University of Ghana http://ugspace.ug.edu.gh GENETIC STUDIES AND MAPPING OF BENEFICIAL ALLELES FROM WILD SPECIES TO IMPROVE AGRONOMIC TRAITS IN CULTIVATED GROUNDNUT (Arachis hypogaea L.) BY AISSATOU SAMBOU (10496579) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT 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, 2017 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. ................................................. Aissatou Sambou (Student) ................................................... Prof. Samuel Kwame Offei (Supervisor) .................................................. Prof. Pangirayi B. Tongoona (Supervisor) .................................................. Dr. Agyemang Danquah (Supervisor) .................................................. Dr. Beatrice E. Ifie (Supervisor) .................................................. Dr. Daniel Fonceka (Supervisor) i University of Ghana http://ugspace.ug.edu.gh ABSTRACT Groundnut (Arachis hypogaea L.) is a legume widely cultivated in tropical and sub-tropical regions. Although Senegal is among the largest African producers, yield has declined over the years. The narrow genetic diversity of the cultivated groundnut due to its monophyletic origin coupled with polyploidization has slowed down its genetic improvement using molecular tools. Wild species of groundnut are important sources of useful alleles that can be used to broaden the genetic base of the cultivated groundnut. However because of their inferior agronomic value, groundnut wild species have been little used to improve agronomic traits. Thus the present study was undertaken to enlarge the genetic diversity of cultivated groundnut and to estimate the relative contribution of the wild species to improve agronomic traits in cultivated groundnut. Specifically the objectives were to: i) identify the main constraints that farmers are facing in groundnut production and their varietal traits preferences; ii) identify promising breeding lines; iii) identify genomic regions that underlie the variation of the observed phenotypes and iv) evaluate the increase of the phenotypic value associated with the alleles of the wild species. A participatory rural appraisal (PRA) was conducted using focus group discussions (FGD) and questionnaires in three regions (Kaolack, Kaffrine and Tambacounda). Furthermore, an advanced backcross BC2F4:6 mapping population was developed at CERAAS (centre d’etudes regional pour l’amelioration de l’adaptation a la secheresse) from a cross between Fleur11 used as recurrent parent and a synthetic allotetraploid donor ISATGR52B (A. duranensis (AA) x A. valida (BB)) 4x. Two hundred BC2F4 families from this cross were genotyped with 128 microsatellite markers (SSRs) at the genotyping lab (CERAAS) and the BC2F4:6 progenies plus Fleur11 were evaluated in four different environments. The PRA study revealed that groundnut remained the main cash crop cultivated by farmers. Farmers were ready to adopt varieties that ii University of Ghana http://ugspace.ug.edu.gh have high pod and haulm yield and large pod and seed size regardless of oil content. The main constraints highlighted by farmers were marketability and access to land. However drought was the constraint that farmers identified as the most important and that could be addressed by breeders. A large phenotypic variation was observed between the advanced backcross (AB) lines. The interaction between genotypes and environments was significant. The analysis performed in each site allowed selecting fifty lines based on the weight attributed to each target trait using the selection index. Among the fifty lines, eleven were constantly present across sites. The GGE biplot identified B7_32_8_1 and Fleur11 as the best performing lines in terms of pod yield. The genotyping analysis conducted on the BC1 progenies permitted the construction of a genetic linkage map which spanned a cumulative length of 1792 cM with an average marker density of 6.4 cM. Three hundred and thirty loci were mapped onto 20 linkage groups (LGs). The quantitative trait loci (QTL) analysis conducted with the genotyping and phenotyping data of the BC2F4:6 families revealed 38 QTL related to agronomical traits. Most of the identified QTL for seed related traits were located on LG B3. For 50% of the identified QTL, the wild alleles contributed positively to the variation of the trait. Five QTL (5) were newly mapped. For 4 out of the 5 QTL, increase of the phenotypic value was associated with the alleles of the wild parent. The QTL for undesirable traits from the wild species (deep pod constriction, pronounced pod beak and spread growth habit) were also mapped and markers linked to them identified. This study has demonstrated the potential of alleles from wild relatives to enrich the genetic base of cultivated groundnut and to improve agronomic traits. iii University of Ghana http://ugspace.ug.edu.gh DEDICATION To my parents Elhadji Ousmane Boute Sambou and Bineta Ndiaye Sambou iv University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENTS I am grateful to the West Africa Agricultural Productivity Programme (WAAPP) for providing me a full scholarship. I thank the staff of West Africa Centre for Crop Improvement (WACCI), University of Ghana, for their support and facilities. I express my sincere acknowledgements to my supervisors at WACCI Prof. Pangirayi Tongoona, Prof. Samuel Kwame Offei, Dr. Beatrice Ifie and Dr. Agyemang Danquah for their constructive suggestions and critics. I am