the
genetic diversity of 25 single genotype landrace accessions collected from the
International Institute of Tropical Agriculture (IITA). The landraces clustered into
two main groups based on their areas of origin.
Massawe et al., (2003) also used RAPD markers on 12 bambara groundnut
landraces with multiple genotypes per landrace. Data from individuals of each
landrace was analysed to determine the level of heterogeneity within the
landraces. AMOVA revealed highly significant variation (P<0.001) among
landraces and also within each individual landrace. The partitioning of total
genetic diversity showed that 71.25 % was explained by the landraces differences
while 28.67% was among individuals within landraces. However, Stadler, (2009)
used Diversity Array Technology markers (DArT) on bambara groundnut and
found that intra-landrace diversity was lower among some landraces than others.
This also shows that some landraces are ‘purer’ than others, which could be a
good basis for selection of pure lines.
6.1.2
Genetic diversity in other leguminous crops
Several authors have reported a general low level of polymorphism among
cultivated peanut germplasm (He et al., 2003; Wang et al., 2007). The narrow
167
gene pool of the cultivated peanut has been attributed to the evolution that
occurred in South America through a limited number of interspecific
hybridizations and polyploidization (Mace et al., 2006) and this has led to limited
genetic diversity of cultivated peanut, through a genetic bottleneck. Gimenes et
al., (2007) observed lower genetic diversity in groundnut (Arachis hypogaea)
when studying 16 accessions of A. hypogaea and 38 accessions of eight other
sections of Arachis using 13 microsatellites markers. They observed mean
polymorphic loci (33%), mean number of alleles (4.02) and mean polymorphic
information content (0.48) and He et al., (2003) recorded a similar number of
alleles per locus on 24 genotypes, when using 19 SSR markers of 4.25 alleles per
locus.
The cultivated chickpea, as a self-pollinated crop with 2n = 2x = 16 shows a lower
genetic diversity as compared to the wild Cicer. Upadhyaya et al., (2008) reported
substantial genetic diversity based on the use of 48 SSR markers to analyse 2915
chickpea accessions (Cicer arietinum). They identified 1683 alleles, with a range
of 14 to 67 alleles per locus and an average of 35. The polymorphic information
content (PIC) ranged from 0.467 to 0.974 with an average of 0.854. These very
high observations of genetic diversity was attributed to the large set of accessions
analysed from the Mediterranean and African regions which are the center of
origin and center of diversity, respectively (Upadhyaya et al., 2008). Castro et al.,
(2011) also observed higher genetic diversity among 32 commercial cultivars of
chickpea using 15 microsatellites markers. They detected a total of 154 alleles,
10.3 mean number of alleles per locus and an average PIC of 0.78
A similar observation was found among 40 genotypes representing seven Cajanus
species which consists of 32 cultivated type and 8 wild forms by Saxena et al.,
(2010). They employed 16 microsatellites, which yielded a total of 72 alleles with
an average of 5.5 alleles per marker in the germplasm. Allele numbers ranged
from 2 to 8, PIC values for these markers ranged from 0.05 to 0.55, with an
average of 0.32 per marker. Higher genetic diversity was observed in the wild
type with a PIC of 0.64 and an average of 5 alleles compared to the cultivated
form with a PIC of 0.15 and an average allele number of 2.08.
168
Yang et al., (2006) used DArT markers in pigeonpea (Cajanus cajan) identified
low levels of genetic diversity cultivated pigeonpea compared to its wild relatives.
They evaluated 20 species of Cajanus, and identified a total of 700 markers that
were polymorphic, but only 64 markers were polymorphic in cultivated
accessions and this indicates the narrow genetic base of cultivated pigeonpea.
In soybean (Glycine max), Liu et al., (2011) observed 250 alleles among 91
accessions at 35 SSR loci, and an average of 7.14 alleles per locus, and an average
PIC of 0.74 in a study conducted in Shaanxi Province of China.
6.1.3
Efficiency of molecular and morphological markers in genetic diversity
estimates
A number of approaches for measuring genetic distance such as the analysis of
morphological characters or molecular markers have been widely used to try to
measure crop diversity. The differences in DNA sequences between individuals
detected when using molecular markers are often more informative compared to
morphological markers (Tanksley et al., 1989). There are several other advantages
for molecular marker application; they are reliable, not influenced by
environmental conditions and are essentially Mendelian markers. In some
instances adequate levels of polymorphism are not available; therefore they can be
limited in the evaluation of genetic diversity (Cupic et al., 2009).Tantasawat et
al., (2010) compared the use of morphological and SSR marker for genetic
diversity and relatedness studies in 17 mungbean (Vigna radiata) and 5 blackgram
(Vigna mungo) accessions. The two species are mainly differentiated by seed
colour, with some differences in seed shape and pod colour. In their findings
morphological characters were not able to differentiate between the two Vigna
species compared to SSR markers which were able to distinguish the two species,
which is an indication that molecular markers can be more effective in
differentiating the two species.
In common bean (Phaseolus vulgaris), Kumar et al., (2009) compared the
morpho-agronomic traits and microsatellites in genetic diversity analysis of 115
common bean. Seventy were Indian landraces, 24 released varieties and 21 exotic
accessions. The Euclidean distance based dendrograms and the PCO were able to
separate varieties from genotypes but based mainly on yield and yield related
169
traits while the microsatellites marker PCO and UPGMA clearly separated the
genotypes into their respective groups. The two marker types were used to
complement each other. However, the use of a Mantel test, revealed a good
correlation between the morpho-agronomic distance and molecular marker genetic
distance estimates (r = 0.876), which indicates that either of the marker can give a
good reflection of genetic estimates from another marker (Kumar et al., 2009),
given this, morphological markers may well be simpler to apply in breeding
situations.
Ntundu et al., (2004) estimated phenotypic distances calculated on 20 quantitative
and 7 qualitative traits, and also genetic distances based on 49 AFLP polymorphic
markers in 100 bambara groundnut single genotype accessions in Tanzania to
determine the relationship between the two markers types. A low correlation of r
= 0.41 was recorded, while the clusters of accessions based on AFLPs compared
well with that based on phenotypic characters.
Gomez et al., (2004) studied the molecular and genetic diversity of common bean
(Phaseolus vulgaris) landraces in Nicaragua using 14 traits measured in 12
individual landraces with seven SSR loci. The use of both morphological and SSR
markers provided complementary information, since the variation at the molecular
level was mostly between and within landraces, but did not reveal consistent
differences
between
ecological
zones,
while
the
phenotypic
variation
corresponded to the ecological zones. The molecular differentiation of the
landraces at FST = 0.34 was due to founder effects, while phenotypic
differentiation was attributed to the effect of adaptation.
In this study a set of molecular data and morphological characters were recorded
on the same landraces with an aim to evaluate the efficiency of these two
techniques in bambara groundnut, so that either morpho-agronomic or DNA
markers could be used, or both as a compliment to one another. The aim of this
study was to assess the genetic diversity of bambara groundnut and to estimate the
genetic correlation between the morphological genetic distance estimates and
molecular (SSR) genetic distance estimates in bambara groundnut landraces.
170
6.2
Materials and methods
6.2.1
Plant Materials used
Thirty five bambara groundnut landraces were selected among the 119 accessions
that were planted in the agronomy bay, listed in (Table 2.1.2.2). Three individuals
were used which makes 105 genotypes per landrace.
6.2.2 Markers used
Twenty microsatellites which showed good amplification and had been previously
shown to be polymorphic were selected from a pool of 75 markers and listed in
Appendix 2.
6.3
Results
6.3.1
Polymorphism of microsatellites in bambara groundnut
A total of 105 genotypes were amplified with 20 microsatellites, a total of 231
alleles were identified with an average of 12 alleles per locus (Table 6.1). The
highest number of alleles was recorded for marker D14 with 29 alleles and the
lowest allele numbers were recorded for marker E7 with 3 alleles. Polymorphic
information content (PIC) ranged from 0.07 to 0.95, (markers D8 and D14,
respectively) with an average of 0.67. The genetic diversity detected using all
microsatellites across the genotypes was high with a range of 0.07 to 0.95 and a
mean of 0.69.
171
Table 6.1: Summary of PowerMarker data analysis for the 35 bambara groundnut
landraces using 20 microsatellites analysis conducted on each of the 105
individual genotypes.
Marker
Primer 1
Primer 7
Primer 10
Primer 15
Primer 16
Primer 19
Primer 21
Primer 23
Primer 30
Primer 31
Primer 32
Primer 33
Primer 37
Primer 44
D8
mBam2co80
D11
D14
D15
E7
Mean
MAJ
0.74
0.61
0.33
0.13
0.41
0.14
0.50
0.76
0.69
0.39
0.17
0.30
0.39
0.51
0.96
0.18
0.30
0.10
0.24
0.67
0.43
GN
6
6
10
20
8
21
7
6
6
9
21
16
14
8
4
16
15
31
12
3
12
SS
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
No.
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
105
AN
6
6
10
20
8
21
7
6
6
9
19
15
12
8
4
16
15
29
11
3
12
He
0.43
0.59
0.78
0.92
0.75
0.92
0.62
0.40
0.49
0.75
0.92
0.84
0.76
0.67
0.07
0.89
0.84
0.95
0.85
0.45
0.69
Ho
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.01
0.03
0.00
0.00
0.00
0.00
0.02
0.01
0.00
0.00
MAF- Major allele frequency
GN-Genotype number observed
No. obs- Number of observations
NA- Allele Number
PIC
0.41
0.55
0.75
0.91
0.72
0.91
0.55
0.38
0.45
0.71
0.91
0.82
0.73
0.64
0.07
0.88
0.82
0.95
0.83
0.36
0.67
F
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.99
0.96
1.00
1.00
1.00
1.00
0.98
0.99
1.00
0.99
SS –Sample size
GD-(Expected heterozygosity)
Het.- Heterozygosity PIC- Polymorphic information content f- Inbreeding coefficient
172
Table 6.2: Intra-landrace diversity among the 35 genotypes conducted on each of
the three genotypes per landrace using 20 SSR markers based on Arlequin version
3.1
Landraces
3Acc 9
4Acc144
6Acc 289
10Acc 1276
20Acc118
30Acc 476
33Acc 484
40Acc 536
45Acc 231
48Acc790
49Acc793
50Acc 792
56Acc 89
60Acc 32
69Acc286
70Acc 329
74Acc335
76Acc390
81Acc385
84Acc696
85Acc 754
88AHM753
90S19-3
91UNIS R
92AHM968
95DODR
99SB4-2
100SB16 A
104S-1913
105MHN black
109BWA1
113BWA5
117VSSP6
118Ramayana
119Hyrid
Mean
No.
observation
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Polymorphic
loci
8
5
5
6
3
5
3
5
6
11
11
13
9
8
8
6
7
8
14
5
11
5
8
14
4
8
6
7
10
6
11
12
2
7
7
8
Average no.
alleles
1.45
1.30
1.30
1.35
1.15
1.25
1.20
1.25
1.35
1.80
1.65
1.80
1.50
1.45
1.40
1.30
1.35
1.40
1.80
1.35
1.60
1.30
1.50
1.95
1.20
1.55
1.35
1.35
1.50
1.35
1.55
1.75
1.10
1.35
1.40
1.43
Ho
0.02
0.03
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.01
0.00
0.00
0.00
0.02
0.00
0.00
0.02
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.00
0.00
0.02
0.00
0.03
0.07
0.01
He
0.22
0.14
0.15
0.17
0.08
0.13
0.09
0.13
0.17
0.35
0.32
0.39
0.25
0.23
0.23
0.16
0.19
0.21
0.39
0.16
0.31
0.14
0.24
0.44
0.11
0.26
0.17
0.19
0.27
0.17
0.29
0.36
0.05
0.18
0.19
0.22
173
All the markers showed an inbreeding coefficient close to 1, and the observed
heterozygosity (Ho) is far lower than the expected heterozygosity (He) and close
to 0, as shown in Table 6.1 and Table 6.2. Each landrace was analysed for intralandrace diversity based on three genotypes sampled. The majority (28) landraces
recorded narrow genetic diversity of an average of 1.5 number of alleles per locus
and below. The more diverse landraces were 113-BWA 5 (Botswana), 48-Acc231
and 50-Acc 792 both from (Kenya), 81Acc385 (Tanzania) and Uniswa Red
(Swaziland) that had at least 1.75 alleles per locus.
Three genotypes from each of the 35 landraces were investigated for genetic
similarity and this made up a total of 105 individual genotypes samples that were
analysed based on the Nei and Li (1979) similarity index. Then a cluster analysis
was conducted based on the similarity matrix produced using UPGMA procedure
on MVSP. The similarity of the genotypes ranges from 0.37 to 0.95 (Figure
6.1.1). Genetically similar genotypes were observed between 20Acc118 from
Côte d Ivoire, while those genetically further from each other were104S-1913
from Namibia and 10Acc1276 from Central African Republic at 0.108.
Forty three points of high bootstrap values more than 50% indicated on the
dendrogram showed that branches are well supported. Among the 43 bootstrap
values 20 were supportingthree individuals within a genotype;this showed that 57
% of genotypes had individuals which are more similar to each other.
174
UPGMA
74
54
63
100
81
68
64
75
86
54
51
93 57
89
96
54
89
97
67
79
60
85
77
55
93
65
81
93
68
83
59
85
53
71
62
0.28
0.4
0.52
0.64
69
67
91
92
100
85
55
0.76
0.88
67
113BWA5BWA
91UNIS RSWA
91UNIS RSWA
119Hyrid
119Hyrid
119Hyrid
118RamayanaIND
118RamayanaIND
118RamayanaIND
95DODRTZA
95DODRTZA
95DODRTZA
91UNIS RSWA
117VSSP6 CMR
117VSSP6 CMR
117VSSP6 CMR
113BWA5BWA
113BWA5BWA
81Acc385TZA
84Acc696ZMB
84Acc696ZMB
84Acc696ZMB
100SB16 ANAM
100SB16 ANAM
100SB16 ANAM
99SB4-2NAM
99SB4-2NAM
99SB4-2NAM
104S-1913NAM
104S-1913NAM
90S19-3NAM
90S19-3NAM
104S-1913NAM
90S19-3NAM
85Acc 754ZMB
48Acc790KEN
88AHM753NAM
88AHM753NAM
88AHM753NAM
109BWA1BWA
109BWA1BWA
109BWA1BWA
81Acc385TZA
81Acc385TZA
92AHM968NAM
92AHM968NAM
92AHM968NAM
105MHN blackNAM
105MHN blackNAM
105MHN blackNAM
85Acc 754ZMB
85Acc 754ZMB
48Acc790KEN
48Acc790KEN
20Acc118CIV
20Acc118CIV
20Acc118CIV
10Acc 1276CAF
10Acc 1276CAF
10Acc 1276CAF
69Acc286NGA
69Acc286NGA
69Acc286NGA
50Acc 792KEN
50Acc 792KEN
49Acc793KEN
49Acc793KEN
49Acc793KEN
45Acc 231GHA
45Acc 231GHA
45Acc 231GHA
56Acc 89MLI
56Acc 89MLI
56Acc 89MLI
50Acc 792KEN
6Acc 289BEN
6Acc 289BEN
6Acc 289BEN
40Acc 536CMR
40Acc 536CMR
40Acc 536CMR
76Acc390SDN
76Acc390SDN
76Acc390SDN
74Acc335NGA
74Acc335NGA
74Acc335NGA
33Acc 484CMR
33Acc 484CMR
33Acc 484CMR
30Acc 476CMR
30Acc 476CMR
30Acc 476CMR
70Acc329NGA
70Acc 329NGA
70Acc 329NGA
60Acc 32NGA
60Acc 32NGA
60Acc 32NGA
4Acc144GHA
4Acc144GHA
4Acc144GHA
3Acc 9NGA
3Acc 9NGA
3Acc 9NGA
1
Nei& Li's Coefficient
Figure 6.1.1: UPGMA dendrogram of 105 bambara groundnut genotypes revealed by UPGMA
cluster analysis of 20 SSR markers based on Nei and Li, 1979 similarity estimate. Bootstrap values
of 1000 replications more than 50% are shown on corresponding nodes.
175
6.3.2
Principal Component Analysis (PCO)
Eigenvalues and the cumulative percentage of the principal component case
scores were used for the analysis of 105 genotypes. Data for the first two axes
accounted for a total variation of 14.95 % (Table 6.3).
Table 6.3: PCO case scores for the population structure of the 105 genotypes
determined from each of the three samples of the 35 bambara groundnut landraces
based on 20 SSR markers
Eigenvalues
Percentage
Cumulative %
Axis
1
1.02
8.69
8.69
Axis Axis Axis Axis Axis Axis Axis Axis Axis
2
3
4
5
6
7
8
9
10
0.73 0.67 0.59 0.53 0.47 0.47 0.42 0.41 0.36
6.27 5.70 5.05 4.50 4.02 4.00 3.60 3.52 3.09
14.95 20.65 25.70 30.19 34.22 38.21 41.81 45.33 48.42
Principle coordinate analysis allowed separation of genotypes mainly based on
their areas of origin (Figure 6.2.1). All the West African landraces were in group
one on the left panel of the diagram together with one landrace from Central
Africa (76Acc390SDN), while the other one (10Acc1276CAF ) grouped with the
Southern African landraces in group 1. The Southern African, Indonesian and
East African landraces are on the right panel of the diagram in group two with the
exception of three individuals of 50Acc792 from Kenya and one individual of
49Acc793 from Kenya
Substantial variation was shown in landraces as they spread on the upper and
lower panel of the diagram from their respective groups. All individuals from the
landraces were able to be uniquely identified by the markers (figure 6.1.1 and
Appendix 9).
176
a)
PCA case scores
0.22
0.18
0.13
Axis 2
0.09
0.04
0.00
- 0.04
- 0.09
- 0.13
- 0.18
- 0.22
-0.22
-0.18
-0.13
Symbol 1
-0.09
-0.04
Symbol 2
0.00
0.04
Axis 1
Symbol 3
0.09
Symbol 4
0.13
0.18
0.22
Symbol 5
b)
Group I
Group II
PCA case scores
0.21
0.17
0.12
Axis 2
0.08
0.04
0.00
-0.04
-0.08
-0.12
-0.17
-0.21
-0.21
-0.17
-0.12
-0.08
-0.04
0.00
0.04
0.08
0.12
0.17
Axis 1
West Africa
Central Africa
East Africa
Southern Africa
Indonesia
Figure 6.2.1: The first two axes of the PCO case scores, generated from the 105 bambara
groundnut genotypes based on 20 SSR markers generated on MVSP, the first Axis accounts for
8.69 % while Axis 2 represent 6.27 % and together explain a cumulative 14.95 % of the molecular
variation. Figure a: shows a PCO plot demarcated on 5 symbols to identify three individuals from
one landrace while Figure b: shows the grouping of the five regions into two major groups. The
two cluster groups were hand drawn on Microsoft Word.
177
0.21
6.3.3
Comparison of SSR and morphological markers
Since accessions are not homogenous as some groups of individuals from
different accessions are more closely related than individuals within an accession
(Figure 6.1.1), 34 lines derived from seed from a single plant were selected to
study variation of morphological and agronomic traits following the IPGRI
descriptors (IITA, BAMNET, 2000). For comparison of molecular (SSR) and
morphological markers, PCO analysis, cluster analysis and correlation matrix was
conducted on both data set based on 20 SSR markers and 37 morpho-agronomic
traits.
6.3.3.1
Principal Component Analysis
The analysis of the percentage variation of principal component scores were used
to reveal the differences between the two marker types. The results are shown in
Table 6.4 and Table 6.5. Looking at the first two Axes suggests that
morphological markers are revealing more variation (22.36 %) as compared to
SSR markers (18.25 %) in the first two axes.
Table 6.4: PCO case scores for the population structure of the 34 bambara
groundnut selected for field studies in Botswana, analyses based on 20 SSR
markers
Eigenvalues
Percentage
Cumulative %
Axis Axis Axis Axis Axis Axis Axis Axis Axis Axis
1
2
3
4
5
6
7
8
9
10
1.16 0.98 0.92 0.72 0.62 0.60 0.58 0.53 0.49
0.45
9.90 8.35 7.85 6.12 5.33 5.09 4.95 4.52 4.22
3.82
9.90 18.25 26.10 32.21 37.54 42.64 47.59 52.11 56.32 60.14
Table 6.5: PCO case scores for the population structure of the 34 bambara
groundnut based on 37 morpho-agronomic characters, from the field experiment
conducted in Botswana
Eigenvalues
Percentage
Cumulative %
Axis Axis Axis Axis Axis Axis Axis Axis Axis Axis
1
2
3
4
5
6
7
8
9
10
2.13 1.38 1.24 1.10 0.99 0.85 0.74 0.70 0.67 0.61
13.56 8.80 7.89 7.00 6.32 5.40 4.72 4.45 4.25 3.89
13.56 22.36 30.25 37.25 43.57 48.97 53.68 58.13 62.38 66.27
178
a) SSR
PCA case scores
0.39
0.32
0.24
Axis 2
0.16
0.08
0.00
-0.08
-0.16
-0.24
-0.32
-0.32
-0.24
West Africa
-0.16
-0.08
Central Africa
0.00
0.08
Axis 1
East Africa
0.16
0.24
Southern Africa
0.32
0.39
Indonesia
b) Morphology
PCA case scores
0.47
0.37
0.28
Axis 2
0.19
0.09
0.00
-0.09
-0.19
-0.28
-0.37
-0.47
-0.47
-0.37
West Africa
-0.28
-0.19
-0.09
Central Africa
Axis 1
0.00
East Africa
0.09
0.19
0.28
Southern Africa
0.37
0.47
Indonesia
Figure 6.3.1: The first two axes of the PCO case scores, generated from the 34 bambara groundnut
landraces using MVSP for figure 6.3.1 (a) SSR marker Axis 1 represent 9.90 % and Axis 2
represent 8.35 %, figure 6.3.1 (b) Morphology marker; Axis 1 represent 13.56 % and Axis 2
represent 8.80 % molecular variation with a cumulative % of 18.25 % and 22.36 % respectively.
179
The principal component analysis for the SSR markers explains 18.25 % of the
variation in the 34 bambara groundnut lines for the first two axes. The genotypes
from Southern Africa and West Africa were clearly defined into two distinct
groups (Figure 6.3.1a). There are some few exceptions of 119Hybrid and two
Namibian lines 99SB4 4 and 100SB 16 A which grouped with the West African
lines and 40Acc536 from Cameroon which grouped with the Southern African
lines. The West African lines on the left panel of the diagram also contain two of
the lines from East Africa and two from Central Africa. Two East African lines
and the Indonesian line are grouped with the Southern African lines which are
scattered in the right pane of the graph in both the lower and upper panel.
Similarly the PCO score for the agronomic data separated the 34 bambara
groundnut lines into two major groups as in the SSR marker PCO score data. The
principle component explained a cumulative percentage for morphological
variation of22.36 % among the lines for the first two axes (Figure 6.3.1 b). There
is a demarcation between the Southern African lines from the West African lines
with the only line 40Acc536 from Cameroon which is morphologically similar to
the Southern African lines. The Southern African lines which grouped with the
West African lines were 119 Hybrid, 92AHM968NAM, and 91UniswaRed SWA
which generally produced narrower leaves with lower leaf width and lower leaf
area observed (Chapter 4 results) which suggests these lines could do well in the
West African environment with a relatively higher amount of rainfall compared to
a Botswanan environment.
Cluster analysis using the UPGMA method based on the Nei’s 1972, clustered the
34 bambara groundnut into four groups (figure 6.4.1a). Cluster one consists of a
total of 15 lines, 13 are from West Africa while two 76Acc390 from Sudan and
10Acc1276 from Central African Republic are from Central Africa. Cluster two
consists of lines from Southern Africa except 81Acc385 from Tanzania. The third
and fourth clusters consists mostly of lines from Southern African, there are also
mixed up with lines from East Africa and Indonesia with an exception of line
74Acc335 from Nigeria. This observation is largely in agreement with the PCO.
The Euclidean distance cluster analysis method based on the 37 agromorphological markers grouped the 34 landraces into three groups (figure 6.4.1b).
180
The landraces are mainly separated based on characters which contribute more
variation in bambara groundnut such as shoot dry weight, pod number, plant
height, seed number and canopy width (Chapter four). Cluster 1 consists of lines
mostly from West Africa with a mixture of Southern Africa lines. These lines
performed poorly in terms of pod yield per plants and lines 70Acc329 from
Nigeria, 50Acc792 from Kenya and 45Acc23 from Ghana, producing no yield at
all. Cluster 2 is a mixture of lines from Southern Africa, West Africa and
10Acc1276 from Central African Republic from Central Africa, these lines had a
higher number of stems per plant, relatively similar plant height and shoot
biomass. Cluster 3 consists of 10 lines which performed relatively well in number
of characters in Southern Africa (Botswana) environment and had lower petioleinternode ratio and high yield such as 88-AHM753 and 90-S19-3 from Namibia,
84Acc696 from Zambia, 81Acc385 from Tanzania and 109Bots1 from Botswana
which produced highest number of pods per plant.
Morphological markers showed that they could to some extent separate landraces
based on their areas of origin, which does reveal the importance of area of origin
on the selection of bambara groundnut. There are some striking similarities
between the SSR marker and the morphological marker cluster analysis; there was
largely a clear demarcation between the Southern African landraces and the West
African landraces.
