University of Ghana http://ugspace.ug.edu.gh
 
 
CHARACTERIZATION OF CULTIVATED GROUNDNUT (Arachis hypogaea L.) AND 
GENETIC ANALYSIS OF LATE LEAF SPOT RESISTANCE IN TOGO 
 
By 
 
 
ESSOHOUNA MODOM BANLA 
(10509160) 
 
 
 
THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL 
FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF DOCTOR OF 
PHILOSOPHY DEGREE IN PLANT BREEDING  
 
 
 
 
WEST AFRICA CENTRE FOR CROP IMPROVEMENT 
COLLEGE OF BASIC AND APPLIED SCIENCES 
UNIVERSITY OF GHANA 
LEGON 
 
 
DECEMBER, 2018 
i 
 
University of Ghana http://ugspace.ug.edu.gh
DECLARATION 
I hereby declare that except for references to works of other researchers, which have been 
duly cited, this work is my original research and that neither part nor whole has been 
presented elsewhere for the award of a degree. 
 
…………………………………… 
Essohouna Modom BANLA 
(Student) 
 
 
…………………………………… 
Prof. Pangirayi TONGOONA 
(Supervisor) 
 
 
…………………………………… 
Prof. Samuel Kwame OFFEI 
(Supervisor) 
 
 
…………………………………… 
Dr. Daniel Kwadjo DZIDZIENYO 
(Supervisor) 
 
 
…………………………………… 
Dr. Haile DESMAE 
(Supervisor) 
  
i 
 
University of Ghana http://ugspace.ug.edu.gh
ABSTRACT 
Though an important crop, the yield of groundnut has been decreasing for years because of a 
lack of a strong breeding programme in Togo. In recent years, efforts have been going on to 
establish a well-functioning breeding programme to develop varieties for improving 
productivity and production of groundnut. The main aims of this study were to: assess farmers’ 
perceptions of production constraints and preference criteria, characterize cultivated groundnut 
varieties and study genetic mechanisms for priority traits. Results will be useful to exploit 
genetic diversity among the cultivated groundnut accessions and to integrate farmers’ 
knowledge and expectations in the development of high yielding and adaptable groundnut 
varieties that are preferred by farmers, marketers, processors and consumers in Togo. The 
investigation on production constraints and farmers’ preferences showed that yield related traits 
such as pod yield (66.66%) and pod size (12.12%) were the most valued selection criteria by 
farmers. According to the farmers, groundnut production is mainly constrained by diseases 
(37.77% of the respondents). Important economic losses and various causes were associated 
with late leaf spot (LLS) disease in the surveyed areas. Diversity assessment of 94 accessions 
from Togo, Senegal and ICRISAT, using agro-morphological and molecular markers, revealed 
that quantitative traits such as LLS incidence, number of pods per plant and yield displayed an 
adequate variability (coefficient of variation > 20%). The most diverse qualitative traits based 
on Simpson index were primary seed colour (0.75), stem pigmentation (0.60), and growth habit 
(0.59). Moreover, principal component analysis underscored quantitative traits such as hundred 
seed weight, days to maturity, and LLS incidence, as the main traits contributing to the 
divergence. Correlation analysis and path analysis showed that the number of pods per plant 
was the main yield-related trait positively affecting yield (PC=0.84; p=0.01). Overall, SNP 
markers revealed high genetic variability in the genotypes and the percentage of heterozygous 
genotypes varied from 0 to 50% for all loci. AMOVA revealed that only 1.1% of the total 
ii 
 
University of Ghana http://ugspace.ug.edu.gh
molecular variance accounted for geographical contribution to the diversity. Cluster analysis 
delineated three clusters harbouring useful alleles and interesting phenotypic features such as 
LLS resistance, a high number of pods per plant and early maturity. Structure co-analysis of 
phenotypic and SNP data showed that differences observed at the phenotypic level are 
underlined by genotypic differences. The phenotypic and genotypic diversity revealed 
appreciable diversity that could be exploited for the identification of parents with preferred 
traits for use in the breeding programme in Togo. The linkage disequilibrium (LD) in Togo 
groundnut collection was found to be high (0.36).  Mean r2 ranged from 0.058 to 0.99 with the 
genomic distance between loci ranging from 0 to 148 Mbp for intrachromosomal linkage 
disequilibrium. The interchromosomal LD was found to be significant at p = 0.01 with a certain 
number of loci located on different chromosomes exhibiting r2 above 0.1. Association study 
revealed that out of 31 agro-morphological traits analyzed, significant associations between 
SNPs and agronomic traits of interest (p ≤ 10-7) were established for nine traits. Chromosomes 
A02, A03, A05 and B10 were significantly associated with LLS resistance with r2 ranging from 
0.50 to 0.57. Hundred seed weight, pod width, pod length, seed length, seed width, growth habit 
and pod reticulation were also associated with various genomic regions. Genetic mechanism 
analysis for LLS using sixteen F2 progenies from full factorial design crosses indicated additive 
gene and maternal effects for LLS resistance with 48,3% and 3.86% of the phenotypic variance 
explained, respectively. Non-additive genetic variance was close to zero for LLS resistance. In 
contrast to LLS resistance, non-additive gene effects played a greater role than additive effects 
for pod weight per plot (21.79% phenotypic variance explained). For the other yield-related 
traits, additive genetic variance was more important than the other variance components. 
Accessions 43AH and ICG 7878 with the highest general combining ability could be considered 
as the best parents for high yielding LLS resistant lines development.  
iii 
 
University of Ghana http://ugspace.ug.edu.gh
DEDICATION 
To my family, brothers, sisters and friends. Your prayers and support made it easier for me to 
persevere to the end. 
  
iv 
 
University of Ghana http://ugspace.ug.edu.gh
ACKNOWLEDGEMENT 
I am sincerely grateful to my supervisors Professor Pangirayi Tongoona, Professor Samuel 
Kwame Offei, Dr. Daniel Kwadjo Dzidzienyo and Dr. Haile Desmae for their guidance and 
valuable inputs into this thesis. I also thank Dr. Beatrice Elohor Ifie, for her priceless support.  
My gratitude to the staff of West Africa Centre for Crop Improvement (WACCI) for the 
opportunity given to me to study in such a prestigious school. I would sincerely like to thank 
the Director, Professor E.Y. Danquah and all his staff for their support during the period of this 
thesis. 
I thank the National agricultural extension agents (ICAT) and all the people who participated 
in this research. Also, I thank my institution (ITRA) for granting me a study leave and for their 
support through WAAPP-Togo. 
For the genotyping part of this work, my special thanks to Dr. Sneller Clay of Integrated 
genotyping Service Support (IGSS) for sponsoring the genotyping of the germplasm collection. 
My gratitude to Dr. Daniel Fonceka from CERAAS (Senegal) for his help. Dr. Fonceka paid 
for the genotyping of F1. 
Special thanks to ICRISAT-WCA, Mali for hosting me during my research. I would like to 
express my gratitude to Dr. Haile and all his team for their support for this thesis achievement. 
I really appreciated the ambiance that prevailed during my stay at ICRISAT-WCA. 
My gratitude to DAAD-German academic exchange for the generous funding I received from 
them in the form of PhD scholarship through the WACCI. 
My wife Nande and my children, Benih and Nice deserve special mention for the sacrifices 
made in order for me to achieve this work. 
v 
 
University of Ghana http://ugspace.ug.edu.gh
TABLE OF CONTENTS 
 
DECLARATION ........................................................................................................................ i 
ABSTRACT ............................................................................................................................... ii 
DEDICATION .......................................................................................................................... iv 
ACKNOWLEDGEMENT ......................................................................................................... v 
TABLE OF CONTENTS .......................................................................................................... vi 
LIST OF FIGURES .................................................................................................................. ix 
LIST OF TABLES .................................................................................................................... xi 
LIST OF ABBREVIATIONS .................................................................................................. xii 
CHAPTER ONE ........................................................................................................................ 1 
1. GENERAL INTRODUCTION .............................................................................................. 1 
CHAPTER TWO ....................................................................................................................... 6 
2. LITERATURE REVIEW ...................................................................................................... 6 
2.1 Origin and biology ........................................................................................................... 6 
2.2 Groundnut cytogenetics .................................................................................................... 7 
2.3 Groundnut genetic resources ............................................................................................ 8 
2.4 Genetic diversity of important traits in groundnut ........................................................... 9 
2.5 Gene action in groundnut ............................................................................................... 10 
2.6 Use of molecular markers for groundnut improvement ................................................. 12 
2.7 Constraints and genetic improvement of cultivated groundnut ..................................... 14 
2.8 Late Leaf Spot disease.................................................................................................... 15 
2.8.1 Origin and occurrence ............................................................................................. 15 
2.8.2 Symptoms ................................................................................................................ 15 
2.8.3 Causal agents ........................................................................................................... 16 
2.8.4 Epidemiology and disease cycle .............................................................................. 17 
2.8.5 Disease control ........................................................................................................ 18 
2.8.6 Resistance to LLS disease ....................................................................................... 19 
2.8.7 Genetic resources for LLS resistance ...................................................................... 19 
2.8.8 Identification and use of molecular markers for LLS .............................................. 20 
2.9 Demand-led breeding in groundnut................................................................................ 21 
2.10 Groundnut production in Togo ..................................................................................... 23 
CHAPTER THREE ................................................................................................................. 25 
3. GROUNDNUT PRODUCTION CONSTRAINTS AND FARMERS’ TRAIT 
PREFERENCES .................................................................................................................. 25 
vi 
 
University of Ghana http://ugspace.ug.edu.gh
3.1 Introduction .................................................................................................................... 25 
3.2 Methodology .................................................................................................................. 26 
3.2.1 Description of the study areas ................................................................................. 26 
3.2.2 Data collection ......................................................................................................... 26 
3.2.3 Data analysis ............................................................................................................ 28 
3.3 Results ............................................................................................................................ 30 
3.3.1 Demographic characteristics of farmers .................................................................. 30 
3.3.2 Practices in groundnut production ........................................................................... 35 
3.3.3 Farm characteristics ................................................................................................. 35 
3.3.4 Cultivated groundnut varieties................................................................................. 38 
3.3.5 Preferred characteristics of groundnut varieties ...................................................... 39 
3.3.6 Perception of farmers on constraints to groundnut production ............................... 42 
3.3.7 Awareness of Late Leaf Spot disease ...................................................................... 44 
3.3.8 Correlation among social and farming system parameters ...................................... 46 
3.4 Discussion ...................................................................................................................... 48 
3.5 Conclusion ...................................................................................................................... 53 
CHAPTER FOUR .................................................................................................................... 54 
4. MOLECULAR AND PHENOTYPIC DIVERSITY OF GROUNDNUT CULTIVARS 
FROM TOGO ...................................................................................................................... 54 
4.1 Introduction .................................................................................................................... 54 
4.2 Materials and Methods ................................................................................................... 55 
4.2.1 Collection of accessions .......................................................................................... 55 
4.2.2 Plant material ........................................................................................................... 58 
4.2.3 Morphological evaluation ........................................................................................ 58 
4.2.4 DNA Extraction and DArT Analysis....................................................................... 61 
4.2.5 Data analysis ............................................................................................................ 61 
4.3 Results ............................................................................................................................ 62 
4.3.1 Morphological diversity .......................................................................................... 62 
4.3.2 Molecular diversity .................................................................................................. 78 
4.4 Discussion ...................................................................................................................... 86 
4.5 Conclusion ...................................................................................................................... 93 
CHAPTER FIVE ..................................................................................................................... 95 
5. ANALYSIS OF LINKAGE DISEQUILIBRIUM AND MARKER-TRAIT 
ASSOCIATION STUDY FOR LLS AND YIELD-RELATED TRAITS .......................... 95 
5.1 Introduction .................................................................................................................... 95 
vii 
 
University of Ghana http://ugspace.ug.edu.gh
5.2 Materials and methods ................................................................................................... 96 
5.2.1 DNA Extraction and DArT Analysis....................................................................... 96 
5.2.2 Data analysis ............................................................................................................ 96 
5.3 Results ............................................................................................................................ 97 
5.3.1 Linkage disequilibrium ............................................................................................ 97 
5.3.2 Genome-wide association ...................................................................................... 101 
5.4 Discussion .................................................................................................................... 103 
5.5 Conclusion .................................................................................................................... 106 
CHAPTER SIX ...................................................................................................................... 108 
6. GENETIC ANALYSIS OF LATE LEAF SPOT RESISTANCE, YIELD AND YIELD 
COMPONENTS ................................................................................................................ 108 
6.1 Introduction .................................................................................................................. 108 
6.2 Materials and Methods ................................................................................................. 109 
6.2.1 Plant materials ....................................................................................................... 109 
6.2.2 Test environments and experimental design ......................................................... 110 
6.2.3 Data collection ....................................................................................................... 110 
6.2.4 Data analysis .......................................................................................................... 112 
6.3 Results .......................................................................................................................... 113 
6.3.1 Variance components in LLS and yield-related traits ........................................... 113 
6.3.2 GCA for LLS and yield-related traits .................................................................... 115 
6.3.3 SCA for LLS and yield-related traits ..................................................................... 118 
6.3.4 Traits correlation for SCA and GCA ..................................................................... 121 
6.3.5 Correlation between SCA and genetic distance .................................................... 122 
6.4 Discussion .................................................................................................................... 123 
6.5 Conclusion .................................................................................................................... 127 
CHAPTER SEVEN ............................................................................................................... 128 
7. GENERAL CONCLUSIONS AND RECOMMENDATIONS......................................... 128 
7.1 General conclusions ..................................................................................................... 128 
7.2 Recommendations ........................................................................................................ 129 
REFERENCES ...................................................................................................................... 131 
APPENDICES ....................................................................................................................... 151 
 
 
  
viii 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF FIGURES 
Figure 3.1 Map of Togo showing the surveyed area ............................................................... 27 
Figure 3.2 Educational qualification. ....................................................................................... 31 
Figure 3.3 Land available for agriculture and land area (ha) used for groundnut production as 
estimated by interviewees ....................................................................................... 35 
Figure 3.4 Source of groundnut seeds (in percentage) ............................................................ 37 
Figure 3.5 Proportion of groundnut sold.................................................................................. 38 
Figure 3.6 Perception on the cause of LLS disease on groundnut (figures are in percentage) 45 
Figure 4.1 Map of Togo showing the collection sites .............................................................. 57 
Figure 4.2 A principal component analysis showing similarity relationships among 92 
accessions based on quantitative traits. . ................................................................. 64 
Figure 4.3 A principal component analysis showing similarity relationships among 92 
accessions based on qualitative traits. . ................................................................... 66 
Figure 4.4 Plot of the direct and indirect effect of pod and seed traits on yield.  .................... 69 
Figure 4.5 Optimal number of clusters, a-K = 2 and b-K = 3 with probability of membership 
of genotypes ............................................................................................................ 70 
Figure 4.6 Dendrogram showing optimal two clusters. ........................................................... 71 
Figure 4.7 Dendrogram showing three clusters. ...................................................................... 72 
Figure 4.8 Dendrogram based on quantitative traits (a) and qualitative traits (b) in diversity 
assessment of 92 groundnut accessions................................................................... 77 
Figure 4.9 Distribution of SNPs based on PIC (a) and MAF (b) ............................................. 78 
Figure 4.10 a-Structure of heterogeneity detected in 92 groundnut accessions; b-Distribution 
of fixation index (Fst) and c-Frequency distribution of total genetic diversity (Ht)
 ................................................................................................................................. 79 
Figure 4.11 Multilocus genetic dissimilarity versus geographic distance.. ............................. 80 
Figure 4.12 A principal component analysis biplot diagram showing similarity relationships 
among populations based on SNP markers ............................................................. 81 
Figure 4. 13 (a) Molecular data dendrogram showing the optimal two clusters; Silhouette 
width of the optimal clusters. .................................................................................. 82 
Figure 4.14 Cluster analysis: two clusters solution (left); three cluster solution (right).......... 83 
Figure 4.15 (A)Dendrogram showing the three clusters based on the marker and phenotypic 
data; B = Optimal number of clusters ..................................................................... 84 
Figure 4.16 Number of alleles vs location sample size. .......................................................... 85 
ix 
 
University of Ghana http://ugspace.ug.edu.gh
Figure 4.17 Agronomic characteristics of the three clusters in the co-analysis of phenotypic 
and molecular data................................................................................................... 86 
Figure 5.1 A principal component biplot diagram showing the three clusters identified by 
structure analysis. .................................................................................................... 99 
Figure 5.2 Linkage Disequilibrium decay in 92 groundnut accessions ................................. 101 
Figure 5.3 Illustration of Manhattan plot depicting strongly associated markers to different 
traits.. ..................................................................................................................... 103 
  
x 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF TABLES 
Table 3.1 Sites of PRA and the number of respondents ....................................................................... 29 
Table 3.2 Ages of groundnut farmers in the surveyed area .................................................................. 31 
Table 3.3 Educational level and household size of groundnut farmers in the surveyed area in Togo. . 33 
Table 3.4 Gender and marital status of groundnut farmers in the surveyed area in Togo .................... 34 
Table 3.5 Farm size, seed sources, cropping system and proportion of groundnut sold ....................... 36 
Table 3.6 Characteristics of cultivated varieties according to farmers in the surveyed area ................ 39 
Table 3.7 Percentage of time preferred characteristic of groundnut varieties were mentioned ............ 41 
Table 3.8 Percentage of times that constraints were mentioned ........................................................... 43 
Table 3.9 Farmers’ perception of LLS .................................................................................................. 44 
Table 3.10 Correlation among social and farming system parameters ................................................. 47 
Table 4.1 Mean square, range and coefficient of variation of quantitative traits .................................. 63 
Table 4.2 Correlation between variables and factors (Eigenvectors).................................................... 65 
Table 4.3 Correlation analysis for quantitative traits ............................................................................ 68 
Table 4.4 Mean of yield-related traits for optimal two clusters ............................................................ 72 
Table 4.5 Ranges of traits in three cluster grouping of the 92 groundnut accessions ........................... 74 
Table 4.6 Qualitative features of k = 3 clusters .................................................................................... 75 
Table 4.7 Pairwise Fst and Nei’s genetic distance value for populations ............................................. 81 
Table 4.8 Pairwise Fst value and Nei’s genetic distance for clusters ................................................... 85 
Table 5.1 Intrachromosomal linkage disequilibrium pattern ................................................................ 98 
Table 5.2 LD characteristics of clusters .............................................................................................. 100 
Table 5.3 Linkage groups for agronomic traits and r2 range ............................................................... 102 
Table 6.1 Parent genotypes characteristics ......................................................................................... 110 
Table 6.2 Description of late leaf spot rating scale ............................................................................. 111 
Table 6.3 Variance component for LLS by sites................................................................................. 113 
Table 6.4 Variance components for yield and yield-related traits across sites ................................... 114 
Table 6.5 General combining ability for LLS and yield-related traits of the genotypes used as male 
parents ............................................................................................................................... 116 
Table 6.6 Female parents GCA for LLS and yield-related traits ........................................................ 117 
Table 6.7 Specific combining ability of cross combination for LLS and Pod traits ........................... 119 
Table 6.8 Specific combining ability of cross combination for yield and seed related traits ............. 121 
Table 6.9 Correlation between traits’ GCA for all cross combination................................................ 122 
Table 6.10 Euclidian distance between parent genotypes based on SNP data .................................... 123 
Table 6.11 Correlation pattern between SCA and the genetic distance between parents. .................. 123 
  
xi 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF ABBREVIATIONS 
AMOVA Analysis of molecular variance 
ANOVA Analysis of variance 
CERAAS Centre d’etude Regionale pour l’Amelioration de l’Adaptation a la secheresse 
FAO  Food and Agriculture Organization 
FGDs  Focus Group Discussion 
GCA  General Combining Ability 
GWAS Genome-Wide Association Study 
IBPGR International Board for Plant Genetic Resources 
ICRISAT International Crops Research Institute for the Semi-Arid Tropics 
ITRA  Institut Togolais de Recherche Agronomique 
LD  Linkage Disequilibrium 
LLS  Late leaf spot  
NC II  North Carolina II design 
NGO  Non-Governmental Organism 
PC  Principal Component 
PCA  Principal Component Analysis 
PRA  Participatory Rural Appraisal 
QTL  Quantitative Trait Loci 
SCA  Specific Combining Ability 
WAAPP West Africa Agriculture Productivity Program 
WACCI West African Centre for Crop Improvement 
 
xii 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER ONE 
1. GENERAL INTRODUCTION 
Groundnut (Arachis hypogaea L.) constitutes one of the most important legumes grown around 
the world (FAOSTAT, 2018). It belongs to the genus Arachis that includes 70 species all nearly 
diploid (Krapovickas & Gregory, 1994; Ferguson et al., 2004). But cultivated groundnut is an 
allotetraploid specie (2n = 4x = 40, genome AABB) divided into six botanical varieties. Based 
on the presence or absence of flowers on the main axis, these varieties are grouped into two 
subspecies, hypogaea and fastigiata. However, all cultivated groundnuts can be grouped into 
four market types based on morphological and agronomical traits: Runner, Virginia, Spanish 
and Valencia. 
The seed, which is an important source of edible oil (48%) and protein (25%), occupies an 
important place in human and animal feeding (Jonnala et al., 2005; McKevith, 2005). Though 
the most abundant proteins in groundnut kernels, arachin and conarachin can be allergenic 
(Latif et al., 2013), groundnut as a source of dietary proteins (Arnarson, 2015), contributes 
significantly to the world food production.  
Groundnut is an excellent cash crop in many developing countries for local and foreign markets 
because of the multiple uses. During the last decade, India had the largest area sown to 
groundnut (6 million ha) while China had the largest production (14.24 million tons) 
(FAOSTAT, 2018). In Africa, important groundnut producing countries are Nigeria and 
Senegal (FAOSTAT, 2018). Though breeding efforts combined with improved agronomic 
practices have led to a yield increase in most groundnut production areas (Janila et al., 2016a), 
some countries, following the example of Togo, are still dealing with low yields. Indeed, 
groundnut, like other crops, is extremely vulnerable to diseases such as Early Leaf Spot (ELS), 
Late Leaf Spot (LLS), rust, rosette, groundnut clump, groundnut mottle and bacterial wilt. 
1 
 
University of Ghana http://ugspace.ug.edu.gh
Among the biotic stresses, LLS and rust are fungal diseases widespread in most tropical 
countries (Sujay et al., 2012).  
To meet the challenge of increasing food demand (Abberton et al., 2016), there is a need to 
develop, through breeding programmes, new and improved cultivars that respond to 
environmental conditions for sustainable food production. The success of a breeding 
programme is dependent on the development of a gene pool with useful genes and a broad 
genetic base (Upadhyaya et al., 2009; Sharma et al., 2013). Despite the availability of elite 
lines in some countries, landraces are still considered to harbor useful genetic variation that can 
be exploited in breeding programme (Dwivedi et al., 2017). Though molecular markers have 
become important for estimating genetic diversity (Milla-Lewis et al. 2010; Roomi et al., 2014; 
Ren et al., 2014; Bhad et al., 2016) and for the stresses management in groundnut (Mishra et 
al., 2015; Kanyika et al., 2015; Janila et al., 2016b, Pandey et al., 2017), morphological 
characters and physiological traits (IBPGR-ICRISAT, 1992) are still important for 
distinguishing groundnut varieties at the farm level and in breeding programmes (Janila et al., 
2013).  
Although the morphological, physiological and agronomic characteristics show significant 
variation in the six botanical varieties, certain studies have documented a low level of genetic 
diversity among cultivated lines at molecular level (Milla, 2003; Seijo et al., 2007; Tang et al., 
2007; Grabiele et al., 2012). Thus, the slow progress in the molecular breeding of cultivated 
groundnut has been mostly attributed to the low level of detectable molecular genetic variation 
using molecular tools such as Random Amplified Polymorphic DNA (RAPD), Amplified 
Fragment Length Polymorphism (AFLP), and Simple Sequence Repeat (SSR) markers 
(Herselman, 2003; Tang et al., 2007). Nowadays, to the detriment of other markers, single 
nucleotide polymorphisms (SNPs) are becoming the markers of choice. They have the 
2 
 
University of Ghana http://ugspace.ug.edu.gh
advantage of being abundant in the genome (Brumfield et al. 2003; Pandey et al., 2014) and 
are amenable to high-throughput analysis enabling large population analysis.  
Though low yielding varieties and LLS have been often reported by farmers and extension 
agents as causing important economic loss in Togo, no scientific study has been carried out to 
ascertain groundnut production constraints. LLS disease is caused by the fungus 
Cercosporidium personatum (Berk and M. A. Curtis) Deighton, and it is encountered 
everywhere groundnut is cultivated (Sujay et al., 2012). Though LLS and ELS disease are 
reported to be the most destructive diseases of groundnuts worldwide (Smith et al., 1992; Liu 
et al., 2013), there is no data on yield loss related to these diseases in Togo. Although LLS 
could be managed by chemical control (Nath et al., 2013), resource-poor farmers cannot afford 
the use of chemical control measures due to lack of resources. As an alternative, many studies 
suggest that the use of host resistance is the most cost-effective and environmentally 
sustainable control measure (USDA, 2015). Recent advances in genetics and genomics provide 
molecular tools to facilitate groundnut breeding especially to exploit LLS resistance within 
cultivated species (Shoba et al., 2012; Sujay et al., 2012). Having said that, the success of the 
exploitation of the host resistance in a breeding programme relies on the nature of gene action 
and the breeding method used for traits improvement. 
Indeed, knowledge of the nature of gene action involved in the expression of the trait to be 
improved is a prerequisite as the breeding method depends mostly on this factor. Then, the 
identification of the source of resistance with good breeding potential is required for the 
development of LLS resistant varieties. The estimation of combining ability helps in the 
establishment of appropriate breeding strategies (Arunachalam et al., 1984; Fasahat et al., 
2016). Mating designs, in addition to establishing the gene action, provide breeders with 
estimates for general combining ability (GCA) and specific combining ability (SCA) (Sprague 
3 
 
University of Ghana http://ugspace.ug.edu.gh
& Tatum, 1942; Zhang et al., 2012; Fasahat et al., 2016).  They are used in groundnut and other 
crops for the identification of the best parents for multiple traits of interest including LLS 
resistance (Wilson et al., 2013; Manivannan, et al., 2008; Mothilal & Ezhi, 2010; Azad et al., 
2014; Patil et al., 2017). Some recent studies have identified general combiners that were used 
as donor parents for the improvement of certain physiological traits and for the development 
of high-yielding LLS resistant genotypes in breeding programmes (Vishnuvardhan et al., 2011; 
John et al., 2012). Most genetic studies of LLS resistance suggest that resistance is controlled 
by several recessive genes (Vasanthi & Reddy, 1997) and additive genetic variance seems to 
play the predominant role in the resistance (Jogloy et al., 1987; Dwivedi et al., 2002). Thus, 
LLS resistance within cultivated species can be exploited by identifying recombinant from 
cross combinations with good combining ability. 
Groundnut is one of the most important legumes grown in Togo. However, while the world 
average yield is between 1.3 t ha-1 and 3.36 t ha-1 (FAOSTAT, 2016), in Togo, the average 
yield of groundnut is about 0.64 t ha-1 (DSID, 2015). Unfortunately, not much efforts have been 
deployed for the development of improved varieties that could boost groundnut production. 
There was no organized breeding programme and few improved varieties introduced into the 
country were made available to farmers. As a result, the yield of groundnut in Togo has been 
decreasing over the years (DSID, 2015).  
Also, despite the economic importance, the existing variability among landraces has not been 
exploited through organized selection to address biotic constraints such as LLS in Togo. In 
fact, due to the absence of an active plant breeding programme for the genetic improvement of 
groundnut, there are no documented records of groundnut accessions grown by farmers in the 
country. Understanding the range of diversity and the genetic structure of the cultivated 
groundnut accessions in Togo is a useful prerequisite for the effective management and use of 
germplasm resources in the groundnut breeding programme. Recently, an interest has been put 
4 
 
University of Ghana http://ugspace.ug.edu.gh
on the development of improved varieties through an organized breeding programme that will 
boost the production of groundnut in Togo. Though no catalogue of the low yield causes is 
available in Togo, late leaf spot has often been reported by the extension agents as the most 
widespread disease. Despite the obvious presence of late leaf spot throughout the growing area, 
a detailed study is required for the confirmation of this constraint. Currently, efforts are going 
on to identifying farmers’ preferences and production constraints, assembling germplasms, 
establishing test environments and identifying target population of environments. As part of 
the efforts, this study focused on understanding farmers’ perceptions of production constraints 
and preference criteria, characterizing cultivated groundnut varieties and understanding genetic 
mechanisms for priority traits to exploit genetic diversity among groundnut accessions and 
farmer knowledge and expectations to develop high yielding groundnut varieties in Togo.  
Specific objectives were to: 
i. identify production constraints, farmers’ perception of leaf spot disease, and their 
preferences of groundnut traits 
ii. characterize groundnut accession using morphological, agronomic and molecular 
markers  
iii. identify genomic regions associated with LLS and yield-related traits and 
iv. determine the mode of inheritance of resistance to groundnut LLS disease.  
5 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER TWO 
2. LITERATURE REVIEW 
2.1 Origin and biology 
To date, nearly eighty-one species of groundnut, both annual and perennial, have been 
described including cultivated groundnut (Valls & Simpson, 2005; Valls et al., 2013). 
Cultivated groundnut is an allotetraploid while nearly all the other Arachis species are diploid 
(2n = 4x = 40, genome AABB) (Krapovickas & Gregory, 1994; Ferguson et al., 2004). The 
diploid progenitors contributed AA (A. duranensis) and BB (A. ipaensis) genomes to the 
cultivated groundnut (Kochert et al., 1996). It is suggested that cultivated groundnut is a result 
of a single hybridization between the diploid progenitors followed by chromosome doubling 
(Kochert et al., 1996; Moretzsohn et al., 2013). A. duranensis is encountered throughout South 
America while A. ipaensis has only been described in Southern Bolivia. For this reason, 
Western Brazil, Bolivia, Paraguay, and Northern Argentina are believed to be the centres of 
diversity (Gregory & Gregory, 1980; Jarvis et al., 2003). In addition, with adaption for survival 
in dry lands, groundnut is believed to have originated in tropical wetland areas (Stalker et al., 
2016).  
Cultivated groundnut is classified into six botanical varieties. Based on the presence or absence 
of flowers and lateral branches on the main axis, likely due to few major genes’ variation 
(Wynne & Coffelt, 1982; Kochert et al., 1996), these varieties are grouped into two subspecies: 
hypogaea (presence) and fastigiata (absence). Subspecies fastigiata includes the varieties 
fastigiata, vulgaris, aequatoriana and peruviana while subspecies hypogaea includes two 
botanical varieties, hypogaea and hirsuta. These botanical types can be distinguished based on 
the plant, pod and seed features (IBPGR-ICRISAT, 1992; Krapovickas & Gregory, 1994). 
However, only three botanical varieties are now widely grown in the Americas, Asia and 
6 
 
University of Ghana http://ugspace.ug.edu.gh
Africa.  These are subsp.  Hypogaea var. hypogaea, subsp. Fastigiata var. fastigiata and var. 
vulgaris (Ferguson et al., 2004). 
Groundnut is adapted to various environmental conditions, between latitudes 40º N and 40º S, 
including soil types, cropping patterns, abiotic and biotic stresses (Simpson et al., 2001; Liao 
& Holbrook, 2007; Mallikarjuna & Varshney, 2014). The optimal growth temperature is 
between 25 and 30ºC. Though groundnut is a self-pollinated crop, up to 10% of cross 
pollination can occur under field conditions when bees’ activity is high (Knauft et al., 1992). 
Following fertilization, the peg carrying the fertilized ovule appears 4 – 10 days and grows to 
enter the soil to a depth of 2-4 cm and develops into pods (Pattee & Mohapatra, 1987; Xi, 
1991). Sixty to eighty days after fertilization, pods should be mature. However, at the harvest, 
it is common to see pods at multiple maturity stages as flowering is indeterminate on 
groundnut. 
2.2 Groundnut cytogenetics 
Tetraploid (2n = 4x =40) and diploid (2n = 2x=20) species are the common species, though not 
the only ones, in the genus Arachis (Krapoviackas et al., 2009). Diploids are likely more 
ancient than tetraploids. Species of section Arachis, including cultivated groundnut, have 
metacentric chromosomes ranging from 1.4 to 3.9 nm. Three genomes have been identified in 
the section Arachis: A genome, B genome and D genome (Lavia & Fernandez, 2008). With a 
very large and complex genome, cultivated groundnut is an allotetraploid species (2n= 4x= 40, 
AABB) which cytologically behaves as a diploid. However, Leal-Bertioli et al. (2015) using 
both genetic and gene expression data, showed that groundnut can display tetrasomic genetics 
recombination and concluded that this genetic behaviour accounts for many of the “off types” 
observed in the fields.  
7 
 
University of Ghana http://ugspace.ug.edu.gh
2.3 Groundnut genetic resources 
Groundnut genetic diversity is large in the centres of origin (Krapovickas & Gregory, 1994). 
Thus, efforts have been deployed for the identification and assembling of cultivated landraces 
in the areas of origin and in other places where groundnut is cultivated (Holbrook & Stalker, 
2003). These accessions, which can be separated based on seed and pod morphology (Stalker 
& Simpson, 1995), are collected from farmers or local markets. The common method of 
groundnut collection has been to gather groundnut samples in small markets and from farmers. 
As farmers grow an array of genotypes, the diversity found in the collected germplasm is 
generally large (Holbrook & Stalker, 2003). When collected from farmers’ fields, a greater 
amount of agronomic data can be obtained.  Because of expeditions and collections in different 
places of the world, many collections are maintained or stored in different places around the 
world. The largest genetic resources collections, including elite germplasms, are maintained in 
gene banks such as International Crop Research Institute for the Semi-Arid Tropics (ICRISAT) 
in India (Upadhyaya et al., 2011). However, groundnut germplasm is stored in many other gene 
banks all over the world (Holbrook, 2001).  
Preservation of groundnut accession is generally simple as seeds can remain viable for more 
than 15 years under optimal conditions (Stalker et al., 2016). A large germplasm is generally 
difficult to phenotype because of the cost associated with evaluating all accessions for traits of 
interest in cultivar development. A core collection has been proposed as an alternative 
(Holbrook, 1999) for handling a reduced number of lines for cultivars development for traits 
of interest such as resistance to LLS (Holbrook & Anderson, 1995).  
In Togo, as a consequence of the lack of a strong breeding programme, there is no groundnut 
collection. Recently, efforts have been put towards establishing a breeding programme that will 
respond to farmers’ expectations. As a starting point, groundnut landraces are being collected 
8 
 
University of Ghana http://ugspace.ug.edu.gh
throughout the country and a core collection, useful for the breeding programme, will be 
established. 
2.4 Genetic diversity of important traits in groundnut  
Genetic variability is the base of the success of a breeding programme targeting the 
improvement of a particular trait. Therefore, knowledge of the genetic variation of germplasm 
is important for improvement, studies and conservation purposes (Falconer & Mackay, 1996). 
Morphological characters and physiological traits have been extensively used for grouping 
groundnut varieties into classes (Upadhyaya et al., 2011).  
A large amount of variation has been observed using these descriptors in different germplasm 
collections (Upadhyaya et al., 2006). Sources of variability for disease resistance have been 
reported for most of the diseases from results of field, laboratory and greenhouse studies (Sujay 
et al., 2012; Tshilenge-Lukanda et al., 2012). Genetic variability has been reported for traits 
such as tolerance to drought, early maturity (Upadhyaya et al., 2006), and tolerance to LLS 
(Izge et al., 2007; Tshilenge-Lukanda et al., 2012). A certain level of variability was reported 
for other traits such as yield attributes (Nigam & Blummel, 2010). However, morphological 
traits are sensitive to environmental variation. In addition, they are not always under genetic 
control. In contrast, genetic markers are reliable to track genomic regions associated with a 
particular trait because they are unaffected by environmental fluctuations.  
Although there is a significant variation in the six botanical varieties for morphological traits 
and physiological traits, certain studies have documented a low level of diversity among 
cultivated groundnut at the molecular level (Herselman, 2003; Tang et al., 2007). The low level 
of diversity in groundnut has been attributed to a combination of three causes which are: 
barriers to gene flow from wild diploid species to cultivated groundnut (Young et al., 1996), 
polyploidization from probably a few numbers of individuals of parental species (Halward et 
al., 1991) and use of few elite breeding lines in breeding programmes (Isleib et al., 2001). 
9 
 