sincerely grateful to my in country supervisor Dr. Daniel Fonceka for the quality of training I received under his supervision. His guidance, support, contributions and critical review had greatly helped me in my research work. I learnt a lot from you Dr. I am indebted to Dr. Ndiaga Cisse, Director of CERAAS for his support. My thanks are extended to the scientific and management staff of CERAAS. CERAAS made it all possible by providing most of the resources that were used to carry out this research. Special thanks to Dr Joel Romaric Nguepjop and Dr Cyril Diatta for their availability, patience and support. I also express my sincere thanks to the groundnut team Hodo Abalo Tossim, Maguette Seye, and Rachelle Yvette Djiboune for their assistance. Big thank you to Alassane, Sane, Fall, Zale, Sonko (Nioro) and Alassane (Nioro). It is a pleasure to express my gratitude to the CERAAS’ PhD students Baye Maguate, Desire Kyky, Jack Akata, Mame Balla, Awa, Oberline and Elisabeth; I have started missing our fruitful discussions and jokes. Success upon you guys. v University of Ghana http://ugspace.ug.edu.gh Big thanks to my classmates of cohort 7 Isata, Dieni, Sanatu, Kumba, Chiaka, Souleymane, Olumide, Ismael, Lydia and Ijeoma. Special thanks to the Senegalese Ghislain and Bathe. You made my stay amazing. Big and special thanks to my family Ta Oumi, Ta Adji, Ndioba, Fatou, Ndeye, Mariama, Awa, Adji, Mareme, Bineta, Mouhamed and Ousmane for your prayers, encouragements and help. To Mariama I just say thank you. I will never forget what you and Mouhamed Ka did for me. I say also thank you to my family in law. I am grateful to everyone who contributed to this work. Finally to my friend, my brother, my husband Cheikh Ahmed Tidiane Sy, thank you. Thank you for everything. vi University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENTS DECLARATION ........................................................................................................................... i ABSTRACT ................................................................................................................................... ii DEDICATION ............................................................................................................................. iv ACKNOWLEDGEMENTS ......................................................................................................... v LIST OF TABLES ........................................................................................................................ x LIST OF FIGURES ..................................................................................................................... xi LIST OF ABBREVIATIONS .................................................................................................... xii CHAPTER ONE ............................................................................................................................. 1 1.0 GENERAL INTRODUCTION .....................................................................................................1 CHAPTER TWO ............................................................................................................................ 5 2.0 LITERATURE REVIEW ..............................................................................................................5 2.1. Origin and distribution ............................................................................................................... 5 2.2. Botany and reproduction features ............................................................................................. 5 2.3. Groundnut production ............................................................................................................... 6 2.4. Nutritional value and utilization of cultivated Groundnut ......................................................... 7 2.5. Systematic and taxonomy .......................................................................................................... 9 2.6. Cytology and genomic relationships in the Arachis genus ....................................................... 10 2.7. Genetic diversity of Arachis ..................................................................................................... 12 2.8. Targeted traits for groundnut improvement ........................................................................... 14 2.9. Potential of wild species in groundnut improvement .............................................................. 15 2.10. Genomic resources in groundnut applicable in marker assisted selection ........................... 17 2.10.1 Molecular markers ........................................................................................................... 17 2.10.2 Genetic linkage map ........................................................................................................ 18 2.11. Exploitation of groundnut wild allelic potential for varietal creation .................................... 19 2.11.1 Interests of integrating molecular breeding approaches ................................................ 19 2.11.2 The concept of advanced Backcross (AB) methodology .................................................. 21 CHAPTER THREE ...................................................................................................................... 24 3.0 ASSESSMENT OF FARMERS’ PRODUCTION CONSTRAINTS AND VARIETAL TRAIT PREFERENCES OF GROUNDNUT IN SENEGAL ...........................................................24 3.1 Introduction .............................................................................................................................. 