181
a) SSR marker dendrogram
3Acc9NGA
4Acc144GHA
6Acc289BEN
10Acc1276CAF
60Acc32NGA
70Acc329NGA
30Acc476CMR
33Acc484CMR
45Acc231GHA
76Acc390SDN
117VSSP6CMR
56Acc89MLI
20Acc118CIV
40Acc536CMR
69Acc286NGA
81Acc385TZA
84Acc696ZMB
99SB4-2NAM
100SB16ANAM
48Acc790KEN
85Acc754ZMB
92AHM968NAM
105MHNblackNAM
109BWA1BWA
88AHM753NAM
95DODRTZA
91UNISRSWA
118RamayanaIND
50Acc792KEN
119Hyrid
74Acc335NGA
90S19-3NAM
104S-1913NAM
113BWA5BWA
54
85
105MHNblackNAMMW
1.01
b)
64
0.82
0.64
Nei’s 1972 coefficient
0.45
I
II
III
VI
0.26
Morphological marker dendrogram
57
105MHNblackNAMMW
88
6.13
5.46
4.79
Euclidean coefficient
4.13
3Acc9NGA
6Acc289BEN
56Acc89MLI
20Acc118CIV
30Acc476CMR
74Acc335NGA
45Acc231GHA
I
70Acc329NGA
117VSSP6CMR
50Acc792KEN
92AHM968NAM
91UNISRSWA
99SB4-2NAM
118RamayanaIND
104S-1913NAM
33Acc484CMR
10Acc1276CAF
119Hyrid
85Acc754ZMB
II
40Acc536CMR
69Acc286NGA
113BWA5BWA
60Acc32NGA
4Acc144GHA
48Acc790KEN
84Acc696ZMB
76Acc390SDN
81Acc385TZA
III
95DODRTZA
109BWA1BWA
88AHM753NAM
100SB16ANAM
105MHNblackNAM
90S19-3NAM
3.46
Figure 6.4.1: Cluster analysis of 34 bambara groundnut analysis with Unweighted pair group method with
arithmetic method (UPGMA) were generated using NTSYS version 2.1, Figure 6.4.1 (a) is SSR marker
dendrogram generated from 20 microsatellites markers based on Nei’s 1972 distance estimates, Figure (b) is a
morphology dendrogram generated on 37 morpho-agronomic traits generated on Euclidean distance
estimates.
182
6.3.4
Genetic distance estimates between landraces
The genetic distance estimates from SSR marker were calculated using the Nei’s
1972 coefficient on Popgene version 1.31. The coefficient ranged from 0.262 to
1.846. The lowest genetic distance was between 30Acc476 and 33Acc48 both
from Cameroon while the highest genetic distance was found between 69Acc286
from Nigeria and 95DodRed from Tanzania.
The lowest genetic distance based on Euclidean was between 45-Acc231 from
Ghana and 70-Acc329 from Nigeria at 12.00, and the highest genetic distance
morphologically was between landraces 10Acc1276 from Central African
Republic and 95DodRed from Tanzania at 49.00.
6.3.5
Correlation between molecular and morphological distance estimates
In carrying out a comparison of the distance estimates between the two marker
types, a correlation between distance estimates matrices established by using
Nei’s1972 coefficient for SSR markers and Euclidean for morpho-agronomic
traits was made using both the Mantel test, Pearson product-moment correlation
coefficient and Spearman (rank) correlations coefficient. Correlation analysis was
conducted for the 35 genotypes that were analysed with 20 SSR markers in the
agronomy bay experiment, and 34 lines that were planted in the field and among
the best 5 lines that were planted in the growth room experiment. A detailed
chronology of how the experiment was conducted is given below in figure 6.5.2.
Highly significant but low correlations were recorded in the agronomy bay (r
=0.139; P <0.006), in the field experiment (r =0.122; P < 0.001) while a relatively
higher correlation (r =0.612; P <0.001), was observed in the controlled growth
room experiment based on Mantel test (Table 6.6).
183
A)
0.17
Genetic
0.10
0.02
0.05
-
0.13
-
- 0.16
-0.09
- 0.03
0.04
0.11
Morphology
B)
Figure 6.5.1: A scatter plot of correlation for morpho-agronomic and molecular marker based on
Pearson, Spearman (rank) and Mantel test, analysis conducted on (A) NTSYS pc version 2.1 and
(B) on SPSS version 16, the morphological markers were based on Euclidean distances estimates
while the molecular marker were on Nei’s 1972 coefficient.
184
Table 6.6: Correlation between morpho-agronomic markers and molecular
markers for the 35 and 34 bambara groundnut genotypes based on 20
microsatellites and 37 morph-agronomic characters, and for 5 lines based on 12
markers and 22 morpho-agronomic characters.
a) Glasshouse
Marker
SSR
Morphology
N
P-value
b) Field experiment
SSR
Morphology
N
P-value
c) Growth room experiment
SSR
Morphology
N
P-value
Pearson correlation
Spearman rank correlation
Mantel test
Morphology
Morphology
Morphology
0.767
0.771
0.139
1
1
1
595
595
595
0.001
0.001
Morphology
Morphology
Morphology
0.112
0.105
0.122
1
1
1
0.006
561
561
561
0.008
0.013
0.030
Morphology
Morphology
Morphology
0.665
0.461
0.612
1
1
1
435
435
435
0.001
0.001
0.001
N = number of values in the matrix
6.3.6
Molecular variance among bambara groundnut landraces.
The partitioning of population diversity within and between populations were
analysed on Analysis of Molecular Variance (AMOVA) based on the two groups
(Figure 6.2.1). Group one consists of all genotypes from West Africa, Central
Africa except one, and four genotypes from East Africa while group two consists
of all genotypes from Southern Africa, all genotypes from Indonesia and most of
the genotypes from East Africa. AMOVA revealed that most of the variation
resides among individuals within populations (87.30 %; P <0.001), there is
significant variation of 11.58% that exists among the two groups, while only 1.12
% is within individual genotypes (Table 6.7). Similar observations were made in
the PCoA figure 6.2.1a and b, where most of the differentiation among genotypes
was observed.
185
Table 6.7 Analysis of molecular variance (AMOVA) for the 105 bambara
groundnut genotypes for the comparison based on the five selected regions,
analysis conducted using Arlequin version 3.5
Source of variation
Among populations
Among individuals within populations
Within individuals
Total
6.3.7
d.f.
1
103
105
209
Sum of squares
88.626
1152.85
7.5
Variance components
0.738Va
5.560Vb
0.017Vc
Percentage of variation P-value
11.58
<0.001
87.3
<0.001
1.12
<0.001
Breeding strategy
From the 119 bambara groundnut accessions that were planted in the agronomy
bay in 2008 season, three individual genotypes from each accession were selected
from the 34 landraces that were analysed with a set of 20 microsatellites markers
and this made up the first selection (Figure 6.5.2 a, b).
Field work was then
conducted on the 34 lines derived from seed from single plants selected from the
previous year’s experiment and planted in the field at Botswana College of
Agriculture, (Botswana) and this made up the second cycle of selection for
bambara groundnut lines (Figure 6.5.2 c). The third cycle of selection was
conducted on the five best lines that were selected from a field experiment in
Botswana, selected after a ranking analysis. A growth room experiment was
conducted for characterisation, evaluation and genetic analysis of these set of
lines. Five individual genotypes from these lines were analysed with a set of 12
markers (Figure 6.5.2 d).
186
A) 119 bambara groundnut accessions were
selected, planted in the agronomy bay
and analysed with a set of 12 SSR
markers
B) 35 landraces were analysed with a set with
20 SSR markers, 3 individuals were selected
per genotyping. ie 105 genotypes
C) The 34 lines were selected for
field work experiment in Botswana
D) 5 Individual genotype from
the best lines were analysed
with a set of 12 markers
Figure 6.5.2: Schematic diagram showing the selection strategy for the three round of
selection of bambara groundnut
The five landraces, 81-Acc385 from Tanzania, 84-Acc696 from Zambia, 88AHM753 and 90S19-3 from Namibia and 109BWA1 from Botswana were
followed through for three generations from the agronomy bay experiment, field
experiment in Botswana and control growth room experiment. Twelve
microsatellites markers were employed in the molecular analysis of the five
landraces in the first season of selection and revealed an average genetic distance
of 0.404 with a range of 0.222 for 88-AHM753 from Namibia to 0.751 for 81Acc385 from Tanzania based on Nei, 1972 genetic distance estimates on Popgene
(Table 6.8).
187
Table 6.8: Mean and range of the genetic distances values for three different
selection cycles of bambara groundnut from single seed descent estimated based
on 12 microsatellites markers using Popgene version1.31 (Yeh and Boyle, 1997).
Selected lines
81-Acc385TZA
84-Acc696ZMB
88-AHM753NAM
90-S19-3NAM
109-BWA1-BWA
N
3
3
3
3
3
First cycle selection
Mean
Ho-He
0.751
0.000-0.356
0.314
0.000-0.222
0.222
0.000-0.267
0.347
0.000-0.311
0.389
0.000-0.311
Range
0.287-1.049
0.206-0.403
0.198-0.248
0.305-0.405
0.331-0.431
Genetic distance estimates
Seecond cycle selection
N
Mean Ho-He
7
0.000 0.000
4
0.000 0.000
7
0.000 0.000
7
0.000 0.000
7
0.000 0.000
Range
0.000
0.000
0.000
0.000
0.000
Third cycle selection
N
Mean Ho-He
6
0.000 0.000
6
0.000 0.000
6
0.000 0.000
6
0.000 0.000
6
0.000 0.000
Range
0.000
0.000
0.000
0.000
0.000
N = Number of individual sample
Variability within bambara groundnut landraces has been reported before, and has
been attributed to a range of causes, from low levels of outcrossing, followed by
the natural development of inbred lines, through to the mixing of seeds during
harvesting or from markets, especially those of same colour (Massawe et al.,
2005; Mayes et al., 2009). As a breeding strategy for inbreeding crops like
bambara groundnut, it is advantageous to obtain pure homozygous lines with good
attributes. As expected, in the second and third round of selection pure lines were
selected through single plants (Table 6.8). There was no observed or expected
heterozygosity in the second and third round of selection. This data strongly
suggests these genotypes are now essentially pure lines or effectively varieties.
The data for the five lines are listed in (Appendix 11).
188
6.4
Discussion
The twenty selected microsatellites markers showed polymorphism among the set
of 105 genotypes of the 35 bambara groundnut landraces. To determine the level
of polymorphism in markers two distinct quantities can be used, that is
heterozygosity and polymorphism information content (Shete et al., 2000).
Markers with polymorphic information content above 0.5 are considered highly
informative (Botstein et al., 1980). The average polymorphic information content
found among the 20 selected markers of 0.67 shows that most of the markers were
highly informative in this material. Approximately 75 % of the markers (15) had
polymorphic information content more than 0.5 while only 15% (5) markers,
marker D8, marker E7, marker 23, marker 21 and maker 30 had lower PIC values
less 0.5.
Bambara groundnut is a self-pollinating crop, so it is not surprising to have shown
an inbreeding coefficient close to 1(Table 6.1). Similar findings in bambara
groundnut were observed by Basu et al., (2007). They found an inbreeding
coefficient 1, among the 8 markers they studied except for two which had
heterozygosity (Ho) of 0.28 and 0.06 against an expected heterozygosity (He) of
0.82 and 0.78, respectively. In other related crops, such as pigeonpea, Kuroda et
al.,(2006) in Japan used 20 microsatellites to study the genetic diversity of 616
individuals of 77 wild soybean (Glycine soja) and 53 varieties of cultivated
soybean (Glycine max). They recorded the expected heterozygosity (He) for wild
soybean of 0.870 and 0.496 for cultivated accessions, and an observed
heterozygosity (Ho) of 0.000 for cultivated and 0.018 for wild accessions,
suggesting that both accessions are predominantly inbreeding. In common bean
(Phaseolus vulgaris), Blair et al., (2009) observed an expected heterozygosity of
0.64 and observed heterozygosity of 0.049 among 604 genotypes analysed using a
set of 36 SSR markers.
To quantify the genetic diversity among the selected bambara groundnut
accessions both molecular analysis and agro-morphological data were analysed by
cluster analysis using the UPGMA method and principal coordinates analysis.
For the cluster analysis of the 105 genotypes 43 nodes had bootstrap values of
more than 50%, which could indicate that higher number of markers may be
189
requested to increase the robustness of the cluster analysis. Twenty nodes were
supporting three individuals within a genotype which also indicate that some
accessions are highly heterogeneous. The results were coherent with the origin or
history of the landraces; in addition all genotypes were uniquely identified, with
highest similarity among genotypes observed among landraces from the same
country.
For the 34 bambara groundnut lines, the molecular marker cluster analysis on
Nei’s 1972 distance estimates and morphological markers based on Euclidean
distances were compared. Molecular markers grouped landraces into four clusters
while morphological markers grouped them into three clusters. Basically there
were some similarities between the clusters produced, with most landraces
grouping based on their areas of origin. Even when there were some mixtures of
landraces within a cluster, most of the landraces found in that particular cluster
had a common area of origin. Similar findings in bambara groundnut have been
observed when using RAPDs by (Massawe et al., 2003) and AFLP by Massawe et
al., (2002), where they showed that landraces were clearly grouped based on their
areas of origin. This is an indication of the importance of adaptation on the
genetic variation in bambara groundnut. A similar pattern of observation was also
seen in the principle coordinate analysis.
One of the best options for crop improvement is through the hybridisation of
genotypes with reasonable genetic distance and desirable agronomic traits
(Parsaeian et al., 2010). The molecular markers identified line 69Acc286 from
Nigeria and line 95DodRed from Tanzania to be genetically far apart, while the
Euclidean distance estimates identified lines 10Acc1276 from Central African
Republic and 95DodRed from Tanzania as dissimilar. However, it is those
landraces that have been found to be agronomically superior that could be used in
a breeding programme as parents, such as S19-3 from Namibia, 76Acc390 from
Sudan and 33Acc484 from Cameroon were some of the lines that produced higher
number of pods per plant in Botswana environment.
The genetically distant lines from Southern Africa and West Africa, which had
been found to cluster in different groups by both morpho-agronomic and SSR
marker type, could significantly lead to an increase in bambara groundnut
190
performance in terms of yield, if the wide cross leads to good transgressive trait
segregation which can be selected from to develop new varieties and possibly
better adapted landraces to both environments (regions).
The correlations observed between morphological and genetic distance estimates
could suggest that SSR markers could be used as substitute for phenotypic
measurement (Ramakrishnan et al., 2004). The highly significant and positive
correlation observed for the agronomy bay experiment, field experiment and in
the controlled growth room data matrices revealed that the morphological genetic
distance could reflect the genetic distance estimates. The positive correlation
found, is of significant importance to plant breeders especially for underutilized
crops with lack of resources since morphological markers are the standard
markers used regularly for crop improvement.
The positive correlation between SSR data marker genetic distance estimates and
morphological markers have been observed in soybean at r = 0.31 (Priolli et al.,
2010). In bambara groundnut, Ntundu et al., (2004) recorded a positive
correlation of r = 0.40 between morphological markers and AFLP among 100
bambara groundnut landraces in Tanzania, possibly due to low heterogeneity
among the Tanzanian landraces.
The close agreement between molecular markers and SSR markers in the
controlled growth room experimentsuggests for genetic diversity studies in
bambara groundnut pure lines or varieties, SSR markers and for morphological
markers were in good correspondence. The observation where there is a
simultaneous increase in the phenotypic distance and molecular distance has been
noted before and termed the ‘triangular relationship’ observed between furthest
points (Burstin and Charcosset, 1997). This phenomenon was only observed in
the correlation conducted in the growth room experiment (Appendix 10.2a).
Molecular markers proved to be more robust and reliable in genetic diversity
analysis as shown in both cluster and PCO analysis where landraces were clearly
defined (Figure 6.1.1 and Figure 6.2.1). A combination of molecular markersand
morphological markers as revealed by high Pearson and Spearman correlations in
the agronomy bay and in the controlled experiment shows that they could be used
191
to compliment the use of morphological markers in identifying landraces with
desirable characters (Table 6.6 and Appendix 10).
AMOVA was conducted to investigate variation between two groups, group one
consists of mainly countries from West Africa and Central Africa while group two
consists mainly of countries from Southern Africa, Indonesia and East Africa.
AMOVA revealed that most of the variation was among individual populations,
and still some significant variation exists among the two groups. This implies that
differentiation is skewed to more variation among genotypes.
In other self-
fertilising leguminous species higher within population variations were reported.
Okori et al., (2005) identified 92.16% within population and 7.82% among
populations using AFLP in 41 pigeonpea (Cajanus cajan) landraces.
A number of factors have been attributed to the genetic diversity in bambara
groundnut. There was a higher within landrace variation observed in this study,
which is most likely due to the fact that bambara groundnut basically exists as
‘inbred lines’ due to its highly inbreeding nature. Farmers or consumers in
different regions prefer specific landraces due to colour type and taste and various
traditional beliefs differ in different regions of Africa (Sesay et al., 2003).
Generally bambara groundnut is planted by small scale farmers with small
hectares and mostly for family consumption, with a little for sale. Thus the
exchange of seeds and their movement to other countries may not be as
pronounced. Similar observation in bambara groundnut by Massawe et al.,(2003),
when using RAPDs identified a significantly higher among landrace 71.25%
compared to 28.67% to difference within individual landraces.
Analysis of landraces based on intra-landrace diversity identified 20% (7)
landraces with at least one genotype different among the three selected genotype
per landrace. The more diverse lines had a recording of more than 1.5 alleles per
locus (Table 6.2).This character of bambara groundnut has implications in pure
line selection for variety development (Basu et al., 2007).
In the selection of pure lines of bambara groundnut 12 microsatellites were
employed among five landraces. In the initial genetic diversity analysis of the
landraces in the first cycle of selection the lowest residual heterogeneity were
192
recorded in lines 84Acc696 (0.222) from Zambia and the highestin81Acc385
from Tanzania (Table 6.8). However, when Stadler, (2009) investigated multiple
samples of six bambara groundnut using DArT markers, S19-3 was identified as
the genetically narrowest landrace. LandraceAHM753 has been used in previous
bambara groundnut projects as core landrace, it has undergone a number of
selections and thus it is relatively pure compared to other lines.
109-BWA1from Botswana, was selected from among four landraces from farmers
and characterised in field experiments based mainly on seed colour, leaf
morphology stability and in other seasons it was selected for grain yield and days
to maturity (Chui et al., 2003). The seeds were generally bulked and no pure line
selection for the landrace was done, hence some relatively higher heterogeneity
were discovered
6.5
Conclusions
In this study the extent of genetic diversity within and among 35 bambara
groundnut accessions from wide geographic range has been investigated. The
application of cluster and PCO analysis revealed that bambara groundnut
individuals mainly grouped based on their area of origin. The genetic distance for
both marker types could be used to identify those landraces that are genetically
distant from each other, and there was a good correspondence between the two
techniques.
Genetic variation in bambara groundnut is significantly higher within individual
landraces compared to among populations, thus a number of landraces could be
identified which are relatively pure for use in the selection as pure lines in
bambara groundnut breeding.
193
CHAPTER SEVEN: General discussions
7.1
Introduction
Although bambara groundnut is cultivated throughout tropical regions of Africa,
and is very important as a food security crop in the sub-Saharan Africa no
concerted effort has been put in place to develop it into improved landraces or
varieties. Farmers rely on landraces which can be inherently low yielding due to
poor physical and genetic quality of seeds, low germination rates and poor crop
management among other constraints. For a comprehensive breeding effort in
bambara groundnut it is necessary that there are available molecular markers to
use in the study of the genetics of bambara groundnut, even if only for quality
control within the breeding programme. Ideally, there must be an understanding
of the genetic diversity and population structure in order to be able to identify
genotypes that could merit further selection and use as varieties and parental
materials. A study of the morphology and genetic diversity of bambara groundnut
would assist in identifying the best method for selecting bambara groundnut for
breeding purposes and to identify accessions for further selection.
In this chapter, the aims of the research work are presented again, before
discussion of the progress and problems encountered. The main areas that
this research investigated are:I
The development of microsatellite markers and their characterisation;
these markers represent an additional tool for use in the genetic analysis of
bambara groundnut, in mating systems studies, genetic diversity studies,
population structure analysis for breeding and for conservation of the crop.
II
Phenotypic diversity and morphological evaluation were conducted to
assess the extent of phenotypic variability and to indicate the genetic advance
possible with the aim of selecting landraces suitable to be grown in the semi-arid
environment of Botswana
III
The application of morphological markers is the standard and most
frequently used tool in genetic diversity studies in underutilised species, including
194
bambara groundnut. For bambara groundnut, the application of molecular markers
is not at an advanced stage and due to technical and costs implications their use is
not yet widespread. Therefore, the comparison of the two techniques could bring
an insight as to which is the most suitable to use and in what circumstances. The
study aimed to investigate the genetic diversity and relationships among the
selected landraces using morphological and SSR markers and also to assess any
correlation between distance estimates based on morphological and molecular
markers.
IV
For a well-defined breeding programme a thorough knowledge of the
population structure of the landraces needs to be understood. The movement of
bambara groundnut landraces from the area of origin to other regions, such as
Southern Africa, East Africa, Central Africa and even Indonesia is likely to have
an impact on population differentiation. The informal movement of seeds material
between farmers within the same countries, between countries and even across
regions also affects the population structure of the crop. However, little
information was known regarding the relationship of the landraces between
regions and countries.
Brief description of the chapter and the main AIM of the experiment
One of the main aims of this study was comparing the utility of SSR and
morphology techniques for genetic diversity analysis of bambara groundnut
landraces and to establish the relationship between the two approaches, if one
exists. The choice to use either morphological markers or molecular markers or
even a combination of the two markers is an important consideration. In practice,
it is often practically difficult and restrictively expensive to use molecular markers
within a developing world context, unless there is a substantial gain to be made in
breeding. The importance of a comparison of different marker systems is to assist
in making informed decisions as to which marker is best to use in germplasm
characterisation and plant breeding. However, the development of an
understanding of breeding systems and germplasm population structures could
allow a more focused breeding effort, even without further application of
molecular markers. Potentially, markers could aid the selection of germplasm for
breeding, quality control within breeding programmes and, potentially direct
195
selection via Marker Assisted Selection (MAS). In this study a number of
experiments were conducted to make this evaluation possible.
7.2
Recap of the study
In the first parts of this thesis (Chapter 2 and 3) a total set of 75 microsatellites
were characterised and used to investigate the genetic diversity of a set of 24
bambara groundnut landrace accessions. The markers were checked for the
presence of null alleles, stutter-bands, spectral overlap and binning was also
conducted as a precautionary measure, to identify suitable markers for further use
(Chapter 2). A set of 68 markers were found to be polymorphic and produced
robust amplification and consistent results. The markers were also compared to a
DArT marker dataset that was previously generated from the same 24 genotypes
(Chapter 3).
In the second part of the thesis (Chapter 4), phenotypic and morpho-agronomic
diversity studies were undertaken on selected landrace accessions planted in the
agronomy bay (greenhouse) in the UK, with a subset later taken for field studies
in Botswana. Several analyses were undertaken to investigate the phenotypic
diversity of the landraces, such as through the generation of Shannon weaver
diversity indices, principal component analysis, cluster analysis and Pearson
coefficient correlation studies.
The third part of the study (Chapter 5) determined the population structure of
bambara groundnut in five regions, four from Africa and one from Asia, using a
set of 12 pre-selected microsatellites developed and characterised in this study
(Chapter 2 and 3 ).
The fourth part of the study (Chapter 6) assessed the use of both molecular and
morphological markers in genetic diversity analysis of bambara groundnut.
Morphological markers characterised in the Agronomy bay, in the field
experiment and controlled growth room experiment were compared with
molecular marker analysis from each respective experiment.
196
7.3
Microsatellites development and characterisation
A set of 68 markers were found to be polymorphic and had consistent
amplification results. These markers represent a new tool for use in the genetic
analysis of bambara groundnut, in mating system studies and for genetic diversity
studies of population structure for breeding and conservation purpose of the crop.
The set of 24 genotypes were used to compare SSR and DArT markers for an
assessment of overall bambara groundnut genetic diversity. The two techniques
have proved useful in the genetic diversity analysis of the selected material. SSR
markers showed slightly higher genetic differentiation between the landraces with
lower similarity coefficients at an average of 0.65 compared to 0.71 for DArT
markers. Similar findings were observed in cassava (Manihot escuenta) by
Hurdato et al., (2008) when comparing the utility of 1000 DArT markers and 36
SSR markers to assess 436 cassava accessions, 155 originally from Africa and
281 from Latin America.
In this study DArT markers appear to provide clearer genetic resolution when
compared to geographical accession origins, as compared to SSR markers. DArT
marker explained a higher percentage of the molecular variation in the first two
PCO axes, show a wider range of genetic distance estimates between genotypes
and more clearly defined grouping of individuals through both PCO and UPGMA
analysis.
A highly significant correlation was found using the Pearson correlation
coefficient (r =0.34), Spearman rank correlation (r = 0.34) and the Mantel test (r =
0.35) between the two techniques. This indicates a reasonable correspondence
between the genetic distance estimates for the two markers types.