University of Ghana http://ugspace.ug.edu.gh
However, some authors have reported that delimitation of varieties on the basis of 
morphological and molecular characters can exhibit a certain degree of similarity (Ferguson et 
al., 2004; Kottapalli et al., 2011). Thus, cultivated accessions of groundnut in the gene banks 
and the elite lines in the breeding programmes are still considered as sources of variability for 
the development of improved varieties. Molecular marker data combined with phenotypic data 
is a powerful tool for grouping genotypes and for selection of progenitors that will generate 
new breeding population (Stalker et al., 2016).  
Despite groundnut’s economic importance, the information on genetic parameters of cultivated 
groundnut required for breeding groundnut for biotic constraints including LLS is lacking in 
Togo. In fact, due to the absence of an active plant breeding programme for the genetic 
improvement of groundnut, there are no documented records of groundnut accessions grown 
by farmers in Togo. 
2.5 Gene action in groundnut 
Type of gene action involved in the expression of a trait is one of the considerations among 
others, which the breeder must look at when starting a breeding programme. Indeed, the success 
of a breeding programme depends on the genetic gain per selection cycle, which in turn depends 
mostly on how genes act or interact in the production of a phenotype (Falconer & Mackay, 
1996). Also, identification and selection of parental lines can be based upon an evaluation of 
the performance of its offspring (Fasahat et al., 2016). Along with the establishment of the 
gene action, combining ability determined through a specific mating design offers the 
possibility to identify and select parent genotypes with good potential (Sprague & Tatum, 
1942). This strategy has been used extensively in the development of improved cultivars of 
many economically important crops including corn, wheat and cotton (Khan et al., 2009; 
Gowda et al., 2010).  
10 
 
University of Ghana http://ugspace.ug.edu.gh
Mating designs are widely used to partition the genetic influences of a genotype into additive 
and non-additive components. According to Sprague and Tatum (1942), general combining 
ability (GCA) is related to the activity of additive genes effects as well as additive × additive 
interactions while specific combining ability (SCA) underlines loci with dominance non-
additive effects and all the epistatic interaction components. For self-pollinated crops such as 
groundnut, a high GCA estimate may result in fewer gene interactions and higher achievement 
in selection (Fasahat et al., 2016). 
In groundnut, likely due to the lack of effective phenotyping tools, few studies on the 
combining ability for traits of interest have been conducted. Also, the first studies carried out 
on groundnut focused mainly on the understanding of the genetic nature of yield-related traits 
(Dwivedi et al., 1989; Anderson et al., 1993a). Dwivedi et al. (1989) found that additive genetic 
effects largely controlled fruit and seed traits and identified parents with good breeding 
potential for pod and seed traits. Gene action was assessed for other traits of interest such as 
disease resistance, drought resistance and qualitative traits. Anderson et al. (1986) attributed 
early and late leaf spot resistance to additive genetic variance and identified best parents for 
incorporating both disease resistance into improved cultivars. Jogloy et al. (2000) ran a 
combining ability study and identified lines suitable for both LLS resistance and agronomic 
traits. Although additive gene action is more important in most of the aforementioned traits, 
non-additive gene effect has been reported for some of the agronomic traits (Azad et al., 2014). 
Because of the allotetraploid and self-pollinated nature of groundnut, the dominance effect 
cannot be exploited in the development of improved varieties as hybridization remains labour 
consuming.  
Different breeding populations can be used for the estimation of the genetic parameters. 
However, F1 hybrids are the most common plant populations used for the estimation of genetic 
parameters, mainly when heterosis is desired (Sprague & Tatum, 1942; Fasahat et al., 2016; 
11 
 
University of Ghana http://ugspace.ug.edu.gh
Patil et al., 2017). But the self-pollinated nature of groundnut coupled with the low number of 
seeds that can be obtained from each cross has forced some breeders to use F2 and even F3 
population in diallel and North Carolina mating designs (Anderson et al., 1986; Jogloy et al., 
2000). One of the major constraints of running North Carolina mating design with F2 generation 
is the sample size which should be as large as possible in order to give accurate estimates of 
the cross means.  
As far as the environment is concerned, a majority of traits are polygenic, therefore highly 
influenced by environment (Piepho et al., 2008). Therefore, the evaluation of the progenies for 
parents’ identification should consider environmental influence. This aspect is really important 
in all breeding programmes as it affects the genetic gain in groundnut and other crops. 
Generally, stable varieties with the least GXE interaction are desirable (Piepho et al., 2008).  
2.6 Use of molecular markers for groundnut improvement 
It is obvious that genetic markers have increased the pace and precision of crops’ genetic 
analysis. As a result, the use of molecular markers has tremendously increased and facilitated 
the development of improved cultivars with multiple traits of interest (Mammadov et al., 2012). 
However, the use of markers in groundnut improvement is in the nascent stage (Stalker et al., 
2016). Microsatellite markers have been considered as the marker of choice in groundnut for 
diverse uses (Gimenes et al., 2007; Pandey et al., 2012).  
Along with diversity purpose, molecular studies were conducted for the understanding of 
genetics underlying agronomic traits in groundnut and the association of markers with traits. 
The first marker-trait association report in an A. hypogaea x A. hypogaea cross was for black 
rot (Cylindrocladium parasiticum) and ELS (C. arachidicola) (Wynne et al., 1979). Since then, 
many studies have been carried out for marker discovery for traits of agronomic importance. 
QTLs for plant architecture, days to flowering, pod and seed shape, and yield-related traits 
12 
 
University of Ghana http://ugspace.ug.edu.gh
were found. Markers were also identified for traits such as pod size and seed size (Gomez et 
al., 2009; Fonceka et al., 2018) and for introgression of wild segments into cultivated 
groundnut (Fonceka et al., 2012). Other positive QTLs for biotic and abiotic stress management 
were recently identified (Mallikarjuna & Varshney, 2014; Stalker et al., 2016). 
However, only recently, the identification of some major QTLs for LLS diseases has been 
reported (Sukruth et al., 2015). Though no marker has proved to be efficient in marker-assisted 
selection for groundnut foliar diseases, DNA markers linked to LLS have been identified 
(Gajjar et al., 2014; Varshney et al., 2014; Kanyika et al., 2015). For many groundnut foliar 
diseases such as leaf spot and viral disease, despite several years of effort, there is no validated 
marker that has proved to be efficient across different breeding programmes. The availability 
of cost-effective genomic resources, such as single nucleotide polymorphism (SNP) and 
genotyping by sequencing (GBS), is expected to speed up markers identification in groundnut 
(Pandey et al., 2012; Mishra et al., 2015; Stalker et al., 2016).  
SNP markers have the advantage of being abundant in the genome (Brumfield et al. 2003) and 
are amenable to high-throughput analysis enabling large population analysis. Also, SNPs 
linked to the coding regions can be more efficiently used in breeding methodologies such as 
diversity study than other markers (Luikart et al. 2003). In addition, SNPs low mutation rate 
often results in more accuracy in the estimation of population structure parameters (Coates et 
al., 2009). Diversity array technology (DArT), one of the GBS tools recently developed 
(Jaccoud et al., 2001), has proved to be efficient to evaluate genetic diversity, relationships and 
population structure in crops such as wheat (Novoselovic´ et al., 2016; Nielsen et al., 2014) 
and bean (Brinez et al., 2012; Nemli et al., 2017). In groundnut, this technology has been 
mainly used in genome-wide association and trait mapping (Pandey et al., 2014; Shasidhar et 
al., 2017). In addition, GBS has the advantage of exhibiting low per sample cost. Once 
additional molecular markers tightly linked to the traits of interest and compatible with high-
13 
 
University of Ghana http://ugspace.ug.edu.gh
throughput platforms are designed, the integration of the marker-assisted selection into 
groundnut breeding programmes will be more significant.  
2.7 Constraints and genetic improvement of cultivated groundnut 
Groundnut breeding history can be traced back to 20th century where the use of improved 
varieties has prevailed in most groundnut production areas (Mallikarjuna & Varshney, 2014). 
Yield has been and is the common target trait for most breeding programmes. Therefore, traits 
affecting yield such as disease resistance and resistance to biotic stresses have been targeted 
for years (Isleib et al., 2001).  
Late Leaf Spot [Phaeoisariopsis personata (Berk & Curtis), Deighton], ELS [Cercospora 
arachidicola Hori], rust [Puccinia arachidis Sperg.], groundnut rosette virus (GRV), peanut 
bud necrosis (PBND), peanut stunt virus (PSV) and bacterial wilt [Ralstonia solanacearum 
E.F. Smith] are the most important foliar diseases on groundnut. However, LLS and ELS are 
considered as the most important biotic constraints in terms of economic losses in tropical areas 
(Shokes and Culbreath, 1997; Dwivedi et al., 2003). Regarding abiotic stresses, drought, low 
soil fertility and low temperature are the most important stresses on groundnut. One of the most 
important food safety issue on groundnut is aflatoxin contamination. Caused by Aspergillus, 
aflatoxin contamination occurs during production, processing, transportation and storage and 
can result in a severe economic loss (Holbrook & Stalker, 2003).  
The aforementioned circumstances force groundnut breeders to develop cultivars with not only 
high yield but also with resistance or tolerance to important constraints. Several researchers 
estimate yield increase because of the use of improved cultivars, to have had an important 
economic impact throughout the world (Mallikarjuna & Varshney, 2014). Though there are 
more cultivars released in developing countries than in developed countries, this does not 
reflect the investment deployed in improved cultivars development. Indeed, most developing 
14 
 
University of Ghana http://ugspace.ug.edu.gh
countries, including Togo, are characterized by diversified production conditions and 
ineffective seed systems. Therefore, the large number of cultivars in groundnut production 
areas in developing countries is more due to the unsystematic agronomic practices as the 
adoption of the released cultivars is uncontrolled. Though many efforts have been deployed for 
the development of high yielding varieties, LLS remains one of the major causes of yield loss 
in groundnut in tropical areas (Pande et al., 2003). 
2.8 Late Leaf Spot disease 
2.8.1 Origin and occurrence 
Late Leaf Spot is one of the most important fungal diseases impacting the production of 
groundnut in tropical areas (Mayee & Datar, 1988). The causal agent probably originated and 
evolved in South America along with its hosts (Subrahmanyam et al., 1995). The disease is 
widespread in tropical countries where it can cause yield loss up to 50 % (Backman & 
Crawford, 1984). Although LLS is commonly present wherever groundnut is cultivated, there 
can be fluctuation in the relative importance (Mayee & Datar, 1988) as LLS severity and the 
incidence are influenced by the climatic conditions. Some varieties could tolerate a low level 
of infection, but all levels of defoliation result in yield loss (Backman & Crawford, 1984). 
Depending on location and environmental conditions, LLS commonly attacks plant of 45-60 
days old.  
2.8.2 Symptoms 
The disease damages plant production by reducing the photosynthetic area, through lesion 
formation and stimulation of leaflet abscission (Subrahmanyam et al., 1984). Ten days 
following the infection, small chlorotic spots are visible on the leaflets. Spots development 
takes about five days from the spot appearance to mature sporulating lesion (Backman & 
Crawford, 1984). The lesions of the LLS are similar to the ones of the ELS. However, many 
15 
 
University of Ghana http://ugspace.ug.edu.gh
characteristics can differentiate between these two diseases. Lesions of LLS disease are usually 
smaller, nearly circular (up to 8 mm in diameter) and darker in colour than those of the ELS 
disease (Mayee & Datar, 1988). Though a distinctive chlorotic halo is often present around 
ELS lesions, the yellow halo is not a good diagnostic tool since similar halo may be found 
around the spot of the LLS on the upper surface. Also, the prominence of the halo is influenced 
by the host-pathogen interaction (Mulder & Holliday, 1974).  
The characters used for the diagnosis of the disease in the field are the colour of the lesions on 
the abaxial leaflet surface and the distribution of fruiting structures. Lesions of LLS are black 
with rough appearance and the fruiting structures are in circular rings on the abaxial surface 
unlike the lesions of ELS which are light brown in colour with fruiting structures randomly 
distributed (McDonald et al., 1985). The most relevant feature of the LLS lesion is the dark 
stroma formed at the conidial state. Lesions of LLS disease may be observed on petioles, stems, 
and pegs. Generally, lesions of LLS on stems vary in shape compared to lesions observed on 
leaves (Mayee & Datar, 1988). Stem lesions are oval to elongate and have more distinct 
margins on petioles and stems than the leaflet lesions. Indeed, in case of severe disease attack, 
chlorotic lesions become necrotic and often coalesce resulting in leaflet shed and drop 
(Subrahmanyam et al., 1984). The observed leaf drop is caused by the change in hormonal 
functions due to the presence of the pathogen. The defoliation occurs generally, first on lower 
leaves and then progresses on upper leaves (McDonald et al., 1985). Disease symptoms 
development is rapid after flowering and during pod formation. 
2.8.3 Causal agents 
Late Leaf Spot is a fungal disease caused by Cercosporidium personatum (Berk. & Curt) 
(Mulder & Holliday, 1974). C. personatum belongs to Mycosphaerellaceae, a family of sac 
fungi and to the phylum Ascomycota, the largest phylum. Ascomycetes are fungi that are 
16 
 
University of Ghana http://ugspace.ug.edu.gh
characterized by the production of microscopic spores inside elongated cells called as 'asci' 
(Crous et al., 2009). Asexual reproduction of the pathogen, which is the dominant form of 
propagation in Mycosphaerellaceae, favour the rapid spread of C. personatum into new areas 
(Videira et al., 2017). The propagation of the pathogen occurs through vegetative reproductive 
spores, the conidia that are dispersed by wind or water. Conidia production is restricted to the 
lower leaf surface where conidiophores develop concentric rings (Backman & Crawford, 
1984). 
C. personatum, the asexual state (Anamorph), is the most common form of the pathogen in 
groundnut production areas (Mayee & Datar, 1988). The existence of the perfect state 
(Teleomorph), called Mycosphaerelia berkeleyii, have been suggested in USA (Mulder & 
Holliday, 1974). However, in tropical areas such as sub-Saharan Africa, the pathogen exists 
only in the asexual state everywhere groundnut is cultivated (Mayee & Datar, 1988).  
2.8.4 Epidemiology and disease cycle 
Conidia produced by the conidiophores on the lower leaves are dispersed by wind, splashing 
water (rain), mechanical dissemination and insects (Damicone & Melouk, 2010). On a new 
plant tissue, conidia germinate and form germ tubes that enter host tissue through open stomata 
(Subrahmanyam et al. 1985). Then, the germ tubes penetrate the host tissue directly through 
the lateral faces of epidermal cells. Unlike the pathogen of the ELS, C. personatum does not 
secrete toxins that kill the host cells but produce digestive enzymes that break down organic 
molecules into smaller ones that are then taken up from the plant cell (Mulder & Holliday, 
1974). Lesions appear as spots and become visible 10 to 14 days after the infection. Later on, 
the pathogen produces conidiophore and conidia for the perpetuation of the disease cycle. The 
new spores produced in spots serve to infect new leaves and plants (Pande et al., 2001).  
17 
 
University of Ghana http://ugspace.ug.edu.gh
After the production and release of the conidia, the germination occurs on newly infected tissue 
within 10 to 14 days when the environmental conditions are adequate (Subrahmanyam et al., 
1995). The optimal conditions required for LLS disease occurrence and development are warm 
temperatures and long periods of high humidity (Damicone & Melouk, 2010). However, the 
pathogen can survive through inoculum reservoir on infected plant debris favouring the 
perpetuation of the disease from season to season (Subrahmanyam et al., 1984, 1995). This 
survival occurs by the means of conidia and mycelium (Mondal & Wahhab, 2001). It is 
suggested that the survival can last from 30 to 60 days on groundnut residue that are submerged 
under the soil but can increase up to 12 months if the residues are stored indoors (Mondal & 
Wahhab, 2001). 
2.8.5 Disease control 
Many chemical substance tests demonstrated that LLS can be effectively controlled by 
fungicides. Indeed, several fungicides that provide excellent control when applied adequately 
are available (Garren & Jackson, 1973; Damicone & Melouk, 2010). However, in addition to 
being harmful to the environment, these fungicides are expensive and developing countries’ 
farmers cannot afford their use.  
Thus, cultural practices and the use of genetic resistance have been adopted in most groundnut 
production areas for sustainable disease control. Cultural practices include the choice of less 
susceptible varieties, adequate irrigation, rotation of crops and management of residues by 
tillage. Rotation of groundnut with other crops and management of groundnut residue help 
delay the onset of the disease and slow its development (Damicone & Melouk, 2010). The most 
important component of the best cultural practices is the use of host plant resistance to 
managing the disease (McDonald et al., 1985). 
18 
 
University of Ghana http://ugspace.ug.edu.gh
2.8.6 Resistance to LLS disease 
Resistance to LLS is complex (Anderson et al., 1993b) as it has several components 
contributing to the resistance. These components include initial infection, lesion size, 
sporulation, and defoliation (Anderson et al., 1993b). According to Subrahmanyam et al. 
(1985), resistance appeared to be associated with small stomatal apertures. After penetration, 
resistance is associated with cell wall thickening around the infection site and with the 
deposition of pectic substances on cell walls and in the intercellular space resulting in the 
difficulties for the pathogen in cell penetration. In response to C. personatum, Pattee & Young 
(1982) found that necrotic defense was operative in resistant cultivars.  
Also, resistance to LLS has been often associated with late maturity and alternative branching 
with dark green leaves. However, a breeding effort through recombination has resulted in 
improved lines carrying different combinations of these features (Stalker et al., 2016). 
2.8.7 Genetic resources for LLS resistance 
Many A. hypogaea have been found to have a desirable level of resistance to LLS disease 
(Branch & Fletcher, 2001). Most of the LLS resistant accessions originated from Bolivia 
(Subrahmanyam et al., 1995; Isleib et al., 2001). It is suggested that the pathogen and its host 
have long been associated in their centres of origin and coevolved, developing complementary 
genetic systems (Branch & Fletcher, 2001). Though complete resistance to LLS has not been 
found in cultivated groundnut (Vishnuvardhan et al. 2011), exploration of resistance gene 
centres provided germplasm for varietal improvement for resistance to LLS diseases. Efforts 
are being made for the incorporation of the high level of LLS resistance found in wild relatives 
into some cultivars (Fonceka et al., 2012; Khera et al., 2018).  
Recently, many efforts have been deployed at ICRISAT to broaden the genetic base of A. 
hypogaea (Malikarjuna & Varshney, 2014) through the development of synthesized tetraploids 
19 
 
University of Ghana http://ugspace.ug.edu.gh
(Shilpa et al., 2013). When some of these lines were crossed with cultivated groundnut, some 
of the progenies had LLS resistance genes. From these results, synthesized tetraploids have 
been considered as a good way of introgressing LLS resistance into cultivated groundnut 
(Shilpa et al., 2013). Thus, breeding progress has been made in the suppression of LLS disease 
but combining a high level of resistance and high yield with acceptable quality traits continues 
to be difficult. 
In addition to the identification of the lines with good breeding potential, most genetic studies 
of LLS resistance suggest that resistance is controlled by several recessive genes (Dwivedi et 
al., 2002) and additive genetic variance seems to contribute predominantly to the resistance 
(Dwivedi et al., 2002; Pasupuleti et al., 2013). It was earlier suggested that the resistance to 
LLS is controlled by more than five genes (Luo et al., 2005). More recently, Leal-Bertioli et 
al. (2009), through a QTL analysis, concluded that LLS resistance is controlled by additive or 
partial dominance gene action. Other QTL studies have enabled the identification of genomic 
regions conferring resistance to LLS and rust (Khedikar et al., 2010; Sujay et al., 2012).  
2.8.8 Identification and use of molecular markers for LLS 
The use of molecular markers in groundnut breeding programmes is slowly increasing. 
Groundnut has been one of the most important orphan crops when talking about a genetic 
advance in crop improvement. Fortunately, recent decades have witnessed the deployment of 
molecular tools to overcome the short comings encountered in conventional breeding (Janila et 
al., 2013).  
Because of the tight link between LLS and rust disease, most of the few studies conducted to 
map the resistance to both diseases have been carried out together (Wang et al., 2011; Sujay et 
al., 2012; Gajjar et al., 2014). For instance, Khedikar et al. (2010) have conducted a QTL 
analysis for rust and LLS using recombinant inbred line (RILs) population and reported only a 
20 
 
University of Ghana http://ugspace.ug.edu.gh
minor QTL for LLS resistance. Similarly, with the objective of identifying major QTL for LLS, 
Sujay et al. (2012) using two RILs population reported two candidate genomic regions 
containing the major QTLs with 10 to 62% phenotypic variance explained. More recently, 
many genomic regions linked to LLS resistance have been reported to be located on linkage 
group A03, A05 and B10 (Han et al., 2018; Agarwal et al., 2018) through QTL analysis or 
Genome Wide Association Study (GWAS). 
In spite of the aforementioned advances, unlike the markers for rust resistance, the validation 
of linked markers for LLS resistance is still awaiting. Thus, studies are still going on for the 
identification of new genomic regions affecting LLS resistance. Their validation will probably 
accelerate the process of introgression of LLS disease resistance into preferred groundnut 
varieties (Gajjar et al., 2014; Stalker et al., 2016). 
2.9 Demand-led breeding in groundnut 
The search for alternative approaches to generating information for the development of new 
technologies in response to the failure of most of the agricultural projects has become crucial 
(Frediani & Boano, 2012). As an alternative, PRA was pioneered by the Plant Science 
programmes of the Overseas Development Institute (PSODI) at the University of Wales in 
Bangor and was based on Witcombe and Joshi’s work on participatory crop improvement in 
Asia (Witcombe et al., 2006). The term participatory rural appraisal has however been slightly 
modified over time and called Client-oriented breeding (COB) or participatory varietal 
selection (PVS) (Witcombe et al., 2006; Soleri & Cleveland, 2004). Participatory Rural 
Appraisal (PRA) enables local farmers, in addition to influencing the research, to carry out 
their own analysis (Abedi & Vahidi, 2011; Sattar et al., 2017). The participation of the most 
important stakeholders ensures the direct involvement of the project beneficiary (Almekinders 
21 
 
University of Ghana http://ugspace.ug.edu.gh
& Elings 2001; Kolech et al., 2017) and in turn, guarantees the adoption and rapid diffusion of 
new varieties (Kraaijvanger et al., 2016; Kolech et al., 2017).  
There are many examples of PRAs that have enabled agricultural project leaders to identify 
and incorporate farmers’ perception in the project development and implementation. For 
instance, Abedi and Vahidi (2011), through a PRA, identified labour and lack of finance as the 
most important constraint to the production of Bambara groundnut in Southern Guinea Savana 
of Nigeria. Seed spoilage of smallholder farmers and trade regulations have been identified as 
very important to ensure the sustainability of groundnut production through PRA study 
conducted by Prasad et al. (2006). In Malawi, a survey on farmers between 2010 and 2013 
enabled Fitzgerald (2015) to make a recommendation for lasting changes in the groundnut 
stakeholder’s life and for the development of sustainable groundnut value chains. With the goal 
to assess the main production constraints of groundnut in Ghana (Oppong-Sekyere et al., 2015), 
the constraints to the production such as drought and diseases were identified by farmers 
through a PRA study. Thus, PRA is not only conducted upstream but also downstream the 
projects, as it offers the opportunity to measure the efficiency of the previous projects in 
addition to the identification of the prevailing constraints that need to be incorporated in the 
new projects.  
The objective of the participatory breeding approaches, since its adoption in many 
development programmes, was to exploit farmers’ perceptions and preferences in setting the 
breeding goals in order to increase the chance of adoption (Morris & Bellon, 2004; Ortiz-
Ferrara et al., 2007; Manzanilla et al., 2011). Also, PRA results helps the breeder in the 
identification of genotypes (parents) to be crossed to generate new breeding populations 
(Bellon & Reeves 2002; Danial et al., 2007).  
22 
 
University of Ghana http://ugspace.ug.edu.gh
Gender differences in farmers preferences were reported in many African countries (Aguilar et 
al., 2014; Ali et al., 2015; Backiny-Yitna & McGee, 2015; Oseni et al., 2015). Thus, as gender 
affects adoption of new technologies (Danial et al., 2007), implementing gender in breeding 
projects will yield greater impact on farmers (Johnson et al., 2016). As a consequence, studies 
for the identification of production constraints and farmers’ preferences, should not ignore the 
role played by men and women (Kristjanson et al., 2017).  
In Togo, no such study has ever been conducted on groundnut for the identification of 
production constraints. The new breeding programme could exploit this tool for more impactful 
outcomes from the breeding products to be developed. 
2.10 Groundnut production in Togo 
Groundnut is mainly cultivated in mixed cropping with cereals in the northern part of Togo. 
However, recent statistics revealed that farmers in the southern part are increasingly interested 
in growing groundnut (DSID, 2015). Both male and female farmers of various ages are 
involved in groundnut production which occurs from June to September.  
Groundnut yield has drastically increased in many parts of the world as a result of improved 
varieties development (Isleib et al., 2001; Wan et al., 2003; Mallikarjuna & Varshney, 2014). 
In Togo, groundnut has been profitable for decades, despite the lack of a strong breeding 
programme probably due to the introduction of improved varieties developed elsewhere. The 
high profitability could also be partly attributed to the crop’s low production inputs and its 
ability to withstand climatic conditions. Indeed, the output/input ratio of groundnut has been 
higher than those of many food crops in Togo (DSID, 2015). Unfortunately, for nearly two 
decades now, low soil fertility, biotic stresses and erratic rainfall in the context of changing 
climate are negatively impacting groundnut production in Togo. As a matter of fact, groundnut 
yield has not been increasing, rather steadily over the years. There is a big yield gap between 
23 
 
University of Ghana http://ugspace.ug.edu.gh
the world average of about 1.7 t ha-1(FAOSTAT, 2018) and the average yield in Togo (0.67 t 
ha-1). Thus, in spite of breeding efforts combined with improved agronomic practices that have 
led to yield increase in most groundnut production areas (Janila et al., 2016b), some countries 
including Togo are still experiencing low yields.  
 
  
24 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER THREE 
3. GROUNDNUT PRODUCTION CONSTRAINTS AND FARMERS’ TRAIT 
PREFERENCES 
3.1 Introduction 
Participatory Rural Appraisal (PRA) has been an alternative approach to generating 
information for the development of new agricultural technologies in response to the failure of 
most of the agricultural projects to meet the need of farmers (Duraiappah et al., 2005; Frediani 
& Boano, 2012). It allows the beneficiary farmers to influence the direction of the research by 
enabling them to carry out their own analysis (Abedi & Vahidi, 2011; Sattar et al., 2017). In 
turn, the diffusion of the new agricultural technologies is insured through the involvement of 
the most important stakeholder (Kraaijvanger et al., 2016; Kolech et al., 2017). 
The objective of the participatory breeding approaches, since their adoption, was to exploit 
farmers’ perceptions and preferences in setting the breeding goals in order to increase the 
likelihood of adoption (Morris & Bellon, 2004; Ortiz-Ferrara et al., 2007; Manzanilla et al., 
2011). Thus, selection of parent genotypes to be crossed for the development of breeding 
populations has been based on PRA results. Some authors reported difference in farmers’ 
preferences and perception in developing countries (Aguilar et al., 2014; Ali et al., 2015; 
Backiny-Yitna & McGee, 2015; Oseni et al., 2015). These authors suggested that 
implementing gender in agricultural project results in greater impact on the project beneficiary 
(Johnson et al., 2016) as gender affects adoption of new agricultural technologies (Doss & 
Morris, 2001; Danial et al., 2007). Thus, studies for the identification of production constraints 
and farmers’ preferences should not ignore differences in the roles played by men and women 
(Kristjanson et al., 2017). 
25 
 
University of Ghana http://ugspace.ug.edu.gh
In Togo, few improved groundnut varieties have been introduced and made available to farmers 
because of the absence of an organized breeding programme. Consequently, yields have been 
decreasing. Though late leaf spot (LLS) and low yielding cultivars have been blamed for the 
decreasing groundnut yield in Togo by extension agents, no study has been carried to ascertain 
production constraints. Therefore, a PRA was carried out in three Northern regions of Togo 
with the aim of identifying groundnut production constraints, assessing farmers’ knowledge of 
LLS, and assessing farmers’ preferred traits. 
3.2 Methodology 
3.2.1 Description of the study areas 
The survey was carried out in Centrale, Kara and Savanes (Table 3.1; Figure 3.1). The study 
area is in Sudano-Sahelian zone (Adewi et al., 2010; Batebana et al., 2015) and experiences 
one rainy season. All the three regions share farming and trading as major socio-economic 
activities. However, Savanes is characterized to a lesser extent by livestock keeping. The 
survey was carried out in nine villages made up of three villages per region. Villages were 
selected on the basis of the importance of groundnut in the farming system and the 
representativeness of farmers using the national statistics data. 
3.2.2 Data collection 
Preliminary data obtained from the Directorate of National Agricultural Statistics (DSID, 2015) 
on groundnut production guided the survey team in the identification of the extension services 
and farmers. Preliminary data included groundnut farming practices and the production 
constraints as reported by the extension agents. The mean number of groundnut farmers was 
73 per village (DSID, 2015), and 20 farmers including males and females were randomly 
selected and interviewed in each village from the list of farmers provided by the extension 
service. In total, 180 groundnut growers were interviewed from the nine villages (Table 3.1). 
26 
 
University of Ghana http://ugspace.ug.edu.gh
Farmers were individually contacted and visited according to their availability. A semi-
structured questionnaire was used for individual interviews (Appendix 4A). 
 
Figure 3.1 Map of Togo showing the surveyed area (ArcGIS 10.2.2) 
27 
 
University of Ghana http://ugspace.ug.edu.gh
One focus group discussion (FG) was also carried out in each village. In addition, two focal 
group discussions were carried in Maritime. These discussions were intended to check the 
assumption that farmers were interested in growing groundnut in the southern Togo. Each focus 
group was made up of 10 males and 10 females, thus 20 farmers per group discussion. To 
ensure that all topics were covered, a checklist (Appendix 4B) was used by the moderator. The 
FGs occurred in the morning and in the afternoon but mostly in the afternoon. 
Because of their familiarity with farmers, village chiefs were of a great help in the organization 
of the FGs and the administration of the questionnaire. Data on the agronomic practices, 
cultivated varieties, the use of groundnut, the preferences and the constraints in groundnut 
production were collected. Farmers’ preferences and their perceptions of constraints were 
assessed using scoring and ranking. Also, data related to farmers’ knowledge and management 
options of diseases, with a focus on LLS, were collected. Along with the farming data, 
demographic information and gender differences were assessed for the perception and 
preferences. Groundnut seed samples were collected from farmers, where possible. 
3.2.3 Data analysis 
The statistical analysis was carried in R version 3.3.1 using the package “surveydata” (Lumley, 
2011). Analysis of variance was carried for quantitative data collected from individual farmer 
interviews. For the presentation of the results, percentage and means were used. The Chi-square 
test was used as statistical test for the comparison of the perceptions between gender and 
between regions. Pearson correlation analysis (Benesty et al., 2009) was carried for the 
determination of the association between social and farming system parameters. 
  
28 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.1 Sites of PRA and the number of respondents  
Region Village Community Location No. of FG No. of 
informants 
Kara Kpoloubal Bassar N 09.77990 1 20 
E 00.62039 
Binadjoub Kabye N 09.56500 1 20 
E 00.69254 
Nampoch Lamba N 09.22001 1 20 
E00.71510 
Savanes Yacle Moba N 10.29908 1 20 
E 00.79510 
Dore Bissa N 10.72191 1 20 
E 00.09526 
Namo Moba N 10.33601 1 20 
E 00.74864 
Centrale Wassarabou Kotocoli N 08.95293 1 20 
E 01.21332 
Atibodo Kabye N 08.95458 1 20 
E 01.24128 
Sonde Kabye N 08.562680 1 20 
E 00.973260 
Maritime Gboto Ewe N 06.67739 1 0 
E 01.53206 
Tabligbo Ewe N 06.588117 1 0 
E 01.499870 
Total 12 7  11 180 
No. of FG = number of focus groups, No. of informants = number of farmers interviewed 
For an in-depth understanding of farmers’ preference and perceptions, content analysis was 
carried out using qualitative data collected at both individual and group level (Noble & Smith, 
2014). The map of Togo showing the PRA sites was drawn in ArcGIS version 10.2.2 (Wang, 
2014) using collected geographic data. 
29 
 
University of Ghana http://ugspace.ug.edu.gh
3.3 Results 
3.3.1 Demographic characteristics of farmers 
• Age of farmers 
Across the surveyed area, the most predominant age group was between 41-50 years (33.89%) 
followed by age group less than 40 years (28.89%) and between 51-60 years (26.11%) while 
age group over 61 years (11.11%) were minorities (Table 3.2). A comparison of the regions 
showed that the most important age group was of farmers aged between 41-50 years both in 
Savanes (31.67%) and Centrale (36.67%) and less than 40 years in Kara (38.33%). The 
statistical analysis of variance revealed a significant difference between regions (P< 0.01). 
Savanes exhibited the highest mean age, while Kara presented the lowest (results not shown). 
No statistical difference was observed at the village level within regions. 
• Educational qualification of groundnut farmers and typical household size  
A large proportion of farmers (42.22%) were illiterate. Only 20.55% and 19.45% of the farmers 
had primary school certificate and secondary school certificate, respectively. The rest of the 
respondents had some form of literacy tuition. Statistical comparison between regions showed 
no significant differences (χ2(6) = 11.79, p > 0.05) though Centrale seemed more educated with 
36.66% illiteracy. With respect to gender, men were more literate than women (χ2(3) = 62.52, p 
= 0.001) with the highest proportion of illiteracy in women in Kara and Savanes (Figure 3.2).  
30 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.2 Ages of groundnut farmers in the surveyed area 
  Age 
Region Village <40 41-50 51-60 >61 
Num.* Perc. Numb. Perc. Numb. Perc. Numb. Perc. 
Kara Kpoloubal 7 35 10 50 2 10 1 5 
Binadjoub 8 40 6 30 4 20 2 10 
Nampoch 8 40 4 20 8 40 0 0 
 Total/Mean 23 38.33 19 33.33 14 23.34 3 5 
Savanes Yacle 4 20 6 30 6 30 4 20 
Dore 3 15 7 35 5 25 5 25 
Namo 4 20 6 30 6 30 4 20 
 Total/Mean 11 18.33 19 31.67 17 28.33 13 21.67 
Centrale Sonde 4 20 10 50 6 30 0 0 
Kitambouli 6 30 6 30 6 30 2 10 
Wassarabou 8 40 6 30 4 20 2 10 
 Total/Mean 18 30 22 36.67 16 26.67 4 6.66 
Grand Total/Mean 52 28.89 61 33.89 47 26.11 20 11.11 
*Num = Number, Perc = Percentage 
  
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Male Female Male Female Male Female
Kara Savanne Centrale
Illiterate Lit.Tuition Primary School Secondary School
 
Figure 3.2 Educational qualification. Axes are in proportion 
 
31 
 
University of Ghana http://ugspace.ug.edu.gh
From the survey results, it was established that the mean family size was: nine (9) in Savanes, 
eight (8) in Kara and seven (7) in Centrale (Table 3.3). Analysis of variance revealed significant 
difference between Savanes and Kara districts (P = 0.05). Chi-square test showed no significant 
difference between gender for the household size.  
• Gender and marital status of farmers 
Sex and marital status of the groundnut growers are summarized in Table 3.4. Overall, 62.22% 
and 37.78% of the respondents representing male and female, respectively, are engaged in 
groundnut production. There was no significant difference between regions on proportion of 
males and females (χ2(2) = 4.30, p > 0.05). However, in Kara, Nampoch exhibited the highest 
proportion of females engaged in groundnut production (80%).  Marital status showed that 
majority of farmers were married (91.11%) with only 5.56% unmarried and 3.33% widowers. 
A significant difference between regions and community on the marital status was observed 
(χ2(4) = 10.19, p = 0.05). Detailed information on the marital status revealed that more than 
half of the married farmers were polygamous. 
 