24 3.2 Methodology ............................................................................................................................ 26 3.2.1 Study area .......................................................................................................................... 26 vii University of Ghana http://ugspace.ug.edu.gh 3.2.2 Sampling procedure ........................................................................................................... 27 3.2.3 Data collection ................................................................................................................... 27 3.2.4 Data analysis ...................................................................................................................... 28 3.3 Results ....................................................................................................................................... 29 3.3.1 Demographic and socio-economic characteristics ............................................................ 29 3.3.2 Groundnut importance and production level .................................................................... 29 3.3.3 Varieties grown and their provenance .............................................................................. 32 3.3.4 Crop management ............................................................................................................. 33 3.3.5 Farmers trait preferences for groundnut variety ............................................................... 34 3.3.6 Groundnut production constraints .................................................................................... 35 3.4 Discussion ................................................................................................................................. 37 3.5 Conclusion ................................................................................................................................ 42 CHAPTER FOUR ......................................................................................................................... 43 4.0 ASSESSMENT OF AGRONOMIC POTENTIAL OF AN ADVANCED BACKCROSS QTL POPULATION AND IDENTIFICATION OF PROMISING BREEDING LINES ..............43 4.1 Introduction .............................................................................................................................. 43 4.2 Materials and Methods ............................................................................................................ 45 4.2.1 Population development ................................................................................................... 45 4.2.2 Study sites and field design ............................................................................................... 47 4.2.3 Data collection ................................................................................................................... 48 4.2.4 Statistical analysis .............................................................................................................. 49 4.3 Results ....................................................................................................................................... 52 4.3.1 Traits variation and correlations ........................................................................................ 52 4.3.2 Performance of the population in the studied environments ........................................... 56 4.3.3 Selection of the promising lines based on IAS ................................................................... 57 4.3.4 Pod yield performance and stability of the selected genotypes ........................................ 57 4.4 Discussion ................................................................................................................................. 61 4.5 Conclusion ................................................................................................................................ 63 CHAPTER FIVE .......................................................................................................................... 64 5.0 QUANTITATIVE TRAIT LOCI (QTL) MAPPING FOR POD YIELD RELATED TRAITS AND USE IN GROUNDNUT ...............................................................................................64 5.1 Introduction .............................................................................................................................. 64 5.2 Materials and Methods ............................................................................................................ 67 5.2.1 QTL mapping ...................................................................................................................... 67 viii University of Ghana http://ugspace.ug.edu.gh 5.2.1.2 Map construction and molecular analysis .................................................................. 67 5.2.1.3 Genetic map and QTL statistical analysis .................................................................... 68 5.2.2 Pyramiding of QTL using CSSLs .......................................................................................... 70 5.2.2.1 Material ...................................................................................................................... 70 5.2.2.2 DNA extraction from seeds ......................................................................................... 