Genetic
distance estimates between genotypes are useful in the selection of crosses that
could generate the best performing lines (Bertan et al., 2010); therefore both
marker types can be useful in the crop improvement of bambara groundnut.
Most of the comparisons between SSR markers and other markers suggest that
SSR markers usually reveal a higher level of polymorphism while, other markers
(for instance AFLP) have a higher marker index (requires less number of primer
combinations to screen whole genome) (Spooneret al., 2005). DArT markers are
197
relatively new markers so comparative data with other marker types is quite
scanty in the literature and in bambara groundnut this comparison had not yet
been conducted. Since both markers effectively differentiated the landraces they
have proved to be useful in genetic diversity studies of bambara groundnut. The
utility of the two marker types could be similar (Hurtado et al., 2008), especially
when more samples and a larger number of markers are employed.
Markers developed in this study will be used to investigate the genetic variability
that exists in bambara groundnut both within and between landraces, and also to
investigate correlations between genetic and morpho-agronomic traits in bambara
groundnut.
A summary of achievements made in the chapters 2 and 3:
Characterisation of 68 SSR markers for bambara groundnut
Establishment of the relationship between and genetic variation detected
by DArT markers and SSR markers in a set of genotypes representative of
the available germplasm
198
7.4
Morphological characterisation
An assessment of morpho-agronomic trait variation of 35 bambara groundnut
lines selected from the agronomy bay experiment (glasshouse) and the subset of
34 lines selected in the field experiment was explored using multivariate analysis
(i.e cluster and principal component analysis), Shannon weaver diversity indices
and Pearson correlation coefficients. These estimates potentially provide useful
information that facilitates the exploitation of bambara groundnut germplasm.
Estimations of heritability and genetic advance were conducted, to allow the use
of selection indices in order to identify the best performing lines.
Multivariate techniques, such as cluster analysis and principal component analysis
are efficient tools used in the estimation of quantitative variation in crops (Rahim
et al., 2008). These techniques have been explored in common bean to group
accessions based on their yield performance (Salehi et al., 2008).
Agro-morphological data for bambara groundnut was standardized to reduce the
effect of scalar differences between traits on correlation estimates of
morphological traits. Tabachnick and Fidel (2007) argue that conversion of data to
standardized scores makes the data independent of scale measurement and sample
sizes used. Correlation coefficient analysis measures the association between
variables, the Pearson correlation coefficient is one of the most commonly used. It
describes the direction and degree to which one variable is linearly related to one
another (Bolboacă and Jäntschi, 2006).
In both the agronomy bay and field experiment a number of strong and positive
correlations were observed within the vegetative characters such as petiole length
and plant height (r = 0.91) and also within the seed yield related traits and seed
number per plant and pod number per plant (r = 0.98). Similar observations were
made by Kobraee et al., (2010), in Chickpea who observed high correlation
between grain yield per plant and plant height (r = 0.827), and a high correlation
between grain yield per plant andplant height at (r =0.813). Farshadfar and
Farshadfar, (2008) in chickpea (Cicer arientinum) where they observed that their
highest correlation were between seed yield per plant and pod number plant (r =
0.78). In pigeonpea, Vange and Moses (2009) recorded their highest correlation
between grain yield and pod dry weight (r = 0.87). This suggests that correlation
199
is also dependent on the characters observed. Correlations found between
vegetative traits and seed yield were generally low even for those which are
statistically highly significant. A higher and significant correlation of r = 0.61 was
observed between pods number per plant to both shoot dry weight and canopy
spread while a slightly higher and significant correlation r = 0.67 was observed
between shoot dry weight and canopy spread in the agronomy bay experiment.
Similar traits recorded lower correlations in the field experiment, pod number per
plant and shoot dry weight had an r = 0.13, while pod number per plant and
canopy spread recorded r = 0.03, shoot dry weight and canopy spread hadr = 0.30,
this observation indicates that selecting for medium canopy spread lines (which
are the bunch and the semi-bunch type) would not significantly increase pod yield
per plant since the variation for this and correlations are very low.
The characters were subjected to Principal component analysis (PCA), using a
correlation matrix, to identify characters showing the highest explanation for the
major variation in morphological traits given in the first eigenvalues which could
be used for characterising bambara groundnut landraces. PCA is an analysis of a
matrix consisting of variances and covariances or correlations among variables to
come up with smaller sets of components that summarise the correlations (Fenty,
2004). The first component extracted accounts for the maximum amount of total
variation observed, subsequently followed by the second principal component
(Jollie, 2002).
The PCs are uncorrelated they are orthogonal, therefore are affected by the sizes
of correlations, and possibly some of the factors that affect correlation too, such as
size of the data sets. The variables which are highly correlated will then tend to be
concentrated into one component. Most of the variation was accounted for by the
vegetative traits and seed yield traits, in the agronomy bay experiment for PC1
and PC2, respectively. Higher loadings of vegetative traits such as shoot dry
weight, leaf area, plant canopy and plant height, together with pod number per
plant were observed both in the agronomy bay and field experiment. Ntunduet al.,
(2006) in a study in 100 bambara groundnut landraces also observed that most of
the most of the vegetative traits were loaded in PC 1 while the seeds characters
are highly loaded in PC 2, which is an indication that similar variation was
observed between these two experiments. Therefore the vegetative, pod and seed
200
yield characters are some important traits useful for the characterisation of
bambara groundnut germplasm.
Landraces were mainly clustered based on the vegetative, pods and seed traits but
the vegetative traits had higher loadings such as shoot dry weight, petiole length
and leaf areaas they appeared in the first principle component (PC1) for both
experiments (Table 4.17 and Table 4.1.8). These traits also differed based on
regions of origin as shown in the cluster analysis in the agronomy bay experiment
(figure 4.2.2). The cluster from the field experiment was slightly different,
because landraces were also grouped based on their seed yield performance as the
latter dominated the loadings in PC2 in the field experiment. Lower yield (number
of seeds per plant and number of pods per plant) were observed in the field
experiment. Overall the principal component and the clusters produced in the two
experiments were slightly different, possibly due to the effect of environmental
conditions in the field compared to the agronomy bay which reduced the variation
expressed in some traits. In addition, the Botswanan material could have reduced
phenotypic and genetic variability since they were selected from single plants,
while the analysis of three individuals in the Agronomy bay experiment from each
landrace almost certainly means that non-identical genotypes were examined.
Knowledge of heritability estimates gives an indication of the expected
performance of progenies (Bertoldo et al., 2010) thus it is a useful parameter for
selection of desirable traits. Numerous indices are at the breeder’s disposal to use
for selection, each with different characteristics (Strefeler and Wehner, 1986). In
this study a simple selection index (SI) was employed to select the best
performing genotypes based on leaf area, shoot dry weight, seed number per plant
and pod number per plant based on the same genotype from three replications.
The genetic advance obtained from the field experiment was used as an economic
weight in the SI and this makes the index more robust Chapter 4, (Table 4.2.3).
The data set for the genotypes were then ranked in such a way that the best lines
were concentrated at the top and the best five performing lines were identified.
Bertoldo et al., (2010) used two selection index methods, the Smith (1936) and
Hazel (1943) method and the one developed by Pesek and Baker (1969) in
common bean to estimates genetic gain among 23 accessions based on 7
201
characters. The Smith (1936) and Hazel (1943) is a linear combination of traits of
economic importance based on the estimation of correlation between characters.
While in Pesek and Baker (1969), the technique replaces the economic weight of
traits by those values determined by the breeders. The selection index (SI) used in
this study has attributes of both techniques as it is a linear combination and also
contains an option of using the economic weight of selected characters of interest
for selection. It has been used successfully in the selection of cassava seedlings by
Ojulong et al., (2010).
Achievements made in the characterisation and evaluation of bambara
groundnut
Shoot dry weight, leaf area, number of pods per plant and number of seeds
per plants have proved to be useful traits for the selection of bambara
groundnut.
The selection index (SI) and Duncan Multiple Range Test identified 5
lines that can potentially be used as varieties in Botswana. These lines
have proved to have potential as they performed well in a Botswana
environment and could be useful as new sources of germplasm for
Botswana.
The morphological markers grouped genotypes which appeared to relate to
their regions of origin, and this was influenced by traits with higher
loading in PCA especially the vegetative traits and seed and pod related
characters.
202
7.5
Genetic diversity based on SSR markers and a comparison with
morphological characters
The comparison of 20 SSR markers for genetic diversity with 37 morphoagronomic characters was conducted in the agronomy bay and in the field
experiment, respectively. The genetic distance estimates for the two techniques
and the dendrograms produced based on the genetic distance estimates differ. The
morphological markers revealed slightly higher genetic distance estimates and
greater genetic diversity, while the molecular marker showed relatively smaller
genetic distance estimates and lower genetic diversity. The principal component
analysis revealed higher explanation of trait diversity in the first two components
for morphological markers (33%) as compared to the genetic (SSR) characters
(19%). This phenomenon was also observed in groundnut by Krishma et al.,
(2004). The use of principle component (PCA) and cluster analysis clearly defined
the 34 landraces tested based on areas of origin, Chapter 6, section 6.3.3.1 (figure
6.3.1a and figure 6.3.1b) with the major groupings based on a separation between
the Southern African and the West African landraces.
However, some landraces were found in the West African cluster but
morphologically were similar to the Southern landraces and vice versa. This could
be an indication that some landraces are adapted to climatic conditions which
occurs in both regions. For example a Sudanese landrace 76Acc390 is ranked
number six in terms of the vegetative and seed yield in a Botswanan environment
(Table 4.2.5) and thus it is grouped together with some highly ranked lines such as
81Acc385 from Tanzania, 90S-19 from Namibia and 109BWA 1 from Botswana.
Selection of bambara groundnut for breeding purposes, based on area of origin
could potentially be misleading as some landraces from different regions can be
morphologically similar. In these cases the use of molecular analysis could be a
better option. For instance, 91UniswaRed from Swaziland and 92AHM968 from
Namibia were grouped together with the West African lines in PCoA, figure
6.3.1a but in the molecular PCoA they are ‘correctly’ placed with the Southern
African lines. Such lines where there is a clear discordance between morphology
and molecular markers could be of interest for breeding for particular
environments, but maintaining maximal genetic diversity between parents.
203
Relatively high and highly significant correlations were detected between the
genetic distance estimates derived from SSR markers (Nei’s 1972) and Euclidean
distances derived from agro-morphological measurements in the agronomy bay
and growth room experiment using the Pearson correlation and Spearman’s rank
correlation. Mantel’s test detected a weak but highly significant correlation (r =
0.139; P< 0.006) in the agronomy bay (glasshouse) experiment, and (r = 0.122; P<
0.03) in the field experiment.
In the growth room experiment the molecular and morphological markers
revealed highly significant correlations that were tested for significance on a twotailed test. The correlation was high for Pearson correlation r = 0.665, and
Mantel’s test r = 0.612 but moderate for Spearman’s rank correlation
r =
0.461.Low but highly significant correlation (r =0.1) were recorded in the field
experiment. The low correlation could reflect low genomic coverage and high
variability of the environment (Cheverud, 1988; Brown, 1997).
The distance estimate methods employed in the correlation tests clearly have
some impact on the final analysis. In comparing the two examples of Euclidean
distances both defined by Pythagoras theorem like simple matching coefficient
and standardized Euclidean measures the two are likely to have a higher
correlation. The simple matching coefficient is a distance measure derived from
calculating the proportion of disagreements as a categorical measure, but as
standardized Euclidean is a transformed variable to have the same variance for
distance estimates. As Pearson product-moment correlation coefficient expects to
find a linear relationship between these two, the fact that one is non-linear would
be expected to reduce the detected correlations. Another effect could be coming
from the inherent differences of the similarity coefficient used in genetic distance
estimates. Ramakrishnan et al., (2004) also observed that the scale of variation
could affect correlation between marker types. The relationship estimated for
phenotypic distances and molecular marker distances sometimes produces
conflicting results. Some researchers report no correlation while other reports
clear correlations (Burstin and Charcosset, 1997)
204
In order to determine the effect of origin of Medicago sativa spp. falcata Li et al.,
(2009), conducted both Pearson product-moment correlation coefficient and a
regression analysis on the genetic distance estimates, phenotypic distance and
geographic distances. Seven SSR markers and 7 morphological characters were
used to compare the two markers and produce an assessment of genetic diversity
of the 12 Medicago sativa spp. falcata. Their study did not reveal any correlation
between genetic distance estimates and phenotypic distance estimates, but a high
correlation (r = 0.78) was found between the morphological distance with
geographic estimates. The lack of correlation between the genetic and the
morphological markers was attributed to low numbers of samples and the narrow
range of altitudes where the samples were collected (Li et al., 2009).
The low association generally observed between morphological distance estimates
and molecular distance estimates has been attributed to a number of possible
reasons. It could be because of the application of different number of markers and
due to the fact that molecular markers target non-adaptive variation while
morphological markers are highly influenced by environment (Veira et al., 2007).
Low levels of genetic variation in the germplasm and sometimes errors in
recording could also contribute to poor correlation between markers (Karuri et al.,
2010; Giancola et al., 2002).
Comparing genetic distance estimates between marker types could reveal how
useful and efficient markers are for plant breeding purposes and identify those that
are most informative (Franco et al., 2001). Mantel tests are particularly useful in
testing genetic distance matrices, as it tests for linear or monotonic independence
between distances (Legendre and Fortin, 2010). In a comparison of the two
techniques, Legendre and Fortin, (2010) established that Pearson correlation has
more power in the analysis of linear relationships between raw data, while
Mantel’s tests have greater power in the analysis of transformed data for distance
estimates.
The statistical tools used for estimating relationships among distance matrix
variables (here, Pearson correlation and the Mantel test) have some instances
where they give contrasting results, therefore further investigation into these two
methods has been recommended (Dutilleul et al., 2000).
205
In addition these methods assume linear relationships; however some evidence
suggests that the relationship between marker distance estimates and phenotypic
distance estimates display a triangular shape. The general trend though is that
lower molecular marker distances are associated with lower phenotypic distances,
but higher molecular marker distances could be associated with either low or high
phenotypic distances (Burstin and Charcosset, 1997)
7.5.1
Pure line selection
The lack of co-dominant markers has hampered the formal assessment of
heterozygosity within bambara groundnut genotypes and this has hampered the
progress in bambara groundnut pure lines selection (Basu et al., 2007) which
could be one of the best options for variety development. Pure line selection could
be a rapid and effective way for bambara groundnut improvement, especially
given that artificial hybridization is difficult in this species (Suwaprasert et al.,
2006; Oyiga et al., 2010).
Twelve microsatellites were employed among the samples of the five landraces.
This set of lines (81-Acc385TZA, 84-Acc696ZMB, 88-AHM753NAM, 90S-193NAM and 109-Bots1-BWA) have previously been identified as the best
performers in the field experiment and were assessed on the basis of pods
numbers per plant, seed numbers per plant, shoot dry weight and leaf area
(Chapter 4). The initial genetic diversity of the first cycle of selection between
three independent seed from each landrace showed a lower average genetic
diversity of 0.222 in84-Acc696 from Zambia and the highest (0.356) for
81Acc385 from Tanzania. However, as all the lines showed no signs of residual
heterozygosity in the second round of selection; these lines could be pronounced
as true varieties.
Achievements made in the comparison of genetic and morphological markers
Substantial genetic and morphological diversity was observed in the
bambara groundnut landraces
206
There was concordance in the genetic differentiation of accessions by both
the morphological and molecular markers based on PCoA and cluster
analysis
The PCoA and cluster analysis largely grouped landraces based on their
areas of origin
All the statistical measures detected relatively high and highly significant
correlations in agronomy bay and growth room but detected low
correlation in the field experiment, with the exception of the Mantel test
which identified a highly significant but weak correlation in agronomy bay
experiment.
AMOVA revealed that the greatest amount of molecular variation was
within landraces as compared toamong regions (identified groups)
The application of microsatellites to confirm the genetic purity of single
pure line selected varieties.
207
7.6
Population structure analysis
In the present study a set of 12 pre-selected SSR markers were employed to
investigate the genetic diversity and population structure among 123 bambara
groundnut accessions, sourced from four regions of Africa, and one region from
Asia (Indonesia). As expected the highest genetic diversity among landraces was
found from West African landraces which is the putative area of origin of
bambara groundnut, Chapter 5, section 5.3.2 (Table 5.2). Principle coordinate
analysis allowed the separation of landraces, based mainly on their areas of origin,
and could identify those genotypes that show ‘transfer’ between the regions. Four
landraces from Southern Africa were clustered with the West Africa lines, and 13
from West Africa showed some grouping with Southern African genotypes. This
potentially indicates a low rate of recent movement of landraces between regions,
as shown by lower levels of between regions variation in the AMOVA at 12 %
Chapter 5, section 5.3.6 (Table 5.6). That much of the variation was between
landraces (85%) might suggest continual movement across regions, with perhaps
some subsequent selection to give the limited molecular variation between
regions. Analysis of Molecular Variance (AMOVA) was employed to partition
the 123 bambara groundnut landraces based on variation among two groups,
whereby group one consists of (West Africa and Central Africa) regions, while
group 2 consists of mainly (Southern Africa, East Africa and Indonesia)regions.
The majority of the molecular variation is mainly found within landraces at 85 %.
This also indicates that the substantial phenotypic diversity observed in bambara
groundnut is also reflected at the genotypic level, an observation not seen as
clearly with groundnut and pigeonpea, which display high levels of phenotypic
diversity but reasonably low genetic diversity.
Population structure analysis was conducted among the 87 landraces that managed
to produce enough pods and seeds to allow reasonable characterisation and these
were co-analysed using the chosen set of 12 microsatellites. Genetic distance
estimates for the 87 bambara groundnut landraces were compared with those
obtained in the analysis of 15 qualitative traits of seed and pods and the two were
tested for potential correlations. The qualitative traits are highly heritable and
208
clearly expressed in all environments (IITA, IPGIR, 2000). Differences in genetic
diversity measures between qualitative and SSR marker data were detected, with a
highly significant negative correlation, albeit very weak, analysed on Pearson and
Spearman correlations, while Mantel’s test recorded a low, negative and nonsignificant correlation, which suggests that the two markers may be giving
different estimates of the diversity present. However, in common bean (Phaseolus
vulgaris), Blair et al., (2009) found some association between SSR marker and
seed size characteristics.
Summary of achievements in population structure analysis of bambara
groundnut
Low levels of genetic differentiation between regions were found (based
on the two groups), but one to note is the non-significant differentiation
between Southern African landraces and the East African landraces, which
could be a transition route for materials from other regions of Africa.
The differentiation of landraces based on their regions shows that
landraces could be traced to their regions of origin relatively easily,
particularly with tools such as PCA.
A relatively low genetic differentiation of Asian landraces from African
landraces could possibly be a suggestion of a recentintroduction to
Indonesia from Africa.
209
7.7
Impact of the findings and future work
This research is the first SSR-based study of genetic diversity among bambara
groundnut from sub-Saharan Africa. Bambara groundnut breeders will benefit
from the knowledge generated in this study, the genetic distances/similarity
estimates for various landraces from different regions and countries are a potential
source of parents for hybridisation breeding of the crop.
The use of morphological and molecular markers provides some complimentary
information, especially where morphological markers fail to differentiate some
landraces. Sometimes it may also not be feasible to undertake a morphological
analysis; therefore the use of molecular markers may be unavoidable.
The use of these techniques is important in crop breeding since markers can be
used in the prediction of variability, estimation of heterosis and for selecting the
best lines for crosses and these may make breeding more efficient and effective.
In this study SSR markers have been found to clearly differentiate landraces based
on areas of their origin, which should assist breeders to track genotypes of interest
and access genetically diverse, but environmentally matched accessions for
breeding.
The techniques that have been used in this study - the application of SSR markers
in pure line selection and the simultaneous use of both agro-morphological marker
and molecular markers - could potentially be employed in other leguminous crop
species. The availability of molecular markers will lead to a further exploitation of
bambara groundnut germplasm and more bambara groundnut varieties will be
developed and this should contribute in poverty alleviation in Sub-Saharan Africa.
A relatively high correlation between morphological and SSR marker suggests a
good congruence between the two techniques, hence SSR markers could be used
to infer morphological diversity in bambara groundnut.
210
7.8
Future work
Establish the minimum number of markers that would be needed to
establish significant correlations between morphological and DNA
markers, if such correlation is currently limited by depth of genome
coverage.
To undertake a study on the selection index (SI) used in this study and
compare it with others used in the literature, which could be used in the
selection of more bambara groundnut varieties.
To conduct a stability analysis on developed varieties in other parts of
Botswana, so as to release them to farmers, with the exception of
109BWA1, this had undergone a number of field trials already.
Conduct a study on the genetic diversity of landraces based on the climatic
zones of Africa which is important to discern the effect of weather
parameters on the genetic diversity of bambara groundnut and adaptation
to local conditions.
Conduct a more detailed comparison of genotypes from various regions,
which would shed more light on the population structure of this landraces.
Increasing the number of landraces per country and including more
countries could give a detailed population structure analysis.
To conduct within and between accession variation with an increased
number of individuals per genotype.
211
REFERENCE
ADENIJI O.T., PETER J.M. and IBRAHIM B. (2008), Variation and
interrelationships
for pod and seed yield characters in bambara
groundnut (Vigna subterranea) in Adamawa State, Nigeria, African
Journal of Agricultural Research 3: 671 - 621
AKANINWOR J.O. and OKECHUKWU P. (2004), Comparative nutrient and
anti-nutrient levels in commercial and formulated weaning mixtures,
BIOKEMISTRI 16: 15 -21
AKBARI M., WENZL P., CAIG V., CARLING J., XIA L., YANG S.,
USZYNSKI G., MOHLER V., LEHMENSIEK A., KUCHEL H.,
HAYDEN M.J., HOWES N., SHARP P., VAUGHAN P., RATHMELL
B., HUTTNER E. and KILIAN A. (2006), Diversityarrays technology
(DArT) for high-throughput profiling of the hexaploid wheat
genome, Theoretical Applied Genetics 113: 1409 -1420
AKKAYA M.S., BHAGWAT A.A.and CREGAN P.B. (1992), Length
polymorphism of simple sequence repeat DNA in soybean, Genetics
132: 1131 -1139
ALLARD R.W. (1960), Principles of plant breeding, John Wiley and sons, New
York
AMADOU H.I., BEBELI P.J., and KALTSIKES P.J. (2001), Genetic diversity in
bambara groundnut (Vigna subterranea L.) germplasm revealed by
RAPD markers, Genome 44: 995 -999
AMARTEIFIO J.O., TIBE O. and NJUGO R.M. (2006), The composition of
bambara groundnut (Vigna subterranea L. Verdc.) grown in
Southern Africa, Africa Journal of Biotechnology 5: 2408 - 2411
AMOS W., HOFFMAN I.J., FRODSHAM A., ZHANG L., BEST S. and HILL
A.V.S. (2007),Automated binning of microsatellites alleles:
problems and solutions,Molecular Ecology Notes 7: 10 - 14
APPA-RAO S., MAZHANI L.M. and ATTERE A.F. (1986), Collecting bambara
groundnut in Botswana, Plant Genetic Resources Newsletter 68: 27- 28
AVILA C.M., SILLERO J.C., RUBIALES D., MORENO M.T. and TORRES
A.M. (2003), Identification of RAPD markers linked to the Urf-1 gene
conferring
hypersensitive resistance against rust (Uromyces viciaefabae) in Vicia Faba L., Theoretical and Applied Genetics 107: 353358
AZAM-ALI S.N., AGUILAR-MANJARREZ J. and BANNAYAN-AVVAL M.
(2001), A globalmapping system for bambara groundnut production, Food
and Agriculture Organization of The United Nations, Rome
AZAM-ALI S.N., SESAY A., KARIKARI S.K., MASSAWE F.J., AGUILARMANJARREZ J., BANNAYAN M, and HAMPSON K.J. (2001),
212
assessing the potential of an underutilized crop- a case study using
bambara groundnut, Experimental Agriculture 37: 433 - 472
BAKER D.C. (1987), Arable farming development priorities in the central
agricultural regions, Botswana: a farming system analysis, PhD thesis,
Michigan States University, USA
BAKER R.J (1986), Selection indices in plant breeding, CRC Press, Boca Raton,
Florida, USA
BAKER R.J. (1974), Selection indexes without economic weights for animal
breeding, Canadian Journal of Animal Science 54: 1 - 8
BALOLE T.V., KARIKARI S.K., KHONGA E.B., RAMOLEMANA G. and
LEGWAILA G. (2003), Effect of earthing on the yield of bambara
groundnut (Vignasubterranea (L.)Verdc.) landraces grown in Botswana.