32 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.3 Educational level and household size of groundnut farmers in the surveyed area in Togo.  
  Qualification Typical 
Household 
Region Village Illiterate Lit. Tuition PS SS Degree 
Num. Perc. Num. Perc. Num. Perc. Num. Perc. Num. Perc. 
Kara Kpoloubal 9 45 6 30 4 20 1 5 0 0 9 
Binadjoub 8 40 2 10 5 25 5 25 0 0 7 
Nampoch 10 50 0 0 8 40 2 10 0 0 6 
 Total/Mean 27 45 8 13.33 17 28.33 8 13.34 0 0 7.33ab** 
Savanes Yacle 12 60 0 0 2 10 6 30 0 0 6 
Dore 3 15 7 35 6 30 4 20 0 0 11 
Namo 12 60 0 0 2 10 6 30 0 0 6 
 Total/Mean 27 45 7 11.66 10 16.66 16 26.68 0 0 8.33a 
Centrale Sonde 9 45 4 20 3 15 4 20 0 0 7 
Attibodo 7 35 6 30 3 15 4 20 0 0 6 
Wassarabou 6 30 7 35 4 20 3 15 0 0 6 
 Total/Mean 22 36.66 17 28.33 10 16.67 11 18.34 0 0 6.33b 
Grand total/mean 76 42.22 32 17.78 37 20.55 35 19.45 0 0 7.33 
*Num = Number, Perc = Percentage; **Means within a column followed by the same letter (s) are not significantly different
33 
 
University of Ghana http://ugspace.ug.edu.gh
 
 Table 3.4 Gender and marital status of groundnut farmers in the surveyed area in Togo 
  Gender  Marital status 
Region Village Male Female  Married Single Widower 
Num. Perc. Num. Perc.  Num. Perc. Num. Perc. Num. Perc. 
Kara Kpoloubal 15 75 5 25  20 100 0 0 0 0 
Binadjoub 13 65 7 35  19 95 1 5 0 0 
Nampoch 4 20 16 80  20 100 0 0 0 0 
 Total/Mean 32 53.33 28 46.67  59 98.33 1 1.67 0 0b 
Savanes Yacle 18 90 2 10  18 90 0 0 2 10 
Dore 12 60 8 40  18 90 2 10 0 0 
Namo 13 65 7 35  15 75 2 10 3 15 
 Total/Mean 43 71.66 17 28.33  51 85 4 6.67 5 8.33a 
Centrale Sonde 16 80 4 20  18 90 2 10 0 0 
Attibodo 11 55 9 45  19 95 0 0 1 5 
Wassarabou 10 50 10 50  17 85 3 15 0 0 
 Total/Mean 37 61.66 23 38.33  54 90 5 8.33 1 1.67b 
Grand Total/Mean 112 62.22 68 37.78  164 91.11 10 5.56 6 3.33 
*Num = Number, Perc = Percentage; **Means within a column followed by the same letter (s) are not significantly different. 
34 
 
University of Ghana http://ugspace.ug.edu.gh
3.3.2 Practices in groundnut production 
Land preparation for groundnut cultivation usually starts around April (in Central), May in 
Kara or June in Savanes with the onset of the rainy season. The main equipment used were 
axes, hoes and tractors. In addition, the use of ox-plows was common in Savanes. Usually, 
sowing is done after plowing and the first rains, and weeding was carried out twice or thrice if 
necessary. All the described activities were carried by both men and women. There was no use 
of inorganic fertilizers on groundnut. However, in mixed cropping with maize or sorghum, 
groundnut can benefit from the fertilizers applied to those cereals. Unlike Kara and Centrale, 
where manure incorporation is not common, Savanes is characterized by the systematic 
incorporation of manure in some areas. 
3.3.3 Farm characteristics  
Groundnut farm size ranged from 0.44 hectare to 1.66 hectares in the surveyed area (Table 
3.5). Statistical difference based on the analysis of variance was observed between regions (P 
= 0.001) and between community (P = 0.01). The largest groundnut plots were observed in 
Savanes (1.35 hectares) and the smallest in Centrale (0.6 hectare) and Kara (0.58 hectare). 
Gender analysis revealed a clear difference in land access as men had larger land than women 
(Figure 3.3). 
8
7
6
5
4
3
2
1
0
Male Female Male Female Male Female
Kara Savanne Centrale
Land Available Grounut Land
 
Figure 3.3 Land available for agriculture and land area (ha) used for groundnut production as estimated by 
interviewees
35 
 
ha 
University of Ghana http://ugspace.ug.edu.gh
Table 3.5 Farm size, seed sources, cropping system and proportion of groundnut sold 
  Farm size (ha)  Seed Sources % M. C.** Sold 
proportion 
Region Village Smallest Mean Largest   Market Saved Others 
Kara Kpoloubal 0.25 0.51 0.75  30.00 60.00 10.00 60.00 60.00 
Binadjoub 0.25 0.71 2.00  45.00 50.00 5.00 95.00 75.5 
Nampoch 0.25 0.52 0.75  05.00 75.00 20.00 80.00 59.2 
 Mean 0.25 0.58b 1.16  26.67 61.67 11.66 78.33 64.90b 
Savanes Yacle 1 1.47 2.00  20.00 65.00 15.00 60.00 93.50 
Dore 0.5 1.12 2.00  25.00 75.00 0.00 7.00 83.75 
Namo 1 1.47 2.00  30.00 70.00 0.00 53.00 93.50 
 Mean 0.83 1.35a 2.00  25.00 70.00 5.00 40.00 90.25a 
Centrale Sonde 0.25 0.77 2.00  85.00 5.00 10.00 35 82.50 
Attibodo 0.25 0.53 1.00  65.00 25.00 10.00 40 93.6 
Wassarabou 0.25 0.49 1.00  45.00 40.00 15.00 30 88.75 
 Mean 0.25 0.6b 1.33  65.00 23.33 11.66 35 88.28a 
Grand Mean 0.44 0.84 1.66  38.89 51.67 9.44 51.11 81.14 
**% M. C.:   % Mixed cropping 
*Means within a column followed by the same letter (s) are not significantly different. 
  
36 
 
University of Ghana http://ugspace.ug.edu.gh
Most of the farmers relied on own saved seeds (51.67% of farmers) for next season planting. 
However, 38.89% of the farmers reported purchasing seeds from the surrounding markets 
(Table 3.5). The remaining farmers (9%) obtained their seeds for planting by the means of 
borrowing or donation from Non-governmental organizations (NGOs). Deutsche Gesellschaft 
für Internationale Zusammenarbeit (GIZ), an NGO in Togo, has been often reported by the 
interviewed farmers as the source of the groundnut seeds. Nonetheless, there were differences 
between regions. The proportion of farmers purchasing seeds from market was higher in 
Centrale than in the other regions (χ2(4) = 32.52, p = 0.001). Also, a slight gender difference 
was observed as women tended to obtain planting materials by means of donation and 
borrowing than men (Figure 3.4).  
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Male Female Male Female Male Female
Kara Savanne Centrale
Market Saved Others
 
Figure 3.4 Source of groundnut seeds (in percentage) 
 
In the surveyed area, groundnut was grown mostly in mixed cropping with cereals such as 
maize, millet and sorghum (Table 3.5). A large proportion of the harvest (81%) is sold at the 
market soon after harvest (Table 3.5). The sold proportion of the harvest was higher in Savanes 
(90%) and Centrale (88%) than in Kara (p = 0.001) where a substantial part is kept for home 
consumption (Figure 3.5). 
37 
 
University of Ghana http://ugspace.ug.edu.gh
 
100
90
80
70
60
50
40
30
20
10
0
Male Female Male Female Male Female
Kara Savanne Centrale
 
Figure 3.5 Proportion of groundnut sold 
3.3.4 Cultivated groundnut varieties 
The rank of groundnut based on the importance in the community varied from one region to 
the other. Groundnut ranked second in Kara and Savanes behind sorghum or maize and third 
in Centrale behind millet and maize. Assessment of the groundnut varieties grown by farmers 
revealed that few local landraces were cultivated in the surveyed area (Table 3.6). For instance, 
no more than one variety has been reported in Centrale in each of the community visited. 
Improved varieties, introduced from foreign countries were reported by farmers. The most 
common improved varieties were T3, SORAD, SOTOCO, RMP12 and ICIAR19B. The most 
widespread improved variety across the regions was SORAD, also called SOTOCO. These 
names given to the same variety were the names of the national companies that carried out the 
distribution of the variety when it was introduced into the country. SORAD was appreciated in 
the past by farmers because of the large pod and seed size. According to the farmers, the variety 
has become low yielding over the years. It is worthy to mention that no local improved varieties 
were reported by the respondents. 
  
38 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.6 Characteristics of cultivated varieties according to farmers in the surveyed area 
Region Villages Varieties Characteristics 
Savanes Yacle Ntifofo  
Koka  
Soulare  
SORAD Big pods and grains 
Dore RMP12 Small pods and grains 
T3  
Namo SORAD Big pods and grains 
Koumongou  
Kara Binadjoub SOTOCO Big pod and grains/ Low yield 
 Tchamba  
Kpoloubal Ngbengbeng Small pods / Low yield 
 Djafo  
 Tchana  
Nampoch Smagbengbe Small grains/Low yield 
 Oukandjassina  
 Tchana  
Centrale Sonde Nale-Nale 2 kernels per pod / Low yield 
Kitambouli Lossoketo Small pods / Low Yield 
Wassarabou SORAD Big pod and grains 
 
3.3.5 Preferred characteristics of groundnut varieties 
Overall, farmers had similar preferences as yield-related traits were the most often mentioned 
by the respondents (Table 3.7). However, a significant difference was observed between 
regions based on chi-square test (χ2(16) = 60.52, p = 0.001). In Kara, pod yield was mentioned 
by 65% of the respondents as the most important. The other preferred traits in Kara were 
drought (13.33%), pod size (6.66%) and taste (6.66%). In Savanes, pod yield (64.44%) 
followed by traits such as pod size (12.22%) and high oil content (6.11%) were emphasized by 
39 
 
University of Ghana http://ugspace.ug.edu.gh
farmers. In addition to pod yield and large pod size which were the most preferred traits, disease 
resistance and a large number of seeds per pod were also indicated as traits of interest by 
farmers in Centrale. Red seed colour (by those who mentioned colour) and early maturity were 
also identified by farmers as important characteristics to have in mind when developing 
improved varieties. Comparison of the preferences between gender showed no significant 
differences. 
40 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.7 Percentage of time preferred characteristic of groundnut varieties were mentioned 
Region Village Large Pod Yield Oil Drought Colour Disease Taste Early Seed/pod 
Size Resistance Maturity 
Kara Kpoloubal 5 55 10 20 10 - - - - 
Binadjoub 5 90 - - - 5 - - - 
Nampoch 10 50 - 20 - - 20 - - 
 Mean 6.66 65 3.33 13.33 3.33 1.66 6.66 - - 
Savanes Yacle 30 60 10 - - - - - - 
Dore - 75 25 - - - - - - 
Namo 25 65 10 - - - - - - 
 Mean 18.33 66.66 15 - - - - -  
Centrale Sonde 35 50 - - - - - 10 5 
Attibodo - 70 - - - 5 - 10 15 
Wassarabou - 65 - - - 25 - - 10 
 Mean 11.66 61.67 - - - 10 - 6.67 10 
Grand mean 12.22 64.44 6.11 4.44 1.11 3.33 2.22 2.22 3.33 
41 
 
University of Ghana http://ugspace.ug.edu.gh
3.3.6 Perception of farmers on constraints to groundnut production 
Generally, diseases and insects were the most widespread constraints limiting groundnut 
production in the surveyed area (Table 3.8). Disease was identified by 37.77% of the 
respondents while insects’ attacks were mentioned by 27.77% of the farmers. Other minor 
constraints mentioned by farmers were the lack of high yielding varieties by 8.33% of farmers 
and Striga by 4.5% of the respondents. 
Though no differences were observed between gender on the perception of the constraints, chi-
square test revealed that the importance of the constraints varied between regions (χ2(18) = 
84.70, p = 0.001). Diseases were more important in Kara and Centrale while insects were 
considered as most important in Savanes. The mention of Striga as one of the most important 
constraints on groundnut occurred only in Kara (13.33% of the respondents). The other minor 
constraints raised by the farmers were the lack of groundnut seeds in Savanes (by 6.66% of the 
respondents) and the lack of high yielding varieties in Centrale (15% of the respondents). 
42 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.8 Percentage of times that constraints were mentioned 
Region Village Disease Insects Striga Drought Low Yield Labour Lack of Flooding Soil Fertility No constraint 
Seed 
Kara Kpoloubal 60 10 20 - - - - - - 10 
Binadjoub 25 35 20 5 10 5 - - - - 
Nampoch 30 20 - 20 20 - - - - 10 
 Mean 38.33 21.66 13.33 8.33 10 1.66 - - - 6.66 
Savanne Yacle 20 60 - 10 - - 10 - - - 
Dore 25 25 - 50 - - - - - - 
Namo 20 60 - 10 - - 10 - - - 
 Mean 21.66 48.33 - 23.33 - - 6.66 -  - 
Centrale Sonde 65 15 - - - - 10 10 - - 
Attibodo 60 15 - - 15 - - - 10 - 
Wassarabou 35 10 - - 30 - - 20 5 - 
 Mean 53.33 13.33 - - 15 - 3.33 10 5 - 
Grand Mean 37.77 27.77 4.44 10.55 8.33 0.55 3.33 3.33 1.66 2.22 
43 
 
University of Ghana http://ugspace.ug.edu.gh
3.3.7 Awareness of Late Leaf Spot disease  
A large majority of farmers were aware of LLS disease (95.55% of the farmers) and the three 
regions exhibited similar awareness of the disease (Table 3.9). Nonetheless, a large proportion 
of farmers ignore the cause of the disease (60.55%) and nearly all the farmers do not have 
measures to control the disease. 
Table 3.9 Farmers’ perception of LLS 
Regions Community % of respondents % of respondents Available control 
knowing LLS knowing the cause measure 
symptoms of LLS 
Kara Kpoloubal 100 0 10 
 Binadjoub 100 10 0 
Nampoch 90 10 0 
Total 96.67 10 0 
Savanes Yacle 100 60 0 
 Dore 100 50 0 
Namo 100 60 0 
Total 100 56.67 0 
Centrale Sonde 95 70 0 
Attibodo 85 30 0 
Wassarabou 90 55 0 
Total 90 51.67 0 
Total 95.55 39.44 1.11 
 
Overall, the disease was associated to pods’ maturity by 27.22% of the respondents while 
7.22% mentioned drought to be the probable cause of LLS (Figure 3.6d). Further investigations 
on the probable cause of the disease reported by farmers, revealed that the perception of the 
44 
 
University of Ghana http://ugspace.ug.edu.gh
cause of LLS disease was not the same between region (χ2(12) = 64.04, p = 0.001). In Centrale 
and Savanes, respectively 48.33% and 30% of the farmers considered LLS symptoms as 
maturity symptoms (Figure 3.6a). Other causes such as drought and sun burn (Sunbeams) were 
mentioned respectively by 20% and 6.65% of the interviewed in Savanes (Figure 3.6b). In 
Centrale, divine punishment was (Div.Punishement) was reported by farmers as one the causes 
among others (Figure 3.6c). 
 
1.67 3.33
1.67 3.33
20
43.33
30
90 (a) Kara (b) Savane
Drought Insects Virus Maturity Unknown 6.67
Drought Maturity Sunbeams Usn known
         
 
7.22 0.55
1.11
27.22
48.33 48.33
60.55
2.22
1.11
(d) Across the three regions 
3.34 Drought Insects
(c) Centrale Virus Maturity
Maturity Div. Punishment Unknown Sunbeams Div. Punishment
Unknown
            
Figure 3.6 Perception on the cause of LLS disease on groundnut (figures are in percentage) 
  
45 
 
University of Ghana http://ugspace.ug.edu.gh
3.3.8 Correlation among social and farming system parameters  
A significant correlation was observed between farmers’ age and farmers’ experience (P = 
0.01). Similarly, a significant correlation was observed between the household size and 
farmers’ age, the land available for agriculture and groundnut farm size, land available for 
agriculture and gender (Table 3.10). 
 
46 
 
University of Ghana http://ugspace.ug.edu.gh
Table 3.10 Correlation among social and farming system parameters  
 Age Household Sex Land Available Education Marital Status Groundnut Farmer 
size Land Exp 
Age  *** * ** NS NS *** *** 
Household size 0.53   NS *** NS *** NS *** 
Sex -0.15 -0.08  *** *** NS *** NS 
  
Land Available 0.30  0.30 0.51  *** NS *** ** 
  
Education 0.008 0.13 0.32 0.29   NS NS NS 
 
Marital Status 0.19  -0.21 0.12 -0.07 -0.11  * ** 
 
Groundnut Land 0.19  0.04 0.36  0.63 0.12 0.15  *** 
  
Farmers’ Experience 0.81  0.43 0.13 0.36  0.09 0.25 0.36   
  
Significance codes: ‘***’ 0.001; ‘**’ 0.01; ‘*’ 0.05, ‘NS’ Non-significant
47 
 
University of Ghana http://ugspace.ug.edu.gh
3.4 Discussion 
A diversity of farming practices was observed in the surveyed area. These practices, though 
characterized by rationality, seemed untidy and unsystematic. The observed practices, in the 
context of the climate change, may partly be the cause of the low yield observed in groundnut 
production in Togo. Moreover, the changing environment calls for developing new technologies 
that adapt to farmers growing conditions. It is well known that the adoption of new agricultural 
technologies is affected by gender and educational level. This study has revealed that most 
groundnut farmers in Togo are illiterate. Hence, any diffusion effort of new varieties should take 
into account the issue in its approach. For instance, farmers in Centrale because of their high 
education level tend to purchase seeds than the other farmers. The tendency to purchase seeds from 
market could be translated as the propensity to adopt new improved varieties if available. 
Preliminary socio-economic data analysis revealed that most of the farmers in Centrale originated 
from Kara and nearly all of them or their parents have migrated in search of agricultural land. This 
assumption was confirmed by this survey as Kara exhibited the smallest land surface available for 
agriculture. 
Also, there was a clear gender bias in land ownership as men had larger lands available for 
agriculture than women. Doss et al. (2015) and Kieran et al. (2015) previously reported similar 
findings. The correlation between gender and land ownership highlight that gender should not be 
ignored when implementing agricultural projects. However, in some areas such as Savanes, the 
gender bias was very low. Financially sound women in Savanes, who are mostly traders, own plots 
as large as men’s plots. In Nampoch, it was reported that, because of the decreasing yield in 
groundnut production, men are shifting to more profitable crops such as soybean and yam. Mainly 
because of the home consumption, and also likely marginalized by the development of cash crops 
48 
 
University of Ghana http://ugspace.ug.edu.gh
(Wooten, 2003), women in that locality are forced to keep growing groundnut. The observed bias 
between gender for access to groundnut seeds and educational level underlines the necessity to 
develop and diffuse breeding products with gender in mind (Kristjanson et al., 2017).  
Except slight differences, farmers’ preferences were very similar across the surveyed areas. 
Generally, yield-related traits such as pod yield and large pod were the most preferred. Also, a 
large number of pods per plant is viewed by farmers as yield selection criterion. Region wise, 
respondents in Kara preferred taste and drought resistant in addition to the yield-related traits. In 
the contrast to the other regions, a substantial proportion of groundnut production is locally 
consumed in Kara. This could explain why taste is a valued trait for farmers in that area. In Savanes 
and Centrale, secondary traits such as disease resistance, number of seeds per pod, and oil content 
were raised as traits to be incorporated into improved varieties. The diversity of farmers 
preferences in Savanes and Centrale could be related to the groundnut uses and the market demand 
because nearly all the production in these areas is sold. In Savanes, groundnut buyers come mainly 
from the surrounding foreign countries such as Ghana and Benin. In contrast, farmers in Centrale 
sell their groundnut to local buyers and processors. In case of home consumption, groundnut is 
mainly used for sauce in the surveyed area. However, boiled and roasted groundnut as well as 
processed into cake are also the common use of groundnut throughout the country. Thus, the 
diversity of the use of groundnut could partly explain the slight differences in farmers’ preferences. 
Overall, farmers face a wide range of constraints in groundnut production. Nonetheless, diseases, 
insects and drought could be considered as the most important factors limiting groundnut 
production according to the farmers. Recently, a survey carried out in Ghana in groundnut 
production revealed similar results (Yussif et al. 2014; Oppong-Sekyere et al., 2015) with drought, 
low yield and diseases raised by farmers as the limiting factors in groundnut production. The 
49 
 
University of Ghana http://ugspace.ug.edu.gh
perceptions of farmers on the importance of the constraints were not the same from one region to 
the other. In fact, insects were considered as the major constraint in groundnut production in 
Savanes. The explanation of this perception could be found in the untidy agronomic practices in 
that area. Harvested groundnut is left in the field for up to fifteen days for drying and facilitating 
of easy pod stripping. Farmers in that area reported that the attacks of the insects occur during the 
period between harvest and pod stripping. This issue could be addressed by adopting best 
agronomic practices. 
Nonetheless, insects’ attacks have been reported during storage in the other regions. This forces 
farmers to sell their groundnut soon after harvest in order to relieve themselves from quality 
maintenance and storage insects’ issues. In addition to the storage issues, farmers often sell their 
groundnut soon after harvest, when the price is low, in order to meet school fees and healthcare 
costs. There is a need to finance the development of groundnut storage facilities. These facilities 
were made available to farmers, through agricultural credits, for other crops such as maize 
(ROPPA, 2013). Extending these credits on groundnut will likely help farmers increase their 
income through the reduction of the loss in storage. Also, such facilities will allow farmers to keep 
their harvest for a higher price in the future. The profitability of such initiative depends also on the 
collaboration between a strong breeding programme, an efficient extension service and the farmers 
for coordinated agricultural activities. Though clearly defined groundnut farmers organization 
does not exist in the surveyed area, the existence of other organizations of which most farmers are 
members could be exploited by the research and the extension service. 
In Kara, Striga has been strikingly mentioned by some farmers as a constraint on groundnut 
production. The perception of striga as a constraint on groundnut may be partly related to the 
cropping system in that region. Indeed, farmers in Kara mostly grow groundnut in mixed cropping 
50 
 
University of Ghana http://ugspace.ug.edu.gh
with other cereal such as maize and sorghum. Thus, Striga hermonthica could have been mistaken 
to affect groundnut. Therefore, cereal breeding programmes in Togo should be aware that 
improved varieties targeting this region should incorporate Striga resistance. 
Farmers are aware of the widespread status of LLS disease in their area. When asked to rank the 
foliar diseases on the basis of the importance, majority of farmers identified LLS as the most 
widespread foliar disease confirming previous reports that spotted LLS as one of the most 
important fungal disease in groundnut production areas (Sujay et al., 2012). The disease severity 
has increased over the years according to the respondents. Farmers in Centrale reported that there 
were local resistant varieties about two decades ago. Because of the changing climate, which 
results often times in the shortening of the rainy season (Adewi et al., 2010; Batebana et al., 2015), 
new early maturing varieties were likely adopted to the detriment of the landraces with long 
maturity period. It is well known that LLS resistance is associated with late maturing landraces 
(Janila et al., 2013). 
The cause of LLS disease seemed unknown to most of the interviewed. To farmers opinion, LLS 
is a natural disease occurring when the pod reached the full maturity. Farmers have been using the 
symptoms as an indication of the pods’ maturity for long, and therefore neglected the development 
of disease control methods. However, farmers reported that the symptoms are no longer used 
nowadays as an indication of pod maturity because of the unmatured pods that were often found 
using LLS as indicator on the currently cultivated varieties. It is worthy mentioning that the same 
belief on LLS has been reported in Congo (Izge et al., 2007). These results underline the ignorance 
of groundnut farmers on the cause of the diseases affecting groundnut production. For instance, 
groundnut rosette has been associated with witchcraft in the surveyed area. Altogether, these 
51 
 
University of Ghana http://ugspace.ug.edu.gh
results suggest that groundnut farmers should be educated on the important diseases through an 
awareness creation. 
Overall, few improved varieties are available for farmers and most farmers rely on the old 
landraces for groundnut production. SORAD, also named SOTOCO, was introduced twenty years 
ago to boost groundnut production. The variety was quickly adopted by farmers because of its 
large pod size and the high yield. This is confirmed by this survey, as SORAD was the most 
widespread variety in the surveyed area. However, though the variety is still cultivated, the 
groundnut farmers complained about its performance. SORAD is a late maturing variety that often 
exhibits low yield in the context of erratic rainfall. Early maturing varieties such RMP12 and T3 
were recently introduced to Togo as alternatives to the existing low yielding varieties. However, 
this survey revealed that few farmers adopted these varieties. The main reproaches to these 
varieties are their small pod and seed size. It could be concluded that farmers in the surveyed area 
would reject improved varieties with small pod and seed size. In fact, difficulty in manual shelling 
and low yield were associated with small pod size. Introduced six years ago from Nigeria, Samnut 
24 is another variety that is currently promoted by the research institute because of the high number 
of pods per plant. Unfortunately, multiple tests and field observations showed that Samnut 24 is 
highly susceptible to early and late leaf spot diseases. Altogether, there was a clearly expressed 
need for higher-yielding varieties that can boost groundnut production in Togo.  
Groundnut breeders in Togo could make use of the opportunity created by the identified challenges 
for the development of improved varieties that will likely meet farmers’ preferences. A 
participatory breeding involving all the stakeholders in groundnut production would probably 
increase the chance of adoption of the new improved varieties. 
52 
 
University of Ghana http://ugspace.ug.edu.gh
3.5 Conclusion 
This study provided a basis for a participatory groundnut breeding programme through the 
identification of the constraints and preferences as well as the characterization of agronomic 
practices in groundnut production in Togo.  This study has revealed that, with few exceptions, 
yield-related traits such as pod yield and pod size were the most important selection criteria of 
farmers. Diseases were considered to be the most important factor limiting groundnut production. 
Among diseases, farmers indicated that LLS was of economic importance. Thus, in Togo, a 
breeding programme on groundnut should consider that groundnut farmers perceive diseases as 
the major limiting factor to production. High yielding groundnut varieties that incorporate traits 
such as large pod size and resistance to LLS are likely to be adopted by groundnut farmers in Togo. 
Thus, the next studies will focus on mining the existing germplasm and breeding tools to use in 
the breeding programme. 
  
53 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER FOUR 
4. MOLECULAR AND PHENOTYPIC DIVERSITY OF GROUNDNUT 
CULTIVARS FROM TOGO 
4.1 Introduction 
The success of a breeding programme is dependent on the development of a gene pool with useful 
genes and a broad genetic base (Upadhyaya et al., 2009; Sharma et al., 2013). Despite the 
availability of elite lines in some countries, landraces are still considered to harbor useful alleles 
that need to be exploited in plant breeding (Dwivedi et al., 2017), justifying the need to collect and 
create a core collection when starting a breeding programme. Genes discovered in diverse 
germplasm resources can be introgressed into farmers’ preferred varieties for the improvement of 
yield and other traits of interest (Cobb et al., 2013).  
Morphological and molecular diversity analyses have been very useful in characterizing 
germplasm collections and identifying potential parents for a breeding programme. Morphological 
characters and physiological traits are visible descriptors that are important for diversity studies. 
Also, morphological descriptors remain important for identifying accessions at the farm level, as 
farmers rely on observable differences to describe landraces. In addition, the knowledge of the 
importance of an agronomic trait, in terms of contribution toward diversity, is becoming very 
crucial in plant breeding (Dwivedi et al., 2017).  
Nowadays, molecular markers have become more important for estimating genetic diversity 
(Milla-Lewis et al. 2010; Roomi et al., 2014; Ren et al., 2014; Bhad et al., 2016). They 
complement morphological markers in a diversity study. Among the available markers, single 
nucleotide polymorphisms (SNPs) are becoming the marker of choice to the detriment of other 
54 
 
University of Ghana http://ugspace.ug.edu.gh
genetic markers. SNPs linked to the coding regions can be more efficiently used in breeding 
methodologies such as diversity study than other markers (Luikart et al. 2003). In addition, SNPs 
low mutation rate often results in more accuracy in the estimation of population structure 
parameters (Coates et al., 2009). DArTseq-based SNPs is one of the low-cost microarrays-based 
methods used for genetic studies including genetic diversity and structure analysis in many crops 
(Jaccoud et al., 2001) that has proven to be efficient.  
As a starting point for the new breeding programme, diversity assessment of the cultivated 
groundnut accessions from Togo, using morphological and molecular marker traits, was the aim 
in the present study. The activities focused on collection of groundnut accessions grown by 
farmers; description of accessions and classification into groups; assessment of the 
interrelationships among accessions, among traits and the estimation of morphological and 
molecular diversity in the collected germplasm. 
4.2 Materials and Methods 
4.2.1 Collection of accessions 
The identification of the collections area was based on the importance of groundnut in local 
farming system and ‘visiting a maximum number of sites’ has been adopted as collection strategy. 
In each region, the collection sites were determined with the aid of the agricultural extension 
agents. A meeting was organized with the extension service to define the best itinerary that would 
facilitate the capture of the diversity of groundnut genotypes grown in the area. Thus, forty-one 
villages from four regions were visited: 16 villages in Savanes, 14 in Kara, 8 in Centrale, and 3 in 
Maritime (Figure 4.1). In each village, groundnut germplasm (seeds and pods) were obtained from 
farmers, mostly in storages or from local markets that are nearly 10 km apart. Shelled groundnut 
samples were mostly found at the market while samples from farmers’ storage were, in a large 
55 
 
University of Ghana http://ugspace.ug.edu.gh
part, unshelled. Geographical coordinates were collected using a GPS instrument. In addition, a 
data form was used to collect data such as name of the locality, agroecological zone, geographical 
zone, local name of the accession, ethnic group and farmer’s name. Overall, 126 groundnut 
samples were collected and put in envelops before their transportation to the Centre de Recherche 
Agronomique de la Savane Seche (CRASS), the arid savanna agronomic research centre in Kara, 
Togo. 
  
56 
 
University of Ghana http://ugspace.ug.edu.gh
 
Figure 4.1 Map of Togo showing the collection sites (Source: Banla, ArcGIS 10.1) 
 
57 
 
University of Ghana http://ugspace.ug.edu.gh
4.2.2 Plant material 
Prior to the morphological characterization, a preliminary field evaluation was carried out for the 
126 collected groundnut samples and 9 lines from Senegal, obtained from the Togolese Agronomic 
Research Institute (ITRA-CRASS). Collected samples were reduced to 94 accessions, for 
facilitating genotyping on standard 96 well plate. Results of preliminary data analysis provided 
information for eliminating the duplicates and facilitated the identification of the 94 genotypes. 
The germplasm collection (94 genotypes) consisting of collected accessions from Togo (n=83), 
lines from Senegal (n=9) and lines from ICRISAT (n=2 only genotyped), representing three 
geographical regions were used in this study. 
4.2.3 Morphological evaluation 
The genotypes were phenotyped during the 2017 and 2018 rainy seasons. The phenotyping trials 
were established in LLS hot spot environments to increase the chance of disease occurrence during 
the growing seasons from March to June at Ativeme (6°25'22" N; 1°6'38" E). Ativeme, with 1270 
mm rainfall annually and 23°C mean temperature, is characterized by a grey sandy loam soil. An 
alpha lattice design with three replications six blocks per replication was adopted as experimental 
design with two checks added to the 94 genotypes. A row of 3 meters containing 10 plants with 
30 cm between plants represented an experimental unit. A spacing of 50 cm between rows and 1 
m between blocks was adopted. The field was kept weed free at an interval of one weeding every 
three weeks. IBPGR-ICRISAT (1992) descriptors were used to assess agro-morphological 
characteristics. Morphological descriptors included: 
- Growth habit: recorded at podding stage as 1 = spreading, 2 = bunch and 3 = erect 
- Main stem: scored as 1 = presence and 2 = absence 
58 
 
University of Ghana http://ugspace.ug.edu.gh
- Number of branches: scored as 1 = primary (n+1), 2 = secondary (n+2) and 3 = Tertiary 
(n+3) 
- Branching pattern: determined on the (n+1) cotyledonary lateral branches as 1 = Alternate 
and 2 = Sequential 
- Stem surface: recorded as 1 = glabrous, 2 = Sub-glabrous and 3 = Moderately Hairy 
- Stem pigmentation: recorded on mature plant as 1 = green, 2 = purple and 3 = mixed 
- Leaflet shape: on fully expanded apical leaflet on the main stem and recorded as 1 = Narrow 
Elliptic, 2 = Oblong Elliptic, 3 = Wide Elliptic and 4 = Obovate 
- Leaflet tip: on fully expanded apical leaflet on the main stem recorded as 1 = acute and 2 
= obtuse 
- Flowers on main axis: recorded on the main stem as 1 = presence or 2 = absence 
- Colour of standard petal marking: colour of markings on the front face of the standard petal 
- Pod beak: recorded on dried pods as 1 = slight, 2 = moderate and 3 = prominent  
- Pod constriction: recorded on dried pods as 0 = none, 1 = slight, 2 = moderate and 3 = 
prominent 
- Pod reticulation: recorded as 0 = none, 1 = slight, 2 = moderate and 3 = prominent  
- Seed colour: recorded two weeks after harvest and complete drying on mature seeds 
The agronomic descriptors included:  
- Plant height (cm): measured from cotyledonary axil up to terminal bud, mean of five plants 
recorded 75 days after planting 
- Plant width (cm): measured at the widest point from branch tip to tip on 10 plants recorded 
60 days after planting  
59 
 
University of Ghana http://ugspace.ug.edu.gh
- Number of primary branches: counted at 60 days after emergence on five plants 
- Leaflet length: (mm): measured on the leaf of the main stem, fully expanded apical leaflet. 
Recorded on five leaflets from five different plants 
- Leaflet width (mm), measured on the leaf of the main stem, fully expanded apical leaflet, 
at widest point. Recorded on five leaflets from five different plants 
- Days to emergence: counted from sowing to 50% emergence 
- Days to maturity: counted from emergence to maturity determined by checking 80% pods 
maturity 
- Days to 50% flowering: counted from emergence to 50 % of plants flowering 
- Pod length (mm): recorded on 10 random mature pods  
- Pod width (mm): measured at the widest point on 10 random mature pods 
- Pod volume (cm3): volume of 10 fresh mature pods measured by water dispersal using a 
graduated test tube 
- Seeds per pod: number of seeds per pod counted on 10 random mature pods  
- Seed length (mm): recorded on 10 random mature seeds  
- Seed width (mm): measured at the midpoint on 10 random mature seeds 
- 100 pod weight (g): weight of 100 random mature pods 
- 100 seed weight (g): weight of 100 random mature wrinkle-free seeds 
- Shelling percentage = Weight of seeds obtained from 100 pods/100 pods weight   
- Pod yield (g) per plant: Mean weight of pods harvested on five random plants.  
Disease scoring for LLS was carried out at 60 days and 80 days after sowing using a 9 points scale 
(Subrahmanyam et al., 1995). Data were collected at the plant level on five plants per plot for all 
variables except for days to emergence, days to maturity and days to 50% flowering. 
60 
 