71 5.2.2.3 F1 control and selection of the F2 lines ....................................................................... 72 5.3 Results ....................................................................................................................................... 73 5.3.1 Marker segregation and polymorphism ............................................................................ 73 5.3.1.1 Linkage map ................................................................................................................ 73 5.3.1.2 QTL detection .............................................................................................................. 75 5.3.1.3 Wild alleles’ introgression rate and examination of the genotype of the selected line 81 5.3.2 Selection of the F2 individuals ............................................................................................ 83 5.4 Discussion ................................................................................................................................. 84 5.5 Conclusion ................................................................................................................................ 88 CHAPTER SIX ............................................................................................................................. 89 6.0 GENERAL CONCLUSIONS AND RECOMMENDATIONS ................................................89 6.1 General conclusions .................................................................................................................. 89 6.2 Recommendations .................................................................................................................... 91 REFERENCES ............................................................................................................................. 92 APPENDICES ............................................................................................................................ 119 ix University of Ghana http://ugspace.ug.edu.gh LIST OF TABLES Table 3.1: Demographic and socio-economic characteristics of the respondents across locations ............30 Table 3.2: Groundnut pod yield estimated as a function of area and NPK fertilizer applied .....................31 Table 3.3: Regression analysis between estimated yield, rate of fertilizer and acreage .............................32 Table 3.4: Farmers preferences for groundnut traits in different communities ..........................................34 Table 3.5: Relative ranking of major groundnut production constraints by community ............................35 Table 4.1: Summary statistics of traits in the studied environments ...........................................................53 Table 4.2: Correlation between studied traits in one site ............................................................................55 Table 4.3: Top selected genotypes across environments based on IAS indices .........................................58 Table 5.1: Details of the CCSLs involved in the development of the pyramiding populations ..................70 Table 5.2: Summary of the detected QTL ...................................................................................................76 Table 5.3: Percentage of introgression of the wild donor into the selected genotypes ...............................81 x University of Ghana http://ugspace.ug.edu.gh LIST OF FIGURES Figure 2.1: Genetic bottlenecks on crop plants during domestication and plant breeding .........................13 Figure 2.2: Comparison of the methods for creating primary and advanced bi-parental mapping populations ...................................................................................................................................................20 Figure 3.1: Location of the study sites on the Senegal map .......................................................................26 Figure 3.2: Cultivated groundnut varieties by community .........................................................................32 Figure 4.1: Description of the development of the advanced backcross QTL (AB-QTL) population .......46 Figure 4.2: Location of the study sites ........................................................................................................47 Figure 4.3: Distribution and variation of studied traits among the BC2F4:6 progenies ...............................52 Figure 4.4: Traits variation among studied sites .........................................................................................56 Figure 4.5: Ranking and stability of selected genotypes based on mean pod yield ....................................59 Figure 4.6: GGE biplot for the “which won where” pattern of genotypes and environments ....................60 Figure 5.1: Description of the way to cut the seed for DNA extraction .....................................................71 Figure 5.2: Genetic map derived from Fleur11 x AvAd population ...........................................................74 Figure 5.3: Location on the genetic map of detected QTL .........................................................................79 Figure 5.4: Graphical representation of the genotype of the introgressed LGs of B7_32_1_8. .................82 Figure 5.5: Selection of the desired F2 individuals based on marker genotype. .........................................83 xi University of Ghana http://ugspace.ug.edu.gh LIST OF