In. Proceedings of the International bambara groundnut Symposium,
Botswana College of Agriculture, Botswana, 8 -12 August 2003 pp153
- 158
BAMB, (2008), Botswana Agricultural Marketing Board, Annual Report, 2008
BASU S., MAYES S., DAVEY M., ROBERTS J.A., AZAM-ALI S.N., MITHEN
R. and PASQUET R.S. (2007), Inheritance of ‘domestication’ traits in
bambaragroundnut (Vigna subterranea (L.) Verdc.) Euphytica 157:
59-68
BASU S., ROBERTS J.A., AZAM-ALI S.N and MAYES S. (2007),
Development of microsatellite markers for bambara groundnut (Vigna
subterranea L. Verdc.) – an underutilized African legume crop species,
MolecularEcology Notes 7: 1326 - 1328
BASU S., ROBERTS J.A., AZAM-ALI S.N. and MAYES S. (2007), Bambara
groundnut. In: Genomic mapping and molecular breeding in plants,
volume 3, Pulses,sugar and tuber crops(Eds. Kole C.), Springer, New
York pp159 - 173
BAYE (2002), Genotypic and phenotypic variability in Vernonia galamensis
germplasm collected from eastern Ethiopia, The Journal of Agriculture
Science 139: 161 - 168
BECHERE E., BELAY G., MITIKU D. and MERKER A. (1996), Phenotypic
diversity of tetraploid wheat landraces from northern and north-central
regions of Ethiopia, Hereditas 124: 165-172
BEEBE E.S., ROJAS-PIERCE M., YAH X., BLAIR M.W., PEDRAZA F.,
MAÑOZ F., TOHME J. and LYNCH J.P. (2006), Quantitative trait loci
for root architecture traits correlated with phosphorus acquisition in
common bean, Crop Science 45: 413 - 423
BERCHIE J.N., SARKODIE –ADDO J., ADU-DAPAAH H., AGYEMANG A.,
ADDY S. ASSARE E. and DONKOR J. (2010), Yield evaluation of
three early maturing bambara groundnut (Vigna subterranea L. Verdc.)
213
Landraces at the CSIR-Crops Research Institute, Fumesua-Kumasi,
Ghana, Journal of Agronomy 9: 175 - 179
BERCHIE J.N., SARKODIE-ADDO J., ADU-DAPAAH H., AGYEMANG A.,
ADDY S., ASARE E. and DONKOR J. (2010), Yield evaluation of the
three early maturing bambara groundnut (Vigna subterranea (L.) Verdc.)
landraces at theCSIR-Crops Research Institute, Fumesua-Kumasi, Ghana,
Journal of Agronomy 9: 111 - 118
BERGMANN F. and RUETZ W. (1991), Isozyme genetic variation and
heterozygosity in random tree samples and selected orchard clones from
the same Norway spruce populations, Forest Ecology and Management 46:
39 - 49
BERTAN I., DE CARVALHO F.I.F., and DE OLIVEIRA C.A. (2010), Parental
selection strategies in plant breeding programs, Journal of Crop Science 4:
211 - 222
BERTOLDO G.J., BARILI D.L., DO VALE M.N., COIMBRA M.L.J.,
STÄHELIN D. and GUIDOLIN F. (2010), Genetic gain in agronomic
traits of common bean inthe region of Planalto Catarinese, Euphytica
171: 381 - 388
BLAIR W.M., DỉAZ M.L., BUENDỉA F.H. and DUQUE M.C. (2009), Genetic
diversity, seed size associations and population structure of common beans
(Phaseolus vulgaris L.), Theoretical and Applied Genetics 119: 955 - 972
BLAIR W.M., HURTADO N., CHAVARRO C.M., MUŇOZ-TORRES C.M.,
GIRALDO M.E.,PEDRAZA F., TOMKIMS J. and WING R. (2011),
Gene-based SSR markers for common bean (Phaseolus vulgaris L.)
derived from root and leaf tissues ESTs: an integration of theBMC series,
BMC Plant Biology 11: 50
BOLBOACĂ S.D. and JÄNTSCHI L. (2006), Pearson versus Spearman,
Kendall’s Tau correlation analysis on structure-activity relationship of
biologic active compounds, Leornado Journal of Sciences 9: 179 - 200
BONIN A., BELLEMAIN E., EIDENSEN P.B., POMPANON F.,
BROCHMANN C. and TABERLET P.(2004), How to track and assess
genotyping errors in population genetic studies, Molecular Ecology
13: 3261 - 3273
BONIN A., EHRICH D. and MANEL S. (2007), Statistical analysis of fragment
length amplified length polymorphism data: a tool box for molecular
ecologistsand evolutionists, Molecular Ecology 16: 3737 - 3758
BOTSTEIN D., WHITE R.L., SKOLNICK M.and DAVIS R.W. (1980),
Construction ofgenetic linkage map in man using fragment length
polymorphism, American of Journal Human genetics 32: 314 -331
BRIÑEZ B., BLAIR M., KILIAN A., CARBONELL S., CHIARATO A. and
RUBIANO L. (2011), A whole genome DArT assay to assess germplasm
collection diversity in common bean, Molecular Breeding 7: 1-13
214
BRINK M. (1997), Rates of progress towards flowering and podding in bambara
groundnut (Vigna subterranea) as a function of temperature and
photoperiod, Annals of Botany 80: 505 - 513
BRINK M., SIBUGA K.P., TARIMO A.J.P. and RAMOLEMANA G.M. (2000),
Quantifying photothermal influences on reproductive development in
bambara groundnut (Vigna subterranea): models and their validation,
Field Crops Research 66: 1-14
BRINK M., COLLINSON S.T. and WIGGLESWORTH D.J (1996), Bambara
groundnut (Vigna subterranea) cultivation in Botswana, Report of
farmer’s survey, pp. 1-52
BRINK M., COLLINSON S.T.andWIGGLESWORTH D.J. (1996),
Characteristics of bambara groundnut cultivation in Botswana, In:
Proceedings of the International Bambara groundnut Symposium, 23 25 July, University of Nottingham, UK pp. 133-142
BROWN R.P. (1997), Testing heterogeneity among phenotypic correlations: a
comparison of methods using Monte Carlo simulations, Genetica 101: 67 74
BURNS M.J., EDWARDS K.J., NEWBURY H.J., FORD-LLOYD B.V. and
BAGGOTT C.D. (2001), Development of simple sequence repeat (SSR)
markers for the assessment of gene flow and genetic diversity in
pigeonpea (Cajanus cajan), Molecular Ecology Notes 1: 283 - 285
BURSTIN J. and CHARCOSSET A. (1997), Relationship between phenotypic
and marker distances: theoretical and experimental investigations,
Heredity 79: 477 - 483
BUSO G.S.C., AMARAL Z.P.S., BRONDANI R.P.V and FERREIRA E. M.
(2006), Microsatellites markers for the common bean (Phaseolus
vulgaris), Molecular Ecology Notes 6: 252-254
CAN D.N. and YOSHIDA T. (1999), Genotypic and phenotypic variances and
covariance in early maturing grain sorghum in a double cropping, Plant
Production Science 2: 67-70
CANNON B.T.,MAY D.G., and JACKSON S.A. (2009), Three sequenced
legume genomes and many crop species: rich opportunities for
translational genomics, Plant Physiology 151: 970 - 977
CASTRO P., MILLÁN T., GIL J., MÉRIDA J., GARCIA M.L., RUBIO J. and
FERNÁDEZ- RAMERO M.D.(2011), Identification of chickpea
cultivars by microsatellites markers, Journal of Agricultural Sciences
149: 451-460
CEBALLOS H., PÉREZ J.C., CALLE F., JARAMILLO G., LENIS J.I.,
MORANTE N. and LÓPES J. (2007), A new evaluation scheme for
cassava breeding at CIAT In: CIAT 2007.Cassava
research
and
development in Asia. Exploring new opportunities for an ancient crop,
pp.125 – 135 (Eds. Howeler H) CIAT, Colombia
215
CHARCOSSET A. and MOREAU L. (2004), Use of molecular marker for the
development of new cultivars and the evaluation of genetic diversity,
Euphytica 137: 81-94
CHARLESWORTH D. (2006), Evolution of plant breeding systems, Current
Biology 16: 726 - 735
CHATFIELD C. and COLLLINS J.A. (1980), Introduction to multivariate
analysis, Chapman and Hall, London, UK.
CHEN J.J, DUAN T, SINGLE R, MATHER K and THOMSON G.(2005), HardWeinberg testing of a single homozygous genotype, Genetics 170: 1439 1442
CHEVERUD J.M. (1988), A comparison of genetic and phenotypic correlations,
Evolution 42: 958 - 968
CHIJIOKE B.O, IFEANYI M.U. and BLESSING C.A. (2010), Pollen behaviour
andfertilisation impairment in bambara groundnut (Vigna subterranea
(L.) Verdc), Journal of Plant Breeding and Crop Science 2: 12 - 23
CHUI J.N., LUZANI T., NKHORI S. and MBULAWA S.T. (2003), Bambara
groundnut selection for various characteristics and response to row
spacing and population density, In: Proceedings of the International
bambara groundnut Symposium, Botswana College of Agriculture,
Botswana 8-12 August, 2003 pp. 297 - 307
CHYBICKI J.I. and BURCZYK J. (2009), Simultaneous estimation of null alleles
and inbreeding coefficient, Journal of Heredity 100: 106 -113
COLLARD B.C.Y., JAHUFER M.Z.Z., BROUWER J.B. and PANG E.C.K.
(2005), An introduction to markers, quantitative trait loci (QTL)
mapping and marker-assisted selection for crop improvement: the basic
concepts, Euphytica 142: 169 -196
COLLINSON S.T., CLAWSON E.J., AZAM-ALI S.N. and BLACK C.R. (1997),
Effects of soil moisture deficits on the water relations of bambra
groundnut (Vigna subterranea L. Verdc.), Journal of Experimental Botany
48: 877-884
COLLINSON S.T., AZAM-ALI S.N., CHAVULA K.M. and HODSON D.A.
(1996), Growth, development and yield of bambara groundnut
(Vigna
subterranea L. Verdc.) in response to soil moisture, Journal of
Agricultural Science 126: 307- 318
COLLINSON S.T., BERCHIE J. and AZAM-ALI S.N. (1999), The effect of soil
moisture on light interception and the conversion coefficient for three
landraces of bambara groundnut (Vigna subterraneaL.Verdc.), Journal
of Agricultural Science 133:151-157
CORNELISSEN E.J.L.R. (2004), Modelling variation in the physiology of
bambara groundnut (Vigna subterraneaL.Verdc.), PhD thesis,
Cranfield University at Silsoe.
216
COUDERT M.J. (1984), Market openings in West Africa for cowpea and
bambara groundnut, International Trade Forum 20: 14 - 29
COULIBALY S., PASQUET R.S., PAPA R.and GEPTS P. (2002), AFLP
organisation analysis of the phonetic organization and genetic diversity of
Vigna unguiculata L. Walp. reveals extensive gene flow between wild
and domesticated types, Theoretical and Applied Genetics 104: 358 - 366
CROUCH H.K., CROUCH J.K., MADSEN S., VUYLSTEKE D.R. and ORTIZ
R. (2000), Comparative analysis of phenotypic and genotypic
diversity among plantain landraces (Musa spp., AAB group),
Theoretical and Genetics 101: 1056-1065
CROUCH J.H. and ORTIZ R. (2004), Applied genomics in the improvement of
crops in Africa, African Journal of Biotechnology 3: 489 - 496
CUC M.L., MACE E.S., CROUCH H.J., QUANG D.V., LONG D.T. and
VARSHNEY K.R.(2008), Isolation and characterisation of novel
microsatellites markers and their application for diversity assessment in
cultivated groundnut (Arachis hypogaea), BMC Plant Biology 8: 55
CUI Z., CATER JR. E.M., BURTON J. W. and WELLS R. (2001), Phenotypic
diversity of modern Chinese and North American soybean cultivars,
Crop Science41: 1954-1967
CUPIC T., TUCAK M., POPOVIC S., BOLARIC S., GRLJUSIC S. and
KOZUMPLIK V. (2009), Genetic diversity of pea (Pisum sativum L.)
genotypes assessed by pedigree, morphological and molecular data,
Journal of Food Agriculture and Environment 7: 343-348
DESWARTE J.C. (2001), Variation in the photosynthetic activity within and
between three bambara groundnut Msc thesis, The University of
Nottingham.
DEWOODY J., NASON J.D. and HIPKINS D.V. (2006), Mitigating scoring
errors in microsatellites data from wild populations, Molecular Ecology
Notes 6: 951 - 957
DIOUF D. and HILU K.W. (2005), Microsatellites and RAPDs markers to study
genetic relationships among cowpea breeding lines and local varieties in
Senegal, Genetic Resources and Crop Evolution 52: 1057 - 1067
DOERGE R.W. (2003), Mapping and analysis of quantitative trait loci in
experimental populations, Genetics 3: 43 - 52
DOI K., KAGA A., TOMOOKA N. and VAUGHAN D.A. (2002), Molecular
phylogeny of genus Vigna subgenus Ceratotropis based on rDNA ITS
and atpB –rbcL intergenic spacer of cpDNA sequences, Genetica 114:
129 -145
DOKU E.V., HAMMONDS T.W. and FRANCIS B.J. (1978), On the composition
of Ghanaian bambara groundnuts (Voandzeia subterranea (L.)
Thours) Tropical Science 20: 263 - 269
217
DOKU E.V and KARIKARI S.K. (1970), Fruit development in bambara
groundnut (Voandzeia subterranea), Annals of Botany 34: 951 - 956
DOKU E.V. and KARIKARI S.K. (1971), Bambara groundnut, Economic Botany
25: 255 - 262
DOYLE J.J. and LUCKOW A. (2003), The rest of the iceberg. Legume diversity
and evolution in a phylogenetic context, Plant Physiology 131: 900 - 910
DURÁN L.A., BLAIR M. W, GIRALDO M. C., MACCHIAVELLI R.,
PROPHETE E., NIN J.C.and BEAVER J. S. (2005), Morphological
and molecular characterization of common bean landraces and cultivars
from Caribbean, Crop Science 45: 1320 - 1328
DURSUN A., HALILOGLU K. and EKINCI M. (2010), Characterization of
breeding lines of common bean as revealed by RAPD and relationship
with morphological traits, Pakistan Journal of Botany 42: 3839 - 3845
DUTILLEUL P., STOCKWELL D.J., FRIGON D. and LEGENDRE P. (2000),
The Mantel test versus Pearson‘s correlation analysis; assessment of
the differences forbiological and environmental studies, Journal of
Agricultural, Biological and Environmental Statistics 5: 131 - 150
EDJE O.T. and SESAY A. (2003), Effect of seed source on performance and
yield of bambara groundnut (Vigna subterranea) landraces, In:
Promoting the conservation and use of underutilized and neglected crops.
9. Proceedings of the workshop on Conservation and Improvement of
bambara groundnut (Vigna subterranea (L.) Verdc.) 14-16 November,
1995, Harare, Zimbabwe, pp 101 – 118 (Eds. Heller J.,
Begemann F.
and Mushonga J.) International Plant Genetic Resources Institute, Rome,
Italy pp.141 - 152
EDWARDS K.J., BARKER J.H., DALY A., JONES C. and KARP A. (1996),
Microsatellites libraries enriched for several microsatellites sequences in
plants, Biotechniques 5: 758 - 760
ELLAH M.M. and SINGH A. (2008), Bambara groundnut (Vigna subterranea L.
Verdc.) yield as influenced by phosphorus and cultivars in the semi-arid
savanna of Nigeria, Journal of Plant Sciences 3: 176 - 181
ELLIS J.R. and BURKE J.M. (2007), EST-SSRs as a resource for population
genetic analyses, Heredity 99: 125 -132
ENDER M. and KELLY J.D ( 2005), Identification of QTL associated with white
mold resistance in common bean, Crop Science 45: 2482 - 2490
ENGELS M.M.J. (1994), Genetic diversity in Ethiopian barley in relation to
altitude,Genetic Resources and Crop Evolution: 41: 67 - 73
ESHGHI R., OJAGHI J. and SALAYEVA S. (2011), Genetic gain through
selection indices in hulless barley, International Journal of Agriculture
and Biology 13: 191 - 197
218
EXCOFFIER L., LAVAL G. and SCHNIEDER S. (2005), Arlequin version 3.1:
An integrated software package for population genetic data analysis.
Evolutionary Bioinformatics Online 1: 47 - 50
EXCOFFIER L. and LISCHER H.E.L. (2010), Arlequin suite version 3.5: A new
series of programs to perform population genetics analyses under linux
and windows, Molecular Ecology Resources 10: 564 - 567
FALCONER D.S. and MACKAY F.C. (1996), Introduction to quantitative
genetics, Longman, England
FALL A.L., BRYNE P.F., JUNG G., COYNE D.P., BRICK M.A. and SWARTZ
H.E. (2001), Detecting and mapping a major locus for fusarium wilt
resistance in common bean, Crop Science 41: 1494 -1498
FAO, (2011), Food and Agriculture Organisation of the United Nations, FAO
Statistical Database and Data set, http://faostat.fao.org/site/291 accessed 1
June 2011
FARSHADFAR M. and FARSHADFAR E. (2008), Genetic variability and path
analysis of Chickpea (Cicer arientinum L.) landraces and lines, Journal
of Applied Sciences 8: 3951-3956
FENTY J. (2004), Analysing distances, The Strata Journal 4: 1- 26
FORNI-MARTINS E.R. (1986), New Chromosome number in the genus Vigna
Savi (Leguminosae-Papilionoideae), Bulletin Nationale Plantentium 56:
129 -133
FRANCO J., CROSSA J., RIBAUT J.M., BETRAN J., WARBURTON M.L. and
KHAIRALLAH M. (2001), A method for combining molecular and
phenotypic attributes for classifying plant genotypes, Theoretical and
Applied Genetics 103: 994 - 952
GAITÁN-SOLÍS E., DUQUE M.C., EDWARDS K.J. and TOHME J. (2002),
Microsatellites repeat in common bean (Phaseolus vulgaris):
Isolation, characterisation, and cross-species amplification in phaseolus,
Crop Science 42: 2128 -2136
GALLEGO J.F., PEREZ A.M., NUNEZ Y. and HILDALGO P. (2005),
Comparison of RADPs, AFLPs and SSR markers for the genetic
analysis of yeast of Saccharomyces cerevisiae, Food Microbiology, 561
-568
GARCIA A.A.F., BENCHIMOL L.L., BARBOSA M.M.A., GERALDI O.I.,
SOUZA JR. L.C. and DE SOUZA A.P. (2004), Comparison of RAPD,
RFLP, AFLP and SSR markers for diversity studies in tropical
maize
inbred lines, Genetics and Molecular Biology 27: 579 - 588
GARCIA R.A.V., RANGEL N.P., BRONDANI C., MARTINS S.W., MELO
C.L., CARNEIRO M.S., BORBA T.C.O. and BRONDANI R.P.V. (2011),
The characterisation of new set of EST-derived simple sequence repeat
219
(SSR) markers as a resource for the genetic analysis of
vulgaris, BMC Genetics12: 41
Phaseolus
GHAFOOR A., AHMAD Z., QURESHI A.S. and BASHIR M. (2001), Genetic
relationship in Vigna mungo (L.) Hepper and V. radiata (L.) R. Wilczek
based onmorphological traits and SDS-PAGE, Euphytica 123: 367-378
GIANCOLA S., POLTRI M.S., LACABATE S.M. and HOPP H.E. (2002),
Feasibility of integration of molecular markers and morphological
descriptors in a real case study of a plant variety protection system for
soybean, Euphytica, 127: 95 - 113
GILBERT J.E., LEWIS R.V., WILKINSON M.J. and CALIGARI P.D.S. (1999),
Developing an appropriate strategy to assess genetic variability in plant
germplasm, Theoretical and Applied Genetics 98: 1125 -1131
GIMENES A.M., HOSHINO A.A., BARBOSA V.G.A., PALMIERI A.D. and
LOPES R.C.(2007), Characterization and transferability of microsatellite
markers of cultivated peanut (Arachis hypogaea), BMC
Plant Biology7:9
GLASZMANN J.C., KILIAN B., UPADHYAYA H.D. and VARSHNEY R.K.
(2010), Accessing genetic diversity for crop improvement, Current
Opinion in Plant Biology 13: 167 -173
GOEL S., RIANA S.N. and OGIHARA Y. (2002), Molecular evolution and
phylogenetic implications of internal transcribed spacer sequences of
nuclearRibosomal DNA in the Phaseolus-Vigna complex, Molecular
Phylogenetics and Evolution 22: 1-19
GOLI A. E., BEGEMANN F., and NG N.Q. (1995), Characterization and
evaluation of IITA’s bambara groundnut collection. In: Promoting the
conservation and use of underutilized and neglected crops. 9.
Proceedings of the workshop on Conservation and Improvement of
bambara groundnut (Vigna subterranea (L.) Verdc.) 14-16 November,
1995, Harare, Zimbabwe, pp 101 – 118 (Eds. Heller J., Begemann F. and
Mushonga J.) International Plant Genetic Resources Institute, Rome,
Italy
GOMEZ O.J., BLAIR W.M., FRANKOW-LINDBERG B.E. and GULLBERG U.
(2004), Molecular and phenotypic diversity of common bean landraces
from Nicaragua, Crop Science 44: 1412 -1418
GONZÁLEZ-CHAVIRA M.M., TORRES-PACHEO I., VILLORDO-PINEDA E.
and GUEVARAGONZALEZ G.R. (2006), DNA markers, Advances
in Agricultural and Food
Biotechnology 6: 99 - 134
GOUDET J. (2001), FSTAT, a program to estimate and test gene diversities and
fixation
indices
(version
2.9.3).
Available
from
http://www.unil.ch/izea/software/fstat.html
GREGORIUS R-H. (2010), Linking diversity and differentiation, Diversity 2:
370 - 394
220
GUPTA P.K. and VARSHNEY R.K. (2000), The development and use of
microsatellite markers for genetic analysis and plant breeding with
emphasis on bread
wheat, Euphtyica 113: 163 - 185
GUPTA P.K., VARSHNEY R.K., SHARMA P.C. and RAMESH B. (1999),
Molecular
markers and their applications in wheat breeding, Plant
breeding 118: 369-390
GUTIERREZ M.C., VAN PATTOS M.C., HUGUET T., CUBERO J.I.,
MORENO M.T. and TORRES A.M. (2005), Cross-species amplification
of Medicago truncatula microsatellites across three major pulse crops,
Theoretical and Applied Genetics 110: 1210 - 1217
HAMRICK J. L. and GODT M.J.W. (1996), Effects of life history traits on
genetic diversity in plants species, Philosophical Transactions: Biological
Sciences351: 1291 - 1298
HANELT P. (2001), Leguminosae subfamily (Fabaceae). In Mansfield’s
Encyclopaedia of Agricultural and Horticultural Crops, 2: 635 -957 (Eds.
Hanelt P.), Springer, Berlin, Germany
HARRIS D and AZAM-ALI S.N (1993), Implications of day length sensitivity in
bambara groundnut (Vigna subterranea L. Verdc.) production in
Botswana, Journal of Agricultural Science 120: 75 - 78
HAZEL L.N. (1943), The genetic basis for constructing selection indexes,
Genetics28: 476 - 490
HE G., MENG R., NEWMAN M., GAO G., PITTMAN R.N. and PRAKASH
C.S. (2003), Microsatellites as DNA markers in cultivated peanut
(Arachis hypogaeaL.), BMC Plant Biology 3: 1-6
HEDRICK P.W. (2005), Genetics of populations, Jones and Bartlett, London,
UK.
HENNINK S. and ZEVEN A.C. (1991), Interpretation of Nei and ShannonWeaver within population variation indices, Euphytica 51: 235 - 240
HEPPER F.N. (1963), Plant of the 1957-58 West Africa expedition II: The
bambara groundnut (Voandzeia subterranea) and Kersting’s groundnut
(Kerstingiella geocarpa) wild in West Africa. Kew Bulletin 16: 395 407
HILL J., BECKER H.C. and TIGERSTEDT (1998), Quantitative and ecological
aspects ofplant breeding, Chapman and Hall, London
HIRST L.G. and ILLAND M. (2001), Automated detection of microsatellite
instability, Molecular Biotechnology 17: 239 - 247
HUBBY J.L. and LEWONTIN R.C. (1966), A molecular approach to the study of
genic heterozygosity in natural populations. I. The number of alleles at
different loci inDrosophila pseudoobscura, Genetics 54: 577 -594
221
HURTADO P., OLSEN K.M., BUITRAGO C., OSPINA C., MARIN J., DUQUE
M., DE VINCENTE C., WONGTIEM P., WENZEL P., KILIAN A.,
ADELEKE M. and FREGENE M. (2008), Comparison of simple
sequence repeat (SSR) and diversity (DArT) markers for assessing genetic
diversity in cassava (Manihot esculenta Crantz), Plant Genetic
Resources: Characterisation and utilization 6: 208 - 214
IDURY R.M. and CARDON L.R. (1997), A simple method for automated allele
binning in microsatellites markers, Genome Research 7: 1104 - 1109
IJAROTIMI S.O. and ESHO R. T. (2009), Comparison of nutritional composition
and anti-nutrient status of fermented, germinated and roasted bambara
groundnut seeds (Vigna subterranea), British Food Journal 111: 376386
IPGRI, IITA, BAMNET (2000), Descriptors for bambara groundnut (Vigna
subterranea), International Plant Genetic Resources Institute, Rome, Italy;
International Institute of Tropical Agriculture, Ibadan, Nigeria;
International Bambara Groundnut Network, Germany.
JACCOUD D., PENG K., FEINSTEIN D. and KILIAN A. (2001), Diversity
Arrays: a solid state technology for sequence information independent
genotyping, Nucleic Acids Research 29: 1 - 7
JACKSON D.S. (2008), Plant responses to photoperiod, New Phytologist, 181:
517 -531
JEUFFROY M-H. and CHABANET C. (1994), A model to predict seed number
per pod from early pod growth rate in pea (Pisum sativum L.) Journal of
Experimental Botany 45: 709 - 715
JOLLIE I.T. (2002), Principal component analysis, Springer, New York.