University of Ghana http://ugspace.ug.edu.gh
4.2.4 DNA Extraction and DArT Analysis 
Leaf samples were collected from three weeks old plants using the LGC plant sample collection 
kit and dried overnight at 35℃ in an oven. Dried samples were then sent to Integrated Genotyping 
Service and Support (IGSS) platform at Biosciences for east and central Africa laboratory at the 
International Livestock Research Institute (BecA/ILRI) for DNA extraction and genotyping using 
Diversity Arrays Technology (DArT) markers. Whole genome genotyping was carried out using 
Genotyping-By-Sequencing (GBS) technology (Elshire et al., 2011). 
For quality control, derived DArTseq SNPs with the following criteria were removed using 
package “dartR” in R statistical software version 3.3.1: SNPs with call rate less than 70%, SNPs 
with PIC less than 0.1, and loci with trimmed sequence tags that were too similar. After filtering 
and quality control, 990 DArTseq SNPs were retained for the diversity study. 
4.2.5 Data analysis 
4.2.5.1 Morphological data analysis 
Analyses for the morphological data were performed in R version 3.1.1 using packages 
“agricolae”, “cluster”, “plyr”, “magrittr”, “Rtsne” and “ggbiplot”. Analysis of variance was run 
for quantitative data after imputation of the missing data (Harrell, 2017) with a linear model 
considering genotypes and blocks as fixed. To handle the data with different units, Gower’s metric 
(Gower, 1971; Pavoine et al., 2009; Henry et al., 2015) and UPGMA clustering (Langfelder et 
al.,2007) were used for determining the structure of germplasm collections (Podani & Shmera, 
2006; Kaufman & Rousseeuw, 2009; Johnson & Wichern, 2014). The average Silhouette method 
in Partitioning Around Medoids (PAM) was used for estimating the optimal number of clusters 
(K) and to estimate the probability of membership (Pedregosa et al., 2011). Cophenetic correlation 
61 
 
University of Ghana http://ugspace.ug.edu.gh
coefficient (CPCC) was used to measure how faithfully the original dissimilarity structure is 
preserved by the dendrogram and as an indicator for the strength of subgroup differentiation (Sokal 
& Rohlf, 1973; Odong et al., 2011). The contribution of individual traits to the diversity was 
assessed through principal component analysis (PCA) (Coghlan, 2013), and Simpson’s diversity 
index (D) (Simpson, 1949) was used to estimate diversity in qualitative traits. The estimation of 
the correlation among variables was carried out using Pearson correlation analysis (Benesty et al., 
2009). Path analysis was carried out for yield-related traits using structural causal models as 
described by Kline (2015) and Byrne (2016) using the R package ‘lavaan’. 
4.2.5.2 Molecular data analysis 
Genetic diversity and analysis of molecular variance (AMOVA) were done in R software version 
3.3.1 using package “dartR”, “StAMPP”, “ggplot2” and “PopGenReport”. The identification of 
clusters of genetically similar genotypes was done using the Bayesian method in R software. The 
optimal number of clusters, K-value, was identified using PAM Silhouette (Pedregosa et al., 2011). 
Principal component analysis (PCA) and Neighbour-joining tree were used to reveal structure in 
the germplasm. The genetic distances among the clusters were calculated as Nei’s minimum 
distance and pairwise Fst (Hedrick, 2005).  
4.3 Results 
4.3.1 Morphological diversity 
4.3.1.1 Analysis of variance in quantitative traits 
The analysis of variance revealed a significant variability in the collected germplasm (p = 0.01) 
for all quantitative traits. Traits of agronomic importance such as LLS, pods per plant, and pod 
yield showed a highly significant variation (Table 4.1).   However, the coefficient of variation was 
very low for most traits. On the 20 quantitative traits, only four (number of lateral branches, LLS, 
62 
 
University of Ghana http://ugspace.ug.edu.gh
number of pods per plant and yield) exhibited a coefficient of variation above 20%. The coefficient 
of variation for days to emergence, leaflet width, pod width, and seed length and width were even 
close to zero. 
Table 4.1 Mean square, range and coefficient of variation of quantitative traits 
 Mean Square  Range  
Traits Genotypes Error      CV (%) 
(d.f=95) 
(d.f= 175) Min Mean Max 
Days Emergence 128.326* 0.023  05.33 6.50 8.66 2.331 
Days Flowering 2233.39** 1.98  23.00 27.19 32.66 5.17 
Days Maturity 26998.5** 2.3  80.00 94.18 120.00 1.592 
Plant Height 159.837*** 33.667  15.84 32.51 48.75 17.849 
Plant Width 154.87** 82.53  39.55 58.69 76.53 15.477 
Num.Lat. Bran. 30.15*** 7.230  05.92 9.67 22.07 27.797 
Leaflet Length 1.7444*** 0.14  03.60 05.51 07.38 6.852 
Leaflet Width 28.4667** 0.037  01.90 3.06 3.87 6.299 
LLS 60 Days 3.3131*** 0.7899  01.66 03.96 05.00 22.423 
LLS 90 Days 59.176*** 0.234  01.8 04.29 06.5 11.264 
Pods / plant 298.35*** 136.73  31.31 54.44 80.05 21.476 
Pod Volume 2.1612*** 0.082  1.98 3.23 6.05 8.889 
Pod Length 0.1452*** 0.014  01.95 02.42 3.28 4.94 
Pod Width 0.190*** 0.001  01.00 01.18 01.40 3.28 
Shelling % 30.687** 4.479  64.85 73.48 88.69 2.880 
Seed Length 5.047** 0.004  01.07 01.29 01.59 5.106 
Seed Width 2.3124** 0.001  00.76 00.87 01.02 3.337 
100 Pod weight 25.3244** 3.619  66.52 73.66 88.69 2.582 
100 Seed Weight 98.740*** 20.290  34.33 45.868 64.666 9.902 
Pod Yield 70.021*** 58.821  01.43 02.62 03.92 22.144 
Significance codes: ‘***’ 0.001; ‘**’ 0.01; ‘*’ 0.05, ‘NS’ Non-significant, Num.Lat. Bran. = number of lateral branc
hes, CV = coefficient of variation, d.f = degree of freedom 
63 
 
University of Ghana http://ugspace.ug.edu.gh
4.3.1.2 Principal component analysis 
Principal component analysis (PCA) underscored traits contributing mostly to the divergence in 
the germplasm. For the quantitative traits, the fist principal component (PC1) accounted for 32.7% 
and the second principal component (PC2) for 15.3% of the variance (Figure 4.2). The PC1 was 
strongly correlated with plant height, LLS at 90 days (LLS.90D), number of lateral branches 
(Lat.Brach) and days to maturity (Days_Mat). Furthermore, based on the loadings, PC1 is 
primarily a measure of contrast between plant height and days to maturity. The PC2 can be viewed 
as a measure of variance in hundred seed weight (Hun.SeedWeight), pod width (Pod.Width), pod 
length (Pod.length), seed length (Seed_Length) and days to flowering (Days_Flow). Both PCs 
accounted for 48% of the total variance. The highest loadings in the third PC were obtained with 
number of pods per plant and yield while PC5 strongly correlated with Days to emergence and pod 
volume (Table 4.2). Also, the first five PCs accounted for 76% of the total variation. 
Origin of lines 
Figure 4.2 A principal component analysis 
showing similarity relationships among 92 
accessions based on quantitative traits. 
Days_Em = days to emergence, Days_Mat = days 
to maturity, Days_Flow = days to flowering, 
Width.Plant = plant width, Plant_H = Plant height, 
Pod.Vol = pod volume, Hun_PW = hundred pods 
weight, Hun.SeedWeight = hundred seed weight, 
Pod.plant = number of pods per plant, SeedP = seed 
per pod, LLS.60D = LLS at 60 days, LLS.80D = 
LLS at 80 days. 
 
  
64 
 
University of Ghana http://ugspace.ug.edu.gh
Table 4.2 Correlation between variables and factors (Eigenvectors) 
 Principals components and variance explained 
Traits PC1 PC2 PC3 PC4 PC5 
(33.68) * (15.7) (10.6) (08.9) (07.10) 
Hundred pods weight 0.124 -0.029 -0.20 0.634 -0.032 
Hundred seeds weight -0.086 0.465 -0.021 0.006 -0.189 
Letteral branches -0.290 -0.145 0.218 0.100 0.059 
Late leaf spot  0.329** 0.067 -0.103 -0.078 -0.056 
Pod volume -0.162 0.174 -0.257 0.121 0.413 
Plant width 0.182 -0.119 0.020 -0.020 0.432 
Number of pods per plant 0.094 0.083 0.619 0.145 0.141 
Pod width -0.205 0.395 -0.065 0.024 -0.005 
Pod length -0.166 0.360 -0.141 0.025 0.307 
Yield per plants 0.127 0.073 0.572 0.259 0.107 
Seed length -0.272 0.294 0.026 0.051 -0.092 
Seed width -0.166 0.308 0.230 0.055 -0.212 
Leaf length 0.263 0.284 -0.030 -0.092 -0.136 
Leaf width 0.299 0.257 0.065 -0.090 0.029 
Days to emergence -0.109 0.035 -0.030 0.052 0.533 
Days to flowering -0.209 -0.220 0.035 -0.107 0.082 
Days to maturity -0.309 -0.138 0.061 0.003 0.021 
Shelling % 0.123 -0.046 -0.175 0.649 -0.101 
*Account of PCs for the variability are in bracket. **Characters that contributed most to the variation 
(>0.3) of the particular component are in bold.  
For the qualitative traits, the first two PCs accounted for nearly 40.4% of the variation (Figure 4.3). 
The first PCA accounted for 28.1% of total variance whereby, presence of flower on main axis 
(FOMA), branching pattern (Branching_P), and growth habit (GrowthH) were the variables that 
contributed most to the divergence. The second component exhibited pod reticulation (Pod_R) and 
leaf shape (Leaf_S) as most important variables contributing to the divergence.  
65 
 
University of Ghana http://ugspace.ug.edu.gh
Origin of lines 
 
Figure 4.3 A principal component analysis showing similarity relationships among 92 accessions based on 
qualitative traits. GrowthH = growth habit, Num_B = number of branching, Branching_P = branching pattern, 
MainS = main stem presence, FOMA = flower on the main axis, Stem_H = stem hairs presence, Stem_P = stem 
pigmentation, Leaf_S = leaf shape, Leaf_T = leaf tip, PoC = pod constriction, PoR = pod reticulation, PoB = pod 
beak, SeedC = seed color. 
 
Overall, the most common qualitative characteristics (in terms of richness) were bunch growth 
habit with main stem somewhat apparent, sub-glabrous stem surface, sequential branching pattern, 
wide elliptic leaflet shape, acute leaflet tip, flower on the main axis, slight pod reticulation, 
moderate pod constriction, slight pod beak and red seed colour type.  
4.3.1.3 Correlations among quantitative traits  
The test for strength and significance of the correlation among quantitative traits revealed that 
there was a significant positive correlation between pod yield and number of pods per plant (r = 
66 
 
University of Ghana http://ugspace.ug.edu.gh
0.95, p = 0.01) and between hundred pod weight and shelling percentage (r = 0.86, p = 0.01). 
However, the correlation of yield with hundred pod weight and hundred seed weight were not 
significant (Table 4.3). Late leaf spot correlated negatively with days to maturity (r = -0.64, p = 
0.01), and positively with plant height (r = 0.55, p = 0.01). Also, a significant negative correlation 
was observed between days to maturity and plant height (r = -0.59, p = 0.01). 
 
67 
 
University of Ghana http://ugspace.ug.edu.gh
Table 4.3 Correlation analysis for quantitative traits 
 HPW HSW Heit LLS PV NPP PW PL Yi SL SW DF DM 
HPW              
HSW -0.05             
Heit 0.10 -0.11            
LLS 0.15** -0.03 0.55**           
PV -0.01 0.14* -0.19** -0.29*          
NPP -0.05 0.01 0.26** -0.02 -0.06**         
PW -0.14* 0.53** -0.30** -0.31** 0.48** -0.22*        
PL -0.05 0.41** -0.15* -0.26** 0.57** -0.06 0.54**       
Yi 0.11 0.03 0.30** 0.03 -0.19* 0.95** -0.13* -0.08      
SL -0.14* 0.45** -0.43** -0.48** 0.31** -0.06 0.60** 0.51** -0.10     
SW -0.14* 0.42** -0.23** -0.29** 0.14* 0.16* 0.45** 0.25** 0.14* 0.60**    
DF -0.17** -0.06 -0.42** -0.38** 0.10 -0.08 -0.03 0.04 -0.14* -0.14* 0.01   
DM -0.27** 0.02 -0.59** -0.64** 0.61** -0.08 0.21** 0.13* -0.12* -0.12* 0.20** 0.48**  
Sh% 0.86** -0.01 0.10 0.18** -0.02 -0.01 -0.13* -0.13* 0.18** -0.17** -0.15* -0.23** -0.17** 
Significance codes: ‘**’ 0.01; ‘*’ 0.05; DM = days to maturity, DF = days to flowering, DE = days to emergence, SL = seed length, PL = pod length, PW = pod 
width, PV = pod volume, Heit = plant height, NPP = Number of pods per plant, HPW = hundred pod weight, HSW = hundred seed weight, Sh% = shelling 
percentage, LLS = late leaf spot at 80 days. 
68 
 
University of Ghana http://ugspace.ug.edu.gh
4.3.1.4 Path analysis for yield-related traits 
Number of pods per plant (NPod.Pl) seemed to have the strongest direct effect on yield (p = 0.05) 
with a path coefficient (Pc) of 0.85.  Though hundred pod weight and hundred seed weight 
contributed to yield, the direct effect was very low and not significant (Figure 4.4). Also, 
comparison of multiple models revealed no indirect effect between pod size and seed size and 
between hundred seed weight and hundred pod weight. Pod length contributed significantly to the 
hundred pod weight. Negative indirect effect was observed between pod size (pod length and 
width) and yield as increasing pod size caused reduction in the number of pods per plant which in 
turn reduced yield. A path coefficient of 0.26 was found for extraneous variables showing that the 
traits considered in this analysis are the ones contributing mostly to the yield. The path coefficient 
for extraneous variable was high for hundred pod weight (Pc = 0.70) and hundred seed weight (Pc 
= 0.64). 
 
Figure 4.4 Plot of the direct and indirect effect of pod and seed traits on yield. Straight arrow represents direct effect. 
Double arrow represents correlation. NPP = number of pods per plants, HPW = hundred pod weight, HSW = 
hundred seed weight, PL = Pod length, PW = pod width, SL = seed length, SW = seed width and Ext = Extraneous 
variables 
69 
 
University of Ghana http://ugspace.ug.edu.gh
4.3.1.5 Population structure and relationship among genotypes 
In the quantitative traits scatter plots, the second PC separated Senegal cultivars from Togolese 
cultivars while PC2 mostly separated local accessions into groups (Figure 4.2). In the qualitative 
data plot, no clear separation was achieved based on the origin (Figure 4.3). However, the first PC 
seemed to identify a group of 10 local accessions as outliers.  
The average Silhouette width of the PAM clustering showed a weak structure (silhouette width < 
0.50) in the studied germplasm. Despite the weak silhouette mean, two clusters (K = 2) appeared 
to be the optimal number of clusters (Appendix 2; Figure 4.5). However, in the search for more 
dissimilarity between genotypes, following the average Silhouette width’s rank of clusters, three 
clusters (K = 3) were also retained. 
a b 
 
Figure 4.5 Optimal number of clusters, a-K = 2 and b-K = 3 with a probability of membership of genotypes 
 
70 
 
University of Ghana http://ugspace.ug.edu.gh
In regards to the optimal two clusters (K=2), a large number of the genotypes (82 lines), including 
lines from Senegal, were grouped in cluster I (Figure 4.6) and the other local accessions were 
grouped in cluster II. The probability of membership of genotypes was 0.46 and 0.52 for cluster I 
and II, respectively when k = 2 (Figure 4.6a). Owing to the low Silhouette width, these probabilities 
can be considered reasonable. Trait means of the optimal two clusters are summarized in Table 
4.4.  A comparison of clusters revealed that cluster II (C2) was mainly grouping late maturing 
genotypes with a low LLS incidence, low hundred pod weight and yield. Conversely, cluster I (C1) 
seemed to be made of early maturing, susceptible to LLS genotypes and exhibited a slightly higher 
yield. 
 
Figure 4.6 Dendrogram showing optimal two clusters. Branches are outlined with respect to the clusters: Cluster I= 
black, cluster II =blue 
  
71 
 
University of Ghana http://ugspace.ug.edu.gh
Table 4.4 Mean of yield-related traits for optimal two clusters  
 Genotypes origin  Traits 
Togo Senegal  LLS Days to Hundred Hundred Yield 
Cluster 
maturity Pod weight seed weight 
C1 72 9  4.49 91.76 73.79 46.07 2.65 
(1.8-6.1)* (80-120) (66.52-81.04) (34.33-66.66) (1.51-3.90) 
C2 10 0  2.60 109.16 72.79 43.97 2.33 
(2-3.5) (90-110) (70.9-75.07) (38.56-49.53) (1.43-3.23) 
*Ranges  
When K=3, the majority of the studied genotypes, including lines from Senegal, were grouped in 
cluster I (C1) as in K = 2 (Figure 4.7). The particularity in K= 3 was cluster 3 (C3), which was 
comprised of 3 local accessions that exhibited quite high probability of membership (0.84). In 
relation to the probability of membership, genotypes in CI, with a score of 0.18 carried the highest 
within cluster diversity (Figure 4.5). 
 
Figure 4.7 Dendrogram showing three clusters. Branches are outlined with respect to the clusters: Cluster I=black, 
cluster II=blue, cluster III=green 
72 
 
University of Ghana http://ugspace.ug.edu.gh
Quantitative and qualitative features of the three clusters are shown, respectively, in Table 4.5 and 
Table 4.5. Cluster III (C3) contained local early maturing genotypes, high LLS incidence, large 
pod volume and a higher hundred pod weight, while in C2 genotypes were essentially late maturing 
with a low incidence of LLS (Table 4.5). Indeed, C2, in marked contrast to cluster I, displayed the 
longest maturity period, the lowest hundred pod weight and the lowest incidence of LLS. The 
majority of the studied genotypes, including lines from Senegal, were grouped in cluster I. 
However, when the number of clusters is brought to K = 9, lines from Senegal were separated into 
a cluster that showed moderate resistance to LLS (result not shown). Overall, the clusters that 
displayed a high number of pods per plant seemed to yield more than the others. Also, clusters that 
tended to be less susceptible to LLS were late maturing. 
  
73 
 
University of Ghana http://ugspace.ug.edu.gh
Table 4.5 Ranges of traits in three cluster grouping of the 92 groundnut accessions 
 Cluster 
Trait C1 C2 C3 
Days Emergence 6.38 6.2 6.66 
Days Flowering 26.87 29.93 24.33 
Days Maturity 92.02 109.16 85 
Plant Height (cm) 33.76 21.46 39.22 
Plant Width (cm) 58.51 55.55 65.54 
Number of Lateral Branches 8.89 15.04 7.07 
Leaflet Length (cm) 5.61 4.42 6.83 
Leaflet Width (cm) 3.13 2.49 3.22 
LLS 90 Days 4.45 2.60 5.16 
Pods / plant 55.69 49.63 43.73 
Pod Volume (cm3) 3.04 3.98 5.05 
Pod Length (cm) 2.39 2.43 2.86 
Pod Width (cm) 1.17 1.20 1.31 
Shelling % 73.80 73.14 76.94 
Number of Seeds per pod 1.7 1.4 3 
Seed Length (cm) 1.27 1.41 1.29 
Seed Width (cm) 0.87 0.92 0.81 
100 Pod weight 73.70 72.79 76.16 
100 Seed Weight 46.01 43.97 45.42 
Yield 2.69 2.32 2.14 
 
74 
 
University of Ghana http://ugspace.ug.edu.gh
Table 4.6 Qualitative features of k = 3 clusters 
Trait Descriptors C1 C2 C3 Simpson Index (H) 
      
Growth Habit Bunch 34.37% 10% 66.66% 0.59 
Erect  65.63% - 33.34% 
Spreading - 90% - 
      
Main stem Non-apparent 08.26% - 25.00% 0.43 
Somewhat apparent 79.74% 33.33% 62.5% 
Apparent 12% 66.67% 12.5% 
      
Stem surface Sub-glabrous 98.30% 100% 100% 0.04 
Moderately Hairy 1.70% - - 
      
Branching Pattern Sequential 100% - 100% 0.22 
Alternate - 100% - 
      
Number of Branches Primary (n+1) 38.00% - 33.33% 0.57 
Secondary (n+2) 62.00% 10% 66.67% 
Tertiary - 90% - 
      
Stem Pigmentation Green 42.16% 60% - 0.60 
Purple - - 100% 
Mixed 57.84% 40% - 
      
Leaflet Shape Narrow Elliptic 18.66% - 100% 0.33 
Oblong Elliptic 16.94% 70% - 
Wide Elliptic 64.40% 10% - 
Obovate - 20% - 
      
Leaflet Tip Acute 93.22% 90% 100% 0.14 
Obtuse 6.78% 10% - 
      
Flower on Main Axis Presence 100% - 100% 0.19 
Absence - 100% - 
 
      
Pod reticulation None 1.69% - 66.66% 0.51 
Slight 66.10% 90% 33.34% 
Moderate 30.50% 10% - 
Prominent 1.69% - - 
      
Pod Constriction None - 30% - 0.55 
Slight 40.68% 60% 100% 
Moderate 59.32% 10% - 
Deep - - - 
 
      
Pod Beak Slight 72.30% 70% 33.33% 0.42 
Moderate 26.50% 30% 66.67% 
Prominent 1.20% - - 
 
      
Primary Seed Colour Pink 1.69% - - 0.75 
Light Red 50.84% - - 
Red 16.94% - - 
Dark Red 25.42% 60% - 
Dark purple 3.38% - 100% 
Variegated 1.69% 40% - 
75 
 
University of Ghana http://ugspace.ug.edu.gh
Most genotypes within the clusters were very similar on qualitative traits bases (Table 4.6). 
Genotypes in C1 were erect types (65.68%), sub-glabrous stem surface (98.30%), sequential 
branching pattern (100%) with flower on the main axis (Fastigiata subspecies). In contrast, C2 
genotypes were spreading types with alternate branching and absence of flowers on the main axis 
(hypogaea subspecies).  The highest uniformity of cluster members was observed in C3. Indeed, a 
total uniformity was observed on growth habit, stem surface, branching pattern, number of 
branches, flowers on the main axis and pod reticulation. In relation to the calculated Simpson 
index, the most diverse traits were primary seed colour (0.75), stem pigmentation (0.60), Growth 
habit (0.59), and number of branches (0.57) (Table 4.5). 
4.3.1.5 Quantitative versus qualitative traits-based Clustering 
To check the collinearity of quantitative traits and qualitative traits in grouping genotypes, two 
trees were generated from both separate data sets. In general, while lines from Senegal clustered 
together based on the quantitative data (Figure 4.8a), the reverse was observed on the qualitative 
trait dendrogram (Figure 4.8b). Also, some genotypes, following the example of accessions 05AH, 
03AH, 07AH, 13AH, were quite similar, when clustered using qualitative traits but distinct on 
quantitative trait bases. Altogether, the low correlation coefficient (r = 0.33) obtained between 
trees based on the quantitative and qualitative traits confirm the spatial divergence observed on the 
dendrograms. 
The evaluation of the performance of hierarchical clustering methods with respect to germplasm 
collections revealed that qualitative traits tree had a higher cophenetic coefficient (0.88) than the 
quantitative traits tree (0.79). 
76 
 
University of Ghana http://ugspace.ug.edu.gh
06AH 54AH 
HG20 HG03 
03AH 61AH 
HG15 HG13 
16AH 46AH 
41AH 40AH 
35AH 18AH 
53AH 37AH 
12AH 33AH 
11AH HG12 
24AH 28AH 
59AH 47AH 
08AH   NAV 
07AH HG02 
62AH ICIAR 
58AH 50AH 
67AH 22AH 
27AH HG20 
38AH HG08 
57AH HG06 
13AH 42AH 
64AH HG06 
31AH 68AH 
45AH HG01 
34AH 38AH 
43AH CS108 
23AH CS056 
HG19 CS060 
55AH HG07 
52AH INC 
05AH 06AH 
HG10 64AH 
15AH 62AH 
22AH CS079 
39AH 24AH 
HG07 HG10 
56AH 31AH 
25AH 15AH 
42AH 01AH 
32AH 35AH 
HG06 05AH 
33AH 03AH 
10AH 07AH 
01AH 13AH 
09AH 67AH 
INC 41AH 
36AH 69AH 
04AH 11AH 
02AH HG16 
HG08 57AH 
HG01 53AH 
ICIAR 59AH 
69AH 55AH 
HG05 52AH 
HG03 16AH 
CS098 58AH 
CS078 32AH 
CS036 27AH 
CS044 12AH 
CS022 HG19 
CS108 34AH 
CS056 04AH 
68AH 36AH 
26AH 08AH 
CS060 21AH 
NAV 10AH 
CS079 CS022 
a HG16 17AH 
37AH 39AH 
28AH 09AH 
61AH HG04 b 
40AH HG15 
46AH 02AH 
51AH CS098 
54AH CS044 
49AH 56AH 
19AH CS078 
HG18 CS036 
17AH 43AH 
47AH 45AH 
HG14 26AH 
66AH 25AH 
50AH 48AH 
HG09 44AH 
HG04 30AH 
21AH 63AH 
14AH 66AH 
HG13 HG04 
18AH 49AH 
HG12 14AH 
48AH 19AH 
44AH 50AH  
63AH 
Figure 4.8 Dendrogram based on quantitative traits (a) and qualitative traits (b) in diversity assessment of 92 
groundnut accessions. Black=lines from Togo, Blue= lines from Senegal 
77 
 
University of Ghana http://ugspace.ug.edu.gh
4.3.2 Molecular diversity 
4.3.2.1 Allelic diversity  
Genotyping of the germplasm collections from Togo, breeding lines from Senegal and 
ICRISAT has allowed the identification of nearly 7072 diverse SNPs. However, only 990 SNPs 
well distributed over the genome with polymorphism information content (PIC) higher than 
0.1 were retained for the diversity analysis (Annex 3). The PIC values ranged from 0.1 to 0.5 
with a mean of 0.24. The distribution of the PIC and of minor allele frequencies (MAF) is 
shown in Figure 4.9. 
a b 
 
Figure 4.9 Distribution of SNPs based on PIC (a) and MAF (b) 
Overall, the percentage of heterozygous genotypes varied from 0 to 50% for all loci.  However, 
less than 10% of the genotypes were heterozygous for more than 30% of the investigated loci 
(Figure 4.10a). The most heterozygous genotype, HG09 carried only 30% homozygous loci. 
The most homozygous genotypes, 39AH, 38AH, 34AH, and 59AH carried less than 5% 
heterozygous loci. Varying proportions of homozygosity were observed in the genotypes from 
Togo unlike in genotypes from Senegal where less than 7% of heterozygous loci was detected 
in all genotypes. The genetic diversity was well distributed across the majority of the loci 
78 
 
University of Ghana http://ugspace.ug.edu.gh
(Figure 4.10C). The test for Hardy Weinberg Equilibrium (HWE) shows that there was a clear 
significant departure from HWE in the population. 
a 
Proportion of genotypes  
b 
c 
 
Figure 4.10 a-Structure of heterogeneity detected in 92 groundnut accessions; b-Distribution of fixation index 
(Fst) and c-Frequency distribution of total genetic diversity (Ht) 
4.3.2.3 Geographical analysis of diversity 
The analysis of molecular variance to partition the total genetic variation within and between 
geographical origin revealed that only 1.1% of the total variation accounted for between 
geographical origin with nearly all the observed diversity being among genotypes. Moreover, 
the isolation by distance analysis of the relationship between genetic dissimilarity (Kosman & 
Leonard, 2005) and geographical diversity showed that the difference between populations was 
more based on their genetic differences than based on their geographical origin (Figure 4.11). 
Also, there was no correlation between genetic diversity and geographical diversity. 
79 
 
University of Ghana http://ugspace.ug.edu.gh
 
Figure 4.11 Multilocus genetic dissimilarity versus geographic distance. The line represents the running 
average. 
 
4.3.2.4 Principal component analysis of populations 
Discriminant analysis of principal components (DAPC) was used for the visualization of the 
emergent population structure and group assignment in the groundnut accessions. DAPC 
revealed that the first two axes, explaining more than 10% of total variation each, were 
adequate for discriminating genotypes. The first two principal components (PC) explained 67% 
of the total variance. The emergent structure is depicted in a two-dimensional scatter plot 
involving investigated groundnut accessions (Figure 4.12). The main result from the scatter 
plot is the extremely low genetic variability of the introduced lines from Senegal compared to 
the cultivated accessions in Togo. Indeed, lines from Senegal formed a compact group but were 
to a large extent overlaid by a group of genotypes from Togo. Genotypes from Togo were the 
most dispersed as they covered the whole range of genetic diversity observed in this study with 
some tendency to form specific subclusters. Though ICG 7878 from ICRISAT was somewhat 
different from the other genotypes, it had some similarities with some lines from Togo. 
80 
 
University of Ghana http://ugspace.ug.edu.gh
 
Figure 4.12 A principal component analysis biplot diagram showing similarity relationships among 
populations based on SNP markers 
The absence of evidence for population structure was confirmed by the low Fst value for the 
relationship between lines from Senegal and lines from Togo. However, the relatively high 
value of Fst between a line from ICRISAT and lines from Senegal confirmed the spatial 
distance observed between them (Table 4.7). 
Table 4.7 Pairwise Fst and Nei’s genetic distance value for populations 
 Togo Senegal ICRISAT 
Togo - 0.034 0.108 
Senegal 0.181 - 0.148 
ICRISAT 0.304 0.471 - 
*The top diagonal is Nei’s Gst; **The bottom diagonal is pairwise Fst 
  
81 
 
University of Ghana http://ugspace.ug.edu.gh
4.3.2.5 Cluster analysis 
PAM Silhouette revealed that two clusters (K = 2) were optimal in both the analysis of marker 
and the co-analysis of marker-phenotypic data with respectively an average silhouette of 0.7 
and 0.69 (Figure 4.13 and 4.15). However, the three-cluster solution (K = 3) was described, to 
understand the characteristic of the three groups, in the co-analysis of phenotypic and 
molecular data because it was the next in rank of the silhouette width. In Figure 4.13, Cluster 
1 grouped majority of lines from Togo and lines from Senegal while cluster 2 grouped 19 lines 
from Togo that clustered with the line from ICRISAT. A probability of membership of 0.71 
and 0.67 was observed respectively for cluster 1 and cluster 2 with quite a high average 
silhouette width (0.7) denoting a strong structure in the collected germplasm. 
 
a 
b 
 
 
Figure 4. 13 (a) Molecular data dendrogram showing the optimal two clusters: Cluster 1 = Black; Cluster 2 = 
Blue; (b) Silhouette width of the optimal clusters. 
82 
 
University of Ghana http://ugspace.ug.edu.gh
 
  
Figure 4.14 Cluster analysis: two clusters solution (left); three cluster solution (right). 
 
Co-analysis of the marker and phenotypic data revealed two optimal clusters that were not 
different from the clusters obtained in molecular data analysis. However, the probability of 
membership of the genotypes to the clusters, 0.70 and 0.66 respectively for cluster 1 (CI) and 
cluster 2 (CII), was slightly lower than the probability in the molecular analysis. Further 
structure analysis of the germplasm revealed that cluster 1 in the two-cluster solution (K=2) 
could be split into two clusters (Figure 4.14). However, the new cluster 1 (CI), made up of lines 
from Togo and lines from Senegal, exhibited a low probability of membership (0.23). The two 
other clusters CII and CIII, made up of a group of local genotypes and the ICRISAT line, 
exhibited a probability of membership of 0.65 (Figure 15Bb). 
83 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
A 
a b 
B 
 
 
Figure 4.15 (A)Dendrogram showing the three clusters based on the marker and phenotypic data: CI=Green; 
CII=Black; CIII= Blue; B = Optimal number of clusters: a = Silhouette width of k = 2, b = Silhouette 
width of k = 3 clusters 
 
A correlation analysis of the marker and phenotypic dendrograms revealed a positive 
correlation (0.46). Analysis of molecular features of the clusters revealed that the distribution 
of private alleles (PA), minor allele frequency (MAF) and richness was not the same across 
clusters (Figure 4.16). Indeed, CII carried the lowest number of private alleles and richness 
while CIII, with the smallest sample size, exhibited the highest number of private alleles (PA), 
minor allele frequencies (MAF) and richness. Altogether, the three clusters were not fixed for 
alternate alleles. There was a significant correlation between allelic richness, number of private 
alleles and total alleles (p = 0.05). 
84 
 
University of Ghana http://ugspace.ug.edu.gh
● cluster3 cluster1 ● 
PA= 160 PA= 36 
MAF=0.45 ● MAF=0.35 Cluster2 
Rich= 1.876 Rich= 1.774 PA= 4 
Ia = 0,143 Ia  = 0,100 MAF=0.35 
  Rich= 1.620 
 Ia  = 0,136  
 
20 25 30 35  
Location sample size 
Figure 4.16 Number of alleles vs location sample size. PA=Number of private alleles; Rich= Mean Allelic 
richness; Ia= Index of association 
Pairwise comparison of genetic distance between clusters using Nei’s Gst and Fst was done 
and a high global Fst (0.54) confirmed the difference observed between clusters. The Fst 
estimates for clusters ranged from 0.399 to 0.67 (Table 4.8) and cluster 3 showed a significant 
level of genetic differentiation (p<0.05).  
Table 4.8 Pairwise Fst value and Nei’s genetic distance for clusters 
 CI CII CIII 
CI  - 0.0029* 0.0309 
CII 0.399** - 0.0375 
CIII 0.563 0.67 - 
*The top diagonal is Nei’s Gst; **The bottom diagonal is pairwise Fst 
Analysis of phenotypic characteristics of the three clusters revealed that there were differences 
in agronomic features. Comparison of agronomic characteristics of clusters revealed that 
genotypes in CI and CII were early maturing and susceptible to LLS (Figure 4.17). In contrast, 
CIII was made up of genotypes that are medium and late maturing with partial resistance to 
LLS disease. Also, CII showed a higher yield than the other clusters. 
85 
 
University of Ghana http://ugspace.ug.edu.gh
 
Figure 4.17 Agronomic characteristics of the three clusters in the co-analysis of phenotypic and molecular data. 
 