ABBREVIATIONS AB: Advanced Backcross AEA: Average Environment Coordinates AFLP: Amplified Fragment Length Polymorphism ANOVA: Analysis of Variance ANSD: Agence National de la Statistique et de la Démographie BIL: Backcross Inbred Line CNRA: Centre National de Recherches Agronomiques CERAAS: Centre d’Etudes Régional pour l’Amélioration de l’Adaptation a la Secheresse CIRAD: Centre International de la Recherche Agricole et du Développement CSSL: Chromosome Segment Substitution Line DAS: Days After Sowing DNA: Deoxyribonucleic Acid EMBRAPA: Empresa Brasileira de Pesquisa Agropecuária ICRISAT: International Crops Research Institute for the Semi-Arid Tropics ISRA: Institut Sénégalais de Recherches Agricoles LG: Linkage Group MARS: Markers Assisted Recurrent Selection MAS: Markers Assisted-Selection NIL: Near Isogenic Line PCR: Polymerase Chain Reaction PRA: Participatory Rural Appraisal PV: Phenotypic Variation xii University of Ghana http://ugspace.ug.edu.gh QTL: Quantitative Trait Loci RFLP: Restriction Fragment Length Polymorphism RIL: Recombinant Inbreed Line SNP: Single Nucleotide Polymorphism SSR: Simple Sequence Repeat WAAPP: West Africa Agricultural Productivity Programme xiii University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE 1.0 GENERAL INTRODUCTION Groundnut or peanut (Arachis hypogaea L.) is an important legume crop grown mostly in tropical and subtropical regions of the world. It is an oilseed crop cultivated in all five continents in about 120 countries. The world production reached 41.3 million tons cultivated on 24.6 million ha in 2015 (FAOSTAT, 2015). The major producing continents are Asia (26 million tons) and Africa (14 million tons) which together account for 23.3 million ha representing 94.71% of the global groundnut cultivated areas in the world. About 60% of the total production is used for oil extraction, while 40% is consumed as food and other utilizations such as self- production of seed (Birthal et al., 2010). The Senegalese groundnut sector is one of the most profitable agricultural sectors in the country. Groundnut is the most important crop in terms of production and harvested areas and it is mainly cultivated under rainfall conditions in the central part of the country called the groundnut basin. Groundnut is a cash crop providing incomes and improving livelihoods of farmers. It represents a strategic crop and its cultivation contributes to the sustainability of the integrated crop-livestock production system, the most predominant one in the semi-arid areas. Although Senegal is among the top five producing countries in Africa, yield has declined over the years. In 2015, the harvested area was 878 659 ha with a total production of 669 329 representing over 50% of the total production of all crops mixed (FAOSTAT, 2015). Nevertheless, the increase of the production has predominantly been caused by an increase of the cultivated area rather than an increase of yield. For instance, groundnut’ acreage increased by 9% in 2013/14, but the yield was lower than the previous season (922 kg/ha versus 943 kg/ha) (FAOSTAT, 2015). The national average yield is approximately 900 kg/ha (ANSD, 2015), which is below the crop yield potential estimated at 5.4 t/ha (Nigam, 2005). With the liberalization of the export market and the ever-growing population, producers have the 1 University of Ghana http://ugspace.ug.edu.gh challenge to produce more on smaller land areas. To meet the increasing demand and to enable farmers to make higher profits from groundnut, more effort is needed in breeding programmes; effort which take into account farmers’ opinions and participation. According to Sanou et al. (2014), breeders have an incomplete understanding of why farmers choose the varieties they grow. Participatory rural appraisal (PRA) is one of the most widely used methods by breeders for identifying farmers’ preferences and integrating their inputs in the varietal development process. According to Witcombe et al. (2005), a successful PRA should provide useful information needed to specify the characteristics in a new variety. In view of some shortcomings in some varieties experienced by farmers, breeders and researchers have to prioritize traits for breeding based on farmers’ inputs. These shortcomings range from economic aspects to demand for more adapted groundnut varieties. Efforts have been made by the national groundnut breeding programme of Senegalese Institute of Agricultural Research (ISRA) towards solving some of the challenges of the past 30 years. These efforts have been rewarded by the release of several varieties (55-437, 73-33, 756-A, 69- 101, 73-28, 78-937, 28-206, Fleur 11). Recently (2016), seven newly developed varieties were released. However, breeders and researchers are facing the narrow genetic variability in the cultivated species which hampers the use of genomic tools to speed up the process of varietal creation. Cultivated groundnut is a self-pollinated crop with an allotetraploid genome structure (AABB) derived from a single natural hybridization between two diploid Arachis species with A and B genomes (Kochert et al., 1996; Seijo et al., 2004). The hybridization process was followed by chromosome doubling, a phenomenon which induced differences in ploidy level and thus isolated cultivated groundnut (tetraploid) from its related wild species which are diploids (Dwivedi & Upadhyaya, 2008; Feng et al., 2012). In addition to ploidy differences, 2 University of Ghana http://ugspace.ug.edu.gh domestication and