JONAH P.M., ADENIJI O.T. and WAMMANDA D.T. (2010), Variability and
genetic correlations for yield and yield characters in some bambara
groundnut (Vigna subterranea) cultivars, International Journal of
Agriculture and Biology 12: 303 - 307
JONAH P.M., ADENIJI O.T. and WAMMANDA D.T. (2010), Genetic
correlations and path analysis in bambara groundnut (Vigna
subterranea), Journal of Agriculture and Social Sciences 6: 1- 5
JØRGENSEN S.T., LUI F., OUÉDRAOGO M., NTUNDU W.H., SARRAZIN J.
and CHRISTIANSEN L.J. (2010), Drought responses of two
bambara groundnut (Vigna subterranea L. Verdc.) landraces collected
from a dry and a humid area of Africa, Journal of Agronomy and Crop
Science 196: 412 - 422
KALINOWSKI T.S. (2004), Counting alleles with rarefaction: private alleles and
hierarchical sampling designs, Conservation Genetics 5: 539 -543
222
KAMPHIUS L.G., WILLIAMS H.A., D’SOUZA K.N., PFAFF T., SINGH B.K.,
OLIVER R.P. and LICHTENZVEIG J. (2007), The Medicago truncatula
reference accession A17 has an aberrant chromosomal configuration,
New Phytologist 174: 229-303
KARIKARI S.K. (2000), Variability between local and exotic bambara groundnut
landraces in Botswana, African Crop Science Journal 8: 145 - 152
KARIKARI S.K. and TABONA T.T. (2004), Constitutive traits and selective
indices of bambara groundnut (Vigna subterranea (L.) Verdc)
landraces for drought tolerance under Botswana conditions, Physics and
Chemistry of the Earth 29: 1029 - 1034
KARIKARI S.K., WIGGLESWORTH D.J., KWEREPE B.C., BALOLE T.V.,
SEBOLAI B. and MUNTALI D.C. (1995), Country report for
Botswana. In: Promoting the conservation and use of underutilized and
neglected crops. 9.
Proceedings of the workshop onConservation
and Improvement of bambara groundnut (Vigna subterranea(L.)
Verdc.) 14-16 November,
1995, Harare, Zimbabwe, pp 11 – 18 (Eds.
Heller J., Begemann F. and Mushonga J.) International Plant Genetic
Resources Institute, Rome, Italy
KARURI H.W., ATEKA E.M., AMATA R., NYENDE A.B., MUIGAI A.W.T.,
MWASAME E. and GICHUKI S.T. (2010), Evaluating diversity among
Kenyan sweet potato genotypes using morphological and SSR
markers, International Journal of Agricultural Biology 12: 33 -38
KHEDIKAR P.Y., GOWDA M.C.V., SARVAMANGALA C., PATGAR K.V.,
UPADADHYAYA H.D. and VARSHNEY R.K. (2010), A QTL study
on late leaf spot and rust revealed one major QTLformolecular
breeding for rust in groundnut (Arachis hypogaea L.), Theoretical and
Applied Genetics 121: 971 - 984
KLODA J.M., DEAN P.D.G., MADDREN C., MACDONALD D.W. and
MAYES S. (2007), Using principle component analysis to compare
genetic diversity across polyploidy levels within plant complexes: an
example from British Restharrows (Ononis spinosa and Ononis repense),
Heredity 100: 253 -260
KOBRAEE S., SHAMSI K., RASEKHI B. and KOBRAEE S. (2010),
Investigation of correlation analysis and relationship between grain
yield and other quantitative traits in Chickpea (Cicer arietinum L.),
African Journal of Biotechnology 9: 2342-2348
KOENIG R. and GEPTS P. (1989), Allozyme diversity in wild Phaseolus
vulgaris: further evidence for two major centres of genetic diversity,
Theoretical and Applied Genetics 78: 809-817
KOUASSI N.J. and ZORO B. I.A. (2009), Effect of sowing density and seedbed
typed on yield and yield components in bambara groundnut (Vigna
subterranea) in woodland savannas of Côte d’Ivore, Experimental
Agriculture 46: 99 -110
223
KOVACH W. (2006), MVSP-Multivariate Statistical Package Version 3.1.
Kovach Computing Services: Anglesey, Wales
KRISHMA G.K., ZHANG J., BURROW M., PITTMAN R.N., LU Y. and
PUPPALA N. (2004), Genetic diversity analysis in Valencia peanut
(Arachis hypogaea L.) using microsatellite
markers,Cellular
and
Molecular Biology Letters 9: 685-695
KUMAR J., CHOUDHARY A.K., SOLANKI R.K. and PRATAP A. (2011),
Towards marker- assisted selection in pulses: a review, Plant breeding
130: 297 -313
KUMAR P., GUPTA V.K., MISRA A.K., MODI D.R. and PANDEY B.K.
(2009), Potential of molecular markers in plant biotechnology, Plant
Omics Journal 2: 141 - 162
KUMAR V., SHARMAR S., SHARMA K.A., SHARMA S. and BHAT V.K.
(2009), Comparative analysis of diversity based on morphoagronomic traits and microsatellite markers in common bean, Euphtyica
170: 249 -262
KURODA Y., KAGA A., TOMOOKA N. and VAUGHAN D.A. (2006),
Population genetic structure of Japanese wild soybean (Glycine soja)
based on microsatellite variation, Molecular Ecology 15: 959 -974
KUSWANTORO H., WIJANARKO A., SETYAWAN D., WILLIAM E.,
DADANG A. and MEJAYA J.M. (2010), Soybean germplasm evaluation
for acid tidal swamp tolerance using selection index, International Journal
of Plant Biology 1: e11
LAZREK F., ROUSSEL V., RONFORT J., CARDINET G., CHARDON F.,
AOUANI M.E. and HUGNET T. (2009), The use of the neutral and nonneutral SSRs to analyse the genetic structure of a Tunisian collection of
Medicago truncatula lines and to reveal association with ecoenvironmental variables, Genetica 135: 391 - 402
LEBERG P.L. (2002), Estimating allelic richness: effect of sample size and
bottlenecks, Molecular Ecology 11: 2445 - 2449
LEGENDRE P. and FORTIN M.J. (2010), Comparison of the Mantel test and
alternative approaches for detecting complex multivariate relationships
in the spatial analysis of genetic data, Molecular Ecology Resources 10:
831 -844
LI P., WANG Y., SUN X.and HAN J. (2009), Using microsatellite (SSR) and
morphological markers to assess the genetic diversity of 12 falcata
(Medicago sativa spp. falcata) populations from Eurasia, African Journal
of Biotechnology 8: 2102 - 2108
LIANG X., CHEN X., HONG Y., LIU H., ZHOU G., LI S. and GUO B. (2009),
Utility of EST-derived SSR in cultured peanut (Arachis hypogaea L.) and
Arachiswild species, BMC Plant Biology 9: 1 - 9
224
LINNEMANN A.R. and AZAM-ALI S.N. (1993), Bambara groundnut (Vigna
subterranea (L.) Verdc). In: underutilised crop series. II. Vegetables and
pulses, pp. 13 -57 (Eds. Williams, J.T), Chapman and Hall, London.
LINNEMANN A. R (1994), Phenological development in bambara groundnut
(Vigna subterranea) at alternate exposure to 12 and 14 h photoperiods,
Journal of Agricultural Science 123: 333 - 340
LINNEMANN A.R and CRAUFURD P.Q. (1994), Effects of temperature and
photoperiod on phenological development in three genotypes of bambara
groundnut (Vigna subterranea), Annals of Botany 74: 675 - 681
LINNEMANN A. R., WESTPHAL E. and WESSEL M. (1995), Photoperiod
regulation of development and growth in bambara groundnut (Vigna
subterranea), Field Crops Research 40: 39 - 47
LIU J., JIAN-PING G., DONG-XU X., XIAO-YANG Z., JING G. and XUXIAO Z. (2008), Genetic diversity and population structure in Lentil
(Lens culinarisMedik) Germplasm detected by SSR markers, Acta
Agronomica Sinica 34: 1901 -1909
LIU M., ZHANG M., JIANG W., SUN G., ZHAO H. and HU S. (2011), Genetic
diversity of Shaanxi soybean landraces based on agronomic traits and
SSR markers, African Journal of Biotechnology 10: 4823 - 4837
LIU K., GOODMAN M., MUSE S., SMITH S.J., BUCKLER E. and DOEBLEY
J. (2003), Genetic structure and diversity among maize inbred lines
as inferred from DNA microsatellites, Genetics 165: 2117 - 2128
LOH P.J., KIEW R., KEE A., GAN H.L. and GAN Y. (1999), Amplified
Fragment Length Polymorphism (AFLP) provides molecular markers for
the identification of Caladium bicolour cultivars, Annals of Botany 84:
155-161
LU J., KNOX M.R., AMBROSE M.J., BROWN J.K.M. and ELLIS T.H.N.
(1996), Comparative analysis of genetic diversity in pea assessed by
RFLP- and PCR-based methods, Theoretical and Applied Genetics 93:
1103-1111
LUI F., CHARLESWORTH D. and KREITMAN M. (1999), The effect of mating
system differences on nucleotides diversity at the phosphoglucose
isomerase locus in the plant genus Leavenworthia, Genetics 151: 353 357
LUI K. and MUSE S. V. (2005), PowerMarker: Integrated analysis environment
for genetic marker analysis. Bioinformatics 21: 2128 - 2129
MACE E.S., PHONG D.T.,UPADHAYAYA H.D., CHANDRA S. and
CROUCH
J.H. (2006), SSR analysis of cultivated (Arachis
hypogaea L.) germplasm resistant to rust and late leaf spot diseases,
Euphytica 152: 317 - 330
225
MACE E.S., RAMIN J-F, BOUCHET S., KLEIN P.E.,KLIEN R.R., KILIAN A.,
WENZEL P., XIA L., HALLORAN K.and JORDAN D.R. (2009), A
consensus genetic map of sorghum that integrates multiple component
maps and high-throughput DiversityArray
Technology
(DArT)
markers, BMC Plant Biology 9: 13
MACIEL F.L., ECHEVERRIGARAY S., GERALD L.T.S. and GRAZZIOTIN G.
F. (2003), Genetic relationships and diversity among Brazilian
cultivars and landraces of common beans (Phaseolus vulgaris L.)
revealed by AFLP markers, Genetic Resources and Crop Evolution 50:
887 - 893
MAGAGULA C.N., MANSUETUS A.B., SESAY A. and KUNENE I.S. (2003),
Yield loss
associated with pests and diseases on bambara groundnut
(Vigna subterranea (L.)Verdc.) in Swaziland. In. Proceedings of the
International bambara groundnut Symposium, Botswana College of
Agriculture, Botswana, 8 -12 August 2003 pp.95- 105
MAKANDA I., TONGOONA P., MADAMBA R., ICISHAHAYO D. and
DERERA J. (2009), Path coefficient analysis of bambara groundnut
pod yield components at four planting dates, Research Journal of
Agriculture and Biological Sciences 5: 287-292
MAKANDA I., TONGOONA P., MADAMBA R., ICISHAHAYO D. and
DERERA J. (2009), Evaluation of bambara groundnut varieties for the
production in Zimbabwe, African Crop Science Journal 16: 175-183
MALIK M.F.A., ASHRAF M., QURESHI A.F. and GHAFOOR A. (2007),
Assessment of genetic variability, correlation and path analyses
for
yield and its components in soybean, Pakistan Journal of Botany 39:
405 - 413
MALIK S.S. and SINGH S.P. (2006), Role of plant genetic resources in
sustainable agriculture, Indian Journal of Crop Science 1: 21 - 28
MANIEE M., KAHRIZI D. and MOHAMMADI R. (2009), Genetic variability of
some morpho-physiological traits in durum wheat (Triticum turgidum
var. durum), Journal of Applied Sciences 7:1383 - 1387
MANTEL N. (1967), The detection of disease clustering and generalised
regression approach, Cancer Research 27: 209 -220
MANTOVANI P., MACCAFERRI M., SANGUITE M.C., TUBEROSA R.,
CATIZONE I., WENZEL P., THOMSON B., CARLING J., HUTTNER
E., DEAMBROGIO E and KILIAN A (2008), An integrated DArT- SSR
linkage map of durum wheat, Molecular Breeding 22: 629 - 648
MAQUET A., BI-ZORO I., DELVAUX M., WATHELET B.,and BAUDOIN
P.J. (1997), Genetic structure of Lima bean base collection using
allozymes markers, Theoretical and Applied Genetics 95: 980 - 991
226
MARAS M., ŠUŠTAR-VOZLIČ J., JAVORNIK B. and MEGLIČ V. (2008), The
efficiency of AFLP and SSR markers in genetic diversity estimation
and genepool classification of common bean (Phaseolus vulgaris L.),
Acta Agriculturae Slovenia 91: 8 - 96
MASI P., ZEULI S.L.P. and DONNI P. (2003), Development and analysis of
multiplex microsatellite marker set in common bean (Phaseolus vulgaris
L.), Molecular Breeding 11: 303 - 313
MASSAWE F.J., MWALE S.S., AZAM-ALI S.N. and ROBERTS J.A. (2005),
Breeding in bambara groundnut (Vigna subterranea (L.) Verdc.):
strategic considerations, African Journal of Biotechnology 4: 463 - 471
MASSAWE F.J., ROBERT J.A., AZAM-ALI S.N. and DAVEY M. R. (2003),
Genetic diversity in bambara groundnut (Vigna subterranea (L.) Verdc)
landraces assessed by Random Amplified Polymorphic DNA
(RADPs) markers, Genetic Resources and Crop Evolution 50: 737-741
MASSAWE F.J., DICKINSON M., ROBERT J.A. and AZAM-ALI S.N. (2002),
Genetic diversity in bambara groundnut (Vigna subterranea (L.)
Verdc) landraces revealed by AFLP markers, Genome 45: 1175 - 1180
MASSAWE F.J., AZAM-ALI S.N. and ROBERTS J.A. (2003), Impact of
temperature on leaf development in bambara groundnut landraces, Crop
Science 43: 1375 -1379
MASSAWE F.J., SCHENKEL W., BASU S. and TEMBA E.M. (2003), Artificial
hybridization in bambara groundnut (Vigna subterranea (L.) Verdc.)In.
Proceedings of the International bambara groundnut Symposium,
Botswana College of Agriculture, Botswana, 8 -12 August 2003 pp. 193 209
MATSCHINER M. and SALZBURGER W. (2009), TANDEM: Integrating
automatic allele binning into genetics and genomics workflows,
Bioinformatics 25: 1982 - 1983
MAYES S., BASU, S., MURCHIE E., ROBERTS J.A., AZAM-ALI S.N.,
STADLER F., MOHLER
V., WENZEL G., MASSAWE F., KILIAN
A., BONIN A., BEENA A. and SHESHSHAYEE M.S.(2009),
BAMLINK – a Cross disciplinary programme to enhance the
role of
bambara groundnut(Vigna subterranea L. Verdc.) for food security
in Africa and India. Acta Horticulturae 806: 137 - 150.
MBEWE D.N., MWALA M.S. and MWILA G.P. (1995), Country report for
Zambia. In: Promoting the conservation and use of underutilized and
neglected crops. 9. Proceedings of the workshop on Conservation and
Improvement of bambara groundnut (Vigna subterranea (L.) Verdc.) 1416 November, 1995, Harare, Zimbabwe, pp. 64 - 67 (Eds. Heller J.,
Begemann F. and Mushonga J.) International Plant Genetic Resources
Institute, Rome, Italy
227
MCCOUCH S.R., CHEN X., PANAUD O., TEMNYKH S., XU Y., CHO G.Y.,
HUANG N., ISHII T. and BLAIR M. (1997), Microsatellite marker
development, mapping and applications in rice genetics and
breeding, Plant Molecular Biology 35: 89 -99
MENDELSOHN R (2000), Climate Change Impacts on African Agriculture,
Ariel Dinar, World Bank, pp. 2 - 25
MICHELMORE R.W. and HUBERT S.H. (1987), Molecular markers for genetic
analysis
of
phytopathogenic
fungi,
Annual
Review
of
Phtyopathology 25: 383 - 404
MILLAN T., CLARKE H.J., SIDDIQUE K.H.M., BUHARIWALLA H.K.,
GAUR P.M., KUMAR J.,
GIL J., KAHL G. and WINTER P. (2006),
Chickpea molecular breeding: New tools and concepts, Euphtyica 147:
81 - 103
MINE O.M., MOTLHABANE L.T. and BATLANG U. (2003), Preliminary
assessment of genetic variation in bambara groundnut in Botswana using
RAPD markers: a case of technology transfer. In: Proceedings of the
International bambara groundnut symposium, Botswana College of
Agriculture, Botswana, 8 -12, August, 2003 pp. 175 - 178
MINISTRY OF AGRICULTURE, GOVERNMENT OF BOTSWANA (1947),
Jugo Beans. In: Review of Crop Experiments in Bechuanaland
Protectorate, Department of Agricultural Research, Botswana pp. 127 128
MISANGU R.N., AZMIO A. and REUBEN S.O.W.M. (2007), Path coefficient
analysis among components of yield in bambara groundnut (Vigna
subterranea L. Verdc) landraces under screen house conditions, Journal
of Agronomy 6: 317 - 323
MKANDAWIRE C.H. (2007), Review of Bambara groundnut (Vigna subterranea
(L.) Verdc.) Production in Sub-Sahara Africa, Agricultural Journal 2:
464 -470
MONDINI L., NOORANI A. and PAGNOTTA M.A. (2009), Assessing plant
genetic diversity by molecular tools, Diversity 1: 19 -35
MONIRIFAR H. (2010), Evaluation of selection indices for alfalfa (Medicago
sativa L.), Notulae Scientia Biologicae 2: 84 - 87
MONTAVANI P., MACCAFERRI M., SANGUINETI M.C.,
CATIZONE I., WENZEL
P., THOMSON B.,
HUTTNER E., DEAMBROGIO E. and KILIAN
Integrated DArT-SSR linkage map of durum
Breeding 22: 629 - 648
TUBEROSA R.,
CARLING J.,
A. (2008), An
wheat,Molecular
MORENO R.R and POLANS N.O. (2006), Development and characterisation of
microsatellites loci in peas, Pisum Genetics 38: 10 -14
228
MUKUMBIRA L.M.(1985), Effects of rate of nitrogen fertilizer and previous
grain legume crop on maize yield, Zimbabwe Agriculture Journal 82:
177 - 179
MULLIS K. (1990), The unusual origin of polymerase chain reaction, Scientific
American April: 56 - 65
MUNTHALI D.C. and RAMORANTHUDI M. (2003), Susceptibility of bambara
groundnut landraces to Hilda patreulis. In. Proceedings of the
International bambara groundnut Symposium, Botswana College of
Agriculture, 8 – 12 August2003 pp. 85-105
MWALE S.S., AZAM-ALI S.N. and MASSAWE F. J. (2007), Growth and
development of bambara groundnut (Vigna subterranea) in response to
soil moisture1.Dry matter and yield, European Journal Agronomy 26:
345 - 353
NATH U.K. and ALAM M.S. (2002), Genetic variability, heritability and genetic
advance of yield and related traits of groundnut (Arachis hypogaea L.),
Journal of Biological Sciences 2: 762 - 764
NEI M. and LI W. (1979), Mathematical model for studying genetic variation in
terms of restriction endonucleases, Proceedings of the National Academic
Science of United States of America 76: 5256 - 5273
NEI M. (1987), Molecular Evolutionary Genetics. Columbia University Press,
New York
NOEPARVAR S., VALIZADEH M., MONIRIFAR H., HAGHIGHI R.A. and
DARBANI B. (2008), Genetic diversity among and within alfalfa
population native to Azerbaijan based on RAPD analysis, Journal of
Biological Research- Thessaloniki 10: 159 -169
NTUNDU W.H., SHILLAH S.A., MARANDU W.Y.F. and CHRISTIANSEN
J.L. (2006), Morphological diversity of bambra groundnut [Vigna
subterranea (L.) Verdc.] landraces in Tanzania, Genetic Resources
and Crop Evolution 53: 367 - 378
NTUNDU W.H., BACH I.C., CHRISTIANSEN J.L. and ANDERSEN S.B.
(2004), Analysis of genetic diversity in bambara groundnut (Vigna
subterranea L. Verdc)landraces using amplified fragment length
polymorphism (AFLP)markers, African Journal of Biotechnology 3: 220
- 225
NUNOME T., NEGORO S., MIYATAKE K., YAMAGUCHI H. and FUKUOKA
H. (2006), A protocol for the construction of microsatellite enriched
genomic library, Plant Molecular Biology Reporter 24: 305 - 312
OBAGWU J. (2003), Evaluation of bambara groundnut (Vigna subterranea
L. Verdc) lines for reaction to Cercospora Leaf Spot, Journal of
Sustainable Agriculture 22: 193 - 100
229
ODENY D.A., JAYASHREE B.,FERGUSON M., CROUCH J. and
GEBHARDT C. (2007),Development, characterisation and utilization of
microsatellites markers in pigeonpea, Plant breeding
126: 130 -136
ODENY D.A., JAYASHREE B., GERBHARDT C. and CROUCH J. (2009),
New microsatellite markers for pigeonpea (Cajanus cajan (L.) millsp.),
BMC Research Notes 2: 35
OFORI K. (1996), Correlation and path-coefficient analysis of components of
seed yield in bambara groundnut (Vigna subterranea), Euphytica 91:
103 -107
OFORI K., KUMANGA F.K. and Tonyigah A. (2010), Morphological
characterization and agronomic evaluation of bambara groundnut (Vigna
subterranea (L.) Verdc.) germplasm, PGR Newsletter, FAOBiodiversity 145: 23 - 28
OHARA M. and SHIMAMOTO Y. (2002), Importance of genetic characterisation
and
conservation of plant genetic resources: The breeding system and
geneticdiversity of wild soybean (Glycine soja), Plant Species
Biology 17: 51-58
OHIOKPEA O. (2003), Food processing and nutrition: A vital link in agricultural
development, Pakistan Journal of Nutrition 2: 204 - 207
OJULONG F.H., LABUSCHAGNE T.M., HERSELMAN L. and FREGENE M.
(2010), Yield traits as selection indices in seedling populations of
cassava, Crop Breeding and Applied Biotechnology 10: 191 -196
OKORI P.W.S. and RUBAIHAYO P.R. (2005), Genetic variability and
relatedness of the Asian and African pigeonpea as revealed by AFLP,
African Journal of Biotechnology 4: 1228 - 1233
OLUKOLU B.A., MAYES S., STADLER F., NG Q.N., FAWOLE I.,
DOMINIQUE D., AZAM- ALI S.N., ABOTT G. A. and KOLE C.
(2011), Genetic diversity in bambara groundnut (Vigna subterranea (L.)
Verdc.) as revealed by phenotypic descriptors and DArT marker
analysis, Genetic Resources and Crop Evolution 59: 347 -358
OMOIGUI L.O, ISHIYAKA M.F, KAMARA A.Y, ALABI S.O and
MOHAMMED S.G (2006), Genetic variability and heritability studies of
some reproductive traits in Cowpea (Vigna unguiculata(L.) Walp.),
African Journal of Biotechnology5: 1191-1195
ONWUBIKO N.I.C., ODUM O.B., UTAZI C.O. and POLY-MBAH P.C. (2011),
Studies on the adaptation of bambara groundnut (Vigna subterranea L.
Verdic.) in Owerri southeastern Nigeria, New York Science Journal 4:
60 - 67
OUEDRAOGO M., OEUDRAOGO J.T., TIGNERE J.B., BALMA D., DABIRE
C.B. and KONATE G. (2008), Characterization and evaluation of
accessions of bambara groundnut (Vigna subterranea L. Verdcourt)
from Burkina Faso, Science and Nature 5: 191 - 197
230
OYIGA B.C and UGURU M.I. (2011) Interrelations among pod and seed yield
traits in bambara groundnut (Vigna subterranea L. Verdc) in the derived
savanna agro-ecology of south-eastern Nigeria under two planting dates,
International Journal of Plant Breeding 5: 106 - 111
OYIGA B.C., UGURU M.I. and ARUAH C.B. (2010), Studies on the floral traits
and their implications on pod and seed yield in bambara groundnut
(Vigna subterranea (L.) Verdc.), Australian Journal of Crop Science 4: 9197
PAINAWADEE M., JOGLOY S., KESMALA T., AKKASAENG C. and
PATANOTHAI A. (2009), Heritability and correlations of drought
resistance traits and agronomic traits in peanut (Arachis hypogaea L.),
Asian Journal of Plant Sciences 8: 325 -334
PARSAEIAN M., MIRLONI A. and SAEIDI G. (2010), Study of genetic
variation in Sesame (Sesamum indicum L.) using agro-morphological
traits and ISSR markers, Russian Journal of Genetics 47: 359- 367
PARZIES H.K., SPOOR W. and ENNOS R.A. (2000), Genetic diversity of barley
landrace accession (Hordeum vulgare ssp. vulgare) conserved for different
length of time in ex situ gene banks, Heredity 84: 476 - 486
PASQUET R.S. (2003), Bambara groundnut and cowpea gene-pool organization
and domestication: In. Proceedings of the International bambara
groundnut Symposium, Botswana College of Agriculture, Botswana,
8- 12, August, 2003 pp.263 - 272
PASQUET R.S., SCHWEDE S. and GEPTS P. (1999), Isozyme diversity in
bambara groundnut, Crop Science 39: 1228-1236
PEJIC I., AJMONE-MARSAN P., MORGANTE M., KOZUMPLICK V.,
CASTGLIONI P., TARAMINO G. and MOTTO M. (1998),
Comparative analysis of genetic similarity among maize inbred lines
detected by RFLPs, RAPDs, SSRs, and AFLPs, Theoretical and
Applied Genetics 97: 1248 -1255
PESEK J. and BAKER R.J. (1969), Desired improvement in relation to selected
indices, Canadian Journal of Plant Science 49: 803 - 804
POLHILL R.M., RAVEN P.H. and STIRTON C.H. (1981), Evolution and
systematics of the legumes, In: Advances in legume systematics, pp. 1
– 26 (Eds Polhill R.M. and Raven P.H), Royal Botanic Garden, Kew,
England.