CI and CII differed mainly on the LLS score, days to flowering and pod volume. Differences 
were also observed between the two clusters based on other phenotypic characteristics such as 
plant height, plant width, pod volume, pod length, seed length, growth habit, number of lateral 
branches, pod reticulation and constriction (result not shown). A quite high cophenetic 
correlation coefficient (0.95) was obtained for the clustering of the marker and phenotypic data. 
4.4 Discussion  
4.4.1 Phenotypic diversity among collected accessions 
The analysis of variance suggests an adequate relative variability for some quantitative traits. 
Indeed, traits such as number of lateral branches, number of pods per plant, pod yield and LLS 
displayed a high coefficient of variation (CV) showing that improvement could be achieved 
86 
 
University of Ghana http://ugspace.ug.edu.gh
through selection for higher productivity. However, the low CV for other traits may suggest a 
directional selection which reduced cover-wide phenotypic variance. Previous diversity studies 
using phenotypic traits reported a low variation in days to emergence, leaflet width, pod width, 
and seed length and width (Upadhyaya et al., 2003; Kushwah, 2011). This uniformity 
underscores the necessity to increase diversity through the incorporation of new alleles into the 
germplasm for these traits’ improvement.  
Correlation-based principal component analysis highlighted that traits of importance such as 
hundred seed weight and resistance to LLS contributed significantly to the diversity. Indeed, 
the higher the trait diversity in a germplasm, the higher the probability of the presence of genotypes 
with particularly important traits. Qualitative trait wise, most traits that showed a high Simpson 
index, correlated with the first two PCs. However, traits of value for farmers such as pod 
constriction, seed length, seed width and primary seed colour (Janila et al., 2016b) which were 
not retained by the first two PCs, exhibited a high Simpson index. Previous diversity studies 
indicated the relatively low importance of these traits as groundnut descriptors (Upadhyaya et 
al., 2003). The high Simpson index for qualitative traits of interest indicates that several 
different trait features with similar abundance are available in the germplasm collection and 
could be used in breeding to respond to farmers’ preferences.  
Nonetheless, the seemingly contrasting result between PCA and Simpson index could be 
explained by the fact that the two approaches do not measure diversity at the same level. While 
Simpson index measures diversity in an individual trait, PC analysis measures contribution of 
an individual trait to the overall diversity. Also, the high loadings observed in the PCs for high 
Simpson index traits suggest a correlation between Simpson index value and trait contribution 
to overall diversity in principal component analysis.  
87 
 
University of Ghana http://ugspace.ug.edu.gh
Population structure analysis of phenotypic traits delineated germplasm accessions into distinct 
groups. However, the silhouette width of PAM cluster showed a low population structure in 
the germplasm collection implying that there were more similarities than differences between 
the collected genotypes. This result was unexpected, because, as there is no modern breeding 
programme in Togo and lines were collected from different areas of the country, one might 
expect more diversity in the germplasm. Artificial selection by farmers over decades to adapt 
to similar environmental constraints across regions may have narrowed the phenotypic 
diversity in cultivated groundnut in Togo. Also, the low number of introduced varieties coupled 
with the free exchange of seeds among farmers could explain the similarity between genotypes.  
Nonetheless, useful differences were conspicuous across clusters within the germplasm 
collections. Overall, the clusters that displayed a high number of pods per plant seemed to yield 
more than those that were characterized by a low number of pods per plant. Also, pod yield 
was strongly correlated with the number of pods per plant suggesting that groundnut pod yield 
is more influenced by the number of pods per plant than the pod weight and seed weight in the 
germplasm. Path analysis confirmed this assumption. This result agreed with the findings of 
Kushwah (2011) and Haghpanah et al. (2018) who suggested the use of the number of pods 
per plant as a selection index for yield improvement. This is interesting for indirect selection, 
as the number of pods per plant is a visible descriptor that could easily be used for selecting 
plants in the field, for yield improvement, without the need for costly equipment. Large pod 
size is another preferred characteristic at farmers’ level. In this study, one cluster exhibited a 
large pod size. However, because of their low number of pods per plant, they were low yielding. 
In addition, pod size did not correlate with seed size and hundred seed weight. Only genotype 
HG11, a long duration variety, combined a high number of pods per plant and a large pod size. 
With the erratic rainfall, a consequence of the changing climate, such genotypes should be 
improved through incorporation of earliness and other beneficial traits.  
88 
 
University of Ghana http://ugspace.ug.edu.gh
Variability was found among clusters in the germplasm for LLS incidence. With few 
exceptions, resistance to LLS was positively associated with late maturity in this study, as 
clusters that showed a low incidence of LLS tended to be late maturing. This tendency was 
confirmed by correlation analysis of quantitative traits. Such a finding could be a genetically 
meaningful trend, but could also be attributed to the fact that disease was scored at the same 
time on all plants, and the incidence of the disease could increase with growth on the late 
maturing plants. However, this confirms the difficulty to find landraces combining LLS 
resistance and early maturity (Janila et al., 2016b). These results imply that the development 
of early maturing LLS resistant lines will necessitate the use of breeding tools such as genetic 
and chromosome shuffling. For this purpose, most resistant and late maturing genotypes could 
be used as parents for incorporation of LLS resistance into early and susceptible varieties.  
Though the quantitative and qualitative traits showed a weak convergence in grouping 
accessions in this study, phenotypic cluster VII and VIII, which exhibited the highest 
probability of membership, grouped genotypes that were homogenous based on both trait types. 
All hypogaea subspecies genotypes (alternate branching and absence of flower on the main 
axis) were grouped in CII while the fastigiata subspecies with sequential branching and 
presence of flower on the main axis (nearly 90 % of the accessions), were scattered in other 
clusters.  
4.4.2 Molecular diversity 
Overall, SNP markers provided estimates of population genetic parameters useful for 
population structure analysis. Principal component analysis showed that the diversity observed 
in the studied germplasm was not associated with the geographical origin. The relatively low 
value of Fst in this study means that there was not much difference between the populations 
based on their geographical origin. Also, the isolation by distance relationship did not 
89 
 
University of Ghana http://ugspace.ug.edu.gh
hold for these populations, as there was no correlation among matrices of genetic (Fst) and 
geographical distances of the populations used in this study. The common alleles between lines 
from Togo and other lines could represent, at least in part, a kind of core allele for groundnut.  
Besides, the common alleles could indicate that lines from Senegal and most lines from Togo 
showed a high level of genetic exchange between them. This result is surprising as one might 
expect a much more genetic difference between breeding lines (Senegal and ICRISAT) and 
cultivated landraces from Togo. However, this could be partly explained by the fact that the 
genetic background of the parents of the breeding lines was very similar to the cultivated 
groundnut landraces in Togo. Also, a small proportion of the collected accessions from Togo 
were introductions from various sources. These lines, mostly those that clustered with lines 
from Senegal, may have a certain parentage with Senegal breeding lines. The ICG 7878 from 
ICRISAT was genetically different from the other genotypes in the scatter plot and the 
difference was even higher with lines from Senegal as evidenced by the Fst value. This result 
suggests that ICG 7878 shares few alleles in common with lines from Senegal. However, 
cluster analysis showed that ICG 7878 could be grouped with some cultivated local Togolese 
accessions. Thus, some of these lines from Togo may originally have come from ICRISAT 
confirming the idea that population structure is associated with origin of germplasm 
(Würschum et al., 2013; Dwivedi et al., 2017). 
Overall, genetic parameters indicated appreciably broad genetic base and diverse nature of the 
germplasm collection evaluated in this study. AMOVA proved that the majority of the genetic 
variation observed was due to differences between individuals. The distribution of the minor 
allele frequency (MAF) was skewed toward the low and high value of MAF with an 
intermediate group of genotypes between the two. Thus, cluster analysis distinguished clusters 
where genotypes, from different geographical origins, shared a similar pattern of variation. The 
90 
 
University of Ghana http://ugspace.ug.edu.gh
cluster of co-analysis of the phenotypic and molecular data was very similar to the clusters of 
the marker data alone.  
A comparison of the resulting clusters distinguished between three groups of genotypes with 
specific diversity accumulation index. The first group (CI) grouped lines from Senegal and 
lines from Togo which carried a medium number of alleles and of private alleles. The second 
group (CII) made up of cultivated landraces from Togo, showed the lowest number of alleles 
and private alleles and carried the lowest proportion of heterozygosity for the highest number 
of accessions. The third group (CIII) with the highest allelic richness and private alleles 
represented lines from Togo and the ICRISAT line. This variation in the number of alleles 
across clusters reflects natural variation and diversity.  
Regarding all diversity criteria, CIII showed the lowest genetic diversity compared to other 
clusters. The low-frequency of private alleles in CII suggests that this group of genotypes may 
have undergone no or very low selection pressure (Zhu et al., 2004). Identification of private 
alleles in a cluster provides specificity of that group of genotypes. Private alleles associated 
with groups of genotypes have been previously reported in groundnut (Kottapalli et al., 2011; 
Varshney et al., 2009) and found to be useful for diversity creation. Thus, genotypes carrying 
these alleles will likely be useful for introducing diversity in the current breeding programme 
in Togo. 
4.4.3 Diversity in co-analysis of phenotypic and marker traits 
The difference observed between the clusters at marker level was associated with phenotypic 
differences in agronomic traits. ICG 7878 was bred for resistance to LLS disease and it is 
interesting to observe that CIII to which it belongs exhibited the lowest disease infection. 
However, genotypes of this cluster tended to be late maturing. Lines from Senegal were 
phenotypically similar mainly in quantitative traits analysis with moderate resistance to LLS. 
As these lines were selected among a high number of genotypes introduced into the country, 
91 
 
University of Ghana http://ugspace.ug.edu.gh
the directional selection based on quantitative traits have probably contributed to the uniformity 
among them. Though qualitative traits showed a certain dissimilarity between them, molecular 
analysis confirmed the similarity of genotypes from Senegal.  
Other agronomic traits of interest such as flowering date, number of pods per plant, pod 
volume, hundred pod and seeds weight were variously associated with the clusters 
distinguished in this study. In addition, general separation of genotypes into subspecies was 
observed in this study as CIII grouped subsp. hypogaea while CI and CII groups were mostly 
from subsp. fastigiata confirming that population structure in groundnut is associated with 
botanical variety types (Wang et al., 2011). However, some genotypes, as evidenced by the 
scatter plot, may contain mixtures of genetic components from different subspecies as a result 
of breeding effort in other countries. 
Low level of detectable genetic polymorphism has been considered as a constraint in 
undertaking molecular breeding in groundnut (Varshney et al, 2009). However, the present 
study reports a high level of genetic dissimilarity in the studied germplasm. For instance, the 
highest dissimilarity was observed between HG06 and HG09, two accessions from Togo. 
HG06 is an early maturing genotype highly susceptible to ELS and LLS, which is widely 
cultivated in the northern regions of Togo. HG09 is a late maturing genotype that is resistant 
to ELS and LLS diseases. These genotypes could be used in the development of breeding 
populations for resistance to foliar diseases. Moreover, such populations can be used for genetic 
mapping and QTL analysis.  
It is worth mentioning that the high diversity observed in the collected germplasm may be 
explained by the lack of breeding programme in Togo which with trait-oriented breeding tends 
to decrease diversity over time (Dwivedi et al., 2017). Overall, SNPs generated by DArTseq 
technology have been efficient in the identification of SNPs that are well distributed throughout 
92 
 
University of Ghana http://ugspace.ug.edu.gh
the groundnut genome and the population structure provided means for exploiting 
heterogeneity potential among genotypes in the germplasm collection. Therefore, clusters 
delineated in this study could guide the development of a core collection useful for the new 
groundnut breeding programme in Togo. 
Overall, the cophenetic coefficient revealed that the clustering method was efficient in this 
study. Also, the clustering techniques proved to be more effective for the marker data and 
qualitative data than the quantitative data as evidenced by the high cophenetic coefficient 
observed in this study for qualitative traits and marker data. Apparently, this is because 
qualitative traits and marker data divide individuals into distinct types with little or no 
intermediate value in contrast to quantitative traits where individual do not fall into sharply 
demarcated groups but form continuous series. In this sense, categorical traits seem more 
effective at capturing variability in trait values than continuous traits. However, qualitative data 
are less used in cluster analysis. The findings of this study indicate that qualitative data could 
be very useful in complementing quantitative data, for grouping genotypes in cluster analysis, 
if the appropriate clustering method is applied. The high cophenetic coefficient, in this study, 
confirms that the choice of UPGMA with Gower metric, as clustering method, was appropriate. 
4.5 Conclusion 
 Phenotypic diversity was low compared to genotypic diversity in this study. Phenotypic traits 
such LLS resistance, days to maturity and hundred seed weight exhibited a high variability. A 
high proportion of SNPs harbouring high genetic variability were detected in this study. The 
allele frequencies distribution found in this study suggests that there was a quite high genetic 
diversity. Altogether, an adequate and useful diversity for breeding purposes was observed in 
the 92 groundnut accessions studied. Diversity observed at the molecular level was associated 
with phenotypic variation, mainly in quantitative agronomic traits suggesting that the generated 
93 
 
University of Ghana http://ugspace.ug.edu.gh
genotypic data may allow capturing the effect of most of the QTLs associated with productive 
traits. The delineation, in this study, of accessions harbouring desired traits/alleles into groups 
is valuable information in the starting of groundnut breeding programme and for improving the 
management and utilization of the collected germplasm in the development of new varieties.  
 
  
94 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER FIVE 
5. ANALYSIS OF LINKAGE DISEQUILIBRIUM AND MARKER-TRAIT 
ASSOCIATION STUDY FOR LLS AND YIELD-RELATED TRAITS 
5.1 Introduction 
Linkage disequilibrium (LD) is one of the statistical methods used to identify loci that are 
associated when conducting population molecular parameters. Thus, knowledge of LD pattern 
along groundnut genome is crucial in diversity study for multiple reasons. Indeed, it is useful 
for the determination of the distribution of crossing-over and regions of genome subjected to 
different selection pressures (Porto-Neto et al., 2014). Also, the success of implementation of 
a large number of breeding methodologies used in genetics nowadays such as genome-wide 
association studies (GWAS), marker-assisted selection (MAS), quantitative trait loci (QTL) 
mapping is dependent on the LD level between markers in the population (Bejarano et al., 
2018). Thus, LD analysis is a step that cannot be ignored as far as GWAS and QTL mapping 
strategies are concerned.  
GWAS is a genome-trait association study in which the whole genome of different individuals 
is assessed to see if there is any significant association between any variant and agronomic 
traits. In contrast to other methods, GWAS as a non-candidate driven approach investigates the 
entire genome for a potential association between SNPs/genomic regions and traits (Pearson & 
Manolio, 2008). It has the advantage of using a broader genetic variation and exploiting 
recombination events from a large numbers of meiosis accumulated throughout the germplasm 
development history (Abdurakhmonov & Abdukarimov, 2008).  
Single nucleotide polymorphisms (SNPs), compared to other markers, have the advantage of 
being abundant in the genome (Brumfield et al. 2003) and are amenable to high-throughput 
analysis. Thus, identification of SNPs linked to the coding regions as result of GWAS could be 
exploited in breeding methodologies such markers assisted selection or genomic selection 
95 
 
University of Ghana http://ugspace.ug.edu.gh
(Luikart et al. 2003, Bush & Moore, 2012). The objectives of this study were to determine the 
level of linkage disequilibrium in the germplasm collection and identify candidate genomic 
regions associated with LLS and yield-related traits. 
5.2 Materials and methods 
Genotypes consisting of collected accessions from Togo (n=83), lines from Senegal (n=9) and 
lines from ICRISAT (n=2), were used in this study. These genotypes are a subset of germplasm 
collection used in the diversity study. 
5.2.1 DNA Extraction and DArT Analysis 
Dried leaf samples were sent to Integrated Genotyping Service and Support (IGSS) platform 
at Biosciences for east and central Africa laboratory at the International Livestock Research 
Institute (BecA/ILRI) for DNA extraction and genotyping using Diversity Arrays Technology 
(DArT) markers as described in the diversity study. Genotyping was carried out using 
Genotyping-By-Sequencing (GBS) technology (Elshire et al., 2011). 
5.2.2 Data analysis 
5.2.2.1 Linkage disequilibrium analysis 
For LD and GWAS analysis, SNPs with unknown genomic positions were removed (Huang et 
al., 2012). Genotypic data were subjected to LD analysis and LD was calculated both at the 
whole germplasm level and within the germplasm for the clusters using a square correlation of 
allele frequencies (r2). LD analysis was done using both web application KDCompute, and R 
statistical software version 3.3.1 (Branca et al., 2011; Esteras et al., 2013; Delourme et al., 
2013; Li et al., 2014; Vos et al., 2017). The r2 value ≥ 0.2 was considered as a threshold to 
declare significant association between loci pairs. 
 
96 
 
University of Ghana http://ugspace.ug.edu.gh
5.2.2.1 GWAS of Agronomic traits 
SNPs with polymorphism information content (PIC) > 0.2 representing 12% of the total 
generated SNPs were selected for the association analysis using general linear model with 
GGT2 software (van Berloo, 2008). The following formula was used for the calculation of the 
association: 
𝑌 = 𝑋𝑎 + 𝑄𝑏 + 𝐾𝑢 + 𝑒 
where: Y represents the phenotype, X represents the genotype, a is the vector of fixed 
systematic effects, Q is the population structure, b is the vector of allele substitution effects of 
the major QTL, K is the relationship between samples, u is the vector of additive genetic effects 
explained by the polygenes, and e the error. Bonferroni multiple test threshold was used for 
significance of the association. The plot of association was done in R software version 3.3.1 
using “qqman” package (Turner, 2014).  
5.3 Results 
5.3.1 Linkage disequilibrium  
Nearly 42% of the analysed SNPs pairs were in linkage disequilibrium (p = 0.001) and r2 ranged 
from 0.058 to 0.99 (Table 5.2). The mean LD for the whole germplasm set was 0.33. However, 
there were large differences in r2 values among SNP pairs and only 13.3% of SNPs pairs 
evaluated exhibited r2 higher than 0.5. 
Intrachromosomal r2 ranged from 0.25 to 0.425 for the entire genome (Table 5.2). Chromosome 
A08, A10 and B10 exhibited the lowest mean r2 (< 0.30) while the highest intrachromosomal 
mean r2 value was observed on chromosome A04 and B05 (> 0.40).  No difference was 
observed between genome A and genome B for the r2 values. 
  
97 
 
University of Ghana http://ugspace.ug.edu.gh
Table 5.1 Intrachromosomal linkage disequilibrium pattern 
A GENOME  B GENOME 
2
Chrom. % LP Mean r2 r2 range  Chrom. % PL Mean r  r2 range 
A01 0.389 0.376 0.061 - 0.998  B01 0.588 0.328 0.059 - 0.999 
A02 0.638 0.329 0.059 - 0.999  B02 0.708 0.383 0.058 - 0.999 
A03 0.573 0.329 0.058 - 0.990  B03 0.601 0.377 0.059 - 0.999 
A04 0.706 0.425 0.060 - 0.990  B04 0.617 0.389 0.058 - 0.999 
A05 0.513 0.338 0.058 - 0.998  B05 0.661 0.414 0.058 - 0.999 
A06 0.692 0.384 0.059 - 0.999  B06 0.550 0.304 0.058 - 0.999 
A07 0.365 0.321 0.059 - 0.999  B07 0.587 0.37 0.058 - 0.99 
A08 0.496 0.25 0.059 - 0.972  B08 0.665 0.371 0.058 - 0.998 
A09 0.758 0.351 0.059 - 0.999  B09 0.501 0.311 0.058 - 0.999 
A10 0.722 0.269 0.052 - 0.999  B10 0.469 0.282 0.059 - 0.990 
Mean A 0.572 0.33 0.058 - 0.99  Mean B 0.582 0.345 0.058 - 0.99 
T-Intra 0.34 0.34 0.058 - 0.99      
T-Inter 0.30 0.18 0.006 - 0.23      
Chrom. = Chromosome, %PL = % loci pair in LD with P < 0.001, T-Intra = Total intrachromosomal, T-Inter = 
Total interchromosomal. 
 
Structure analysis revealed three distinguishable clusters in the entire germplasm (k=3).  Three 
clusters were confirmed by principal component analysis with a low level of admixture 
between clusters (Figure 5.1). The squared allele frequencies (r2) were assessed for all pairwise 
SNP combinations and Index of association (Ia) was used as a measure of multi-locus linkage 
disequilibrium comparison between clusters. Monomorphic loci in clusters were removed prior 
to the LD computation. The highest LD was observed in CIII (0.63) while the lowest was 
exhibited by CI (0.39). For CI, CII, and CIII respectively 3.85%, 6.26% and 18.33% of the 
SNP pairs were in LD with r2 value above 0.5 (Table 5.2).  
98 
 
University of Ghana http://ugspace.ug.edu.gh
 
Figure 5.1 A principal component biplot diagram showing the three clusters identified by structure analysis. 
 
These observations were confirmed by the significant index of association (Ia) in the entire 
germplasm collection (P = 0.001). CIII showed the highest Ia (0.143) while CI exhibited the 
lowest (0.10). The distribution of the LD in clusters followed the same trend observed in Ia 
showing a certain correlation. The plot of r2 as a function of genetic distance revealed no 
significant difference in LD decay between clusters. 
  
99 
 
University of Ghana http://ugspace.ug.edu.gh
Table 5.2 LD characteristics of clusters 
 CI CII CIII Overall 
Proportion of polymorphic loci (%) 0.68 0.45 0.85 1 
Proportion of Significant LD (p = 0.001) 16.04 % 18.04% 30.82% 58.21% 
Mean LD 0.39 0.50 0.63 0.33 
Range 0.15-0.99 0.14-0.99 0.27-0.99 0.058-0.99 
LD>0.5 3.85% 6.26% 18.73% 13.30 
PIC 0.25 0.27 0.24 0.24 
 
The plot of LD over the genetic distance in base pairs (Mbp) showed a decay that appeared to 
be low (Fig 11). Furthermore, the distance at which half of the maximum LD has decayed was 
quite high (113.14 Mbp). In this study, the genomic distance at which LD value (r2) decreases 
below the threshold (0.2) was around 140–145 Mbp (Figure 5.2). No significant difference in 
LD decay was observed between genome A and B as well as between chromosomes. 
 
100 
 
University of Ghana http://ugspace.ug.edu.gh
Figure 5.2 Linkage Disequilibrium decay in 92 groundnut accessions; Blue line= trend line for the mean LD; 
Green line= half decay; red line= r2 threshold for LD 
5.3.2 Genome-wide association  
Calculated r2 revealed that among 31 phenotypic traits scanned, a significant association was 
found with only 13 traits (Table 5.3). Nearly 56 SNP peaks in total reached the optimal P value 
(-log(P) = 7.21). Most of the significant markers (p< 10-5) showed r2 below 0.5. However, SNPs 
with high correlation (r2 > 0.50) were found to be associated with LLS resistance. These SNPs 
were mainly located on chromosome A02, A03, A05 and B10. Thus, in this study, LLS 
resistance was mainly associated with the A than the B genomes in groundnut. 
  
101 
 
University of Ghana http://ugspace.ug.edu.gh
Table 5.3 Linkage groups for agronomic traits and r2 range 
Traits No. of Sig. SNPs Chromosome r2 range 
LLS 12 A02; A03; A05; B10* 0.50-0.57 
100Seed Weight 36 A07; A08; B10 0.22-0.39 
Pod width 16 A08; B04; B05; B6 0.30-0.42 
Pod length 14 A07; B05 0.25-0.31 
Seed width 18  B05 0.21-0.31 
Seed length 36 A03; A04; A08; A10; B03; B04; B05; B06; B10 0.30-0.56 
Growth Habit 122 A01; A04; A06; A07; A08; A09; A10; B07; B08 0.30-0.46 
Pod Reticulation 35 A01; A02; A07; A08; B02; B07; B08; B10 0.30-0.42 
Seed Color 33 A04; A08; B01; B02; B07  0.30-0.41 
*Previously reported genomic region are in bold, No. of Sig. SNPs = Number of significant associations 
SNPs with high association to specific chromosomes were discovered for other traits such as 
growth habit, pod width, pod length and seed length. SNPs on chromosome A07, A08 and B10 
were found to be significantly associated with hundred seed weight with r2 ranging from 0.22 
to 0.39. Among the chromosomes associated to hundred seed weight, chromosome A07 
harboured the highest number of SNPs (6 SNPs). Pod length and pod width were significantly 
associated, respectively with A07 and B05 (0.25 ≤ r2 ≤ 0.31) and with A08, B04, B05, B6 (0.30 
≤ r2 ≤ 0.42). Also, pod width was found to be more associated with B genome while pod length 
was significantly associated with both A and B genomes. Seed coat colour was found to be 
specifically associated with chromosome A04, A08, B01, B02 and B07 with r2 ranging from 
0.30 to 0.41.  
Various genomic regions were significantly associated with growth habit, pod reticulation, and 
seed width (Figure 5.3). Also, seed length and seed width seemed to be more associated with 
the B genome than the A genome in this study. Growth habit exhibited the highest number of 
102 
 
University of Ghana http://ugspace.ug.edu.gh
SNPs significantly associated with up to nine chromosomes that mostly belong to the A 
genome. 
 
Figure 5.3 Illustration of Manhattan plot depicting strongly associated markers to different traits. The 
Bonferroni multiple test threshold is shown by a red line. Associated significant SNPs are at the top 
of the red line. Chromosomes were numbered as 1 – 20 where the A genome is represented by 1-10 
and the B genome is represented as 11-20. 
 
5.4 Discussion 
Knowledge of the structure and linkage disequilibrium in a germplasm collection is a 
prerequisite for the trait association study. In this study, structure analysis revealed three 
distinguishable clusters sharing similar molecular features. The presence of three clusters in 
the germplasm collection was further confirmed by principal component analysis with a low 
level of admixture between clusters.  A large number of SNP pairs were found to be in linkage 
disequilibrium in the studied germplasm with an r2 value above 0.3. The high proportion of 
103 
 
University of Ghana http://ugspace.ug.edu.gh
SNPs with high values of minor allele frequency (MAF) (>0.4) found in the diversity study 
may be related with the high LD (r2>0.3) observed in this study. This confirms the idea that 
LD distance increase with the increase of MAF (Yan et al., 2009). 
The plot of r2 as a function of the distance showed a decay. However, the half decay position 
(Branca et al., 2011; Vos et al., 2017) in this study was high confirming the high level of 
linkage disequilibrium observed in this study. Similar high LD values have been reported on 
groundnut germplasm collection from 48 countries using SSR and SNP markers by Pandey et 
al. (2014). Also, previous reports of LD study in groundnut mini core using SSR markers, 
showed that a large number of marker pairs were in linkage equilibrium (Belamkar et al., 2011). 
Though the difference may be partly due to the population studied, in comparison to the LD 
reported by these authors, in this study a higher LD in germplasm collection from Togo with 
the SNP markers was detected. Besides, the low decay observed in this study could partly be 
explained by the self-pollinated nature of groundnut (Vos et al., 2017) and the limited 
occurrence of recent genetic introgression events.  
Linkage disequilibrium useful for association study was found to be very high in this 
population (143 Mb). However, several studies reported similar findings for highly self-
pollinating crops (Abdurakhmonov & Abdukarimov, 2008; Agrama & Eizenga, 2008). In 
addition, the single domestication event (single wild progenitor has generated all the cultivated 
groundnut), probably resulting in fewer alleles being passed on to subsequent generations, may 
have induced the high LD observed in groundnut (Krapovickas & Gregory, 1994).  
Overall, population structure influenced LD in this study, as the mean r2 varied between 
clusters identified. Cluster III, consisting of groundnut accessions from Togo and a line from 
ICRISAT, exhibited the highest LD while cluster II, made up of lines from Togo and from 
Senegal, showed the lowest LD. Belamkar et al. (2011) reported similar findings with SSR 
104 
 
University of Ghana http://ugspace.ug.edu.gh
markers in the USA and suggested that any association study using the studied population 
should consider population structure.  Significant low LD between unlinked loci was found in 
this study. This confirmed the existence of other LD generating factors suggested by previous 
studies (Stich et al., 2007). The existence of distinct clusters is one factor that could explain 
the LD pattern in this study as relatedness could generate LD between unlinked loci 
(Malysheva-Otto et al., 2006). However, the low level of LD between unlinked loci suggests 
that an association study would be effective with the present germplasm. In addition, the high 
LD observed for the intrachromosomal loci pairs suggests that few markers are required for the 
association mapping (Abdurakhmonov, & Abdukarimov, 2008). 
For a more accurate estimate of the association, a highly significant threshold (r2 ≥ 0.2) was set 
in this study. Indeed, using r2 = 0.2 as a threshold (Delourme et al., 2013; Li et al., 2014), 
linkage disequilibrium decay showed that 143 Mbp is the approximate extent of LD distance 
measure useful for association study in the collected germplasm.  
GWAS yielded useful information that could be exploited for the improvement of multiple 
agronomic traits of interest. Many SNPs (56) were found to be associated with useful 
agronomic traits such as LLS resistance, Pod size, seed size, growth habit and seed coat colour 
suggesting there is a group of QTLs controlling these traits. The association analyses revealed 
up to 12 SNPs significantly associated with LLS resistance. Some of the genomic regions 
described in this study for LLS resistance genes on chromosome A03, A05 and B10 are well 
known. Recently, Zhang et al. (2017) in a GWAS study reported genes encoding a protein that 
is involved in LLS resistance to be distributed mainly on chromosome A03. Many QTLs are 
reported to be located on linkage group A05 and B10 (Agarwal et al., 2018). Also, the A 
genome seems to harbour more SNPs for LLS resistance than the B genome in this study as 
suggested by Leal-Bertioli et al. (2015). However other genomic regions such as the SNPs 
located on chromosome A02 for LLS are new. LLS is one of the major foliar diseases 
105 
 
University of Ghana http://ugspace.ug.edu.gh
worldwide causing significant yield loss. Thus, the identification of new genomic regions 
controlling resistance is worthy of further investigation.  
Along with LLS, many SNPs were found to be associated with yield component traits. Among 
the significant linkage group associated to pod size, only A08 and B05 (for pod width) and B10 
(pod length) have been previously reported to be associated with pod width and pod length 
(Fonceka et al., 2012). Also, new SNPs were found to be associated with seed width and seed 
length and were located on linkage groups different from those reported by previous studies 
(Fonceka et al., 2012; Pandey et al., 2014). In addition, significant associations were also 
identified for hundred seed weight (36 SNPs), growth habit (122 SNPs), pod reticulation (32 
SNPs) and seed colour (33 SNPs). Overall, and at the best of our knowledge, most of the 
identified marker-trait association in this study are new. Thus, further analysis of agronomic 
traits using SNPs identified in this study could allow the identification of candidate genes for 
the future breeding programmes. The association study confirmed that GWAS is a useful 
method for the analysis of complex agronomic traits (Fahrenkrog et al., 2017).  
5.5 Conclusion 
A high LD has been observed in this study with the germplasm collection and was found to be 
consistent with other LD reported on highly self-pollinating crops. Several marker-trait 
associations were found for LLS resistance, hundred pod weight, pod length, pod width, seed 
length, seed width, pod reticulation, seed colour and growth habit. For LLS resistance, six new 
SNPs (100027806, 100015404, 100024878, 100052500, 100005206 and 100002029) were 
detected on chromosome A02. The identification of SNPs and linkage groups harbouring useful 
genes in this study could play a key role in the identification of candidate genes and the 
development of markers for the future breeding programme in Togo. These markers could be 
106 
 
University of Ghana http://ugspace.ug.edu.gh
deployed in marker-assisted selection for the development of high yield LLS resistant 
genotypes adapted to farmers growing conditions in Togo and worldwide. 
  
107 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER SIX 
6. GENETIC ANALYSIS OF LATE LEAF SPOT RESISTANCE, YIELD AND 
YIELD COMPONENTS 
6.1 Introduction  
Nature of gene action involved in the expression of a trait, is one of the considerations among 
others that the breeder must look at when starting a breeding programme. Indeed, the success 
of a breeding programme depends on the genetic gain per selection cycle which in turn depends 
mostly on how genes act or interact in the production of a phenotype (Falconer & Mackay, 
1996). Another important step in plant breeding programmes is the identification of best 
performing lines to use in future crosses. Selection of parental lines can be based upon an 
evaluation of the performance of their offspring (Fasahat et al., 2016).  
In parallel with establishing the gene action, combining ability determined through a specific 
mating design offers the possibility to select parents that would generate improved progenies 
(Sprague & Tatum, 1942). To predict the response to the selection, breeders rely on additive 
genetic variance. Thus, partitioning the phenotypic variation into genotypic and environmental 
variance and further into additive and non-additive is crucial (Ortiz & Golmirzaie, 2002). Also, 
the estimation of maternal effect is important as it influences the evolution of progenies based 
on the correlation between maternal and offspring traits (Räsänen & Kruuk, 2007).  
Full factorial design (also called North Carolina II design) is one of the best methods for 
simultaneous estimation of additive, non-additive and maternal variances that explain 
phenotypic traits (Neff & Pitcher 2005; Neff et al. 2011). The objectives of the present study 
were to i) determine the type of gene action conditioning late leaf spot (LLS) resistance and 
yield-related traits and ii) identify parents with a good combining ability to initiate groundnut 
breeding programme in Togo.  
108 
 
University of Ghana http://ugspace.ug.edu.gh
6.2 Materials and Methods 
6.2.1 Plant materials 
Eight parent genotypes were crossed in a full factorial mating design (North Carolina II design). 
Four farmers’ varieties from Togo, 38AH, 09AH, 68AH and 43AH were used as males and 
four genotypes 12CS_22 and 12CS_36 from Senegal and ICG 7878 and ICGV 02271 from 
ICRISAT were used as females. The farmers’ local varieties were highly susceptible to LLS 
and exhibited small pods and seed size. The lines from Senegal were moderately resistant to 
LLS but with large pod and seed size. In contrast to local genotypes from Togo, ICRISAT 
genotypes were resistant to LLS with medium pod and seed size (Table 6.1). F1 hybrids were 
developed at ICRISAT-WCA in Bamako. Male parents were planted in one row of 40 plants 
per male parent two weeks before female parents. Four rows of 4 m long and 30 cm between 
hills and 1 m between rows were adopted for female parents. Hand emasculation in the 
afternoon and pollination early in the morning were carried out by skilled workers at 
International Crops Research Institute for the Semi-Arid Tropics (ICRISAT, Bamako) under 
the supervision of the researcher and precaution were taken to avoid contamination. F1 seeds 
were then planted with both parents in unreplicated trial during the off season to check the 
hybridity based on morphological traits such us plant height, leaf color and shape and pod traits. 
Selected hybrids were then sent to CERAAS for genotyping using a set of polymorphic 
markers. Confirmed hybrids were then advanced to F2 population for multilocation trials. 
  