breeding ultimately resulted in narrowing the genetic variability of groundnut. Selection has resulted in crops that are able to feed the world but at the price of reducing their genetic diversity compared to their wild relatives (Ross-Ibarra et al., 2007; van Heerwaarden et al., 2011). This phenomenon is often called “domestication bottleneck” (Tanksley & McCouch, 1997). Fortunately, groundnut wild species are rich in valuable alleles that could be used to enhance the genetic base of the cultivated groundnut. Many sources of disease resistance have been reported in wild Arachis species (Abdou et al., 1971; Subrahmanyam et al., 1983; Garcia et al., 1996; Pande et al., 2001; Leal-Bertioli et al., 2015). While groundnut wild species are being extensively used as sources of biotic resistance (Leal-Bertioli et al., 2009, 2010, 2015), they have been little utilized to improve agronomic traits like yield related traits. The inferior phenotypic appearance of wild species associated with their long growth cycle could explain to some extend the low rate of utilization of wild species to improve complex traits (Pandey et al., 2012). Nevertheless, many studies conducted on several crops have demonstrated the potential of wild species in improving complex traits with the help of molecular markers (Fulton et al., 1997; Grandillo & Tanksley, 2003; Huang et al., 2004). In groundnut, Fonceka et al. (2012) showed that the positive effect of half of the detected QTL for drought and yield related traits were brought by the alleles from wild relatives. In this dynamic process of varietal creation using wild species, the Centre d’Etudes Regional pour l’Amélioration de l’Adaptation à la Sécheresse (CERAAS) in collaboration with other research institutes such as the Centre National de Recherche Agricole (CNRA), the Centre de cooperation Internationale en Recherche Agronomique pour le Développement (CIRAD), the Empresa Brasileira de Pesquisa Agropecuaria (EMBRAPA) and the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) have developed a research programme 3 University of Ghana http://ugspace.ug.edu.gh to widen the genetic basis of domesticated groundnut. Various types of interspecific genetic material ranging from a library of chromosome segment substitution lines (CSSLs) to the advanced backcross QTL (AB-QTL) populations, pyramiding populations and Near Isogenic Lines (NILS) (Fonceka et al., 2012; CERAAS, 2014 not published data) have been developed or are under development. This present work aimed at exploring and exploiting the potential of wild groundnut species to improve yield related traits in cultivated groundnut Arachis hypogaea. The objectives used to: • identify farmers’ groundnut trait preferences and their major production constraints, • diversify the primary gene pool through introgression of beneficial alleles from wild species to cultivated groundnut, • identify lines that could be used as breeding lines in the breeding programme, • identify genomic regions (QTL) associated with the variation of useful pod yield related traits, • determine the genetic contribution of the wild genome in the variation of yield related traits; and • pyramid two wild fragments underlying QTL for pod and seed size, and biomass. 4 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO 2.0 LITERATURE REVIEW 2.1. Origin and distribution The history of the origin of groundnut traces a path back over several continents and finally ends up in America. The discovery of 7,600 years old fossilized groundnut hulls in the Ñanchoc Valley of Peru by archaeologists indicated that Arachis hypogea plant originated from South America in a region covering southern Bolivia - northern Argentina - northern Paraguay and western region of Matto Grosso Brazil (Hammons, 1994). Moreover, most of the wild species are found in that region. According to Hammons. (1994), groundnut distribution throughout the world is believed to has occurred in 15th century through Portuguese explorers when they arrived in Brazil and carried groundnuts from South America to Western Africa. The crop was then quickly accepted by African growers because of its similarity to the native Bambara groundnut (Voandzeia subterranea (L.). The Spanish then took them to west coast of South America, Philippines, China and various islands in the western Pacific. Later in 1700s the crop was taken to Europe. Introductions of groundnut in North America are presumed to have been from northern Brazil and the West Indies, where they are believed to have been loaded onto ships carrying slaves from Africa to the New World. 2.2. Botany and reproduction features Arachis hypogaea is an annual and self-fertilized crop. Its flowers are perfect / hermaphroditic containing both male and female reproductive organs. Although appearing as and referred to as a nut, it is the underground pod of a legume rather than a true nut. As a legume, it is a member of the family Fabaceae. The leaves are opposite and pinnate with four leaflets (two opposite pairs; no terminal leaflet). Groundnut displays a geocarpy feature (Rao, 1994). The inflorescences are 5 University of Ghana http://ugspace.ug.edu.gh born in the axils of leaves of primary and secondary branches. The inflorescences are simple or compound and each comprises up to 3 flowers. One flower per inflorescence opens per day and last for just one day. The colour of the flowers varies from orange to yellow depending on the cultivar. Groundnut flowers contain 10 anthers containing stamens. The stigmas