POLHILL R.M. and RAVEN P.H. (1981), Advances in legumes systematics, part
1, Royal Botanic Gardens, Kew.
POMPANON F., BONIN A., BELLEMAIN E. and TABERLET P. (2005),
Genotyping errors: causes, consequences and solutions, Genetics 6: 847
- 859
POULTER N.H. and CAYGILL J.C. (1980), Vegetable milk processing and
rehydration characteristics of bambara groundnut (Voandzeia
231
subterranea (L.) Thouars),Journal of the Science of Food and
Agriculture 31: 1158 - 1163
POWELL W., MORGANTE M., ANDRE C., HANAFEY M., VOGEL J.,
TINGEY S. and RAFALSKI A. (1996), The comparison of RFLP,
RAPDs, AFLP and SSR (microsatellite) markers for germplasm
analysis, Molecular Breeding 2:225 -238
PRIOLLI G.B.H.R., PINHEIRO B.J., ZUCCHI I.M., BAJAY M.M. and VELLO
A. N. (2010), Genetic diversity among Brazilian soybean cultivars based
on SSR loci and pedigree data, Brazilian Archives of Biology and
Technology 53: 519 - 531
PUTTHA R., JOGLOY S., WONGKAEW S., SANITCHON J., KESMALA T.
and PATANOTHAI A. (2008), Heritability, phenotypic and
genotypic correlation of Peanut bud necrosis virus resistance and
agronomic traits of peanut, Asian Journal of Plant Science 7: 276 - 283
QADIR S.A., DATTA S., SINGH N.P. and KUMAR S.(2007), Development of
highly polymorphic markers for chickpea (Cicer arietinum L.) and their
use as parental polymorphism, Indian Journal of Genetics 67: 329 -333
RAHIM A.M., MIA A.A., MAHMUD F. and AFRIN K.S. (2008), Multivariate
analysis in some mungbean (Vigna radiata L. Wilczek) accessions on the
basis of agronomic traits, American-Eurasian Journal of
Scientific
Research 3: 217 - 221
RAINA S.N., KOJIMA T., OGIHARA Y., SINGH K.P. and DEVAURUMATH
R.M. (2001), RAPDs and ISSR fingerprints as useful genetic markers
for analysis of genetic diversity, varietal identification, and phylogenetic
relationships in peanut (Arachis hypogaea) cultivars and wild
species, Genome 44: 763 -772
RAMAKRISHNAN P.A., MEYER E.S., WATERS J., STEVENS R.M.,
COLEMAN E.C. and FAIRBANKS D.J. (2004), Correlation between
markers and adaptively significant variation in Bromus tectorum
(Poaceae), an inbreeding annual grass, American Journal of Botany,
91: 797 - 803
RAMAN R., LUCKETT D.J. and RAMAN H. (2008), Estimating the genetic
diversity in Albus Lupin (Lupinus albus L.), using DArT and genetic
markers. In: Lupins for Health and Wealth’ Proceedings of the 12th
International Lupin Conference, 14 – 18 September 2008, pp. 236 – 241
(Eds. Palta J.A. and Berger J.B.), Fremantle, Western Australia,
International Lupin Association, Canterbury, New Zealand.
RAMOLEMANA G.M., KWEREPE B.C., BALOLE T.V., KHONGA E.B. and
KARIKARI S.K.(2003), Farmer and consumer perception on the
bambara groundnut (Vigna subterranea) ideotype in Botswana, In:
Proceedings of the
International bambara groundnut Symposium,
Botswana College of Agriculture, Botswana, 8 -12 August 2003 pp.17 24
232
RAUF S., JAIME A., DA SILVA T., KHAN A.A. and NAVEED A. (2010),
Consequences of plant breeding on genetic diversity, International Journal
of Plant Breeding 4:1 - 21.
RAVI K., VADEZ V., ISOBE S., MIR R.R., GUO Y., NIGAM S.N., GOUDA
M.V.C., RADHAKRISHMAN T., BERTIOLE D.J., KNAPP S.T. and
VARSHNEY R.K. (2011), Identification of several small main-effect
QTLs and a large number of epistatic QTLs for drought tolerance related
traits in groundnut (Arachis hypogaea L), Theoretical and Applied
Genetics 122: 1119 - 1132
REUSCH B.T. (2001), New markers- old questions: population genetics of
sea grasses, Marine Ecology Progress Series 211: 261-274
RITSCHEL P.S., DE LIMA LINS T.C., TRISTAN R.L., BUSO G.S.C., BUSO
J.A. and FERREIRA M.E (20040, Development of microsatellites markers
from enriched genomic library for genetic analysis of melon (Cucumis
melo L.), BMC Plant Biology 4: 9
ROBERTSON A. (1961), Inbreeding in artificial selection programmes, The
International Journal for Genetics and Genomics Research 2: 189 - 194
ROBERTSON A. and HILL W.G. (1984), Deviations from Hardy-Weinberg
proportions:sampling variables and use in estimation of inbreeding
coefficient, Genetics 107: 703 - 718
ROHLF F.J. (2000), NTSYS-pc Numerical Taxonomy and Multivariate Analysis
system version 2.1, Applied Biostatistics, New York
ROVEN S. and SKALETSKY H.J. (2000), Primer3 on WWW for general users
and for biologists programmers. In: Bioinformatics Methods and
Protocols: Methods in Molecular Biology, pp. 365 - 386 (Eds
Krawetz S. and Misener S.) Humana Press, Totowa
ROY D.(2000), Plant breeding, analysis and exploitation of variation, Alpha
Science International, UK.
RYCHLIK W., SPENCER W. and RHOADS R.E. (1990), Optimization of the
annealing temperature for DNA amplification in vitro, Nucleic Acids
Research 18: 6409 - 6412
RYMER D. P., MORRIS E. C. and RICHARDSON B.J. (2002), Breeding
systems and population genetics of the vulnerable plant Dillwynia
tennuifolia(Fabaceae), Austal Ecology 27: 241 - 248
SAIKI R.K., SCHARF S., FALOONA F., MULLIS K.B., HORN G.T., ERLICH
H.A. and ARNHEIM N. (1985), Enzymatic amplification of beta- globin
genomic sequences and restriction site analysis for diagnosis of sickle cell
anaemia, Science 230: 1350 -1354
SALEHI M., TAJIK M. and EBADI A.G. (2008), The study of relationships
between different traits in common bean (Phaseolus vulgaris L.) with
233
multivariate statistical methods, American Journal of Agriculture and
Environmental Science 3: 806-809
SANDAL N., KRUSSELL L., RADUTOIU S., OLBRYT M., PEDROSA A.,
STRACKE S., SATO S., KATO T., TABATA S., PARNISKE M.,
BACHMAIR A., KETELSEN T. and STOUGAARD J. (2002), A
genetic linkage map of the model legume Lotus japonica and
strategies for fast mapping of new loci, Genetics 161: 1678 -1683
SANTANA C.Q., COETZEE M.P.A., STEENKAMP E.T. and MLONYENI X.
O. (2009), Microsatellites discovery by deep sequencing of enriched
genomic libraries, BioTechniques 46: 217 - 223
SANTRA K.D., TEKEOGLU M., RATNAPARKHE M., KAISER J.W. and
MUEHLBAUER J.F. (2000), Identification and mapping of QTLs
conferring resistance to Aschochyta Blight in chickpea, Crop Science
40: 1606 -1612
SATO S., ISOBE S. and TABATA S. (2010), Structural analyses of the genomes
in legumes, Current Opinion in Plant Biology 13: 146 - 152
SAXENA R.K., PRATHIMA C., SAXENA K.B., HOISINGTON D.A., SINGH
N.K. and VARSHNEY R.K. (2010), Novel SSR markers for
polymorphism detection in pigeonpea (Cajanus spp.), Plant Breeding
129: 142 - 148
SCHNEIDER A., WALKER S.A., SAGAN M., DUC G., ELLIS T.H.N. and
DOWIE J.A. (2002), Mapping of the nodulation loci sym9 and sym10,
Theoretical and Applied Genetics 104: 1312 -1316
SCHUELKE M. (2000), An economic method for the fluorescent labelling of
PCR fragments. Nature Biotechnology 18: 233-234
SEEHALAK W., SOMTA P., SOMMANAS W. and SRINIVES P. (2009),
Microsatellites markers for mungbean developed from sequence
database, Molecular Ecology Resources 9: 862 - 864
SEMAGN K., BJØRNSTAD Å. and NDJIONDJOP M.N. (2006), An overview of
molecular marker methods for plants, African Journal of Biotechnology
5: 2540 -2568
SEMAGN K., BJØRNSTAD Å. and XU Y. (2010), The genetic dissection of
quantitative traits in crops, Electronic Journal of Biotechnology 13: 1 45
SESAY A. (2009), Influence of flooding on bambara groundnut (Vigna
subterraneaL.) germination: effect of temperature, duration and timing,
African Journal of Agricultural Research 4: 100 - 106
SESAY A., COLLINSON S.T. and AZAM-ALI S.N. (1996), Where are we now
with bambara groundnut? In: Proceedings of the international Bambara
groundnut symposium, University of Nottingham, UK 23 – 25 July, 1996
pp.215 - 228
234
SESAY A., MAGAGULA C.N. and MANSUETUS A.B. (2008), Influence of
sowing date and environmental factors on the development and yield of
bambaragroundnut (Vigna subterranea) landraces in a sub-tropical region,
Experimental Agriculture 44: 167 - 183
SESAY A., EDJE O.T. and MAGAGULA C.N. (2003), Working with farmers on
the bambara groundnut (BAMFOOD) research project in Swaziland,
In Proceedings of the International bambara groundnut Symposium,
Botswana College of Agriculture, Botswana 8 - 12 August 2003 pp.3 - 15
SETHY K.N., SHOKEEN B. and BHATIA S. (2003), Isolation and
characterisation of sequence-tagged microsatellite site markers in
chickpea (Cicer arietinumL.), Molecular Ecology Notes 3:428 - 430
SHARMA C.P., GROVER A. and KAHL G. (2007), Mining microsatellites in
eukaryotic genome, TRENDS in Biotechnology 25: 490 - 498
SHARMA S., SHARMA S., KOPISCH-OBUCH F.J., KEIL T., LAUBACH E.,
STEIN N., GRANER A. and JUNG C. (2011), QTL analysis of rootlesion nematode resistance in barley: 1. Pratylechus neglectus Theoretical
and Applied Genetics 122: 1321 -1330
SHETE S. (2003), Uniformly minimum variance unbiased estimation of gene
diversity, Journal of Heredity 94: 421- 424
SHETE S., TIWARI H. and ELSTON R.C. (2000), On estimating the
heterozygosity and polymorphisminformation
content
value,
Theoretical Population Biology, 57: 265 -271
SHI C., NAVABI A. and YU K. (2011), Association mapping of common
bacterial blight resistance QTL in Ontario bean breedingpopulations,
BMC Plant Biology 11: 52
SHI J., LI R., QUI D., JIANG C., LONG Y., MORGAN C., BACROFT I., Z
HAO J. and MENG J. (2009), Unraveling the complex trait of crop yield
with quantitative trait loci, Genetics 182: 851 - 861
SINGH R.K. and CHADHARY S.D. (1985), Biometrical methods in quantitative
genetic analysis, Kalian publishers, New Delhi
SINGH S.P. and MUÑOZ C.G. (1999), Resistance to common bacterial blight
among Phaseolus species and common bean improvement, Crop Science
39: 80 -89
SINGRUN C. and SCHENKEL W. (2003), Fingerprinting of bambara groundnut
germplasm with molecular marker, In: Proceedings of the International
Bambara groundnut Symposium, International Cooperation with
Developing Countries, Botswana college of Agriculture, Botswana 8-12
August, 2003, Gaborone, Botswana pp.161 - 170
SIOL M., PROSPERI J.M., BONNIN I. and RONFORT J. (2008), How
multilocus genotypic pattern helps to understand the history of selfing
235
populations: a case study of in Medicago truncatula, Heredity 100: 517 525
SMARTT J. (1985), Evolution of grain legumes. II. Old and New world pulses of
lesser economic importance, Experimental Agriculture 21: 1 - 18
SMITH H.F. (1936), A discriminant function for plant selection, Annual Eugenic
7: 240 - 250
SOMTA P., SOMMANAS W. and SRINIVES P. (2009), Molecular diversity
assessment
of AVRDC – The World Vegetable Center elite – parental
mungbeans, Breeding Science 59: 149 -157
SONGOK S., FERGUSON M., MUIGAI A.W. and SILIM S. (2010), Genetic
diversity in pigeonpea (Cajanus cajan (L.) Millsp.) landraces as
revealed by simple sequencerepeat markers, African Journal of
Biotechnology 9: 3231 - 3241
SPOONER D., VAN TREUREN R. and DE VINCENTE M.C. (2005), Molecular
markers for genebank management. IPGRI Technical Bulletin NO
10. International Plant Genetic Resources, Institute, Rome, Italy
STADLER F. (2009), Analysis of differential expression under water-deficit
stress and genetic diversity in bambara groundnut (Vigna subterranea
(L.) Verdc.) using novel high-throughput technologies, PhD thesis,
TECHNISCHE UNIVESITÄT MÜNCHEN (TUM)
STRUSS D. and PLIESKE J. (1998), The use of microsatellites markers for
detection of genetic diversity in barley population, Theoretical and
Applied Genetics 97: 308 - 315
STODART J.B., MACKAY M. and RAMADAN H. (2005), AFLP and SSR
analysis of
genetic diversity among landraces of bread wheat
(Triticum aestivum L. em Thell) from different geographic regions,
Australian Journal of Agricultural Research 56: 691 - 697
STODART J.B., MACKAY M.C. and RAMADAN H. (2007). Assessment of
molecular diversity in landraces of bread wheat (Triticum aestivum L.)
held in an exsitu collection with Diversity Array Technology
(DArT),
Australian Journal of Agricultural Research 58: 1174 -1182
STREFELER M.S. and WEHNER T.C. (1986), Comparison of six methods of
multiple-trait selection for fruit yield and quality traits in three freshmarket cucumber populations, Journal of the American Society for
Horticultural Science 111: 792 - 798
SUWANPRASERT J, TOOJINDA T., SRINIVES P. and CHANPRAME S.
(2006), Hybridization technique
of
bambara
groundnut
(Vigna
subterranea), Breeding Science 56: 125 - 129
SWAMY M.B.P., UPADHAYAYA H.D., GOUDAR P.V.K., KULLAISWAMY
B.Y. and SINGH S. (2003), Phenotypic variation for agronomic
236
characters in a groundnut core collection for Asia, Field Crops
Research 84: 359 -371
SWANEVELDER C.J. (1998), Bambara- Food for Africa bambara groundnut
(Vigna Subterranea), National Department of Agriculture, Republic of
South Africa.
TABACHNICK B.G. and FIDEL S. L. (2007), Using multivariate statistics,
PearsonInternational Edition, Boston
TANTASAWAT P., TRONGCHUEN J., PRAJONGJAI T., THONGPAE T.,
PETKHUM C., SEEHALAK W. and MACHIKOWA T. (2010), Variety
identification and genetic relationships of mungbean and blackgram in
Thailand based on morphological characters and ISSR
analysis,
African Journal of Biotechnology 9: 4452 -4464
TARDIN D.F., PEREIRA M.G., GABRIEL A.P.C., DO AMARAL JÚNIOR A.T.
and DE SOUZA FILHO A. (2007), Selection index for and molecular
markers in reciprocal recurrent selection in maize, Crop Breeding and
Applied Biotechnology 7: 225 - 233
TAUTZ D. (1989), Hypervariability of simple sequences as a general source for
polymorphic DNA markers, Nucleic Acids Research 17: 6463 - 6471
TIBE O., AMARTEIFIO J.O. and NJOGU R.M. (2007), Trypsin inhibitor activity
and condensed tannin content in bambara groundnut (Vigna subterranea
(L.) Verdc) grown in Southern Africa, Journal of Applied Science and
Environmental Management 11: 159 - 164
TOSTI N. and NEGRI V. (2005), On-going on-farm microevolutionary process
in neighbouring Cowpea landraces revealed by molecular markers,
Theoretical and Applied Genetics 110: 1275 - 1283
TSILO T.J., HARELAND G.A., SIMSEK S., CHAO S. and ANDERSON J.A.
(2010), Genome mapping of kernel characteristics in hard red
spring
wheat breeding lines, Theoretical and Applied Genetics 121: 717 - 730
UGURU M.I. and EZEH N. E. (1997), Growth, nodulation and yield of bambara
groundnut (Vigna subterranea (L.) Verdc.) on selected Nigerian soils,
Journal of Science of Food and Agriculture 73: 377 - 382
UPADHYAYA H.D. (2003), Phenotypic diversity in groundnut (Arachis
hypogaeaL.) core collection assessed by morphological and
agronomic evaluations, Genetic Resources and Crop Evolutions 50:
539 - 550
UPADHYAYA D. H., DWIVEDI L.S., BAUM M., VARSHNEY R.K.,
UDUPA M.S., GOWDA L.L.C., HOISINGTON D. and SINGH S. (2008),
Genetic structure, diversity and allelic richness in composite
collection and
reference set in chickpea (Cicer arientum L.) BMC Plant Biology 8: 106
237
VAN OOSTERHOUT C., HUTCHISON W. F., WILLS D.P.M. and SHIPLEY P.
(2004), Micro- Checker: software for identifying and correcting
genotyping errors in microsatellites data, Molecular Ecology Notes 5:
535 - 538
VANGE T. and MOSES E. O. (2009), Studies on genetic characteristics of
pigeonpea germplasm at Otobi Benue State of Nigeria, World
Agricultural Sciences 5: 714 -719
VARSHNEY R.K., BERTIOLI J.D., MORETZSOHN M.C., VADEZ V.,
KRISHNAMURTHY L., ARUNA R., NIGAM S.N., MOSS B.J.,
SEETHA K., RAVI K., HE G., KNAPP S.J. and HOISINGTON D.A.
(2009), The first SSR-based genetic linkage map for cultivated
groundnut (Arachis hypogaea L.), Theoretical and Applied Genetics 118:
729 - 739
VARSHNEY R.K., HORRES R., MOLINA C., NAYAK S., JUNGMANN R.,
SWANY R., WINTER P., JAYASHREE B., KAHL G. and
HOISINGTON D.A, (2007), Extending the repertoire of microsatellite
markers for genetic linkage mapping and germplasm screening in
chickpea, Semi- Arid Tropic Journal 5: 1- 3
VARSHNEY R.K., PENMETSA R.V., DUTTA S., KULWAL P.L., SAXENA
R.K., DATTA S.,
SHARMA T.R., ROSEN B., CARRASQUILLAGARCIA N., FARMER A.D., DUBEY A., SAXENA B.K.,
GAO
J., FAKRUDIN B., SINGH M.N., SINGH B.P., WANJARI K.B., YUAN
M.,
SRIVASTAVA R.K., KILIAN A., UPADHYAYA H.D.,
MALLIKARJUNA N.,TOWN C.D., BRUENING G.E.,
HEG., MAY
G.D., MCCOMBIE R., JACKSON A.S., SINGH N.K. and
COOK D.R.
(2009), Pigeonpea genomics initiative (PGI): an
international
effort
to
improve crop productivity (Cajanus cajanL.),
Molecular Breeding 26: 393
- 408
VARSHNEY R.K., CLOSE T.J., SINGH N.K., HOISINGTON D.A and COOK
D.R. (2009), Orphan legume crops enter the genomics era! Current
Opinion in Plant Biology 12: 202 - 210
VELASQUEZ V.L. and GEPTS P. (1994), RFLP diversity of common bean
(Phaseolus vulgaris) in its centres of origin, Genome 37: 256 -263
VIJAYKUMAR A., SAINI A. and JAWALI N. (2009), Phylogenetic analysis of
subgenus Vigna species using nuclear ribosomal RNA ITS: evidence of
hybridisation among Vigna unguiculata subspecies, Journal of Heredity
2: 177 - 178
VEIRA A.E., DE CARVAHLO F.I., BERTAN I., KOPP M.M.,ZIMMER D.P.,
BENIN G., DE SILVA G.A.J., HARTWIG I., MALONE G. and DE
OLIVIERA C.A (2007), Association between genetic distances in wheat
(Triticum aestivum L.) as estimated by AFLP and morphological
markers, Genetics and molecular Biology 30: 392-399
238
VOS P., HOGERS R., BLEEKER M., REIJANS M, VAN DE LEE T., HORNES
M., FRIJTERS A., POT J., PELEMAN J. and KUIPER M. (1995),
AFLP: a new technique for DNA fingerprinting, Nucleic Acids
Research 11: 4407 - 4014
WANG L.M.,BARKLEY A.N., GILLASPIE A.G and PEDERSON A.G. (2008),
Phylogenetic relationships and genetic diversity of the USDA Vigna
germplasm collection revealed by gene-derived markers and
sequencing, Genetics Research 90: 467 -480
WANG T.C., YANG D.X., CHEN X.D., YU L.S., LIU Z.G., TANG Y.Y. and
XU Z.J. (2007), Isolation of simple sequence repeat from groundnut,
Electronic Journal of Biotechnology 10: 3
WANG X., RINEHART T.A., WADL P.A., SPIERS J.M., HADZIABDIC D.,
WINDHAM M.T., and TRIGIANO R.N. (2009), A new electrophoresis
technique to separate microsatellite alleles, African
Journal
of
Biotechnology 8: 2432 - 2436
WAGNER A.P., CREEL S. and KALINOWSKI S.T. (2006), Estimating
relatedness and relationships using microsatellites loci with null alleles,
Heredity 97: 336 -345
WAMBETE J. and MPOTOKWANE S. (2003), Investigating opportunities for
bambara groundnut in the development of weaning foods. In.
Proceedings of the
International bambara groundnut Symposium,
Botswana College of Agriculture, 8 – 12, August 2003 pp.273 - 279
WASIKE S., OKORI P. and RUBAIHAYO P.R. (2005), Genetic variability and
relatedness of the Asian and African pigeonpea as revealed by AFLP,
African Journal of Biotechnology 4: 1228 - 1233
WEIR B. S. (1996), Genetic Data Analysis II, Methods for Discrete Population
Genetic Data, Sinauer Associates, Inc Publishers, Massachusetts, USA.