109 
 
University of Ghana http://ugspace.ug.edu.gh
Table 6.1 Parent genotypes characteristics 
Genotype Status Pedigree Growth habit LLS Pod and seed 
resistance size 
38AH Male Accession (Togo) Erect Low Small 
09AH Male Accession (Togo) Erect Low Small 
43AH Male Accession (Togo) Erect Low Small 
69AH Male Accession (Togo) Bunch Moderate Large 
12CS_22 Female Fleur11 x AiAB Bunch Moderate Medium 
12CS_36 Female Fleur11 x AiAB Bunch Moderate Medium 
ICG 7878 Female Accession (North Carolina) Bunch High Medium 
ICGV 02271 Female TAG 24 x ICGV 86031 Bunch High Medium 
 
6.2.2 Test environments and experimental design 
Sixteen cross combinations F2 populations and their parents were evaluated using an alpha-
lattice design in four agroecological environments representing the major ecological zones of 
groundnut production: Talo_Kara (9.546° N, 1.193° E), Tantigou_Dapaong (10.8733° N, 
0.2010° E), Sotouboua_Central (8.565° N, 0.977° E) and Mission Tove_Maritime (6.322°N, 
1.126°E).  A plot of four rows of 5 m length containing 33 plants per row with 15 cm between 
plants represented an experimental unit. A spacing of 50 cm between rows and 1 m between 
blocks was adopted. The field was kept weed free at an interval of one weeding every three 
weeks by hoeing. 
6.2.3 Data collection  
Disease scoring for LLS was done at 60 days and 90 days after sowing using a 9 points scale 
(Subrahmanyam et al., 1995) on individual F2 plants as described in Table 6.2. In parallel with 
establishing the gene action for the improvement of LLS, observation on yield parameters such 
as pod weight (weight/plant, 100 pods), number of pods per plant, were recorded on 120 F2 
110 
 
University of Ghana http://ugspace.ug.edu.gh
plants per cross combination. Hundred seed weight was collected on a plot basis. Pod length, 
pod width, seed length, and seed width were collected on plot basis on 30 randomly selected 
pods and seeds.  
Table 6.2 Description of late leaf spot rating scale 
Leaf Description Disease 
spot severity 
score (%) 
1 No disease  0 
2 Lesions largely on lower leaves; no defoliation 1-5 
3 Lesions largely on lower leaves; very few lesions on middle leaves; 6-10 
defoliation of some leaflets evident on lower leaves 
4 Lesions on lower and middle leaves, but severe on lower leaves; 11-20 
defoliation of some leaflets evident on lower leaves 
5 Lesions on all lower and middle leaves; over 50% defoliation of 21-30 
lower leaves 
6 Lesions severe on lower and middle leaves; lesions on top leaves 31-40 
but less severe; extensive defoliation of lower leaves; defoliation 
of some leaflets evident on middle leaves 
7 Lesions on all leaves but less severe on top leaves; defoliation of 41-60 
all lower and some middle leaves 
8 Defoliation of all lower and middle leaves; lesions severe on top 61-80 
leaves and some defoliation of top leaves evident 
9 Defoliation of almost all leaves leaving bare stems; some leaflets 81-100 
may be present, but with severe leaf spot 
 
  
111 
 
University of Ghana http://ugspace.ug.edu.gh
6.2.4 Data analysis 
6.2.4.1 Variance component estimates 
Phenotypic variances were partitioned into additive, non-additive and maternal effect using the 
Restricted Maximum Likelihood (REML) model (Neff & Pitcher 2005; Bates et al., 2015) in 
“fullfact” R package as follow: 
𝑌𝑖𝑗𝑘𝑙𝑚𝑛 = 𝜇 + 𝐵𝑖 + 𝐿𝑗 + 𝑀𝑘 + 𝐹𝑙 + 𝐼𝑚 + 𝑒𝑖𝑗𝑘𝑙𝑚𝑛 
where Yijklmn is the phenotypic response, µ is the mean value, Bi is the effect of block, Lj is the 
effect of location, M  is the kth effect of male, F  is the lth effect of female, I  is the mthk l m  effect 
of interaction between male and female and eijklmn is the random residual. The random and 
fixed effect of all the factors were estimated. 
The significance for block, location, male, Female and the interaction was estimated using the 
likelihood ratio test with ‘observLmer3’ function in the “fullfact” R package (Houde & Pitcher, 
2016) for the normally distributed traits (Appendice 3). The phenotypic variance was then 
apportioned into male (Vm), female (Vf) and the interaction between male and female (VI). The 
calculation of additive (VA), non-additive (VNA) and maternal variance (VM) was as follows: 
VA= 4Vm; VNA= 4VI; VM = Vf -Vm (Lynch & Walsh, 1998). 
6.2.4.2 Combining ability estimates 
The estimation of GCA and SCA was done using excel and package “lme4” in R software 
version 3.3.1. The following model was used in R: 
1
𝑋𝑖𝑗𝑘𝑙𝑚𝑛 = 𝜇 + 𝑔𝑖 + 𝑔𝑗 + 𝑠𝑖𝑗 + ( )𝛴𝑏 𝑘
𝑒𝑖𝑗𝑘𝑙𝑚𝑛 
where 𝜇 = the population mean; 𝑔𝑖 = the general combining ability effect of the ith parent; 𝑔𝑖 
= the general combining ability effect of the jth parent; 𝑠𝑖𝑗 = the specific combining ability 
112 
 
University of Ghana http://ugspace.ug.edu.gh
effect of the cross between ith and jth parents; 𝑒𝑖𝑗𝑘𝑙𝑚𝑛= the environmental effect associated 
with ijkth observation; b = number of blocks/replication. Using parent’s genotyping data, 
Euclidian genetic distance was computed between the 8 parents for the evaluation of the 
association between genetic distance and SCA (Hedrick, 2005). 
6.3 Results 
6.3.1 Variance components in LLS and yield-related traits 
A significant random effect for location (8.09%) and significant interactions between the 
environment and variance components for LLS (Table 6.3) and most yield-related traits was 
observed. Additive gene effect played the major role in all the three locations, but the 
estimation of non-additive and maternal effect varied between environments for LLS 
resistance. The maternal effect was higher in Kara whereas non-additive effect was not 
detected. Analysis of variance component across environment for LLS resistance showed that 
additive genetic explained the largest part of the phenotypic variance observed (87.84%) while 
non-additive variance was very low (4.3). Maternal variance (8.6) contributed to the rest of the 
phenotypic variance. With the addition of the residual, the variance components sum up to 
more than 100% showing that there may be epistasis. 
Table 6.3 Variance component for LLS by sites 
 Variance components (%) 
Sites VI Vm Vf VA VNA VM 
Savanes 1.2 21.9 13.3 87.84 4.83 -8.6 
Kara 0.00 18.13 0.00 72.52 0.00 -18.26 
Sotouboua 1.1 22.3 14.5 89.2 4.4 -7.3 
Mission_Tove 0.8 15.3 15.2 61.2 3.1 -0.1.1 
VI, Vm and Vf represent the phenotypic variance components (%) male x female interaction, male and female 
effects, respectively. VA, VNA and VM represent additive genetic, non-additive genetic and maternal effect variance 
components, respectively. 
113 
 
University of Ghana http://ugspace.ug.edu.gh
 
Additive genetic variance seemed to prevail over non-additive variance for all the other traits 
except for pod weight per plant (Table 6.4). Non-additive genetic variance played a major role 
(21.79% of the phenotypic variance explained) in the expression of the pod weight per plant. 
For the number of pods per plants, additive genetic variance explained 7.3% of the total 
phenotypic variance while non-additive (0.56%) and maternal effect (-1.82%) were very low. 
Table 6.4 Variance components for yield and yield-related traits across sites 
 Variance components (%) 
Traits VL VI Vm Vf VA VNA VM 
LLS 9.25 2.35e-3 10.87 19.73 43.48 9.4E-3 3.86 
Pods / plant 53.7 0.142 1.82 0.00 7.30 0.56 -1.82 
Pod length 1.49 0.00 0.48 46.04 1.93 0.00 45.56 
Pod width 14.83 0.00 0.13 63.66 0.53 0.00 63.72 
Seed length 24.8 0.00 0.51 16.4 2.07 0.00 15.8 
Seed width 10.54 0.00 10.02 1.01 4.05 0.00 9.01 
100 Pod weight 14.30 2.22e-13 2.34 2.21 9.38 0.00 -0.12 
100 Seed weight 10.6 0.00 3.21 11.66 12.84 7.15E-15 8.44 
Pod weight/plant 41.01 5.44 1.42 0.64 5.69 21.79 -0.78 
Shelling % 69.92 1.67 0.79 2.60 3.18 6.69 1.81 
VL, VI, Vm and Vf represent the phenotypic variance components (%) explained by location, male x female 
interaction, male and female effects, respectively. VA, VNA and VM represent additive genetic, non-additive genetic 
and maternal effect variance components, respectively. 
Overall, LLS resistance, pod length, pod width, seed length, seed width and hundred seed 
weight showed a high additive effect. For Pod length, pod width, seed length, seed width pod 
weight per plant and shelling percentage, the additive effect was significant but low. Posotive 
maternal effects were observed for the LLS, pods/plant, pod width, seed size (length and 
114 
 
University of Ghana http://ugspace.ug.edu.gh
width), hundred seed weight and shelling percentage. In contrast, hundred pod weight and pod 
weight per plant showed a negative maternal effect. 
6.3.2 GCA for LLS and yield-related traits 
Significant interactions were observed between the environments and GCA in the number of 
pods per plant, pod length, pod width, seed length, seed width, hundred pod weight, and 
shelling percentage and at a lesser extent in LLS incidence (result not shown). GCA in Kara 
and Sotouboua showed a high correlation for most traits while Sotouboua and Mission_Tove 
showed a weak but significant correlation (p = 0.05). Estimates of GCA effect across 
environment for the tested genotypes are presented in Table 6.4 and 6.5. As suggested by 
Bookmyer et al. (2009) and Mukankusi et al. (2011), on the basis of the LLS scale where less 
disease incidence is associated with the lowest value, negative combining ability effects are 
preferable for disease resistance. 
  
115 
 
University of Ghana http://ugspace.ug.edu.gh
Table 6.5 General combining ability for LLS and yield-related traits of the genotypes used as 
male parents 
GCA of male parents  
Traits 
43AH 68AH 09AH 38AH Significance 
LLS -0.137  -0.007 0.039 0.1042  * 
(±0.046) (±0.047) (±0.046) (±0.047) 
Pod/Plant 0.164  0.311  0.005 -0.481 * 
(±9.4E-3) (±9.3E-3) (±9.3E-3) (±9.3E-3) 
Pod Length 0.026  -0.037  -0.005  0.016 NS 
(±0.009) (±0.009) (±0.009) (±0.009) 
Pod Width 0.018  0.006  -0.011  -0.002  * 
(±0.004) (±0.004) (±0.003) (±0.003) 
Seed Length -0.007  -0.016  0.021  0.001 NS 
(±0.007) (±0.007) (±0.007) (±0.007) 
Seed Width 0.018  -0.007  -0.005  -0.006  * 
(±0.004) (±0.004) (±0.004) (±0.004) 
Hund PW 1.082 2.613 -2.01 -1.685 * 
(±0.126) (±0.127) (±0.166) (±0.124) 
Hund SW -0.427 -1.263  0.385 1.304 * 
(±0.080) (±0.080) (±0.082) (±0.081) 
PW/Plant 0.280 0.577  -0.12  -0.737  
* 
(±0.264) (±0.264) (±0.264) (±0.264) 
 
Shelling % 1.657 -0.757 0.045 -0.944 * 
(±0.619) ( ±0.620) (±0.620) (±0.620) 
The best High GCA were highlighted (in bold) and ranked (rank in parenthesis). Significance codes: ‘***’ 
0.001; ‘**’ 0.01; ‘*’ 0.05, ‘NS’ Non-significant; Hund PW = hundred pod weight, Hund SW = hundred seed 
weight, PW/Plant = pod weight per plant; SD = Standard deviation 
Among male parents, 43AH showed a slightly higher GCA effect in the desired direction for 
LLS resistance (Table 6.5). Among the female parents, ICG 7878 exhibited the highest GCA 
effect in the desired direction (Table 6.6). Thus, these two parents were good general combiners 
for LLS resistant genotypes’ development. Line 43AH exhibited, along with LLS resistance, 
good GCA effect for yield-related traits such as number of pods per plant, pod width, hundred 
pod weight, and shelling percentage (Table 6.4 and 6.6). The best general combiner for LLS 
116 
 
University of Ghana http://ugspace.ug.edu.gh
resistance, ICG 7878, exhibited negative GCA for most yield-related traits such as hundred 
pod weight and hundred seed weight. 
Table 6.6 Female parents GCA for LLS and yield-related traits 
GCA of female parents 
Traits 12CS_22 12CS_36 ICGV 02271 ICG 7878 Significance 
LLS 0.771  0.641 0.215  -1.627 ** 
(±0.146) (±0.145) (±0.145) (±0.145) 
Pod/Plant 0.54 0.151 -0.849 0.158 NS 
(±0.187) (±0.186) (±0.186) (±0.186) 
Pod Length -0.01 0.115 -0.065 -0.038 NS 
(±0.036) (±0.036) (±0.036) (±0.036) 
Pod Width -0.02 -0.116 0.054(2) 0.08(1) * 
(±0.027) (±0.026) (±0.026) (±0.026) 
Seed L 0.011 0.044 -0.06 0.005 * 
(±0.019) (±0.018) (±0.018) (±0.019) 
Seed W 0.006 -0.027 0.024 -0.002 * 
(±0.008) (±0.008) (±0.008) (±0.008) 
Hund PW 1.646 -1.794 8.556 -8.407 ** 
(±3.333) (±3.313) (±3.323) (±3.325) 
Hund SW 2.051(2) 0.571 2.461(1) -5.083 ** 
(±1.571) (±1.571) (±1.571) (±1.571) 
PW/Plant 0.657(1) 0.289(2) -0.365 -0.581 * 
(±0.199) (±0.198) (±0.199) (±0.198) 
Shelling % 0.764(2) 1.436(1) 0.0301 -2.229 * 
(±0.760) (±0.761) (±0.762) (±0.762) 
The best High GCA were highlighted (in bold) and ranked (rank in parenthesis). Significance codes: ‘***’ 0.001; ‘**’ 0.01; ‘*’ 0.05, ‘NS’ 
Non-significant; Hund PW = hundred pod weight, Hund SW = hundred seed weight, PW/Plant = pod weight per plant 
Genotypes such as 12CS_22 and 12CS_36 which did not show significant negative GCA effect 
for LLS resistance, are good combiners for yield related traits such as pod weight per plant and 
shelling percentage. The other two male parents, 09AH and 38AH displayed positive GCA 
effect for only seed length and hundred seed weight. No significant GCA effect was observed 
for pod length, seed length on the male side, and the number of pods per plant and pod length 
on the female side. 
117 
 
University of Ghana http://ugspace.ug.edu.gh
6.3.3 SCA for LLS and yield-related traits 
Multi-environment SCA analysis revealed a significant interaction between environment and 
SCA for yield-related traits (result not shown). The interaction was even high with the ranking 
of SCA changing across different environments. Estimation of SCA across environments 
revealed positive and negative cross combinations. For LLS resistance, eight hybrids exhibited 
negative SCA effects and the top three were 12CS_22x38AH, ICG 7878x68AH, ICGV 
02271x43AH (Table 6.7). These cross combinations had as parent combinations resistant x 
susceptible, resistant x moderately resistant, and moderate resistant x susceptible confirming 
the non-additive gene action observed in the variance component analysis. Most of the 
combinations with LLS resistance generally displayed an undesirable SCA effects for pod/plant 
and pod weight per plant, except 12 CS_36x38AH (Table 6.7). 
  
118 
 
University of Ghana http://ugspace.ug.edu.gh
Table 6.7 Specific combining ability of cross combination for LLS and Pod traits 
Traits 
Cross combination LLS Pod/Plant PW/Plant Pod Length Pod Width 
12CS_22 x 43AH 0.2106 1.231(2) 0.809(2) -0.006 -5E-04 
(±0.048) (±0.261) (±0.266) (±0.039) (±0.002) 
   
12CS_36 x 43AH 0.0069 -0.547 -0.228 -0.014 -0.007 
(±0.047) (±0.262) (±0.266) (±0.039) (±0.005) 
   
ICGV 02271 x 43AH -0.16(3) -0.519 0.2025 -0.009 -0.012 
(±0.047) (±0.263) (±0.267) (±0.038) (±0.006) 
   
ICG 7878 x 43AH -0.058 -0.165 -0.784 0.0285 0.019 
(±0.047) (±0.262) (±0.266) (±0.040) (±0.006) 
   
12CS_22 x 68AH -0.104 -0.054 -0.872 0.034 -0.014 
(±0.047) (±0.262) (±0.266) (±0.041) (±0.006) 
12CS_36 x 68AH 0.1736 1.279(1) 0.640(4) 0.0137 0.022 
(±0.047) (±0.262) (±0.266) (±0.038) (±0.021) 
ICGV 02271 x 68AH 0.1921 -0.498 -0.176 0.0073 0.0049 
(±0.048) (±0.262) (±0.266) (±0.041) (±0.002) 
ICG 7878 x 68AH -0.262(2) -0.727 0.408(5) -0.055 -0.012 
(±0.048) (±0.262) (±0.266) (±0.041) (±0.266) 
12CS_22 x 09AH 0.1829 -0.332 1.114(1) 0.031 0.020 
(±0.047) (±0.262) (±0.266) (±0.039) (±0.021) 
12CS_36 x 09AH -0.095 -0.832 -1.078 0.032 0.0004 
(±0.048) (±0.261) (±0.266) (±0.038) (±0.003) 
ICGV 02271 x 09AH -0.113 0.613(3) -0.09 -0.031 -0.011 
(±0.047) (±0.262) (±0.266) (±0.037) (±0.002) 
ICG 7878 x 09AH 0.0255 0.550 0.0543 -0.032 -0.009 
(±0.048) (±0.262) (±0.266) (±0.039) (±0.004) 
12CS_22 x 38AH -0.289(1) -0.845 -1.051 -0.058 -0.005 
(±0.048) (±0.262) (±0.267) (±0.040) (±0.003) 
12CS_36 x 38AH -0.086 0.099 0.666(3) -0.032 -0.016 
(±0.046) (±0.263) (±0.267) (±0.039) (±0.021) 
ICGV 02271 x 38AH 0.081 0.405 0.0638 0.032 0.018 
(±0.048) (±0.261) (±0.266) (±0.037) (±0.021) 
ICG 7878 x 38AH 0.294 0.342 0.3208 0.058 0.0028 
(±0.047) (±0.262) (±0.266) (±0.390) (±0.0018) 
Significance * * ** NS NS 
The top High SCA were highlighted (in bold) and ranked (rank in parenthesis). Significance codes: ‘***’ 0.001; 
‘**’ 0.01; ‘*’ 0.05, ‘NS’ Non-significant; PW/Plant = pod weight per plant 
Other interesting cross combination were: 12CS_22 x 43AH, ICGV 02271 x 09AH, ICG 7878 
x 09AH for the number of pods per plants; 12CS_22 x 09AH for pod weight per plant; ICG 
7878 x 43AH for shelling percentage; ICGV 02271 x 43AH for hundred seed weight and 
119 
 
University of Ghana http://ugspace.ug.edu.gh
hundred pods weight (Table 6.8). Also, these crosses were mostly between high GCA x high 
GCA or high GCA and low GCA for the different traits. However, another cross between low 
GCA x low GCA, following the example of ICGV 02271 x 09AH, exhibited a high SCA effect 
for the number of pods per plant. Altogether, these results underlined the implication of non-
additive gene action in the yield-related trait expression, even though additive genes play the 
major role. Specific combining ability for pod length, pod width, seed length and seed width 
displayed no statistical difference between cross combinations (Table 6.7 and 6.8). 
  
120 
 
University of Ghana http://ugspace.ug.edu.gh
Table 6.8 Specific combining ability of cross combination for yield and seed related traits 
Traits 
Crosses 100 PodW 100 SeedW Shelling % Seed Length Seed Width 
12CS_22 x 43AH 1.6035 -1.446 -1.187 -5E-05 0.007 
(±1.783) (±4.4E-8) (±0.710) (±0.012) (±0.075) 
12CS_36 x 43AH -8.2078 -0.259 0.587 0.0043 -0.0143 
(±1.720) (±4.4E-8) (±0.720) (±0.013) (±0.076) 
ICGV 02271 x 43AH 5.321(1) 2.993(1) -1.21 -0.013 0.023 
(±1.746) (±4.4E-8) (±0.720) (±0.012) (±0.076) 
ICG 7878 x 43AH 1.2838 -1.289 1.81(1) 0.0083 -0.015 
(±1.720) (±4.3E-8) (±0.720) (±0.013) (±0.077) 
12CS_22 x 68AH -1.6436 -0.795 1.739(2) -0.008 -0.0171 
(±1.735) (±4.3E-8) (±0.740) (±0.012) (±0.076) 
12CS_36 x 68AH 1.3563 0.3344 -1.552 -0.004 0.012 
(±1.720) (±4.4E-8) (±0.720) (±0.011) (±0.073) 
ICGV 02271 x 68AH 1.279 -0.256 -0.449 0.002 0.0047 
(±1.720) (±4.1E-8) (±0.700) (±0.011) (±0.076) 
ICG 7878 x 68AH -0.9918 0.716(4) 0.262 0.011 5E-05 
(±1.820) (±4.2E-8) (±0.710) (±0.012) (±0.076) 
12CS_22 x 09AH 3.3693(3) 1.452(2) 1.337(4) -0.004 0.009 
(±1.952) (±4.2E-8) (±0.730) (±0.013) (±0.076) 
12CS_36 x 09AH 2.445 0.2903 -0.351 -0.009 -0.0005 
(±1.920) (±4.4E-8) (±0.720) (±0.013) (±0.071) 
ICGV 02271 x 09AH -3.2738 -1.834 0.254 0.044 -0.015 
(±1.820) (±4.2E-8) (±0.720) (±0.012) (±0.076) 
ICG 7878 x 09AH -2.54 0.0921 -1.239 -0.031 0.0062 
(±1.820) (±4.4E-8) (±0.770) (±0.010) (±0.066) 
12CS_22 x 38AH -3.3292 0.789(3) -1.889 0.012 0.0013 
(±1.820) (±4.4E-8) (±0.750) (±0.012) (±0.007) 
12CS_36 x 38AH 4.407(2) -0.366 1.317(5) 0.010 0.0023 
(±1.710) (±4.3E-8) (±0.740) (±0.011) (±0.078) 
ICGV 02271 x 38AH -3.3258 -0.904 1.406(3) -0.033 -0.0123 
(±1.700) (±4.4E-8) (±0.730) (±0.014) (±0.076) 
ICG 7878 x 38AH 2.248(5) 0.481(5) -0.834 0.012 0.009 
(±1.820) (±4.3E-8) (±0.720) (±0.012) (±0.076) 
 
Significance * * ** NS NS 
Top three top SCA were highlighted (in bold) and ranked (rank in parenthesis). Significance codes: ‘***’ 0.001; 
‘**’ 0.01; ‘*’ 0.05, ‘NS’ Non-significant; 100 SeedW = hundred seed weight; 100 PodW = hundred pod weight 
 
6.3.4 Traits correlation for SCA and GCA  
GCA effect correlation analysis of LLS and other agronomic traits revealed a positive 
correlation between LLS and hundred seed weight (r = 0.9; p = 0.01), LLS and shelling 
121 
 
University of Ghana http://ugspace.ug.edu.gh
percentage (r =0.73; p = 0.05) (Table 6.9). Conversely, LLS exhibited negative GCA effect 
correlation with pod width (r =-0.71; p = 0.05). Positive correlation was also observed between 
other pairs traits such as number of pods per plant and pod weight per plant (r = 0.72; p = 0.05), 
pod length and seed length (r = 0.76; p = 0.05), hundred pods weight and hundred seed weight 
(r = 0.73; p = 0.04). Surprisingly, negative correlation for GCA effects between pod length and 
pod width (r= -0.88; p = 0.01), seed length and seed width (r = -0.80; p = 0.05) were observed. 
Table 6.9 Correlation between traits’ GCA for all cross combination 
 LLS PPP PL PW SL SW HPW HSW PWP SH 
LLS           
PPP 0.01          
PL 0.39 0.26         
PW -0.71* -0.31 -0.88**        
SL 0.11 0.61 0.76* -0.69       
SW -0.09 -0.36 -0.65 0.67 -0.80*      
HPW 0.61 -0.41 -0.29 0.01 -0.69 0.57     
HSW 0.90** -0.37 0.14 -0.41 -0.19 0.22 0.73*    
PWP 0.53 0.72* 0.22 -0.48 0.22 -0.09 0.31 0.19   
SH 0.73* 0.18 0.59 -0.66 0.21 0.08 0.43 0.59 0.61  
LLS=late leaf spot, PPP=Number of pods per plant, PL=pod length, PW=pod width, SL=Seed length, SW=Seed 
width, HPW= 100 pod weight, 100Seed weight, PWP=pod weight per plant, and SH = Shelling % 
6.3.5 Correlation between SCA and genetic distance 
Genetic distances between parent genotypes were calculated based on SNP data (Table 6.10) 
and a correlation analysis of SCA and genetic distance revealed no significant correlation 
between SCA and genetic distance. Thus, for most of the traits’ SCA varied independently 
from the genetic distance. However, some traits such as pod weight per plant (PWP) and 
shelling percentage (SH) seemed to exhibit high SCA when the genetic distance between 
parents increases (Table 6.11). 
  
122 
 
University of Ghana http://ugspace.ug.edu.gh
Table 6.10 Euclidian distance between parent genotypes based on SNP data 
 09HA 38AH 43AH 68AH 12CS_22 12CS_36 ICG 7878 ICGV 02271 
09AH 0.00        
38AH 9.63 0.00       
43AH 9.64 2.59 0.00      
68AH 11.59 12.01 12 0.00     
12CS_22 11.65 12.12 12.07 5.89 0.00    
12CS_36 11.53 11.82 11.89 5.33 4.6 0.00   
ICG 7878 15.54 15.54 15.47 14.57 14.67 14.57 0.00  
ICGV 02271 13.54 14.63 14.23 13.82 14.53 14.22 10.21 0.00 
 
Table 6.11 Correlation pattern between SCA and the genetic distance between parents. 
 LLS PP PWP PL PW HPW HSW SH SL SW GD 
LLS            
PP 0.56*           
PWP 0.51* 0.49*          
PL 0.63** 0.16 -0.03         
PW 0.57* 0.25 0.18 0.59*        
HPW 0.47 0.42 0.23 0.18 0.62**       
HSW 0.24 0.31 0.09 0.15 0.66** 0.41      
SH -0.39 0.10 0.10 -0.36 -0.49* -0.08 -0.37     
SL -0.33 -0.34 -0.36 -0.01 -0.18 -0.06 -0.27 0.02    
SW 0.65** 0.28 0.10 0.51* 0.91** 0.66** 0.74** -0.66** -0.22   
GD -0.06 -0.22 0.05 -0.20 -0.11 -0.13 -0.23 0.22 -0.39 -0.17  
LLS=late leaf spot, PP=Number of pods per plant, PL=pod length, PW=pod width, SL=Seed length, SW=Seed width, 
HPW= 100 pod weight, 100Seed weight, PWP=pod weight per plant, SH= Shelling % and GD= Genetic distance 
6.4 Discussion 
Determination of the mode of inheritance of agronomic traits helps adopt better planning and 
execution of the breeding programme. Nature of gene action controlling agronomic traits is 
crucial as the breeding method and response to a selection from breeding population depends 
on it. In this study, the additive genetic effects were higher than the non-additive and maternal 
effects for most traits.  Indeed, LLS resistance, pod length, pod width, seed length, seed width 
and hundred seed weight showed a high additive gene action. Other traits such us number of 
123 
 
University of Ghana http://ugspace.ug.edu.gh
pods per plant, hundred pods weight, and shelling percentage showed significant but low 
additive gene action. This result agrees with previous studies that reported that LLS resistance 
and yield-related traits are controlled by additive gene action (Dwivedi et al.,1989; Pasupuleti 
et al., 2013). However, some authors reported a predominance of non-additive gene effect for 
pod yield and related traits using line by tester (Manivannan et al., 2008; Mothilal & Ezhil, 
2010; Patil et al., 2017) or diallel population (Azad et al., 2014).  
In this study, non-additive gene effect prevailed over additive gene effect only for pod weight 
per plant. The discrepancy between these results and those of other researchers highlights the 
importance of GxE in the estimation of the genetic variance components for agronomic traits. 
In addition, the estimates of heritability for LLS resistance in this study was much higher than 
those previously reported. Also, the estimates of the additive gene effects for yield-related traits 
were either lower or higher than those reported in previous studies. The differences in additive 
genetic effects between this study and previous studies may be due to the differences in the 
populations used or difference in the experimental designs and analytical methods. It is known 
that strong selection reduces additive genetic variance, thus populations subjected to different 
selection pressure can exhibit a different distribution of genetic variance components.  
Although gene action has been investigated in various agronomic traits of interest, few studies 
focused on the estimation of maternal effects. In this study, the maternal effect was 
significantly high for LLS resistance and pod size (length and width) but small for most of the 
yield-related traits. It is worthy to mention that maternal effect has been previously reported 
for LLS resistance (Singh et al.; 1997; Pasupuleti et al., 2013). The highly significant maternal 
effect in LLS resistance suggests that the choice of the resistance donor as female parent for 
the development of the breeding population is the best approach. Also, the hundred pods weight 
and pod weight per plant showed significant female effect than male effect. Overall, the 
maternal effect was lower than additive gene effect for all traits suggesting that maternal effect, 
124 
 
University of Ghana http://ugspace.ug.edu.gh
though significant, contributes less than additive genes to the phenotypic expression of the 
agronomic traits studied. The observed differences between traits in the magnitude of the 
additive, non-additive and maternal effect is not surprising as different traits can exhibit a 
different level of heritability (He et al., 2017). In this study, LLS resistance, pod size and seed 
size exhibited higher additive gene action than the other agronomic traits whereas pod weight 
per plant and shelling outturn had a higher non-additive genetic effect than the other traits. 
These differences in gene action indicate that the agronomic traits studied will have different 
responses to selection. 
The success of a breeding programme relies upon the accuracy in the choice of the parent 
genotypes, and combining ability has been considered as an efficient tool for the identification 
of the best parents (Fasahat et al., 2016). In this study, combining ability analysis has allowed 
the identification of parents for LLS resistance and yield-related traits for utilization in genetic 
improvement of groundnut. The GCA estimate revealed genotypes such us 43AH and ICG 
7878 as the best general combiners for LLS resistance. Interestingly, 43AH also exhibited good 
GCA effect for yield-related traits such as pod width, hundred pod weight, pod weight per 
plant, and shelling percentage. These two genotypes could be used as sources of LLS resistance 
for the development of resistant inbred lines. However, the best general combiner for LLS 
resistance, ICG 7878 seems to carry negative genes for most of the yield component traits. 
Thus, the use of such genotypes for the introgression of resistant genes in a breeding 
programme should be done using a breeding methodology that allows getting rid of these 
negative alleles.  
In addition, males’ GCA effect did not show a significant difference for the number of pods 
per plant, pod length and seed width. These genotypes are preferred by farmers because of their 
earliness and relatively high number of pods per plant. This could explain the similarity of the 
GCA effect observed. In contrast, 12CS_22 and 12CS_36 which share common genetic origin 
125 
 
University of Ghana http://ugspace.ug.edu.gh
(common parents) displayed different combining ability effects for all the traits except LLS 
and seed length. This result implies that similar genetic background does not always infer 
similar combining ability.  
High SCA effects were obtained for LLS, number of pods per plant, pod weight per plant, 
hundred seed weight, hundred pod weight and shelling percentage. These SCA were obtained 
by high x high, high x low, and low x low GCA effect parents. The pedigree breeding method 
should be efficient for the development of high yielding varieties using crosses such ICGV 
02271 x 43AH and ICG7878 x 68AH as the crosses involved parents with high GCA and 
displayed high SCA effect for LLS resistance. In that case, the hope is to get good transgressive 
segregants as SCA cannot be exploited in the development of pure lines. 
It is evident that a majority of quantitative characters, in particular yield-related traits, show 
varying degree of association among themselves. Even if the combining ability is carried out 
for a number of traits simultaneously, there was a difficulty in identifying best parents or cross 
combinations for a number of traits. For that purpose, correlation analysis of GCA and SCA 
effects among evaluated traits was used in this study. Various degree of association both 
desirable and undesirable were observed between LLS and yield-related traits for combining 
ability. GCA effect for LLS resistance, hundred seed weight and shelling percentage were 
correlated in this study. Conversely, LLS exhibited negative GCA effect correlation with pod 
width. Other yield-related traits exhibited a positive correlation between themselves. GCA and 
SCA positive correlations for different traits suggest that these traits could be improved 
together through selection using parents with good GCA for at least one of the correlated traits.  
Significant interactions between environment and GCA in yield related traits indicate that 
developing breeding products that target specific environments would be more effective. It has 
been often suggested that the use of extremely divergent parents results in high SCA. In this 
study, only three traits out of eleven seem to confirm that assumption. The overall result of the 
126 
 
University of Ghana http://ugspace.ug.edu.gh
analysis of the correlation between molecular divergence and SCA suggests that the use of 
extremely divergent parents to obtain heterotic combinations for agronomic traits of interest in 
groundnut does not hold for this study. 
6.5 Conclusion 
The results suggest that additive gene action predominates over the other types of gene action 
for LLS resistance and most yield-related traits. Maternal effect seems to play an important 
role in the phenotypic variance of LLS resistance, pod and seed size. Furthermore, selection 
could be more effective when based on performance in specific environment as there was a 
significant environment by GCA effect. For that purpose, genotypes such us 43AH and 
ICG7878 are the best parents for LLS resistance and yield-related traits improvement.  
  
127 
 
University of Ghana http://ugspace.ug.edu.gh
CHAPTER SEVEN 
7. GENERAL CONCLUSIONS AND RECOMMENDATIONS 
7.1 General conclusions 
The results of the present study would help in guiding the new breeding programme from 
setting the breeding goals and strategies to the product delivery. The PRA survey revealed that, 
with few exceptions, farmers in Togo face similar production constraints. Overall, information 
on the farming practices, constraints in groundnut production, and farmers preferred 
characteristics provide the basis for a participatory breeding programme. The breeding 
programme on groundnut should consider foliar diseases, mainly LLS, as a major constraint. 
In the present state, high yielding groundnut varieties with large pod size and resistance to LLS 
are likely to be adopted by groundnut farmers.  
The diversity analysis of the cultivated groundnut cultivars in Togo revealed an appreciable 
diversity for phenotypic traits as well as molecular markers. Traits of interest such LLS 
resistance, days to maturity and hundred seed weight showed significant variation within the 
germplasm collection. At the molecular level, many SNPs harbouring large genetic variability 
were found in the groundnut collection. Besides, the diversity observed at the molecular level 
was associated with phenotypic variation, particularly for quantitative agronomic traits. The 
identification of three groups of accessions harbouring desired traits/alleles is valuable 
information for starting the groundnut breeding programme. 
The GWAS analysis revealed several new marker-trait associations for LLS resistance, 
hundred pod weight, pod length, pod width, seed length, seed width, pod reticulation, seed 
colour and growth habit. The identification of SNPs and linkage groups harbouring useful 
genes is worthy of further investigation. These SNPs could play a key role in the development 
of markers for the future breeding programme in Togo.  
128 
 
University of Ghana http://ugspace.ug.edu.gh
Most of the traits mentioned by farmers are quantitative traits with a complex mode of 
inheritance. In this study, eight parents selected based on farmers preferences were evaluated 
together with their progenies for combining ability for LLS resistance and yield-related traits. 
The results indicated that additive gene effect is predominant over other variance components 
except for pod weight per plant. Thus, applying selection for these traits in later generations 
would be efficient in biparental breeding populations. In addition, GCA effect estimation has 
identified ICG 7878 and 43AH as the best combiner for LLS resistance. The significance of 
genotypes x environment in this study for most yield-related traits, suggests that selection for 
specific environment would be efficient.  
7.2 Recommendations 
Based on the results of this study, the following points are important on the way forward for 
developing improved groundnut varieties and increasing their adoption: 
- An awareness creation to educate farmers on the important diseases and how they can 
be managed for higher productivity at harvest. 
- The establishment of a groundnut seed system to respond to the clearly expressed 
demand. 
- The new groundnut breeding programme should exploit the genetic diversity observed 
in groundnut collection from Togo. For the other traits that showed low diversity, it is 
recommended to introduce genotypes harboring these traits into the germplasm 
collection. 
- Some SNPs identified in this study are new and further investigation are recommended 
for the confirmation of these genomic regions associated with useful traits. 
- The breeding populations generated from ICG 7878 x 43AH and ICG 7878 x 68AH 
need to be further evaluated for the development of high yielding LLS resistant lines. 
129 
 
University of Ghana http://ugspace.ug.edu.gh
However, because of the negative yield-related alleles carried by ICG 7878 the newly 
developed varieties should contain a low proportion of ICG 7878 genetic background. 
- For LLS resistant lines development, it is preferable to use the resistant donor as the 
female parent in order to take advantage of the maternal effect. 
- Selection for yield should be applied in specific environments for more efficiency as 
significant GxE was observed in this study. 
- The use of single seed descent or pedigree as breeding strategies could be employed 
when targeting LLS resistance and most of the yield components due to additive gene 
effect. 
- The use of the number of pods per plant as visual selection criteria would be important 
for rapid and low-cost selection of high yielding varieties. 
  