are generally the same size or slightly smaller than the anthers. Pollen matures approximately 6 - 8 h before anthesis, which occurs within a few hours after sunrise. Generally, only one flower at an axil reaches anthesis on any given day. Pollination occurs at the same time that anthesis occurs. Enzymes expressed on the surface of the stigma to facilitate the adhesion of pollen degrade within 8 hours after anthesis (Rao, 1994). That is why it is important for practical aspects and success in crosses, to make the pollination in early hours of the morning. After pollination, a pedicel elongates to form a peg which will bear the pod. 2.3. Groundnut production Groundnut is an important legume crop grown in more than 100 countries under different agro- ecological environments in the world. It was cultivated on 25.4 million ha worldwide with a production of 45.6 million tons in 2013 (FAOSTAT, 2015). For the periods 2003–2013, the total groundnut production increased with an annual production increasing from 0.4% in 2010 to 2.05% in 2013, and this could be attributed to increase in both yield (1%) and area under cultivation (1.04%) (FAOSTAT, 2015). To meet the demands of the ever growing world population, the production has to increase considerably. Asia and Africa account for more than 90% of global groundnut area where it is cultivated mostly under rainfed conditions by small scale farmers. China is the largest producer (16.9 million tons) followed by India (9.4 million), Nigeria (3 million tons) and USA (1.9 million tons) (FAOSTAT, 2015). Africa accounts for 6 University of Ghana http://ugspace.ug.edu.gh 25.5% (12 million hectares) of the total world production and ranks as the second most important groundnut producing continent in terms of acreage and production after Asia. Groundnut production in Africa has grown significantly since 1990 and West African countries have contributed greatly (Revoredo & Fletcher, 2002). The high production attributed to groundnut in Africa is predominantly due to increase in cultivated area (12 million ha, 11.6 million tons) rather than the yield. Africa, although is the second largest producer has the lowest average yields per hectare (1 t/ha) compared to Asia (3 t/ha) and America (4 t/ha). When comparing the average yield obtained between 1972 and 1975 to that obtained between 1996 and 2000, Asia with 74.9% had the highest yield increase. For the same periods, the increase was 24.9% for America and 15.2% for Africa (Revoredo & Fletcher, 2002). The yields in major producing countries vary from just over 1 t/ha in Africa to 3 t/ha in Asia; 4 t/ha in America and can attain 6 t/ha in Israel and even 8 t/ha in Cyprus in Europe (FAOSTAT, 2015). The differences between developed and developing countries in terms of yield reflect the differences in the production systems of the crop. Groundnut cultivation is mechanized and very intensive with high level of inputs in developed countries compared to developing countries like Senegal where it is grown with low inputs under rainfed conditions by poor farmers. 2.4. Nutritional value and utilization of cultivated Groundnut Groundnut provides income to farmers. It is also used as food through consumption of the kernels and provides nutritious haulms to livestock. It is a crop qualified as “gold” in Senegal. Groundnut is a rich source of energy composed of oil (48 - 50%) and protein (25 - 28%) in the kernels providing 564 kcal of energy from 100 g of seeds (Jambunathan, 1991). Groundnut is mainly used for oil production but 40% of the global production is consumed as food or other uses (Birthal et al., 2010). In addition, groundnut seeds contain many nutrients such as minerals, 7 University of Ghana http://ugspace.ug.edu.gh antioxidants, and vitamins and are rich in mono-unsaturated fatty acids (Jambunathan, 1991; Savage, 1994). Jiang et al. (2002) demonstrated in their clinical study that nuts consumption in general and more particularly groundnut butter reduces the incidence of type 2 diabetes in women. This can be explained by the richness in unsaturated fatty acids of groundnut which may be beneficial for glucose and insulin homeostasis. Moreover, the same authors found gain in regular nut consumption to avoid increasing caloric intake and therefore can be recommended as a replacement for red meats consumption. Being a legume, groundnut is used in rotation with mainly cereals in Africa to help improve soil fertility by fixing nitrogen (N) and leaving behind organic matter. Despite its high economical, nutritional and therapeutic values, groundnut consumption is no longer without risk in human health. Several clinical studies reported the side effects of groundnut. Food allergy is a common problem and groundnut has been cited among the eight foods that cause more than 90% of food allergies (Metcalfe, 1996). These reactions are caused by the presence of allergenic peptides Ara h 1, 2, 3 and 6 in groundnut seeds (Burks et al., 1997; Rabjohn et al., 1999; Koppelman et al., 2010). However it has been reported by clinical and epidemiology studies that high levels of environmental (cutaneous) exposure to groundnut during infancy may promote sensitization (Fox, 2009). Another continuing challenge in groundnut production and marketing is the occurrence of the aflatoxin produced by the mold Aspergillus flavus. However, the research community is working to solve the problem and a sensitization is being done at producers’ level to reduce the occurrence of the fungal infestation by good drying and storage practices. 