WEIR B.S. (1990), Genetic data analysis: methods for discrete population
geneticdata, Sinauer: Massachusetts
WEIR B.S. and COCKERAM C. (1984), Estimating F-statistics for the analysis
of population structure, Evolution 38: 1358 - 1370
WENZL P, CARLING J, KUDRNA D, JACCOUD D, HUTTER E,
KLEINHOFS A and KILIAN A
(2004), Diversity Arrays Technology
(DArT) for whole-genome profiling of barley, Proceedings of the
National Academy of Sciences of the United States of America 101: 9915
- 9920
WIGGLESWORTH D.J. (1996), The potential for genetic improvement of
bambara groundnut (Vigna subterranea (L.) Verdc.) Botswana, In
Proceedings of the International Bambara groundnut Symposium,
University of Nottingham, UK, 23- 25 July, 1996 pp.181 - 191
239
WILLIAM J.G.K., KUBELIK A.R., LIVAK K.J., RAFALSKI J.A and TINGEY
S.V (1990), DNA polymorphism amplified by arbitrary primers are
useful as genetic markers, Nucleic Acids Research 18: 6531 - 6535
WRIGHT S. (1978), Evolution and genetics of population no. 14, variability
within and among natural populations. University of Chicago press,
Chicago
www.bioversityinternational.org African Vigna. Accessed 12.4.2011
www.icrisat.org Accessed 21.4.2011
www.iita.org) Accessed 21.4.211
www.spgrc.org.zm accessed 1 June, 2011
XU Y., MCCOUCH S.R. and ZHANG Q. (2005), How can we use genomics to
improve cereals with rice as a reference genome? Plant Molecular
Biology 59: 7 - 26
YAGHATIPOOR A. and FARSHDFAR E. (2007), Non-Parametric estimation
and component analysis of phenotypic stability in chickpea (Cicer
arietinumL.), Pakistan Journal of Biological Sciences 10: 2646 - 2652
YANG S., PANG W., ASH G., HARPE J., CARLING J., WENZEL P.,
HUTTNER E., ZONG X.and KILIAN A. (2006), Low level of genetic
diversity in cultivated pigeonpea compared to its wild relatives revealed
by diversity arrays technology, Theoretical and Applied Genetics 113:
585 - 595
YANG S.S., VALDÉS-LÓPEZ O., XU W.W., BUCCIARELLI B., GRONWALD
J .W., HERNÁNDEZ G., and VANCE P.C. (2010), Transcript profiling
of common bean (Phaseolus vulgaris L.) using the GeneChip®Soybean
genome array: optimizing analysis by masking biased probes, BMC Plant
Biology 10: 85
YAP V.M. and NELSON R.J. (1996), WinBoot: A program for performing
bootstrap analysis of binary data to determine the confidence limits
of UPGMA based dendrograms, International Rice Institute (IRRI),
Manila, Philippines
YU K, PARK S.J, POYSA V. and GEPTS P. (2000), Integration of simple
sequence repeat (SSR) markers into a molecular linkage map of
common bean (Phaseolus vulgaris L.), The Journal of Heredity 91: 429 424
YU W-J., DIXIT A., MA H-K., CHUNG W-J. and PARK J-Y. (2009), A study on
relative abundance, composition and length of microsatellites in18
underutilized crop species, Genetic Resources and Crop Evolution 56: 237
- 246
240
ZANE L., BARGELLONI L. and PATARNELLO T. (2002), Strategies for
microsatellites isolation: a review, Molecular Ecology 11: 1 – 16
ZANNOU A., KOSSOU D.K. AHNCHÉDÉ A., ZOUNDJIHÉKPON J.,
AGBICODO E., STRUIK P.C and SANNIN A (2008), Genetic variability
of cultivated cowpea in Benin assessed by random amplified polymorphic
DNA, African Journal of Biotechnology 7: 4407 - 4417
ZEVEN A. C. (1998), Landraces: A review of definitions and classifications,
Euphytica 104: 127 – 139
ZHANG Y.L., MARCHAND S., TINKER N.A.and BELZILE F. (2009),
Population structure and linkage disequilibrium in barley assessed by
DArT markers, Theoretical and Applied Genetics 119: 43 - 52
ZHU H., CHOI H., COOK. D. and SHOEMAKER C.R. (2005), Bridging model
and crop legumes through comparative genomics, Plant Physiology,
137: 1189 -1196
ZHURAVLEV N.Y., REUNOVA G.D., KATS L.I., MUZAROK T.I. and
BONDAR A. A.(2010), Genetic variability and population structure
of endangered Panax ginseng in Russia Primorye, Chinese Medicine 5:
21
241
APPENDICES
Appendix 1: Preparations of standard solutions
1.0M Tris pH 8.0
12.11 g Tris base dissolved in 1.0M HCl until pH 8.0
Final volume is adjusted to 100 mL with water
0.5 EDTA dissolved in 75 mL water
18.61 g EDTA dissolved in 75 mL water
2 g of NaOH pellets dissolved
pH adjusted to 8.0 with 1.0M NaOH solution
Final volume adjusted to 100 mL with water
5.0M NaCl
29.22 g NaCl dissolved in 70mL water
Final volume adjusted to 100 mL
50 x TAE DNA Electrophoresis buffer
242 g Tris base
57.1 mL glacial acetic acid
100 mL 0.5M EDTA pH 8.0
Final volume brought to 1 Litre
1 x TAE Buffer
20 mL 50x TAE buffer
Final volume adjusted to 1 litre
5 x TBE DNA Electrophoresis buffer
54 g Tris base
27.5 g Boric acid
20 mL 0.5M EDTA pH 8.0
Final volume brought to 1Litre
0.5 x TBE Buffer
100 mL 0.5x TBE buffer
Final volume adjusted to 1Litre
6 x loading buffer (for DNA gels)
To make 30% glycerol (15 mL + 35 MQ water)
0.025 g Bromophenol Blue
0.025g Xylene Cyanol
242
Appendix 2: List of characterised 75 primers used in bambara groundnut
diversity, and list of primer combinations used in multiple experiments.
75 microsatellites characterised on 24 genotypes selected based on Singruin and Schenkel, (2003)
Marker
PRIMER1F
PRIMER1R
PRIMER2F
PRIMER2R
PRIMER3F
PRIMER3R
PRIMER4F
PRIMER4R
PRIMER5F
PRIMER5R
PRIMER6F
PRIMER6R
PRIMER7F
PRIMER7R
PRIMER8F
PRIMER8R
PRIMER9F
PRIMER9R
PRIMER10F
PRIMER10R
PRIMER11F
PRIMER11R
PRIMER12F
PRIMER12R
PRIMER13F
PRIMER13R
PRIMER14F
PRIMER14R
PRIMER15F
PRIMER15R
PRIMER16F
PRIMER16R
PRIMER17F
PRIMER17R
PRIMER18F
PRIMER18R
PRIMER19F
PRIMER19R
PRIMER20F
PRIMER20R
PRIMER21F
PRIMER21R
PRIMER22F
PRIMER22R
PRIMER23F
PRIMER23R
PRIMER24F
PRIMER24R
PRIMER25F
PRIMER25R
PRIMER26R
PRIMER26R
PRIMER27F
PRIMER27R
PRIMER28F
PRIMER28R
PRIMER29F
PRIMER29R
PRIMER30F
PRIMER30R
PRIMER31F
PRIMER31R
PRIMER32F
PRIMER32R
PRIMER33F
PRIMER33R
Sequence
AACTTGCCATACGTGGAAGG
ACACGCTGCATAATTCACCA
CGTGGATACCCATACCGTCT
TAAGTCCATTTTGTCCGATTGA
TGATGAATGAATGCAAAGTAAGA
TTGGCTCATTGCCTAGTTCA
CATTGTCTCTGCCACCATTTT
CAGACTGGGATTTGCATGTG
CTGCTGTGGTGAGCTTTTGT
CTCCTTGCAGCTAAGCGTCT
TACGGTCCTACACGGGAAAC
ACCTGTCCAGCCGCAATTA
GTAGGCCCAACACCACAGTT
GGAGGTTGATCGATGGAAAA
GGAAGAGTGCGTTTTGGTGT
CTGTGTGGACCCCAGAAAAT
CCAGGAGTGAGGAGTGAGAAA
ATGCATTTTCAGGGTCCAAG
TCAGTGCTTCAACCATCAGC
GACCAAACCATTGCCAAACT
TGGAGGTGGAAATGATAACG
TCCACCTTCACCTGCACT
GTCTTGCAGGAAGGTTCAGC
CAGATTACACACGCGCACTT
CATTGCACGTCATAGAATTTGG
GGGTGAACTACACCACCTTCA
TGGTGGTAGAGAATTGGAGGA
CACACAGAAACACAAACACAGC
AGGAGCAGAAGCTGAAGCAG
CCAATGCTTTTGAACCAACA
CCGGAACAGAAAACAACAAC
CGTCGATGACAAAGAGCTTG
CAAAGCAACACAAACGATGG
ATAACCATTGGCCGATTGAC
TCTGCCACATTTCGCATAAG
CGCTTCAAATCCGATGTTCT
AGGCAAAAACGTTTCAGTTC
TTCATGAAGGTTGAGTTTGTCA
CCCTTCACATACACTTAAGAACCA
CCTCTTCCACGAGAACAAGC
CAAACTCCACTCCACAAGCA
CCAACGACTTGTAAGCCTCA
TCCCAAAATGGGACCAACTA
ATCCGACTGATTAAGCCTAAAA
CAGTAGCCATAATTTGCTATGAACA
CGAATCACCATTCAATACGC
TTGGGTTGAATGGAAGTATGAA
CAGAAGATCCCTTTCGACCA
GCTGGAACTGATCCACCTTT
ATGTAGCAGTGCCACCAACA
CGCTCATTTTAACCAGACCTC
CAAACAAACCAACGGAATGA
ACACCGCCATCATGAGATTT
CATTTCAGGATTTGGGAGGA
CAATGCTTCAACCATCAACC
AGTGTATGGATGCCCAGACC
TCTGACGCAAGCAAGAAGAA
GGTTCGATCGGAAATCTGAA
AATGCAAGATTTTGGCTTGG
CCCACTCAAACCATACACCA
GCTAAGGTGGAGTGGTGGAA
CAATCATCTTTTGCGCTTCA
TTCACCTGAACCCCTTAACC
AGGCTTCACTCACGGGTATG
ACGCTTCTTCCCTCATCAGA
TATGAATCCAGTGCGTGTGA
o
Primer length Tm ( C)
20
59.0
20
59.0
20
46.3
22
46.3
23
59.0
20
59.0
21
57.6
20
57.6
20
59.0
20
59.0
20
59.0
19
59.0
20
55.3
20
55.3
20
57.6
20
57.6
21
57.6
20
57.6
20
55.3
20
55.3
20
59.0
18
59.0
20
57.6
20
57.6
22
55.3
21
55.3
21
57.6
22
57.6
20
55.3
20
55.3
20
57.6
20
57.6
20
55.3
20
55.3
20
55.3
20
55.3
20
55.3
22
55.3
24
59.0
20
59.0
20
57.6
20
57.6
20
55.3
22
55.3
25
55.3
20
55.3
22
55.3
20
55.3
20
59.0
20
59.0
21
57.6
20
57.6
20
60.2
20
60.2
20
47.7
20
47.7
20
55.3
20
55.3
20
59.0
20
59.0
20
57.6
20
57.6
20
57.6
20
57.6
20
57.6
20
57.6
12 markers used in (123)
population structute
analysis Chapter five
20 markers used in (105)
genetic diversity
analysis Chapter six
Marker 1
12 markers used in (5)
genetic anlaysis Chapter
six
Marker 7
Marker 7
Marker 7
Marker 10
Marker 15
Marker 15
Marker 15
Marker 16
Marker 16
Marker 16
Marker 19
Marker 19
Marker 19
Marker 21
Marker 23
Marker 23
Marker 23
Marker 30
Maker 31
Marker 32
Marker 33
Marker 33
Marker 33
243
Appendix 2: Continued
75 microsatellites characterised on 24 genotypes based on Singruin and Schenkel, (2003)
Marker
Sequence
PRIMER34F
PRIMER34R
PRIMER35F
PRIMER35R
PRIMER36F
PRIMER36R
PRIMER37F
PRIMER37R
PRIMER38F
PRIMER38R
PRIMER39F
PRIMER39R
PRIMER40F
PRIMER40R
PRIMER41F
PRIMER41R
PRIMER42F
PRIMER42R
PRIMER43F
PRIMER43R
PRIMER44F
PRIMER44R
PRIMER45F
PRIMER45R
PRIMER46BF
PRIMER46BR
PRIMER47F
PRIMER47R
PRIMER48F
PRIMER48R
G33AB4-D1F
G33AB4-D1R
G111AB4-D2F
G111AB4-D2R
G185AB4-D3F
G185AB4-D3R
G194AB4-D4F
G194AB4-D4R
G196AB4-D5F
G196AB4-D5R
G278AB4-D6F
G278AB4-D6R
G331AB4-D7F
G331AB4-D7R
G372AB4-D8F
G372AB4-D8R
G11-9-B2-D9F
G11-9-B2-D9R
G174B2-D10F
G174B2-D10R
G180B2-D11F
G180B2-D11R
G240-7-B2-D12F
G240-7-B2-D12R
G326B2-D13F
G326B2-D13R
G240-9-B2-D14F
G240-9-B2-D14R
G358B2-D15F
G358B2-D15R
mBam2co80
mBam2co80
E1F
E1R
E2F
E2R
E3F
E3R
E4F
E4R
E5F
E5R
E6F
E6R
E7F
E7R
E9F
E9R
E10F
E10R
E11F
E11R
E12F
E12R
CATGTTGAAACCCGATGTCC
ACCTCCTGGTGCATCTATGG
CGTGCCTACCTTCGACTACC
CGGTGGAAACTCCGATTAAA
CGAAAGAACTTGACAGGCAGA
TCAGCAGAATGATCCTCCAA
CCGATGGACGGGTAGATATG
GCAACCCTCTTTTTCTGCAC
TCACACTTGCAATGGTGCTT
TCGTTGTTTCTCTTTTCATTGC
TCGTACCGAATCACCATTCA
CAGTAGCCATAATCTGCTATGAACA
TGGACCATACCCATCTTCAAA
TCAGGGACATTACCCAGACC
ACACCGCCATCATGAGATTT
CAGGATTTGCGAGGAGAGAG
CCTTTCAGCTTCTCCAAACG
TCAACCCACACAGAATCGAA
ACTTGATGCTACCGAGAGAGAG
AGGCTCCAACAATGCGATAG
TGTGGGCGAAAATACACAAA
TCGTCGAATACCTGACTCATTG
CGTGGATACCCATACCGTCT
AAGTCCATTTTGTCCGATTGA
TTTGTCCGGTTCAACTGAATTA
TTGAAGATGGGTATGGTCCAC
ACCCATTGCACGTCATAGAA
GGGTGAACTACACCACCTTCA
TACCTGCATTCGGGACAGTT
TTCACTCTTTCTTGATCACATGC
TGCTTCTTCAAGGAGGAAGTAAGT
ACAAACATACGCACAACAGAGAAT
AGGTTATGAGGTAAGCATTTCAGG
TCAGATTGCATAATTTGCTTGATT
CTCCACTCCACAAGCAATAAACTA
CCATTTGTAAACCAACGACTTGTA
CCCTTCAACCCTAGTTGAGATAGA
TCCTATTTCTTTCGGCATATTTTT
CCACGTTCTGGTTGTGAGTAGATA
GTGCTTTCAGACCATTACTTGCTT
TGGTTTTATAAATTGGGATTTTGG
ACCTATAATTCACGCACACACG
TCTTCTTTATTGGTGGACCATACC
AAAACCAAGGACACAAATTCTAGC
GCATCTTTACAGCAAGAGTTTCAA
TGGATCTTCCTCATTGCAGTATAA
ATCAAAATCAAGCAAATGAGA
ACCTTTTACGCTCATTTTAACCAG
GTTTTAGGATCAAATGGTTTTGGA
TGCCTTTTATAATGATGTGCATTC
GAGGAAATAACCAAACAAACC
CTTACGCTCATTTTAACCAGACCT
TTTTGTTGTTGTATGAATCCAGTG
CCTCATCAGACGCTCATCATT
AGAGGTGGAGGGGTTGGAT
CCTCAATAGCTGAATCCATTTCTC
GAACGAAGCCAGGATAATGATAGT
CGAAAGCGACAACTCACTACTAAA
TGACGGAGGCTTAATAGATTTTTC
GACTAGACACTTCAACAGCCAATG
GAGTCCAATAACTGCTCCCGTTTG
ACGGCAAGCCCTAACTCTTCATTT
TGTTGTGTCAACAAATTAAGATGAG
ATGCTTCAAACTGTCCCTGA
CATGTTCGTAATGATTTGAAGTGTT
GCCAAAACAATATCTTCAAGAGG
GGACGGAGTCCTTCAAACAA
CCTTGTGCATACCCATAGTATCC
CATGGCGAAGGAGGGCAGCGA
AGCGATTACTGGGGTTGAGA
CATGGAGTGCTATGTGGTGAT
ATACGGTTGTGGCAGTGTCC
CATGGACGAGGGATTAGCGCAG
CCCTAGCCAAATGACCTACC
CATGATTTGTTGTGATGATGAT
AACAACAAATGTACCAAAGAATCG
CATGAGAAGGCCTTCTGATGAT
CCACAAGTTCTTTTTATTCCCTTC
CATGACTTTCTTCATTGGT
TGCATTCCAATTAAATTCATAACAA
CATGACCACCAGAGAAGATGT
ATTCAGAATCCTCAAC
CATGAAGGCGGAGACGGCGG
CATGACCACCAGAGAAGATGT
12 markers used in (123) 20 markers used in (105)
population structure
genetic diversity
o
Primer length
Tm ( C)
20
20
20
20
21
20
20
20
20
22
20
25
21
20
20
20
20
20
22
20
20
22
20
21
22
21
20
21
20
23
24
24
24
24
24
24
24
24
24
24
24
22
24
24
24
24
21
24
24
24
21
24
24
21
19
24
24
24
24
24
24
24
25
20
25
23
20
23
21
20
21
20
22
20
22
24
22
24
19
25
21
16
20
21
57.6
57.6
55.3
55.3
55.3
55.3
55.3
55.3
53.3
53.3
55.3
55.3
55.3
55.3
59
59
57.3
57.3
57.6
57.6
59.7
59.7
55.3
55.3
51.4
51.4
59
59
59
59
59.0
59.0
59.7
59.7
49.4
49.4
55.3
55.3
49.4
49.4
57.6
57.6
45.5
45.5
59.0
59.0
53.3
53.3
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
59.0
47.7
47.7
59.0
59.0
47.7
47.7
57.6
57.6
51.4
51.4
47.7
47.7
51.4
51.4
51.4
51.4
51.4
51.4
47.7
analysis Chapter five
analysis Chapter six
12 markers used in (5)
genetic analysis
Chapter six
Marker 37
Marker 37
Marker 37
Marker 44
Marker 44
Marker 44
Marker D8
Marker D11
Marker D11
Marker D11
Marker D14
Marker D14
Marker D14
Marker D15
mBam2co80
mBam2co80
mBam2co80
Marker E7
Marker E7
Marker E7
47.7
53.3
53.3
244
Appendix 3: Estimated repeat length of alleles and adjustment for the
characterisation of 75 markers used in the analysis of 24 landraces
Primer 1
Estimated repeat
length
0.030
Primer 2
0.935
1.116
Primer 3
0.035
0.968
Primer 4
0.275
0.975
Primer 5
0.375
0.774
Primer 6
0.705
0.787
Primer 7
0.530
0.774
Primer 8
0.220
0.774
Primer 9
0.455
0.774
Primer 10
0.080
0.986
Primer 11
0.615
0.795
Primer 12
0.810
0.774
Primer 13
0.530
0.824
Primer 14
0.200
0.964
Primer 15
0.920
1.011
Primer 16
0.007
0.981
Primer 17
0.925
0.950
Primer 18
0.215
0.981
Primer 19
0.465
0.829
Primer 20
0.205
0.957
Primer 21
0.710
0.781
Primer 22
-0.035
1.060
Primer 23
-0.030
1.036
Primer 24
1.010
0.971
Primer 25
0.550
1.069
Primer 26
0.580
0.802
Primer 27
0.305
0.929
Primer 28
0.045
1.008
Primer 29
0.340
0.774
Primer 30
0.045
0.991
Primer 31
0.170
1.003
Landraces
Adjustment
1.032
245
Appendix 3(continued)
Primer 32
Estimated
repeat
length
0.030
Primer 33
0.980
1.031
Primer 34
0.665
0.777
Primer 35
0.460
0.903
Primer 36
-0.005
1.080
Primer 37
0.605
0.785
Primer 38
0.390
0.889
Primer 40
0.335
0.922
Primer 41
0.900
0.999
Primer 42
0.030
0.985
Primer 43
0.100
0.981
Primer 44
0.125
1.013
Primer 45
0.790
1.161
Primer 48
0.100
0.921
D1
-0.075
1.031
D2
0.130
0992
D3
0.680
0.774
D4
0.405
0.9089
D5
0.365
0.957
D6
0.115
0.975
D7
0.745
0.775
D8
0.105
1.079
D9
0.160
0.975
D10
0.385
0.842
D11
0.165
0.981
D12
0.035
0.985
D13
-0.020
1.100
D14
-0.015
1.027
D15
-0.005
1.021
E1
0.220
0.806
E3
0.565
0.775
E5
0.520
0.775
E7
0.440
0.866
E9
0.040
1.031
E10
0.160
0.957
E11
0.170
0.879
mBam2co80
0.345
0.859
Landraces
Adjustment
0.985
246
Appendix 4: A comparison of Nei and Li, (1979) similarity estimates for DArT marker (upper) and SSR markers (bottom) matrices calculated
using MVSP version 3.1 for the 24 bambara groundnut landraces.