  
130 
 
University of Ghana http://ugspace.ug.edu.gh
REFERENCES 
Abberton, M., Batley, J., Bentley, A., Bryant, J., Cai, H., Cockram, J., Costa de Oliveira, A., 
Cseke, L.J., Dempewolf, H., De Pace, C. and Edwards, D., 2016. Global agricultural 
intensification during climate change: a role for genomics. Plant Biotechnology 
Journal, 14(4), 1095-1098. 
Abdurakhmonov, I.Y. and Abdukarimov, A., 2008. Application of association mapping to 
understanding the genetic diversity of plant germplasm resources. International 
Journal of Plant Genomics, 08,1-18. 
Abedi, M. and Vahidi, F., 2011. The importance of Participatory Rural Appraisal (PRA) in 
Research. Journal of Applied Environmental and Biological Sciences, 1(8), 264-267. 
Adewi, E., Badameli K.M.S, and Dubreuil. V., 2010. Evolution of rainy seasons potentially 
useful for Togo on the 1950-2000 period. Climatology, 7, 89-108. 
Agarwal, G., Clevenger, J., Pandey, M.K., Wang, H., Shasidhar, Y., Chu, Y., Fountain, J.C., 
Choudhary, D., Culbreath, A.K., Liu, X. and Huang, G., 2018. High‐density genetic 
map using whole‐genome resequencing for fine mapping and candidate gene 
discovery for disease resistance in peanut. Plant Biotechnology Journal. 16(11), 1954-
1967. 
Agrama, H.A. and Eizenga, G.C., 2008. Molecular diversity and genome-wide linkage 
disequilibrium patterns in a worldwide collection of Oryza sativa and its wild 
relatives. Euphytica, 160(3), 339-355. 
Aguilar, A., Carranza, E., Goldstein, M., Kilic, T. and Oseni, G., 2014. Decomposition of 
gender differentials in agricultural productivity in Ethiopia. Available: 
https://elibrary.worldbank.org/doi/abs/10.1596/1813-9450-6764. Accessed on 10th 
December 2016. 
Ali, D., Bowen, D., Deininger, K. and Duponchel, M., 2015. Investigating the gender gap in 
agricultural productivity: evidence from Uganda. The World Bank. 
Almekinders, C.J.M. and Elings, A., 2001. Collaboration of farmers and breeders: Participatory 
crop improvement in perspective. Euphytica, 122(3), 425-438. 
Anderson, W., Wynne, J.C., Green, C.C. and Beute, M.K., 1986. Combining ability and 
heritability of resistance to early and late leafspot of peanut. Peanut Science, 13(1), 
10-14. 
Anderson, W.F., Fitzner, M.S., Isleib, T.G., Wynne, J.C. and Phillips, T.D., 1993a. Combining 
ability for large pod and seed traits in peanut. Peanut Science, 20(1), 49-52. 
Anderson, W.F., Holbrook, C.C. and Brenneman, T.B., 1993b. Resistance to Cercosporidium 
personatum within peanut germplasm. Peanut Science, 20(1), 53-57. 
Arnarson A. 2015. Peanuts 101: Nutrition Facts and Health Benefits. Extracted from 
https://www.healthline.com/nutrition/foods/peanuts, data:25/02/2018. 
131 
 
University of Ghana http://ugspace.ug.edu.gh
Arunachalam, V., Bandyopadhyay, A., Nigam, S.N. and Gibbons, R.W., 1984. Heterosis in 
relation to genetic divergence and specific combining ability in groundnut (Arachis 
hypogaea L.). Euphytica, 33(1), 33-39. 
Azad, M., Kalam, A., Hamid, M., Rafii, M.Y. and Malek, M.A., 2014. Combining ability of 
pod yield and related traits of groundnut (Arachis hypogaea L.) under salinity stress. 
The Scientific World Journal, 7(3), 1-7. 
Backiny-Yetna, P. and McGee, K., 2015. Gender differentials and agricultural productivity in 
Niger. (World Bank Policy Research Working Paper No. 7199). Washington, DC: The 
World Bank. 
Backman, P.A. and Crawford, M.A., 1984. Relationship between yield loss and severity of 
early and late leafspot diseases of peanut. Phytopathology, 74, 1101-1103. 
Batebana, K., Ogwang, B. A., Sein, Z. M. M., Ogou, F. K., Ongoma, V., and Ngarukiyimana, 
J. P., 2015. Rainfall Characteristics over Togo and their related Atmospheric 
circulation Anomalies. Journal of Environmental and Agricultural Sciences, 5, 34-48. 
Bates, D., Kliegl, R., Vasishth, S. and Baayen, H., 2015. Parsimonious mixed models. arXiv 
preprint arXiv, 1506, 04967. 
Bejarano, D., Martínez, R., Manrique, C., Parra, L.M., Rocha, J.F., Gómez, Y., Abuabara, Y. 
and Gallego, J., 2018. Linkage disequilibrium levels and allele frequency distribution 
in Blanco Orejinegro and Romosinuano Creole cattle using medium density SNP chip 
data. Genetics and Molecular Biology, 41(2), 426-433. 
Belamkar, V., Selvaraj, M.G., Ayers, J.L., Payton, P.R., Puppala, N. and Burow, M.D., 2011. 
A first insight into population structure and linkage disequilibrium in the US peanut 
minicore collection. Genetica, 139(4), 411. 
Bellon, M.R. and Reeves, J., 2002. Quantitative analysis of data from participatory methods 
in plant breeding. CIMMYT. Available: 
https://repository.cimmyt.org/xmlui/bitstream/handle/10883/909/448231.pdf?sequen
ce=1.  Accessed on 10th December 2017. 
Benesty, J., Chen, J., Huang, Y. and Cohen, I., 2009. Pearson correlation coefficient. In Noise 
reduction in speech processing. Springer, Berlin, Heidelberg, 1-4. 
Bhad, P.G., Mondal, S. and Badigannavar, A.M., 2016. Genetic diversity in groundnut (Arachis 
hypogaea L.) genotypes and detection of marker trait associations for plant habit and 
seed size using genomic and genic SSRs. Journal of Crop Science and Biotechnology, 
19(3), 203-221. 
Bookmyer JM, Bonos SA, and Meyer WA., 2009. Inheritance characteristics of brown patch 
resistance in tall fescue. Crop Science, 49(6), 2302-2308. 
Branca, A., Paape, T.D., Zhou, P., Briskine, R., Farmer, A.D., Mudge, J., Bharti, A.K., 
Woodward, J.E., May, G.D., Gentzbittel, L. and Ben, C., 2011. Whole-genome 
132 
 
University of Ghana http://ugspace.ug.edu.gh
nucleotide diversity, recombination, and linkage disequilibrium in the model legume 
Medicago truncatula. Proceedings of the National Academy of Sciences, 8(12), 
e201104032. 
Branch, W.D. and Fletcher, S.M., 2001. No-pesticide preliminary yield trials in peanut. Peanut 
Science, 28(1), 21-24. 
Brinez, B., Blair, M.W., Kilian, A., Carbonell, S.A.M., Chiorato, A.F. and Rubiano, L.B., 2012. 
A whole genome DArT assay to assess germplasm collection diversity in common 
beans. Molecular Breeding, 30(1), 181-193. 
Brumfield, R.T., Beerli, P., Nickerson, D.A. and Edwards, S.V., 2003. The utility of single 
nucleotide polymorphisms in inferences of population history. Trends in Ecology & 
Evolution, 18(5), 249-256. 
Bush, W.S. and Moore, J.H., 2012. Genome-wide association studies. PLoS Computational 
Biology, 8(12), e1002822. 
Byrne, B.M., 2016. Structural equation modeling with AMOS: Basic concepts, applications, 
and programming. 3rd Edition, New York, Routledge, 460. 
Coates, B. S., Sumerford, D. V., Miller, N. J., Kim, K. S., Sappington, T. W., Siegfried, B. D., 
& Lewis, L. C., 2009. Comparative performance of single nucleotide polymorphism 
and microsatellite markers for population genetic analysis. Journal of Heredity, 
100(5), 556-564. 
Cobb, J.N., DeClerck, G., Greenberg, A., Clark, R. and McCouch, S., 2013. Next-generation 
phenotyping: requirements and strategies for enhancing our understanding of 
genotype–phenotype relationships and its relevance to crop improvement. Theoretical 
and Applied Genetics, 126(4), 867-887. 
Coghlan, A., 2013. A little book of R for multivariate analysis. Wellcome Trust Sanger 
Institute, Cambridge, UK, 1-47. 
Crous, P.W., Summerell, B.A., Carnegie, A.J., Wingfield, M.J., Hunter, G.C., Burgess, T.I., 
Andjic, V., Barber, P.A. and Groenewald, J.Z., 2009. Unravelling Mycosphaerella: do 
you believe in genera?. Persoonia, 23, 99. 
Damicone, J.P. and Melouk, H.A., 2010. Foliar Diseases of Peanut. Division of Agricultural 
Sciences and Natural Resources, Oklahoma State University. Available: 
http://osufacts.okstate.edu. Accessed on 13th February 2018. 
Danial, D., Parlevliet, J., Almekinders, C. and Thiele, G., 2007. Farmers’ participation and 
breeding for durable disease resistance in the Andean region. Euphytica, 153(3), 385-
396. 
Delourme, R., Falentin, C., Fomeju, B.F., Boillot, M., Lassalle, G., André, I., Duarte, J., 
Gauthier, V., Lucante, N., Marty, A. and Pauchon, M., 2013. High-density SNP-based 
133 
 
University of Ghana http://ugspace.ug.edu.gh
genetic map development and linkage disequilibrium assessment in Brassica napus L. 
BMC Genomics, 14(1), 120. 
Direction des Statistiques Agricoles de l’Information et de la Documentation (DSID)., 2015. 
Rapport d’enquete. 
Direction des Statistiques Agricoles de l’Information et de la Documentation (DSID)., 2014. 
Principales caractéristiques de l’agriculture togolaise, 4ème recensement national de 
l’agriculture 2011-2014 / volume 4: module de base. 
Doss, C. R., and Morris, M. L., 2001. How does gender affect the adoption of agricultural 
innovations? The case of improved maize technology in Ghana. Agricultural 
Economics, 25(1), 27–39. 
Doss, C.R., Kovarik, C., Peterman, A., Quisumbing, A. and van den Bold, M., 2015. Gender 
inequalities in ownership and control of land in Africa: Myths versus reality. 
Agricultural Economics, 46(3), 403-434. 
Duraiappah A.K. Roddy, P. and Parry, J.E., 2005. Have participatory approaches increased 
capabilities?. International Institute for Sustainable Development. Winnipeg, 
Manitoba, Canada, 34. 
Dwivedi S.L, Pande S, Rao J.N, and Nigam S.N., 2002. Components of resistance to late leaf 
spot and rust among interspecific derivatives and their significance in a foliar disease 
resistance breeding in groundnut (Arachis hypogaea L.). Euphytica, 125(1), 81-88. 
Dwivedi, S.L., Crouch, J.H., Nigam, S.N., Ferguson, M.E. and Paterson, A.H., 2003. Molecular 
breeding of groundnut for enhanced productivity and food security in the semi-arid 
tropics: opportunities and challenges. Advances in Agronomy, 80, 153-221. 
Dwivedi, S.L., Scheben, A., Edwards, D., Spillane, C. and Ortiz, R., 2017. Assessing and 
exploiting functional diversity in germplasm pools to enhance abiotic stress adaptation 
and yield in cereals and food legumes. Frontiers in Plant Science, 8, 1461. 
Dwivedi, S.L., Thendapani, K. and Nigam, S.N., 1989. Heterosis and combining ability studies 
and relationship among fruit and seed characters in peanut. Peanut Science, 16(1), 14-
20. 
Elshire, R.J., Glaubitz, J.C., Sun, Q., Pol, J.A., Kawamoto, K. and Buckler, E.S., 2011. A 
robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. 
PLoS ONE, 6, e19379. 
Esteras, C., Formisano, G., Roig, C., Díaz, A., Blanca, J., Garcia-Mas, J., Gómez-Guillamón, 
M.L., López-Sesé, A.I., Lázaro, A., Monforte, A.J. and Picó, B., 2013. SNP 
genotyping in melons: genetic variation, population structure, and linkage 
disequilibrium. Theoretical and Applied Genetics, 126(5), 1285-1303. 
Fahrenkrog, A.M., Neves, L.G., Resende, M.F., Vazquez, A.I., Campos, G., Dervinis, C., 
Sykes, R., Davis, M., Davenport, R., Barbazuk, W.B. and Kirst, M., 2017. Genome‐
134 
 
University of Ghana http://ugspace.ug.edu.gh
wide association study reveals putative regulators of bioenergy traits in Populus 
deltoides. New Phytologist, 213(2), 799-811. 
Falconer, D.S. and Mackay, T.F.C., 1996. Introduction to quantitative genetics. Longman, 
Essex, UK. Introduction to quantitative genetics. 4th ed. Longman, Essex, UK. 
FAOSTAT (2018) Available: http://faostat.fao.org/. Accessed on 10th December 2018. 
Fasahat, P., Rajabi, A., Rad, J.M. and Derera, J., 2016. Principles and utilization of combining 
ability in plant breeding. Biometrics & Biostatistics International Journal, 4(1), 1-24. 
Ferguson, M.E., Bramel, P.J. and Chandra, S., 2004. Gene diversity among botanical varieties 
in peanut (Arachis hypogaea L.). Crop Science, 44(5), 1847-1854. 
Fitzgerald G., 2015. The Production of Ready to Use Therapeutic Food in Malawi. Smallholder 
farmers’ experience with groundnut production Results from a four year livelihoods 
analysis in Malawi’s Central Region. University College Cork. Available 
http://www.validnutrition.org/wp-content/uploads/2015/03/The-Production-of-
RUTF-in-Malawi-Part-1-Smallholder-famers-experience-with-groundnut-
production-.pdf Accessed on 17th July 2016. 
Fonceka, D., Tossim, H.A., Rivallan, R., Vignes, H., Faye, I., Ndoye, O., Moretzsohn, M.C., 
Bertioli, D.J., Glaszmann, J.C., Courtois, B. and Rami, J.F., 2012. Fostered and left 
behind alleles in peanut: interspecific QTL mapping reveals footprints of 
domestication and useful natural variation for breeding. BMC Plant Biology, 12(1), 
26. 
Fonceka, D., Tossim, H.A., Nguepjop, J.R., Bertioli, D.J., Leal-Bertioli, S., Jackson, S.A., 
Rami, J.F. and Ozias-Akins, P., 2018. Could small seeded wild relatives of cultivated 
peanut be used to increase the size of peanut seeds? [W661]. 
Frediani A.A. & Boano C., 2012. Processes for Just Products: The Capability Space of 
Participatory Design. In: Oosterlaken I., van den Hoven J. (eds) The Capability 
Approach, Technology and Design. Philosophy of Engineering and Technology, vol 
5. Springer, Dordrecht. 
Gajjar, K.N., Mishra, G.P., Radhakrishnan, T., Dodia, S.M., Rathnakumar, A.L., Kumar, N., 
Kumar, S., Dobaria, J.R., Kumar, A., 2014. Validation of SSR markers linked to the 
rust and late leaf spot diseases resistance in diverse peanut genotypes. Australian 
Journal of Crop Science, 8, 927-936.  
Garren, K.H. and Jackson, R.H., 1973. Peanuts-Culture and Uses. American Peanut Research 
and Education Society,37, 57-65. 
Gimenes, M.A., Hoshino, A.A., Barbosa, A.V., Palmieri, D.A. and Lopes, C.R., 2007. 
Characterization and transferability of microsatellite markers of the cultivated peanut 
(Arachis hypogaea). BMC Plant Biology, 7(1), 9. 
135 
 
University of Ghana http://ugspace.ug.edu.gh
Gomez Selvaraj, M., Narayana, M., Schubert, A.M., Ayers, J.L., Baring, M.R. and Burow, 
M.D., 2009. Identification of QTLs for pod and kernel traits in cultivated peanut by 
bulked segregant analysis. Electronic Journal of Biotechnology, 12(2), 3-4. 
Gowda, M., Kling, C., Würschum, T., Liu, W., Maurer, H.P., Hahn, V. and Reif, J.C., 2010. 
Hybrid breeding in durum wheat: heterosis and combining ability. Crop Science, 
50(6), 2224-2230. 
Gower, J. C., 1971. A general coefficient of similarity and some of its properties. Biometrics 
27, 857–871. 
Grabiele, M., Chalup, L., Robledo, G. and Seijo, G., 2012. Genetic and geographic origin of 
domesticated peanut as evidenced by 5S rDNA and chloroplast DNA sequences. Plant 
Systematics and Evolution, 298(6), 1151-1165. 
Gregory, M.P. and Gregory, W.C., 1980. Exotic germ plasm of Arachis L. interspecific 
hybrids. Journal of Heredity, 70(3), 185-193. 
Haghpanah M., Hassanzadeh A., Mirabadi A. Z., Foroozan K., and Talaee S., 2018. Evaluation 
of the relationship between yield and yield components by sequential path analysis in 
peanut (Arachis hypogaea L.) genotypes. Iranian Journal of Crop Sciences. 19(2): 
168-179. 
Halward, T.M., Stalker, H.T., Larue, E.A. and Kochert, G., 1991. Genetic variation detectable 
with molecular markers among unadapted germ-plasm resources of cultivated peanut 
and related wild species. Genome, 34(6), 1013-1020. 
Han, S., Yuan, M., Clevenger, J.P., Li, C., Hagan, A., Zhang, X., Chen, C. and He, G., 2018. 
A SNP-based linkage map revealed QTLs for resistance to early and late leaf spot 
diseases in peanut (Arachis hypogaea L.). Frontiers in Plant Science, 9, e1012. 
Harrell F.E., 2017. Miscellaneous. Version 4.0-3, Availble: https://cran.r-
project.org/web/packages/Hmisc/Hmisc.pdf. Date:14/12/2017: Accessed on 17th 
January 2016. 
He, X., Houde, A.L.S., Pitcher, T.E. and Heath, D.D., 2017. Genetic architecture of gene 
transcription in two Atlantic salmon (Salmo salar) populations. Heredity, 119(2), 117. 
Hedrick, P.W., 2005. A standardized genetic differentiation measure. Evolution, 59(8), 1633-
1638. 
Henry, D., Dymnicki, A.B., Mohatt, N., Allen, J. and Kelly, J.G., 2015. Clustering methods 
with qualitative data: a mixed-methods approach for prevention research with small 
samples. Prevention Science, 16(7), 1007-1016. 
Herselman, L., 2003. Genetic variation among Southern African cultivated peanut (Arachis 
hypogaea L.) genotypes as revealed by AFLP analysis. Euphytica, 133(3), 319-327. 
Holbrook, C. C., 1999. Testing and utilization of a core collection for the U. S. germplasm 
collection of peanut. p. 68–73. In: R. C. Johnson and T. Hodgkin (ed.), Core 
136 
 
University of Ghana http://ugspace.ug.edu.gh
collections for today and tomorrow. International Plant Genetic Resources Institute, 
Rome, Italy. 
Holbrook, C. C., 2001. Status of the United States germplasm collection of peanut. Peanut 
Science, 28, 84-89. 
Holbrook, C.C. and Anderson, W.F., 1995. Evaluation of a core collection to identify resistance 
to late leafspot in peanut. Crop Science, 35(6), 1700-1702. 
Holbrook, C.C. and Stalker, H.T., 2003. Peanut breeding and genetic resources. Plant Breeding 
Reviews, 22, 297-356. 
Houde, A.L.S. and Pitcher, T.E., 2016. fullfact: an R package for the analysis of genetic and 
maternal variance components from full factorial mating designs. Ecology and 
Evolution, 6(6), 1656-1665. 
Huang, B.E., George, A.W., Forrest, K.L., Kilian, A., Hayden, M.J., Morell, M.K. and 
Cavanagh, C.R., 2012. A multiparent advanced generation inter‐cross population for 
genetic analysis in wheat. Plant Biotechnology Journal, 10(7), 826-839. 
IBPGR, ICRISAT. 1992. Descriptors for groundnut. Rome, Italy and Patancheru, AP, India: 
Int. Board of Plant Genetic Resources and Int. Crops Res. Inst. For the Semi-Arid 
Tropics. Extracted from 
https://www.researchgate.net/profile/Abdallah_Bari/publication/311512128_Descrip
teurs_pour_l%27arachide/links/58529f2208ae7d33e01aa9ac/Descripteurs-pour-
larachide.pdf.  Accessed on 24th January 2015. 
Isleib, T.G., Holbrook, C.C. and Gorbet, D.W., 2001. Use of plant introductions in peanut 
cultivar development. Peanut Science, 28(2), 96-113. 
Izge, A.U., Mohammed, Z.H. and Goni, A., 2007. Levels of variability in groundnut (Arachis 
hypogaea L.) to cercospora leaf spot disease implication for selection. African 
Journal of Agricultural Research, 2(4), 182-186. 
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(4), 25-35. 
Janila, P., Nigam, S.N., Pandey M.K., Nagesh, P. & Varshney, R.K., 2013. Groundnut 
improvement: use of genetic and genomic tools. Frontiers in Plant Genetics and 
Genomic, 4(23), 1-16. 
Janila, P., Pandey, M.K., Shasidhar, Y., Variath, M.T., Sriswathi, M., Khera, P., Manohar, S.S., 
Nagesh, P., Vishwakarma, M.K., Mishra, G.P. and Radhakrishnan, T., 2016a. 
Molecular breeding for introgression of fatty acid desaturase mutant alleles (ahFAD2A 
and ahFAD2B) enhances oil quality in high and low oil containing peanut genotypes. 
Plant Science, 242, 203-213. 
137 
 
University of Ghana http://ugspace.ug.edu.gh
Janila, P., Variath, M.T., Pandey, M.K., Desmae, H., Motagi, B.N., Okori, P., Manohar, S.S., 
Rathnakumar, A.L., Radhakrishnan, T., Liao, B. and Varshney, R.K. 2016b., Genomic 
tools in groundnut breeding program: status and perspectives. Frontiers in Plant 
Science, 7, 289. 
Jarvis, A., Ferguson, M.E., Williams, D.E., Guarino, L., Jones, P.G., Stalker, H.T., Valls, J.F., 
Pittman, R.N., Simpson, C.E. and Bramel, P., 2003. Biogeography of wild Arachis. 
Crop Science, 43(3), 1100-1108. 
Jogloy, S., Kormsa-art, T., Wongkaew, S. and Lertrat, K., 2000. Combining ability of late 
leafspot resistance and agronomic traits in peanut (Arachis hypogaea L.). 
Songklanakarin Journal of Science and Technology, 27(2), 213-221. 
Jogloy, S., Wynne, J.C. and Beute M.K., 1987. Inheritance of late leaf spot resistance and 
agronomic traits in peanut. Peanut Science, 14, 86-90. 
John, K., Reddy, P.R. and Reddy, P.H., 2012. General and specific combining ability estimates 
of physiological traits for moisture stress tolerance in groundnut (Arachis hypogaea 
L.). International Journal of Applied Biology and Pharmaceutical Technology, 2(4), 
1-12. 
Johnson, N.L., Kovarik, C., Meinzen-Dick, R., Njuki, J. and Quisumbing, A., 2016. Gender, 
assets, and agricultural development: Lessons from eight projects. World 
Development, 83, 295-311. 
Johnson, R. A., and Wichern, D. W. 2014. Applied multivariate statistical analysis, 4, New 
Jersey: Prentice-Hall. 
Jonnala, R. S., Dunford, N. T., and Chenault, K., 2005. Nutritional composition of genetically 
modified peanut varieties. Journal of Food Science, 70(4). 
Kanyika, B.T.N., Lungu, D., Mweetwa, A.M., Kaimoyo, E., Njung'e, V.M., Monyo, E.S., 
Siambi, M., He, G., Prakash, C.S., Zhao, Y. and de Villiers, S.M. 2015. Identification 
of groundnut (Arachis hypogaea) SSR markers suitable for multiple resistance traits 
QTL mapping in African germplasm. Electronic Journal of Biotechnology, 18(2), 61-
67. 
Kaufman, L., and Rousseeuw, P. J., 2009. Finding groups in data: an introduction to cluster 
analysis. John Wiley & Sons, 334, 234. 
Khan, N.U., Hassan, G., Marwat, K.B., Kumbhar, M.B., Khan, I., Soomro, Z.A., Baloch, M.J. 
and Khan, M.Z., 2009. Legacy study of cottonseed traits in upland cotton using 
Griffing’s combining ability model. Pakistan Journal of Botany, 41(1), 131-142. 
Khedikar, Y.P., Gowda, M.V.C., Sarvamangala, C., Patgar, K.V., Upadhyaya, H.D. and 
Varshney, R.K., 2010. A QTL study on late leaf spot and rust revealed one major QTL 
for molecular breeding for rust resistance in groundnut (Arachis hypogaea L.). 
Theoretical and Applied Genetics, 121(5), 971-984. 
138 
 
University of Ghana http://ugspace.ug.edu.gh
Khera, P., Pandey, M.K., Mallikarjuna, N., Sriswathi, M., Roorkiwal, M., Janila, P., Sharma, 
S., Shilpa, K., Sudini, H., Guo, B. and Varshney, R.K., 2018. Genetic imprints of 
domestication for disease resistance, oil quality, and yield component traits in 
groundnut (Arachis hypogaea L.). Molecular Genetics and Genomics, 294(2), 365-
378. 
Kieran, C., Sproule, K., Doss, C., Quisumbing, A., and Kim, S. M., 2015. Examining gender 
inequalities in land rights indicators in Asia. Agricultural Economics, 46(S1), 119-
138. 
Kline, R.B., 2015. Principles and practice of structural equation modeling. Guilford 
Publications, 534. 
Knauft, D.A., Chiyembekeza, A.J., and Gorbet, D.W., 1992. Possible reproductive factors 
contributing to outcrossing in peanut (Arachis hypogaea L.). Peanut Science, 19, 29-
31. 
Kochert, G., Stalker, H.T., Gimenes, M., Galgaro, L., Lopes, C.R. and Moore, K., 1996. RFLP 
and cytogenetic evidence on the origin and evolution of allotetraploid domesticated 
peanut, Arachis hypogaea (Leguminosae). American Journal of Botany, 83(10), 1282-
1291. 
Kolech, S.A., De Jong, W., Perry, K., Halseth, D. and Mengistu, F., 2017. Participatory Variety 
Selection: A Tool to Understand Farmers′ Potato Variety Selection Criteria. Open 
Agriculture, 2(1), 453-463. 
Kosman, E. and Leonard, K.J., 2005. Similarity coefficients for molecular markers in studies 
of genetic relationships between individuals for haploid, diploid, and polyploid 
species. Molecular ecology, 14(2), pp.415-424. 
Kottapalli, P., Upadhyaya, H.D., Kottapalli, K.R., Payton, P., Dwivedi, S., Burow, M., David, 
K.O., Sanogo, S. and Puppala, N., 2011. Population structure and diversity in Valencia 
peanut germplasm collection. Crop Science, 51(3), 1089-1100. 
Kraaijvanger, R., Veldkamp, T. and Almekinders, C., 2016. Considering change: Evaluating 
four years of participatory experimentation with farmers in Tigray (Ethiopia) 
highlighting both functional and human–social aspects. Agricultural Systems, 147, 38-
50. 
Krapovickas, A. and Gregory, W.C., 1994. Taxonomia del genero" Arachis (Leguminosae)". 
Bonplandia, 8, 1-186. 
Krapovickas, A., Vanni, R.O., Pietrarelli, J.R., Simpson, C.E., 2013. The peanut landraces from 
Perú. Bonplandia, 22, 19-90. 
Krapovickas, A., Vanni, R.O., Pietrarelli, J.R., Williams, D.E., and Simpson, C.E., 2009. Las 
Razas de Maní de Bolivia. Bonplandia, 18, 95-189. 
139 
 
University of Ghana http://ugspace.ug.edu.gh
Kris-Etherton, P. M., Hu, F. B., Ros, E., and Sabaté, J., 2008. The role of tree nuts and peanuts 
in the prevention of coronary heart disease: multiple potential mechanisms. The 
Journal of Nutrition, 138(9), 1746-1751. 
Kristjanson, P., Bryan, E., Bernier, Q., Twyman, J., Meinzen-Dick, R., Kieran, C., Ringler, C., 
Jost, C. and Doss, C., 2017. Addressing gender in agricultural research for 
development in the face of a changing climate: where are we and where should we be 
going?. International Journal of Agricultural Sustainability, 15(5), 482-500. 
Kushwah, A., 2011. studies on genetic diversity of groundnut (Arachis hypogaea l.) using 
morphological and molecular markers (Doctoral dissertation). Accessed on 
http://krishikosh.egranth.ac.in/bitstream/1/96503/1/MSc%20Th%204075%20Kushw
ah%20A.pdf. 
Langfelder, P., Zhang, B. and Horvath, S., 2007. Defining clusters from a hierarchical cluster 
tree: the Dynamic Tree Cut package for R. Bioinformatics, 24(5), 719-720. 
Latif, S., Pfannstiel, J., Makkar, H. P. S., and Becker, K., 2013. Amino acid composition, 
antinutrients and allergens in the peanut protein fraction obtained by an aqueous 
enzymatic process. Food Chemistry, 136(1), 213-217. 
Lavia, G.I. and Fernández, A., 2008. Genome size in wild and cultivated peanut germplasm. 
Plant Systematics and Evolution, 272(1-4), 1-10. 
Leal-Bertioli, S., Shirasawa, K., Abernathy, B., Moretzsohn, M., Chavarro, C., Clevenger, J., 
OziasAkins, P., Jackson, S., and Bertioli, D., 2015. Tetrasomic recombination is 
surprisingly frequent in allotetraploid Arachis. Genetics, 199, 1093-1105. 
Leal-Bertioli, S.C., José, A.C.V., Alves-Freitas, D.M., Moretzsohn, M.C., Guimarães, P.M., 
Nielen, S., Vidigal, B.S., Pereira, R.W., Pike, J., Fávero, A.P. and Parniske, M., 2009. 
Identification of candidate genome regions controlling disease resistance in Arachis. 
BMC Plant Biology, 9(1), 112. 
Li, T. Y., Brennan, A. M., Wedick, N. M., Mantzoros, C., Rifai, N., and Hu, F. B., 2009. 
Regular consumption of nuts is associated with a lower risk of cardiovascular disease 
in women with type 2 diabetes. The Journal of Nutrition, 139(7), 1333-1338. 
Li, X., Han, Y., Wei, Y., Acharya, A., Farmer, A.D., Ho, J., Monteros, M.J. and Brummer, 
E.C., 2014. Development of an alfalfa SNP array and its use to evaluate patterns of 
population structure and linkage disequilibrium. PLoS One, 9(1), e84329. 
Liao, B., and Holbrook, C., 2007. Groundnut. In Genetic Resources, Chromosome 
Engineering, and Crop Improvement Series: Oil-seed crops, Vol. 4. Edited by R.J. 
Singh. CRC Press, Inc., Boca Raton, Fla, 51–87. 
Liu, Z., Feng, S., Pandey, M.K., Chen, X., Culbreath, A.K., Varshney, R.K. and Guo, B., 2013. 
Identification of expressed resistance gene analogs from peanut (Arachis hypogaea 
L.) expressed sequence tags. Journal of Integrative Plant Biology, 55(5), 453-461.  
140 
 
University of Ghana http://ugspace.ug.edu.gh
Luikart, G., England, P.R., Tallmon, D., Jordan, S. and Taberlet, P., 2003. The power and 
promise of population genomics: from genotyping to genome typing. Nature Reviews 
Genetics, 4(12), 981. 
Lumley, T., 2011. Complex surveys: a guide to analysis using R (Vol. 565). John Wiley & 
Sons. Available: https://faculty.washington.edu/tlumley/old-survey/survey-wss.pdf. 
Data:04/05/2016. Date: 13/02/2016. Accessed on 14th June 2017. 
Luo, M., Dang, P., Bausher, M.G., Holbrook, C.C., Lee, R.D., Lynch, R.E. and Guo, B.Z., 
2005. Identification of transcripts involved in resistance responses to leaf spot disease 
caused by Cercosporidium personatum in peanut (Arachis hypogaea). 
Phytopathology, 95(4), 381-387. 
Lynch, M., and Walsh, B., 1998. Genetics and analysis of quantitative traits. Sunderland, MA, 
Sinauer, 1, 535-557. 
Mallikarjuna, N. and Varshney, R.K., 2014. Genetics, genomics and breeding of peanuts. CRC 
Press. 
Malysheva-Otto, L.V., Ganal, M.W. and Röder, M.S., 2006. Analysis of molecular diversity, 
population structure and linkage disequilibrium in a worldwide survey of cultivated 
barley germplasm (Hordeum vulgare L.). BMC Genetics, 7(1), 6. 
Mammadov, J., Aggarwal, R., Buyyarapu, R. and Kumpatla, S., 2012. SNP markers and their 
impact on plant breeding. International Journal of Plant Genomics, 12, 1-11. 
Manivannan, N., Muralidharan, V. and Mothilal, A., 2008. Combining ability analysis in 
groundnut (Arachis hypogaea L.). Madras Agricultural Journal, 95(1-6), 14-17. 
Manzanilla, D.O., Paris, T.R., Vergara, G.V., Ismail, A.M., Pandey, S., Labios, R.V., 
Tatlonghari, G.T., Acda, R.D., Chi, T.T.N., Duoangsila, K. and Siliphouthone, I., 
2011. Submergence risks and farmers’ preferences: implications for breeding Sub1 
rice in Southeast Asia. Agricultural Systems, 104(4), 335-347. 
Mayee, C.D. and Datar, V.V., 1988. Diseases of groundnut in the tropics. Review of Tropical 
Plant Pathology, 5, 85-118. 
McDonald, D., Subrahmanyam, P., Gibbons, R.W. and Smith, D.H., 1985. Early and late leaf 
spots of groundnut. Information Bulletin no. 21. 
McKevith, B., 2005. Nutritional aspects of oilseeds. Nutrition Bulletin, 30(1), 13-26. 
Milla, S.R., 2003. Relationship and Utilization of Arachis Germplasm in Peanut Improvement 
(Ph.D. dissertation). North Carolina State University, Raleigh. Extracted from 
https://repository.lib.ncsu.edu/bitstream/handle/1840.16/3932/etd.pdf?sequence=1&i
sAllowed=y. Accessed on 21 December 2016. 
Milla-Lewis, S. R., Zuleta, M., and Isleib, T. G., 2010. Assessment of genetic diversity among 
US runner-type peanut cultivars using simple sequence repeat markers. Crop Science, 
50(6), 2396-2405. 
141 
 
University of Ghana http://ugspace.ug.edu.gh
Mishra, G.P., Radhakrishnan, T., Kumar, A., Thirumalaisamy, P.P., Kumar, N., Bosamia, T.C., 
Nawade, B. and Dobaria, J.R., 2015. Advancements in molecular marker development 
and their applications in the management of biotic stresses in peanuts. Crop 
Protection, 77, 74-86. 
Mondal, M.R.I. and Wahhab, M.A., 2001. Production technology of oilcrops. Oilseed 
Research Centre, BARI, Joydebpu, Gazipur, 111. 
Moretzsohn, M.C., Gouvea, E.G., Inglis, P.W., Leal-Bertioli, S.C., Valls, J.F. and Bertioli, 
D.J., 2013. A study of the relationships of cultivated peanut (Arachis hypogaea) and 
its most closely related wild species using intron sequences and microsatellite markers. 
Annals of Botany, 111(1), 113-126. 
Morris, M.L. and Bellon, M.R., 2004. Participatory plant breeding research: opportunities and 
challenges for the international crop improvement system. Euphytica, 136(1), 21-35. 
Mothilal, A. and Ezhil, A., 2010. Combining ability analysis for yield and its components in 
groundnut (Arachis hypogaea L.). Electronic Journal of Plant Breeding, 1(2), 162-
166. 
Mukankusi C, Derera J, Melis R., Gibson P.T., and Buruchara R., 2011. Genetic analysis of 
resistance to Fusarium root rot in common bean. Euphytica, 182, 11-23. 
Mulder, J.L. and Holliday, P., 1974. CMI descriptions of pathogenic fungi and bacteria. CMI 
Descriptions of Pathogenic Fungi and Bacteria, 42, 411-420. 
Nath, B.C., Singh, J.P., Srivastava, S. and Singh, R.B., 2013. Management of late leaf spot of 
groundnut by different fungicides and their impact on yield. Plant Pathology Journal, 
12, 85-91. 
Neff, B. D., and T. E. Pitcher., 2005. Genetic quality and sexual selection: an integrated 
framework for good genes and compatible genes. Molecular Ecology, 14,19-38. 
Neff, B. D., S. R. Garner, and T. E. Pitcher., 2011. Conservation and enhancement of wild fish 
populations: preserving genetic fitness versus genetic diversity. Canadian Journal of 
Fisheries and Aquatic Sciences, 68, 1149-1154. 
Nemli, S., Aşçioğul, T.K., Ateş, D., Eşiyok, D. and Tanyolac, M.B., 2017. Diversity and 
genetic analysis through DArTseq in common bean (Phaseolus vulgaris L.) 
germplasm from Turkey. Turkish Journal of Agriculture and Forestry, 41(5), 389-
404. 
Nielsen, N.H., Backes, G., Stougaard, J., Andersen, S.U. and Jahoor, A., 2014. Genetic 
diversity and population structure analysis of European hexaploid bread wheat 
(Triticum aestivum L.) varieties. PLoS One, 9(4), e94000. 
Nigam, S.N. and Blummel, M., 2010. Cultivar-dependent variation in food-feed-traits in 
groundnut (Arachis hypogaea L.). Animal Feed Science and Technology, 10, 39-48. 
142 
 