DodR
DodC
AS17
DipC
SwaziRed
Tiganicuru
Ramayana
LunT
VSSP6
Nav 4
Nav Red
MHNblack
S19/3
S19-3
Uniswa
SB16 5A
AHM968
NAM 1761/3
Malawi 3
Tvsu 569
Tvsu 610
Tvsu 747
GabC
Tvsu 999
DodR
1
0.71
0.67
0.68
0.66
0.68
0.69
0.63
0.62
0.57
0.68
0.70
0.68
0.67
0.64
0.59
0.64
0.62
0.70
0.70
0.61
0.66
0.61
0.64
DodC
0.79
1
0.75
0.77
0.68
0.68
0.77
0.69
0.69
0.54
0.63
0.65
0.67
0.68
0.64
0.64
0.76
0.69
0.77
0.77
0.57
0.67
0.71
0.71
AS17
0.73
0.83
1
0.74
0.67
0.63
0.74
0.70
0.64
0.58
0.60
0.65
0.69
0.65
0.68
0.66
0.76
0.58
0.65
0.75
0.56
0.68
0.74
0.62
DipC
0.74
0.80
0.80
1
0.68
0.62
0.73
0.64
0.60
0.51
0.62
0.65
0.72
0.68
0.62
0.63
0.70
0.60
0.71
0.72
0.56
0.68
0.75
0.71
SwaziRed Tiganicuru Ramayana LunT
0.76
0.57
0.67
0.53
0.86
0.61
0.80
0.54
0.92
0.56
0.81
0.54
0.83
0.58
0.75
0.55
1
0.57
0.80
0.53
0.61
1
0.59
0.63
0.73
0.67
1
0.56
0.59
0.67
0.68
1
0.59
0.69
0.66
0.65
0.50
0.57
0.54
0.56
0.62
0.65
0.64
0.67
0.64
0.60
0.64
0.63
0.65
0.63
0.71
0.61
0.66
0.67
0.72
0.60
0.55
0.62
0.58
0.60
0.57
0.66
0.62
0.62
0.60
0.65
0.70
0.66
0.60
0.66
0.66
0.59
0.62
0.60
0.76
0.63
0.64
0.71
0.73
0.70
0.53
0.63
0.61
0.57
0.58
0.61
0.66
0.62
0.61
0.59
0.65
0.61
0.63
0.59
0.72
0.62
VSSP6
0.50
0.65
0.66
0.62
0.62
0.57
0.66
0.52
1
0.59
0.66
0.63
0.64
0.68
0.60
0.83
0.65
0.62
0.60
0.69
0.61
0.63
0.64
0.62
Nav 4
0.51
0.60
0.63
0.63
0.61
0.70
0.58
0.59
0.67
1
0.63
0.57
0.61
0.57
0.60
0.56
0.57
0.50
0.52
0.52
0.49
0.57
0.55
0.45
Nav Red MHNblack S19/3
0.49
0.71
0.73
0.57
0.80
0.83
0.59
0.83
0.91
0.59
0.81
0.83
0.58
0.82
0.89
0.68
0.57
0.56
0.57
0.77
0.82
0.61
0.51
0.55
0.63
0.62
0.67
0.74
0.66
0.62
1
0.55
0.57
0.77
1
0.84
0.73
0.75
1
0.67
0.68
0.87
0.58
0.64
0.68
0.67
0.60
0.70
0.61
0.65
0.71
0.57
0.58
0.56
0.63
0.68
0.69
0.65
0.59
0.69
0.62
0.54
0.60
0.64
0.67
0.66
0.63
0.60
0.73
0.61
0.65
0.66
S19-3
0.74
0.82
0.89
0.83
0.87
0.59
0.81
0.55
0.66
0.61
0.57
0.82
0.95
1
0.70
0.69
0.70
0.57
0.67
0.67
0.68
0.63
0.70
0.70
Uniswa SB16 5A AHM968NAM 1761/3Malawi 3 Tvsu 569 Tvsu 610 Tvsu 747
0.72
0.69
0.70
0.70
0.72
0.50
0.48
0.72
0.84
0.82
0.79
0.79
0.83
0.59
0.53
0.80
0.90
0.87
0.87
0.79
0.83
0.59
0.53
0.85
0.84
0.84
0.85
0.83
0.83
0.59
0.55
0.83
0.88
0.84
0.86
0.80
0.84
0.59
0.51
0.85
0.57
0.56
0.58
0.60
0.57
0.71
0.72
0.60
0.82
0.81
0.81
0.74
0.86
0.63
0.55
0.83
0.59
0.53
0.55
0.54
0.57
0.57
0.61
0.53
0.67
0.66
0.69
0.65
0.66
0.66
0.54
0.63
0.63
0.63
0.65
0.66
0.61
0.73
0.59
0.59
0.58
0.55
0.60
0.58
0.57
0.75
0.68
0.56
0.80
0.81
0.81
0.84
0.82
0.60
0.48
0.80
0.89
0.88
0.90
0.83
0.83
0.62
0.53
0.85
0.89
0.83
0.88
0.82
0.82
0.63
0.52
0.85
1
0.87
0.87
0.82
0.86
0.59
0.54
0.84
0.66
1
0.84
0.80
0.82
0.62
0.54
0.79
0.67
0.65
1
0.83
0.83
0.65
0.55
0.85
0.56
0.56
0.66
1
0.84
0.63
0.55
0.82
0.62
0.57
0.74
0.67
1
0.59
0.51
0.88
0.58
0.67
0.74
0.66
0.72
1
0.62
0.61
0.59
0.58
0.61
0.57
0.60
0.59
1
0.54
0.60
0.65
0.64
0.60
0.67
0.67
0.61
1
0.61
0.66
0.74
0.58
0.71
0.72
0.62
0.74
0.57
0.63
0.69
0.62
0.76
0.67
0.67
0.69
GabC
0.72
0.80
0.89
0.86
0.87
0.57
0.82
0.55
0.69
0.64
0.60
0.83
0.91
0.89
0.88
0.86
0.99
0.83
0.83
0.64
0.54
0.85
1
0.73
247
Tvsu 999
0.69
0.82
0.85
0.84
0.84
0.60
0.79
0.53
0.66
0.64
0.61
0.82
0.85
0.85
0.84
0.81
0.84
0.85
0.87
0.63
0.54
0.86
0.84
1
Appendix 5: Mean values for the characters of the 35 landraces grown in the
agronomy bay experiment (UK)
Landraces
DAE
DAF
LNO
SPRD
LL
LW
LA
PHT
ITN
PTL
PITN
PTLL
3Acc9NGA
9
39
82
12
7
3
2956
28
1
13
9
2
4Acc144NGA
10
41
49
34
10
4
2953
34
4
15
4
2
6Acc289BEN
12
41
57
13
8
3
2587
28
2
14
6
2
10Acc1276CAF
12
50
67
20
9
4
4316
38
3
22
8
3
20Acc118BFA
10
39
54
12
9
3
2512
30
2
13
7
2
30Acc476CMR
10
45
62
14
9
4
3391
30
1
15
11
3
33Acc484CMR
11
46
68
14
9
3
3470
32
2
16
10
2
40Acc563CMR
9
47
82
16
9
4
4896
33
2
14
8
2
45Acc231GHA
9
40
72
19
8
3
3318
24
1
14
10
2
48Acc790KEN
10
43
205
23
8
4
10369
36
2
18
8
3
49Acc793KEN
8
41
50
22
8
3
2281
34
3
16
7
3
50Acc792ZWE
12
49
49
12
8
3
1887
28
2
14
8
2
56Acc89MLI
10
39
35
11
8
3
1357
29
1
14
10
2
60Acc32NGA
10
42
54
14
8
3
2613
31
2
14
7
2
69Acc286NGA
10
47
47
11
9
4
2953
32
1
15
10
2
70Acc329NGA
11
45
84
12
8
3
4049
31
2
15
8
2
74Acc335NGA
10
43
52
12
8
4
2944
32
2
16
8
1
76Acc390SDN
13
41
39
9
9
3
1955
32
1
14
14
2
81Acc385TZA
8
38
104
27
8
4
6713
35
3
18
7
3
84Acc696ZMB
11
43
150
23
9
4
9939
38
3
21
8
4
85Acc754ZMB
8
45
198
25
8
3
9997
40
3
22
8
3
88-AHM753NAM
10
47
81
23
8
4
4347
32
3
16
5
2
90-S19-3NAM
9
39
55
19
9
4
3364
30
2
16
7
2
91-UNISRSWA
10
39
105
16
9
4
6059
36
2
17
9
3
92-AHM968NAM
10
39
79
22
8
4
4103
36
3
20
7
2
95-DODRTZA
10
39
68
34
8
4
3822
36
4
17
4
2
99-SB4-2NAM
10
38
116
32
8
4
6500
35
3
19
6
2
100-SB16ANAM
10
41
116
28
8
4
6429
36
3
19
8
2
104-S-1913NAM
10
39
65
21
9
4
4240
32
3
16
7
2
105-MHNblackAM
9
41
89
32
10
4
6794
45
5
25
6
2
109-BOTS1
11
41
42
11
9
4
2393
34
2
16
11
3
113-BOTS5
8
41
83
28
9
3
4557
39
2
19
9
2
117-VSSP6CMR
11
46
51
19
9
3
2704
29
2
13
6
2
118-Ramayana-IND
12
47
109
33
10
5
9583
39
3
22
7
3
119-Hybrid
10
43
62
13
10
4
4751
39
2
22
13
3
DAE: days to emergence, DAF: days to 50% flowering, LNO: number of leaves per plant, SPRD: plant spread/canopy, LL:
leaflet length, LW: leaflet width, LA: leaf area, PHT: plant height, ITN: Internode length, PTL: petiole length, PITN:
petiole-internode ratio, PTL: petiolule length, PNL: penduncle length, STEM: number of stems, DAM: days to maturity,
SDW: shoot dry weight, POD: pod number per plant, PDW:Pod dry weight, PODL: pod lenght, PODW:pod width, SNO:
seed number plant, SL: seed length, SW: seed width, SWE: seed weight
248
Appendix 5: continued
Landraces
PNL
STEM
DAM
SDW
POD
PDW
PODL
PODW
SNO
SL
SW
SWE
3Acc9NGA
2
8
152
17
36
32
19
12
54
10
8
20
4Acc144NGA
3
7
160
30
46
35
19
12
50
12
8
16
6Acc289BEN
3
9
151
11
17
58
21
10
25
10
7
8
10Acc1276CAF
2
7
161
36
46
56
20
13
52
11
8
21
20Acc118BFA
2
12
156
17
34
43
21
12
43
11
8
21
30Acc476CMR
2
9
153
24
54
48
21
13
60
11
8
26
33Acc484CMR
3
10
152
23
50
27
18
11
56
10
7
22
40Acc563CMR
3
11
155
34
20
27
21
15
25
12
10
13
45Acc231GHA
2
7
151
36
40
40
18
12
33
10
8
13
48Acc790KEN
2
14
160
68
106
34
20
11
105
11
8
32
49Acc793KEN
3
12
156
20
52
24
18
13
55
11
9
25
50Acc792ZWE
2
6
154
15
33
20
19
12
38
10
9
14
56Acc89MLI
2
7
148
10
28
19
21
11
40
11
9
21
60Acc32NGA
3
8
153
20
22
45
20
12
30
11
9
10
69Acc286NGA
2
6
152
13
10
36
16
10
15
9
7
3
70Acc329NGA
2
8
151
29
11
20
19
12
12
9
7
4
74Acc335NGA
3
7
156
23
20
26
21
14
26
12
9
11
76Acc390SDN
2
10
139
11
30
12
18
11
39
10
8
16
81Acc385TZA
2
9
160
43
86
17
20
13
102
12
9
46
84Acc696ZMB
2
14
143
54
35
4
20
11
50
10
8
13
85Acc754ZMB
2
11
160
76
106
7
18
12
94
11
8
29
88-AHM753NAM
2
12
160
34
121
17
16
11
133
10
8
36
90-S19-3NAM
2
9
143
22
48
21
15
11
48
10
8
21
91-UNISRSWA
2
14
160
53
105
60
19
13
107
11
9
49
92-AHM968NAM
2
10
161
36
87
23
18
13
96
10
8
32
95-DODRTZA
4
14
160
34
79
44
21
13
86
13
10
47
99-SB4-2NAM
2
10
157
47
100
47
15
11
105
10
8
29
100-SB16ANAM
2
12
155
48
67
27
17
12
64
11
8
25
104-S-1913NAM
2
13
161
37
92
68
18
12
103
11
9
46
105-MHNBlackNAM
3
14
160
63
81
47
22
14
76
13
9
34
109-BOTS1
2
8
156
21
47
60
21
13
53
12
9
33
113-BOTS5
2
10
160
32
84
39
18
12
77
10
9
33
117-VSSP6CMR
3
9
160
21
31
20
18
12
30
10
8
13
118-Ramayana-IND
3
8
161
81
48
27
18
14
50
11
9
19
119-HYBRID
2
11
161
30
31
17
18
11
40
11
9
13
DAE: days to emergence, DAF: days to 50% flowering, LNO: number of leaves per plant, SPRD: plant spread/canopy, LL: leaflet length,
LW: leaflet width, LA: leaf area, PHT: plant height, ITN: Internode length, PTL: petiole length, PITN: petiole-internode ratio, PTL: petiolule
length, PNL: penduncle length, STEM: number of stems, DAM: days to maturity, SDW: shoot dry weight, POD: pod number per plant,
PDW:pod dry weight, PODL: pod lenght, PODW:pod width, SNO: seed number plant, SL: seed length, SW: seed width, SWE: seed weight
249
Appendix 6: Mean values for the characters of the 34 lines grown in the field
experiment (Botswana).
Lines
DAE
DAF
LNO
SPRD
LL
LW
3Acc9NGA
16
57
208
27
6
2
4Acc144NGA
15
53
85
49
8
6Acc289BEN
16
51
113
19
10Acc1276CAF
18
66
67
20Acc118BFA
16
54
30Acc476CMR
17
33Acc484CMR
LA
PHT
ITN
PTL
PITN
PTLL
4329
25
2
14
6
2
3
2878
26
5
13
4
2
7
3
3391
23
2
12
7
2
15
7
3
2157
26
2
102
6
2
164
17
6
2
3988
25
2
12
8
2
58
108
19
7
2
2861
26
2
12
7
1
16
62
119
21
6
2
2811
28
2
14
8
2
40Acc563CMR
16
63
119
22
7
2
3614
29
2
16
10
2
45Acc231GHA
15
60
73
14
6
2
1446
22
1
10
9
2
48Acc790KEN
13
65
142
23
7
3
4586
29
1
15
10
2
50Acc792ZWE
14
61
147
25
6
2
3059
23
2
13
6
1
56Acc89MLI
16
52
105
13
4
2
1614
23
1
12
10
1
60Acc32NGA
14
51
125
23
6
2
2583
23
2
10
6
1
69Acc286NGA
11
59
57
29
10
4
4319
33
2
17
8
2
70Acc329NGA
18
59
23
17
5
2
387
25
1
11
8
2
74Acc335NGA
17
61
126
21
7
3
3786
26
2
12
7
1
76Acc390SDN
14
51
153
29
8
3
5986
32
2
16
10
2
81Acc385TZA
13
63
174
31
6
3
5437
28
3
17
7
2
84Acc696ZMB
12
56
121
26
8
3
4343
30
2
16
9
3
85Acc754ZMB
14
64
143
21
7
3
4098
32
2
18
11
3
88-AHM753NAM
14
54
118
26
7
3
3879
27
5
15
5
2
90-S19-3NAM
17
53
98
33
8
3
4019
30
3
16
6
2
91-UNISRSWA
15
53
88
18
7
2
2442
30
2
13
6
2
92-AHM968NAM
16
54
82
26
7
2
2062
26
2
14
9
2
95-DODRTZA
13
47
104
50
7
3
3152
32
4
16
5
2
99-SB4-2NAM
17
58
99
22
7
3
3256
29
2
15
8
2
100-SB16ANAM
17
56
95
25
7
3
3428
32
3
15
5
2
104-S-1913NAM
14
48
87
22
6
3
2715
28
3
13
5
1
105-MHNBlackNAM
14
64
94
24
8
3
3306
35
3
17
7
2
109-BOTS1
13
54
124
25
7
3
3982
28
2
14
7
2
113-BOTS5
15
47
120
18
8
2
3865
33
2
18
7
2
117-VSSP6CMR
16
64
86
30
7
2
2252
25
2
12
6
2
118-Ramayana-IND
16
61
97
19
6
2
2609
28
2
10
4
1
119-HYBRID
16
54
83
17
7
2
2383
30
2
16
8
2
DAE: days to emergence, DAF: days to 50% flowering, LNO: number of leaves per plant, SPRD: plant spread/canopy, LL: leaflet length,
LW: leaflet width, LA: leaf area, PHT: plant height, ITN: Internode length, PTL: petiole length, PITN: petiole-internode ratio, PTL: petiolule
length, PNL: penduncle length, STEM: number of stems, DAM: days to maturity, SDW: shoot dry weight, POD: pod number per plant,
PDW:pod dry weight, PODL: pod lenght, PODW:pod width, SNO: seed number plant, SL: seed length, SW: seed width, SWE: seed weight
250
Appendix 6: Continued
Lines
PNL
STEM
3Acc9NGA
3
8
4Acc144NGA
2
9
6Acc289BEN
2
10Acc1276CAF
DAM
SDW
POD
138
36
7
138
31
8
8
140
29
9
1
8
144
45
20Acc118BFA
2
5
127
30Acc476CMR
1
7
33Acc484CMR
2
40Acc563CMR
PDW
PODL
PODW
SNO
SL
SW
3
16
10
7
10
7
2
19
19
11
9
10
7
13
4
20
9
7
11
8
2
8
5
25
11
8
10
7
2
22
11
10
18
12
12
12
9
7
138
38
13
7
18
11
14
11
8
4
8
144
27
23
16
21
12
22
11
9
8
2
11
*
45
*
*
*
*
*
*
*
*
45Acc231GHA
1
7
*
19
*
*
*
*
*
*
*
*
48Acc790KEN
2
8
155
61
11
4
16
10
11
11
8
3
50Acc792ZWE
1
12
*
48
*
*
*
*
*
*
*
*
56Acc89MLI
1
8
134
22
6
3
17
11
6
9
7
1
60Acc32NGA
2
11
127
39
12
7
23
13
13
13
10
4
69Acc286NGA
1
15
*
16
*
*
*
*
*
*
*
*
70Acc329NGA
1
7
*
14
*
*
*
*
*
*
*
*
74Acc335NGA
2
6
139
23
11
10
22
12
10
12
8
4
76Acc390SDN
2
7
142
83
10
7
18
11
12
12
8
4
81Acc385TZA
3
9
146
103
10
2
16
8
8
9
7
1
84Acc696ZMB
2
6
133
51
36
10
20
11
33
12
9
8
85Acc754ZMB
88AHM753NAM
2
10
143
72
17
7
19
10
20
10
7
4
2
8
136
47
35
13
14
9
31
10
8
9
90-S19-3NAM
3
8
129
48
68
28
19
11
66
12
9
18
91-UNISRSWA
92AHM968NAM
3
9
134
37
9
8
16
12
9
9
8
6
2
6
127
25
21
7
15
10
18
9
9
5
95-DODRTZA
3
9
155
38
5
3
16
9
5
11
9
2
99-SB4-2NAM
2
11
130
29
21
8
16
10
21
11
7
6
100-SB16ANAM
3
6
139
45
19
12
19
11
18
12
9
6
104-S-1913NAM
105MHNBlackNAM
2
7
134
21
23
13
15
9
22
10
8
10
2
8
127
44
9
4
25
9
9
13
8
4
109-BOTS1
2
14
139
70
19
9
19
11
18
13
9
6
113-BOTS5
2
8
127
50
5
2
17
10
5
10
7
1
117-VSSP6CMR
118-RamayanaIND
2
6
*
27
*
*
*
*
*
*
*
*
2
7
133
32
8
5
15
11
7
11
8
4
119-HYBRID
2
9
136
46
11
7
19
12
12
12
9
3
DAE: days to emergence, DAF: days to 50% flowering, LNO: number of leaves per plant, SPRD: plant spread/canopy, LL: leaflet length,
LW: leaflet width, LA: leaf area, PHT: plant height, ITN: Internode length, PTL: petiole length, PITN: petiole-internode ratio, PTL: petiolule
length, PNL: penduncle length, STEM: number of stems, DAM: days to maturity, SDW: shoot dry weight, POD: pod number per plant,
PDW:pod dry weight, PODL: pod lenght, PODW:pod width, SNO: seed number plant, SL: seed length, SW: seed width, SWE: seed weight
251
SWE
Appendix7: Range of classes for the quantitative traits used for both the
glasshouse and the field experiment
Characters
Days to emergence
Days to 50% flowering
Leaf number
Canopy spread
leaflet length
Leaflet width
Leaf area
Plant height
Petiole
Internode
Pet-internode ratio
Petiolule
Peduncle
Stem number
Days to maturity
Shoot dry weight
Pod number
Pod dry weight
Pod length
Pod width
Seed number
Seed length
Seed width
Seed weight
<7
<30
<50
<11
<7
<3
<30
<<25
<13
<1
<7
<1
<1
<7
<100
<15
<20
<10
<15
<10
<20
<9
<7
<10
Range of class
8-10
31-35
51-100
11-20
8-10
4-5
31-45
26-30
2-3
14-16
8-9
2-3
2-3
8-10
101-115
16-30
21-40
11-20
16-20
11-14
21-40
10-14
8-10
11-15
>11
36-40
101-150
21-30
>11
>6
46-55
31-35
>4
17-19
>10
>4
>4
>11
116-130
31-45
41-60
21-30
>21
>15
41-60
>15
>11
16-20
41-45
151-200
31-40
>46
>201
>41
56-65
36-40
>66
>41
20-22
>23
131-145
46-60
61-80
31-40
>145
>61
>80
>41
61-80
>81
21-30
>31
252
Appendix 8: Hardy Weinberg Equilibrium (HWE) and the exact p-values
estimated using PowerMarker (Version 3.25)
Marker
X2 value
X2d.f.
Exact p-value
Primer 7
984.00
36
0.0000
Primer 15
1633.06
120
0.0000
Primer 16
1095.73
45
0.0000
Primer 19
3321.00
378
0.0000
Primer 23
615.00
15
0.0000
Primer 33
1599.00
91
0.0000
Primer 37
1599.23
120
0.0000
Primer 44
615.00
15
0.0000
mBam3co18
2706.00
253
0.0000
Primer D11
1968.00
136
0.0000
Primer D14
3690.00
465
0.0000
Primer E7
482.33
10
0.0000
253
Appendix 9: Cluster analysis, genetic similarity among the 105 bambara
groundnut genotypes, analysis using 141 variables and 105samples/cases
UPGMA Nei and Li’s coefficient
Node
Group 1
Group 2
1
20Acc118CIV
20Acc118CIV
2
117VSSP6 CMR
117VSSP6 CMR
3
6Acc 289BEN
4
Simil.
Objects in group
0.95
2
0.923
2
6Acc 289BEN
0.9
2
92AHM968NAM
92AHM968NAM
0.9
2
5
60Acc 32NGA
60Acc 32NGA
0.9
2
6
Node 2
117VSSP6 CMR
0.883
3
7
4Acc144GHA
4Acc144GHA
0.878
2
8
Node 1
20Acc118CIV
0.875
3
9
30Acc 476CMR
30Acc 476CMR
0.872
2
10
33Acc 484CMR
33Acc 484CMR
0.872
2
11
40Acc 536CMR
40Acc 536CMR
0.872
2
12
85Acc 754ZMB
85Acc 754ZMB
0.865
2
13
Node 10
33Acc 484CMR
0.861
3
14
Node 4
92AHM968NAM
0.849
3
15
84Acc696ZMB
84Acc696ZMB
0.842
2
16
99SB4-2NAM
99SB4-2NAM
0.842
2
17
70Acc 329NGA
70Acc 329NGA
0.829
2
18
100SB16 ANAM
100SB16 ANAM
0.821
2
19
88AHM753NAM
88AHM753NAM
0.821
2
20
105MHN blackNAM
105MHN blackNAM
0.821
2
21
90S19-3NAM
104S-1913NAM
0.821
2
22
119Hyrid
119Hyrid
0.81
2
23
45Acc 231GHA
45Acc 231GHA
0.8
2
24
76Acc390SDN
76Acc390SDN
0.8
2
25
88AHM753NAM
Node 19
0.79
3
26
74Acc335NGA
74Acc335NGA
0.789
2
27
69Acc286NGA
69Acc286NGA
0.789
2
28
118RamayanaIND
118RamayanaIND
0.789
2
29
40Acc 536CMR
Node 11
0.785
3
30
105MHN blackNAM
Node 20
0.785
3
31
3Acc 9NGA
3Acc 9NGA
0.78
2
32
Node 9
30Acc 476CMR
0.779
3
33
Node 3
6Acc 289BEN
0.775
3
34
Node 23
45Acc 231GHA
0.775
3
35
4Acc144GHA
Node 7
0.771
3
254
Appendix 9 Continued
Node
Group 1
Group 2
Simil.
Objects in group
36
10Acc 1276CAF
10Acc 1276CAF
0.757
2
37
81Acc385TZA
81Acc385TZA
0.75
2
38
118RamayanaIND
Node 28
0.749
3
39
Node 17
70Acc 329NGA
0.74
3
40
90S19-3NAM
104S-1913NAM
0.737
2
41
Node 18
100SB16 ANAM
0.734
3
42
76Acc390SDN
Node 24
0.732
3
43
90S19-3NAM
Node 21
0.727
3
44
56Acc 89MLI
56Acc 89MLI
0.722
2
45
109BWA1BWA
109BWA1BWA
0.718
2
46
119Hyrid
Node 22
0.718
3
47
95DODRTZA
95DODRTZA
0.718
2
48
Node 26
74Acc335NGA
0.711
3
49
84Acc696ZMB
Node 15
0.71
3
50
69Acc286NGA
Node 27
0.7
3
51
Node 32
Node 13
0.696
6
52
10Acc 1276CAF
Node 36
0.693
3
53
Node 31
3Acc 9NGA
0.69
3
54
Node 40
Node 43
0.687
5
55
99SB4-2NAM
Node 16
0.683
3
56
48Acc790KEN
48Acc790KEN
0.667
2
57
49Acc793KEN
49Acc793KEN
0.667
2
58
113BWA5BWA
113BWA5BWA
0.667
2
59
Node 45
109BWA1BWA
0.658
3
60
95DODRTZA
Node 47
0.649
3
61
50Acc 792KEN
50Acc 792KEN
0.632
2
62
60Acc 32NGA
Node 5
0.625
3
63
Node 54
104S-1913NAM
0.624
6
64
56Acc 89MLI
Node 44
0.602
3
65
48Acc790KEN
85Acc 754ZMB
0.6
2
66
Node 49
81Acc385TZA
0.597
4
67
Node 33
50Acc 792KEN
0.583
4
68
Node 46
91UNIS RSWA
0.58
4
69
Node 56
Node 12
0.579
4
70
Node 35
Node 62
0.574
6
71
Node 55
Node 41
0.569
6
255
Appendix 9 Continued
Node
Group 1
Group 2
72
Node 69
Node 30
73
Node 57
74
Simil.
Objects in group
0.56
7
49Acc793KEN
0.556
3
91UNIS RSWA
113BWA5BWA
0.55
2
75
Node 60
Node 38
0.545
6
76
Node 51
Node 48
0.537
9
77
Node 53
Node 70
0.533
9
78
Node 34
Node 73
0.533
6
79
Node 68
Node 74
0.521
6
80
Node 25
Node 65
0.517
5
81
Node 67
Node 64
0.514
7
82
Node 61
Node 50
0.507
5
83
Node 77
Node 39
0.505
12
84
Node 37
Node 59
0.505
5
85
Node 76
Node 42
0.499
12
86
Node 72
Node 14
0.48
10
87
Node 58
Node 6
0.48
5
88
Node 85
Node 29
0.474
15
89
Node 80
Node 63
0.47
11
90
Node 81
Node 78
0.47
13
91
Node 83
Node 88
0.466
27
92
Node 75
Node 79
0.462
12
93
Node 86
Node 84
0.457
15
94
Node 66
Node 87
0.449
9
95
91UNIS RSWA
Node 92
0.441
13
96
Node 91
Node 90
0.432
40
97
Node 89
Node 71
0.43
17
98
Node 96
Node 82
0.412
45
99
Node 94
Node 95
0.406
22
100
Node 93
Node 97
0.403
32
101
Node 100
Node 99
0.395
54
102
Node 98
Node 52
0.376
48
103
Node 102
Node 8
0.376
51
104
Node 103
Node 101
0.371
105
256
Appendix 10: Scatter plots for morpho-agronomic markers on (Euclidean
distance estimates) and molecular markers on (Nei’s 1972) conducted using
Mantel’s test on NTSYS, Pearson correlation and Spearman’s rank correlations on
SPSS in the Agronomy bay and controlled growth room experiment: Appendix
10.1 and 10.2.
a)
0.13
Genetic
0.07
0.00
-0.06
- 0.12
-0.16
- 0.10
- 0.04
0.03
0.09
Morphology
b)
Appendix 10.1: A scatter plot for Agronomy bay experiment on34 morpho-agronomic markers
and 20 SSR molecular marker, a) Mantel’s test (r = 0.139; P<0.006) b) Pearson correlation (r
=0.767: P<0.001) and Spearman rank correlation (r=771; P<0.001).
257
a)
0.27
Genetic
0.19
0.11
0.03
- 0.05
- 0.08
-0.03
0.03
0.09
Morphology
0.14
b)
Appendix 10.2: A scatter plot for controlled growth experiment on 22 morpho-agronomic markers
and 12 SSR molecular marker, a) Mantel’s test (r = 0.612; P<0.001) b) Pearson correlation (r
=0.665; P<0.001) and Spearman rank correlation (r=0.461; P<0.001).
258
Appendix 11: Mean for the phenotypic measures of 5 lines from the controlled
growth room experiment
DAE: days to maturity, LNO: number of leaves per plant, LL: leaflet length, LW: leaflet width, LA: leaf area, PHT: plant height, ITN:
Internode length, PTL: petiole length, PITN: petiole-internode ratio, PNL: penduncle length,SDW: shoot dry weight, POD: pod number per
plant, PDW:pod dry weight, PODL: pod lenght, SNO: seed number plant, SL: seed length,SWE: seed weight
259
Odireleng Molosiwa