University of Ghana http://ugspace.ug.edu.gh
Noble, H., and Smith, J., 2014. Qualitative data analysis: a practical example. Evidence-based 
Nursing, 17(1), 2-3. 
Novoselović, D., Bentley, A.R., Šimek, R., Dvojković, K., Sorrells, M.E., Gosman, N., 
Horsnell, R., Drezner, G. and Šatović, Z., 2016. Characterizing Croatian wheat 
germplasm diversity and structure in a European context by DArT markers. Frontiers 
in Plant Science, 7, 184. 
Odong, T. L., Van Heerwaarden, J., Jansen, J., van Hintum, T. J., and Van Eeuwijk, F. A. 2011. 
Determination of genetic structure of germplasm collections: are traditional 
hierarchical clustering methods appropriate for molecular marker data?. Theoretical 
and Applied Genetics, 123(2), 195-205. 
Oppong-Sekyere, D., Akromah, R., Akpalu, M. M., Ninfaa, A. D., Nyamah, E. Y., Braimah, 
M. M., and Salifu, A. R. S., 2015. Participatory Rural Appraisal of Constraints to 
Groundnut (Arachis hypogaea L.) Production in Northern Ghana. International 
Journal of Current Research and Academic Review, 3(10), 54-76. 
Ortiz, R. and Golmirzaie, A., 2002. Hierarchical and factorial mating designs for quantitative 
genetic analysis in tetrasomic potato. Theoretical and Applied Genetics, 104(4), 675-
679. 
Ortiz-Ferrara, G., Joshi, A.K., Chand, R., Bhatta, M.R., Mudwari, A., Thapa, D.B., Sufian, 
M.A., Saikia, T.P., Chatrath, R., Witcombe, J.R. and Virk, D.S., 2007. Partnering with 
farmers to accelerate adoption of new technologies in South Asia to improve wheat 
productivity. Euphytica, 157(3), 399-407. 
Oseni, G., Corral, P., Goldstein, M., and Winters, P., 2015. Explaining gender differentials in 
agricultural production in Nigeria. Agricultural Economics, 46(3), 285–310. 
Pande, S., Rao, J.N., Upadhyaya, H.D. and Lenne, J.M., 2001. Farmers' participatory integrated 
management of foliar diseases of groundnut. International Journal of Pest 
Management, 47(2), pp.121-126. 
Pande, S., Bandyopadhyay, R., Blümmel, M., Rao, J.N., Thomas, D. and Navi, S.S., 2003. 
Disease management factors influencing yield and quality of sorghum and groundnut 
crop residues. Field Crops Research, 84(1-2), pp.89-103. 
Pandey, M.K., Monyo, E., Ozias-Akins, P., Liang, X., Guimaraes, P., Nigam, S.N., Upadhyaya, 
H.D., Janila, P., Zhang, X., Guo, B., Bertioli, D.J., Michelmore, R., and Varshney, 
R.K., 2012. Advances in Arachis genomics for peanut improvement. Biotechnology 
Advances, 30, 639-651. 
Pandey, M.K., Upadhyaya, H.D., Rathore, A., Vadez, V., Sheshshayee, M.S., Sriswathi, M., 
Govil, M., Kumar, A., Gowda, M.V.C., Sharma, S. and Hamidou, F., 2014. 
Genomewide association studies for 50 agronomic traits in peanut using the ‘reference 
set’comprising 300 genotypes from 48 countries of the semi-arid tropics of the world. 
PLoS One, 9(8), e105228. 
143 
 
University of Ghana http://ugspace.ug.edu.gh
Pandey, M.K., Wang, H., Khera, P., Vishwakarma, M.K., Kale, S.M., Culbreath, A.K., 
Holbrook, C.C., Wang, X., Varshney, R.K. and Guo, B., 2017. Genetic dissection of 
novel QTLs for resistance to leaf spots and tomato spotted wilt virus in peanut 
(Arachis hypogaea L.). Frontiers in Plant Science, 8, 25. 
Pasupuleti, J., Ramaiah, V., Rathore, A., Rupakula, A., Reddy, R.K., Waliyar, F. and Nigam, 
S.N., 2013. Genetic analysis of resistance to late leaf spot in interspecific groundnuts. 
Euphytica, 193(1), 13-25. 
Patil, K.S., Kenchanagoudar, P.V. and Kulmi, M.R.M., 2017. Combining Ability Studies for 
Yield and Its Component Traits in Groundnut (Arachis hypogaea L.). International 
Journal of Current Microbiology and Applied Science, 6(8), 336-342. 
Pattee, H.E. and Mohapatra, S.C., 1987. Anatomical changes during ontogeny of the peanut 
(Arachis hypogaea L.) fruit: mature megagametophyte through heart shaped embryo. 
Botanical Gazette, 148, 156-164.  
Pattee, H.E. and Young, C.T., 1982. Peanut science and technology. American Peanut Research 
and Education Society. Available: http://www.nal.usda.gov/ Accessed on 14th 
December 2016.  
Pavoine, S., Vallet, J., Dufour, A.B., Gachet, S. and Daniel, H., 2009. On the challenge of 
treating various types of variables: application for improving the measurement of 
functional diversity. Oikos, 118(3), 391-402. 
Pearson, T.A. and Manolio, T.A., 2008. How to interpret a genome-wide association study. 
Jama, 299(11), 1335-1344. 
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., 
Prettenhofer, P., Weiss, R., Dubourg, V. and Vanderplas, J., 2011. Scikit-learn: 
Machine learning in Python. Journal of Machine Learning Research, 12, 2825-2830. 
Piepho, H.P., Möhring, J., Melchinger, A.E. and Büchse, A., 2008. BLUP for phenotypic 
selection in plant breeding and variety testing. Euphytica, 161(1-2), 209-228. 
Podani, J. and Schmera, D., 2007. How should a dendrogram‐based measure of functional 
diversity function? A rejoinder to Petchey and Gaston. Oikos, 116(8), 1427-1430. 
Porto-Neto, L.R., Kijas, J.W. and Reverter, A., 2014. The extent of linkage disequilibrium in 
beef cattle breeds using high-density SNP genotypes. Genetics Selection Evolution, 
46(1), 22. 
Prasad, V L and Bezkorowajnyj, P G and Nigam, S N and Hanson, J and Romney, D (2006) 
Participatory Varietal Selection to Multiple Actor Orientation–Case study of groundnut in 
Anantapur, Andhra Pradesh. In: International Conference on social science perspectives 
in agricultural Research and development, February 15-18, 2006, New Delhi, India. 
Räsänen, K. and Kruuk, L.E.B., 2007. Maternal effects and evolution at ecological time‐scales. 
Functional Ecology, 21(3), 408-421. 
144 
 
University of Ghana http://ugspace.ug.edu.gh
Ren, X., Jiang, H., Yan, Z., Chen, Y., Zhou, X., Huang, L., Lei, Y., Huang, J., Yan, L., Qi, Y. 
and Wei, W., 2014. Genetic diversity and population structure of the major peanut 
(Arachis hypogaea L.) cultivars grown in China by SSR markers. PLoS One, 9(2), 88-
91. 
Roomi, S., Sabiha, B., Iqbal, A., Suleman, M., Muhammad, I., Zia, M.A., Ahmad, M.Z., 
Rashid, F., Ghafoor, A. and Tabbasam, N., 2014. SSR based genetic diversity analysis 
in a diverse germplasm of groundnut (Arachis hypogaea L.) from Pakistan. Australian 
Journal of Crop Science, 8(1), 55. 
ROPPA. 2013. Ten years after Maputo Declaration on Agriculture and Food Security: An 
assessment of Progress in West Africa, Case of Togo. Available: http://www.roppa-
afrique.org/IMG/pdf/togo_rapport_final-kf.pdf. Accessed on 15th December 2017. 
Ros, E., 2010. Health benefits of nut consumption. Nutrients, 2(7), 652-682. 
Sales, J. M., and Resurreccion, A. V., 2014. Resveratrol in peanuts. Critical Reviews in Food 
Science and Nutrition, 54(6), 734-770. 
Sattar, R. S., Wang, S., Muqadas, M., and Ashraf, M. F., 2017. Qualitative and quantitative 
approaches to study adoption of sustainable agricultural practices: A research-note on 
mixed method approach. International Journal of Agricultural Extension and Rural 
Development, 5(2), 539-544. 
Sebei, K., Gnouma, A., Herchi, W., Sakouhi, F., and Boukhchina, S., 2013. Lipids, proteins, 
phenolic composition, antioxidant and antibacterial activities of seeds of peanuts 
(Arachis hypogaea L.) cultivated in Tunisia. Biological Research, 46(3), 257-263. 
Seijo, J.G., Lavia, G.I., Fernández, A., Krapovickas, A., Ducasse, D.A., Bertioli, D.J., 
Moscone, E.A., 2007. Genomic relationships between the cultivated peanut (Arachis 
hypogaea L.) and its close relatives revealed by double GISH. American Journal of 
Botany, 94, 1963-19710. 
Sharma, S., Upadhyaya, H.D., Varshney, R.K. and Gowda, C.L.L., 2013. Pre-breeding for 
diversification of primary gene pool and genetic enhancement of grain legumes. 
Frontiers in Plant Science, 4, 309. 
Shasidhar, Y., Vishwakarma, M.K., Pandey, M.K., Janila, P., Variath, M.T., Manohar, S.S., 
Nigam, S.N., Guo, B. and Varshney, R.K., 2017. Molecular Mapping of Oil Content 
and Fatty Acids Using Dense Genetic Maps in Groundnut (Arachis hypogaea L.). 
Frontiers in Plant Science, 8, 794. 
Shilpa, K., Sunkad, G., Srinivasu, K., Marri, S., Padmashree, K., Jadhav, D.R., Sahrawat, K.L. 
and Mallikarjuna, N., 2013. Biochemical composition and disease resistance in newly 
synthesized amphidiploid and autotetraploid peanuts. Food and Nutrition Sciences, 
4(2), 169-176. 
145 
 
University of Ghana http://ugspace.ug.edu.gh
Shoba, D., Manivannan, N., Vindhiyavarman, P. and Nigam, S.N., 2012. SSR markers 
associated for late leaf spot disease resistance by bulked segregant analysis in 
groundnut (Arachis hypogaea L.). Euphytica, 188(2), 265-272. 
Shokes, F. M., and A. K. Culbreath., 1997. Early and late leaf spots. p. 17–20. In: N. 
KokalisBurelle, D. M. Porter, R. Rodriguez-Kabana, D. H. Smith, and P. 
Subrahmanyam (eds.), Compendium of peanut diseases. 2nd ed. APS Press, American 
Phytopathological Society, St. Paul, MN. 
Simpson, C., Starr, J., Cunfer, B.M., Johnson, J.W., Buntin, G.D., Barnett, R.D., Roberts, J.J., 
Morris, C.F., Konzak, C.F., Allan, R.E. and Quick, J.S., 2001. Registration of 
‘COAN’ peanut. Crop Science, 41, 918. 
Simpson, E. H., 1949. Measurement of diversity. Nature, 163, 688. 
Singh A.K., Mehan V.K., and Nigam S.N., 1997. Source of Resistance to Groundnut Fungal 
and Bacterial Wilt Diseases: an Update and Appraisal. Information bulletin Nο. 50, 
ICRISAT, Patancheru, India. 
Smith D.H., Pauer G.D.C., and Shokes F.M., 1992. Plant disease of international importance. 
Diseases of Vegetables and Oil seed. p.232. 
Sokal, R. R., and Rohlf, F. J. 1973. The comparison of dendrograms by objective methods. 
Taxon, 11(2), 33-40. 
Soleri, D. and Cleveland, D.A., 2004. Farmer selection and conservation of crop varieties. 
Encyclopedia of Plant and Crop Science, 2004, 433-438. 
Sprague, G.F. and Tatum, L.A., 1942. General vs. Specific Combining Ability in Single 
Crosses of Corn 1. Agronomy Journal, 34(10), 923-932. 
Stalker, H. T., and C. E. Simpson., 1995. Genetic resources in Arachis. Advances in Peanut 
Science, 8, 14-53. 
Stalker, H.T., Tallury, S.P., Seijo, G.R. and Leal-Bertioli, S.C., 2016. Biology, speciation, and 
utilization of peanut species. In Peanuts Genetics, Processing, and Utilization Ch. 2s 
(27-66). AOCS Press. 
Stich, B., Melchinger, A.E., Piepho, H.P., Hamrit, S., Schipprack, W., Maurer, H.P. and Reif, 
J.C., 2007. Potential causes of linkage disequilibrium in a European maize breeding 
program investigated with computer simulations. Theoretical and Applied Genetics, 
115(4), 529-536. 
Subrahmanyam, A.M., Ghanekar, B.L., Nolt, B.L., Reddy, D.V.R. and McDonald, D., 1985. 
Resistance to groundnut diseases in wild Arachis species [a review article]. In 
International Workshop on Cytogenetics of Arachis, Patancheru, AP (India), 31 Oct-
2 Nov 1983. ICRISAT, 49-55. 
Subrahmanyam, P., McDonald, D., Waliyar, F., Reddy, L.J., Nigam, S.N., Gibbons, R.W., Rao, 
V.R., Singh, A.K., Pande, S., Reddy, P.M. and Rao, P.S., 1995. Screening methods 
146 
 
University of Ghana http://ugspace.ug.edu.gh
and sources of resistance to rust and late leaf spot of groundnut. Information Bulletin 
no. 47. International Crops Research Institute for the Semi-Arid Tropics. 
Subrahmanyam, P., Williams, J.H., McDonald, D. and Gibbons, R.W., 1984. The influence of 
foliar diseases and their control by selective fungicides on a range of groundnut 
(Arachis hypogaea L.) genotypes. Annals of Applied Biology, 104(3), 467-476. 
Sujay, V., Gowda, M.V.C., Pandey, M.K., Bhat, R.S., Khedikar, Y.P., Nadaf, H.L., Gautami, 
B., Sarvamangala, C., Lingaraju, S., Radhakrishan, T. and Knapp, S.J., 2012. 
Quantitative trait locus analysis and construction of consensus genetic map for foliar 
disease resistance based on two recombinant inbred line populations in cultivated 
groundnut (Arachis hypogaea L.). Molecular Breeding, 30(2), 773-788. 
Sukruth, M., Paratwagh, S.A., Sujay, V., Kumari, V., Gowda, M.V.C., Nadaf, H.L., Motagi, 
B.N., Lingaraju, S., Pandey, M.K., Varshney, R.K. and Bhat, R.S., 2015. Validation 
of markers linked to late leaf spot and rust resistance, and selection of superior 
genotypes among diverse recombinant inbred lines and backcross lines in peanut 
(Arachis hypogaea L.). Euphytica, 204(2), 343-351. 
Tang, R., Gao, G., He, L., Han, Z., Shan, S., Zhong, R., Zhou, C., Jiang, J., Li, Y. and Zhuang, 
W., 2007. Genetic diversity in cultivated groundnut based on SSR markers. Journal 
of Genetics and Genomics, 34(5), 449-459. 
Thompson, S.K., 2012. Stratified Sampling, Wiley, Hoboken, 139-156. 
Tshilenge-Luk, L., Nkongolo, K.K.C., Narendrula, R., Kalonji-Mbuyi, A. and Kizungu, R.V., 
2012. Molecular characterization of groundnut (Arachis hypogaea L.) accessions 
from a gene pool: Application of gamma ray radiations. Journal of Plant Breeding 
and Crop Science, 4(11), 175-183. 
Turner, S.D., 2014. qqman: an R package for visualizing GWAS results using QQ and 
manhattan plots. BioRxiv, e005165. 
Upadhyaya, H.D., Dwivedi, S.L., Ambrose, M., Ellis, N., Berger, J., Smýkal, P., Debouck, D., 
Duc, G., Dumet, D., Flavell, A. and Sharma, S.K., 2011. Legume genetic resources: 
management, diversity assessment, and utilization in crop improvement. Euphytica, 
180(1), 27-47. 
Upadhyaya, H.D., Ortiz, R., Bramel, P.J. and Singh, S., 2003. Development of a groundnut 
core collection using taxonomical, geographical and morphological descriptors. 
Genetic Resources and Crop Evolution, 50(2), 139-148. 
Upadhyaya, H.D., Pundir, R.P.S., Dwivedi, S.L. and Gowda, C.L.L., 2009. Mini Core 
Collections for Efficient Utilization of Plant Genetic Resources in Crop Improvement 
Programs. Information Bulletin No. 78. International Crops Research Institute for the 
Semi-Arid Tropics. 
147 
 
University of Ghana http://ugspace.ug.edu.gh
Upadhyaya, H.D., Reddy, L.J., Gowda, C.L.L. and Singh, S., 2006. Identification of diverse 
groundnut germplasm: Sources of early maturity in a core collection. Field Crops 
Research, 97(2-3), 261-27. 
Valls, J.F. and Simpson, C.E., 2005. New species of Arachis (leguminosae) from Brazil, 
Paraguay and Bolivia. Bonplandia, 7, 35-63. 
Valls, J.F., DA Costa, L.C. and Custodio, A.R., 2013. A novel trifoliolate species of Arachis 
(Fabaceae) and further comments on the taxonomic section Trierectoides. 
Bonplandia, 12, 91-97. 
van Berloo, R., 2008. GGT 2.0: versatile software for visualization and analysis of genetic data. 
Journal of Heredity, 99(2), 232-236. 
Varshney, R.K., Mahendar, T., Aruna, R., Nigam, S.N., Neelima, K., Vadez, V. and 
Hoisington, D.A., 2009. High level of natural variation in a groundnut (Arachis 
hypogaea L.) germplasm collection assayed by selected informative SSR markers. 
Plant Breeding, 128(5), 486-494. 
Varshney, R.K., Pandey, M.K., Janila, P., Nigam, S.N., Sudini, H., Gowda, M.V.C., Sriswathi, 
M., Radhakrishnan, T., Manohar, S.S. and Nagesh, P., 2014. Marker-assisted 
introgression of a QTL region to improve rust resistance in three elite and popular 
varieties of peanut (Arachis hypogaea L.). Theoretical and Applied Genetics, 127(8), 
1771-1781.  
Vasanthi, R.P. and Raja Reddy, C., 1997. Inheritance of testa colour and resistance to late leaf 
spot and rust in groundnut (Arachis hypogaea L.). Journal of Oilseeds Research, 14, 
244-248. 
Videira, S.I.R., Groenewald, J.Z., Nakashima, C., Braun, U., Barreto, R.W., de Wit, P.J. and 
Crous, P.W., 2017. Mycosphaerellaceae–Chaos or clarity?. Studies in Mycology, 87, 
257-421. 
Vishnuvardhan, K.M., Vasanthi, R.P. and Reddy, K.H., 2011. Combining ability of yield, yield 
traits and resistance to late leaf spot and rust in groundnut. Journal of SAT Agricultural 
Research, 9, 6. 
Vos, P.G., Paulo, M.J., Voorrips, R.E., Visser, R.G., van Eck, H.J. and van Eeuwijk, F.A., 
2017. Evaluation of LD decay and various LD-decay estimators in simulated and 
SNP-array data of tetraploid potato. Theoretical and Applied Genetics, 130(1), 123-
135. 
Wan, S.B., Zhang, J.M., Zhang, X.Y., Feng, H.S., Yu, S.L., Li, T.S. and Liao, B.S., 2003. 
Peanut cultivation science in China. Shanghai Science and Technology Press, 
Shanghai, 531. 
Wang, M.L., Sukumaran, S., Barkley, N.A., Chen, Z., Chen, C.Y., Guo, B., Pittman, R.N., 
Stalker, H.T., Holbrook, C.C., Pederson, G.A. and Yu, J., 2011. Population structure 
148 
 
University of Ghana http://ugspace.ug.edu.gh
and marker–trait association analysis of the US peanut (Arachis hypogaea L.) mini-
core collection. Theoretical and Applied Genetics, 123(8), 1307-1317. 
Wang, Q., 2014. Introduction to ArcGIS 10.2. Retrieved from 
http://www.lib.uwaterloo.ca/locations/umd/digital/documents/IntrotoArcGIS10.2-
June2014.pdf. Date:06/07/2017. Accessed on 16th July 2017. 
Wilson, J.N., Baring, M.R., Burow, M.D., Rooney, W.L. and Simpson, C.E., 2013. Diallel 
anaysis of oil production components in peanut (Arachis hypogaea L.). International 
Journal of Agronomy, 9,13. 
Witcombe, J.R., Gyawali, S., Sunwar, S., Sthapit, B.R. and Joshi, K.D., 2006. Participatory 
plant breeding is better described as highly client-oriented plant breeding. II. Optional 
farmer collaboration in the segregating generations. Experimental Agriculture, 42(1), 
79-90. 
Wooten, S., 2003. Women, men, and market gardens: Gender relations and income generation 
in rural Mali. Human Organization, 62(2), 166-177. 
Würschum, T., Langer, S.M., Longin, C.F.H., Korzun, V., Akhunov, E., Ebmeyer, E., 
Schachschneider, R., Schacht, J., Kazman, E. and Reif, J.C., 2013. Population 
structure, genetic diversity and linkage disequilibrium in elite winter wheat assessed 
with SNP and SSR markers. Theoretical and Applied Genetics, 126(6), 1477-1486. 
Wynne J.C. and Coffelt T.A., 1982. Genetics of Arachis hypogaea. In: Patee HE, Young CT 
(eds) Peanut science and technology. American Peanut Research and Education 
Society, Yoakum, Texas, 50-94. 
Wynne, J.C., Mozingo, R.W. and Emery, D.A., 1979. Registration of NC 7 Peanut1 (Reg. No. 
22). Crop Science, 19(4), 563-563. 
Xi, X-Y, 1991. Development and structure of pollen and embryo sac in peanut (Arachis 
hypogaea L.). Botanical Gazette, 152, 164-172. 
Yan, J., Shah, T., Warburton, M.L., Buckler, E.S., McMullen, M.D. and Crouch, J., 2009. 
Genetic characterization and linkage disequilibrium estimation of a global maize 
collection using SNP markers. PloS one, 4(12), e8451. 
Young, N.D., Weeden, N.F. and Kochert, G., 1996. Genome mapping in legumes (Family 
Fabaceae). Genome Mapping in Plants, 32, 211-277. 
Yussif, I., Kwoseh C., Osman M., and Acheremu K., 2014. Farmers’ Perception and Farming 
Practices on the Effect of Early and Late Leaf Spots on Groundnut Production in 
Northern Ghana. Journal of Biology, Agriculture and Healthcare, 4(19), 22-29. 
Zhang, X., Zhang, J., He, X., Wang, Y., Ma, X. and Yin, D., 2017. Genome-wide association 
study of major agronomic traits related to domestication in peanut. Frontiers in Plant 
Science, 8, 1611. 
149 
 
University of Ghana http://ugspace.ug.edu.gh
Zhang, Y., Kang, M.S. and Lamkey, K.R., 2005. Diallel-Sas05. Agronomy Journal, 97(4), 
1097-1106. 
Zhu, L.C., Smith, C.M., Fritz, A., Boyko, E.V. and Flinn, M.B., 2004. Genetic analysis and 
molecular mapping of a wheat gene conferring tolerance to the greenbug (Schizaphis 
graminum Rondani). Theoretical and Applied Genetics, 109(2), 289-293. 
 
  
150 
 
University of Ghana http://ugspace.ug.edu.gh
APPENDICES 
 
Appendix 1 Optimal number of clusters in phenotypic data 
 
 
Appendix 2 SNPs distribution over chromosomes 
 
151 
 
University of Ghana http://ugspace.ug.edu.gh
Appendix 3 Distribution of the evaluated traits in combining ability study 
 
 
 
 
  
152 
 
University of Ghana http://ugspace.ug.edu.gh
Appendix 4. PRA Questionnaire 
A-Individual questionnaire 
PRA QUESTIONNAIRE: Groundnut’s Farmer 
Nom de 
l’enquêteur………………………………………Date………………………………. 
I. IDENTIFICATION 
1.1.Caractéristiques sociodémographiques 
Région : Préfecture: Village: 
Nom du répondant : 
Age : Sexe :         Masculin                         Féminin 
Niveau d’éducation ? 1= Aucun ;  2 = Alphabétisé ; 3 = Primaire 4 = 
Secondaire ; 5 = Supérieur 
Situation matrimoniale ? 1= Marié ; 2= Veuf (ve) ; 3= Divorcé ; 4= 
Célibataire 
Nombre de personne en charge ? 1= Enfants…….….; 2 = Adultes………… ;  
 
1.2. Activités agricole 
Taille de l’unité de production Droit de propriété terrien 
1 Surface total disponible :………….. ha  Privée ……………….………ha 
2 Surface exploitée :……….................ha  Familiale………………….....ha 
3 Surface en jachère :………………....ha  Communale………………….ha 
4 Surface cultivée d’arachide…………ha  Gouvernemental……………..ha 
5 Depuis quand êtes-vous agriculteur ?   
6 Etes-vous agriculteur en plein temps ou en  Plein temps 
temps partiel ?  Temps partiel 
7 Si temps partiel, combien de quel  Moins de la moitié 
proportion de votre temps consacrez-vous  Moitié 
à l’agriculture ?   Plus de la moitié 
8. Citez 5 cultures cultivées les deux dernières années par ordre d’importance, et les 
personnes impliquées dans la production 
 Culture Superficies Genre Rendement Proportion vendue 
1      
153 
 
University of Ghana http://ugspace.ug.edu.gh
2      
3      
4      
5      
 
1.3. Main d’œuvre 
a- De quelle main d’œuvre disposez-vous pour les activités agricoles ? 
 Travailleurs en temps plein Travailleurs en temps plein 
Age Masculin Féminin Masculin Féminin 
Familiale Loué Familiale Loué Familiale Loué Familiale Loué 
(ouvrier) (ouvrier) (ouvrier) (ouvrier) 
         
         
         
         
 
b- De quels types main d’œuvre louée employé vous chaque saison ? 
1= Occasionnelle   2= Permanente   3= Entraide 
villageoise 
c- Pour quelles activités employez-vous de la main d’œuvre loué ? 
 Activité Occasionnelle Permanente Entraide 
villageoise 
1 Labour    
2 Semi    
3 Sarclage    
4 Récolte    
5 Post récolte    
6 Autres (Spécifier)    
 
II. Production d’arachide 
2.1. Système de production d’arachide 
1. Quelle place occupe la culture l’arachide 1 = Important ; 2 = Non négligeable ; 3 = 
dans vos activités ? 
négligeable 
2. Quels équipements utilisez-vous pour la  
culture d’arachide ? 
3. Quel sont les périodes de semi ?  
4. Quel sont les périodes de récoltes ?  
5. Utilisez l’engrais pour la culture  
d’arachide ? Quel type d’engrais ? 
154 
 
University of Ghana http://ugspace.ug.edu.gh
6. Combien de sarclage effectuez-vous du  
semi à la récolte de l’arachide ? 
7. A quel moment effectuez-vous le  
sarclage ? 
8. Cultivez-vous l’arachide en culture ou  1= Culture pure ;    2= Culture associé 
associée ? 
9. Etes –vous membre d’un groupement 1 = Oui 
agricole ? 
2 = Non 
10. Quel type de groupement ?  
11. Depuis quand êtes-vous membre ?  
12. Quels problèmes discutez-vous dans le  
groupement ? 
 
2.2. Production d’arachide au cours des deux dernières campagnes 
Cultures Superficie emblavée Rendements 
2013/2014 2014/2015 2013/2014 2014/2015 
Arachide     
Culture associé     
 
III. Existence de contraintes liées à la production et barrières à la commercialisation 
3.1.Rencontrez-vous des difficultés dans la 
production d’arachide?................................................. 
Citez-les par ordre d’importance : 
Contraintes Causes Perte de rendements Moyens de lutte 
estimée 
    
    
    
    
    
 
3.2.Quelles sont les variétés cultivées ? Par ordre de préférence. 
Variétés Qualités Défauts 
   
   
   
   
 
3.3. Quels sont les variétés les plus exigeantes (main d’œuvre et intrants) 
155 
 
University of Ghana http://ugspace.ug.edu.gh
Ordre Variétés Expliquer les raisons 
1   
2   
3   
4   
 
3.4. Quel couleur préférez-vous pour les graines d’arachides ? 
1= Rouge   2= Blanchâtre        3= Pas de préférence  4= Autres (Spécifier) 
3.5. Mode et système de production 
Activités Périodes Main d’œuvre impliqué 
Hommes Femmes Enfants 
Labour     
Semi     
Sarclage     
Récolte     
Transport     
Séchage     
Décorticage     
Tri     
Vente     
Autres     
(Spécifier) 
 
3.6. Caractéristiques désirées par ordre de préférence pour une variété d’arachide : 
Ordre Caractéristiques par ordre de préférence 
1  
2  
3  
 
3.7.Comment sélectionnez-vous les semences d’arachides pour la saison suivante ? 
…………………………………………………………………………………………………
…………………………………………………………………………………………………
…………………… 
3.8. Rencontrez-vous des problèmes dans l’obtention des semences 
d’arachides ?............................. 
Si oui, 
lesquels ?......................................................................................................................................
.... 
156 
 
University of Ghana http://ugspace.ug.edu.gh
…………………………………………………………………………………………………
…………. 
 
3.9. Quelles  insectes / maladies rencontrez-vous sur vos cultures d’arachides ? Par ordre 
d’importance. 
Maladies / Insectes Causes Perte de rendements 
estimée 
   
   
   
 
3.10. Comment lutez-vous contre ces insectes /maladies ? 
Maladies / Insectes Moyens de luttes 
  
  
  
  
 
3.11.  Connaissez-vous la « Cercosporiose » (montrer une photo des symptômes) 
?............ 
- Quel est le nom local ?................................................................................................ 
- Quels est la cause, selon vous ?.................................................................................. 
- Comment est transmit la maladie selon vous ?........................................................... 
- Estimer les pertes liées à la cercosporiose d’arachide ? 
1= Faible (moins de 20%) ;      2= Modéré (21-41%) ;     3 = Elevé (plu de 50%) ;     Total 
(100%) 
- Quels moyens de luttes employées 
vous?....................................................................... ; 
 
3.12.  Connaissez-vous des variétés non sensibles à la « Cercosporiose » ? Citez-les. 
…………………………………………………………………………………………………
…………………………………………………………………………………………………
…………………………………………………………………………………………………
…………………………………………………………………………………………………
………………………………………… 
3.13.  Selon vous, quel est l’évolution de la maladie dans votre zone ces dernières 
années? 
157 
 
University of Ghana http://ugspace.ug.edu.gh
1=  Progresse (augmente)                   2= Reste constante     3= 
Diminue 
3.14. Quels problèmes rencontrez-vous pendant le stockage et la conservation de 
vos récoltes d’arachides ? Par ordre d’importance. 
Problèmes Causes Perte de rendements 
estimée 
   
   
   
   
 
3.15.  Existe-t-il des obstacles à la vente de l’arachide dans votre 
zone ?.................................. 
Enumérez-
les :…………………………………………………………………………………... 
…………………………………………………………………………………………………..
. 
IV. Revenus liée à la production d’arachide au cours des deux dernières années 
4.1. Quelles affectations aviez fait de vos récoltes d’arachide précédentes? 
  AFFECTATION  
Année Production Proportion Proportion vendue Autres (préciser)…. 
totale consommée  (Fraction) ….……………………. 
(fraction) ….……………………. 
2013/2014     
2014/2015     
 
- Qui décide de l’affectation des récoltes d’arachides ?................................. 
- Quand vendez-vous vos récoltes d’arachides ?............................................ 
- Vendez-vous l’arachide en coque ou en graines ?....................................... 
 
4.2. Quelles sont vos revenus ? 
Année Production Coût de production Prix moyen de Revenu 
totale vente 
2013/2014     
2014/2015     
 
V. Appui conseil technique 
5.1. Avez-vous accès aux services d’appui-conseil technique ?          1= Oui         2 = Non 
5.2. Si oui, auprès de quel(s) structure(s) et pour quelles activités ? 
158 
 
University of Ghana http://ugspace.ug.edu.gh
…………………………………………………………………………………………………
…………………………………………………………………………………………………
………………………………………………………………………………………………… 
5.3.A quel moment de la saison, recevez-vous la visite des techniciens ? 
 
Opération Nombre de visite 
Labour  
Semi  
Sarclage  
Récolte  
 
Avez-vous eu à visité une parcelle de démonstration ?........................................ 
 
VI. Accès au crédit agricole 
Etes-vous membre d’une microfinance ? 1= Oui                    2 = Non 
Si oui, quelle microfinance ?  
Depuis quand êtes-vous membre ?  
Obtenez-vous des crédits pour la production  
agricoles ? 
Obtenez-vous des crédits pour la production  
d’arachide ? 
Type de crédit ? 1 = crédits espèces ;    2 = crédits 
intrants 
Types d’intrants ? 1= Semences ; 2 = Engrais ; 3 = 
Semences et engrais ; 4 = Prdts 
phytosanitaires 
Le crédit est-il adapté ? Justifier.  
 
VII. Avez-vous un commentaire ou des suggestions liés à la production agricole et 
surtout à la culture d’arachide ? 
 
  
159 
 
University of Ghana http://ugspace.ug.edu.gh
B-Focus group discussion 
PRA QUESTIONNAIRE : Focus group discussion 
Nom de 
l’enquêteur………………………………………Date………………………………. 
Région : Préfecture : Village : 
 
CHECK LIST OF THE FOCUS GROUP DISCUSSION 
1. Production végétale 
 - Citez 5 cultures majeures cultivées dans votre zone par ordre de d’importance 
2. Identification et caractérisations de variétés cultivées d’arachides 
- Combien de variété d’arachides cultivez-vous ? 
 - Pouvez-vous décrire chacune d’elle ? 
- Pouvez-vous classer les variétés par ordre d’importance ? 
3. Disponibilité et gestion des semences 
- Comment sélectionnez-vous les semences pour la saison suivante ? 
- Comment stocker vous l’arachide? 
- Avez-vous des problèmes de stockage/ conservation des semences ? 
 - Comment gerez-vous les problèmes de stockage/ conservation des semences 
?  
- D’où proviennent vos semences ? 
- Eprouvez-vous des difficultés à obtenir des semences d’arachide ? 
4. Système de culture 
- Quelle est la taille moyenne d’une parcelle d’arachide dans votre zone ? 
- Quels types d’équipement utilisez-vous pour la culture d’arachide ? 
- Pouvez-vous décrire comment vous cultivez l’arachide ? 
 - Quelle sont les périodes de semi d’arachide ? 
- A quelles périodes récoltez-vous l’arachide ? 
- Appliquez-vous des engrais ? De quel type d’engrais s’agit-il ? 
- Combien de fois sarclez-vous vos champs d’arachide ? 
5. Contraintes à la production d’arachide 
160 
 
University of Ghana http://ugspace.ug.edu.gh
- Pouvez-vous citer les insectes et maladies rencontrés sur vos cultures 
d’arachide ? 
- Comment luttez-vous contres ces contraintes ? 
 - Quelles contraintes abiotiques rencontrez-vous dans la culture d’arachide ? 
- Pouvez-vous classer les contraintes par ordre d’importance ? 
- Connaissez-vous la « Cercosporiose d’arachide » ? Cause ? Moyen de lutte ? 
- Existe-t-il des variétés résistantes à la « Cercosporiose » 
6. Consommation d’arachide 
- Comment consommez-vous l’arachide ? 
 
- Nourrissez-vous vos troupeaux avec les fanes d’arachide ? 
7. Commercialisation de l’arachide 
- Comment vendez-vous vos récoltes d’arachide ? 
 
- Avez-vous des difficultés à vendre vos récoltes ? Citez-les. 
8. Préférence des producteurs d’arachide 
- Que souhaitez-vous avoir comme caractéristiques dans une variété 
 
d’arachide ? 
- Avez-vous des commentaires ou des suggestions concernant la production 
 
d’arachide ? 
 
 
161