8 University of Ghana http://ugspace.ug.edu.gh 2.5. Systematic and taxonomy Groundnut belongs to the tribe Aeschynomeneae, subtribe Stylosantinae, family Leguminosae which is divided into 3 sub-families: Mimosoideae, Caesalpinoideae and Papilionoideae. The later includes groundnut and most of legume crops of high economic value. In 1994, Krapovickas & Gregory published a revised taxonomic of the genus Arachis which recognizes 80 species divided into 9 sections: Caulorrhizae, Erectoides, Extranervosae, Heteranthae, Procumbentes, Trierectodes, Triseminatae, Rhizomatosae and Arachis. The distinctions of the various sections are based on morphological characters, life cycle attributes, eco-geographic distribution, cross compatibility, cytological study as well as chromatographic and antigenic reactions (Krapovickas & Gregory, 1994; Stalker, 1997; Valls, 2005). Section Arachis is the most widely distributed. It includes the cultivated groundnut (A. hypogaea) species and 30 wild species. In this section, 26 species are diploid (2n = 2x = 20), three (03) are aneuploid or dysploid (2n = 2x = 18) and two (02) species, A. hypogaea and A. monticola are tetraploid (2n = 4x = 40) (Smartt & Stalker, 1982). On the basis of morphological features, Arachis hypogaea was classified into two (02) subspecies hypogaea and fastigiata. Subspecies hypogaea have a long life cycle, no flowers on the central stem, and regularly alternating vegetative and reproductive side stems. Subspecies fastigiata have a shorter life cycle, flowers on the central stem, and reproductive and vegetative stems distributed in a disorganized way. Arachis hypogaea ssp hypogaea is classified into two botanical varieties (var. hypogaea and var. hirsuta) while ssp fastigiata is further divided into four botanical varieties (var. vulgaris, var. fastigiata, var. peruviana, and var. aequatoriana) (Krapovickas & Gregory, 1994). These botanical varieties correspond to different commercial types and differ from each other by plant characteristics and 9 University of Ghana http://ugspace.ug.edu.gh their growth habit; spreading in hypogaea and erect in fastigiata (Krapovickas & Gregory, 1994). 2.6. Cytology and genomic relationships in the Arachis genus Arachis hypogaea is an amphidiploid (2n = 4x = 40) with a large and complex genome (2.8 Gb) (Berthioli et al., 2016). Chromosome morphology was reported by Husted. (1936) and Stalker & Dalmacio. (1986). Mendes. (1947) reported the number of chromosome of the diploid species (2n = 2x = 20). It is only in 2005 that Peñaloza & Valls. (2005) discovered the dysploid species (2n = 2x = 18) in some Arachis species. The genus is composed mostly of diploid species, but tetraploids exist in sections Arachis and Rhizomatosae, and aneuploid species in sections Arachis and Erectoides. Polyploidy evolved independently in sections Arachis and Rhizomatosae (Smartt & Stalker, 1982). In the genus Arachis, the largest section is the Arachis section which includes the cultivated species, one other tetraploid (A. monticola), 26 diploid (2n = 2x = 20), and three aneuploid species (2n = 2x = 18). These species are annual or perennial and are highly variable especially within the annual species. The first published interspecific hybridization attempt in the genus was between the tetraploids A. hypogaea (section Arachis) and A. glabrata (section Rhizomatosae) (Hull & Carver, 1938). No hybrids were obtained from the crosses. Following this first experiment, an extensive hybridization programme was undertaken by Gregory & Gregory. (1979). These authors used a collection of 91 Arachis and reported the cross-compatibility among the species of the genus. Their findings indicated that hybridization between species within the same section was more successful than crosses between sections. Based on their results, the membership of a given species to a section was performed by analysis of the fertility of the interspecific hybrids (Gregory & Gregory, 1979). However, reproductive isolation mechanisms and pollen infertility 10 University of Ghana http://ugspace.ug.edu.gh developed by a number of species and accessions within the same species have hindered classification efforts. Several attempts were then performed using A. hypogaea and various species of Arachis section. However, because of differences in ploidy level, most of the crosses failed. Besides the classification based on chromosome number and hybrids fertility analysis, there is a classification based on karyotype morphology. Husted. (1936) identified one pair of smaller chromosome (‘A’ chromosome) in species of Arachis and one chromosome pair harbouring a secondary constriction (‘B’ chromosome) in the same section. Later studies reported on one hand that hybrids between species having the small chromosome pair (A chromosome) were partially to fully fertile and will produce F2 seeds; on the other hand, hybrids between the species with A genome and the species harbouring a secondary constriction (B chromosome) were sterile (Stalker & Simpson, 1995). Thus, the terminology ‘A’ and ‘B’ genome has been used to describe the two cytological groups in section Arachis. Because the cultivated groundnut has one pair of smaller chromosomes and one pair of chromosomes with a constriction, it was described as an allotetraploid species with AABB genome (Husted, 1936). In order to better understand and clarify the phylogenetic relationships among groundnut species, molecular studies have been undertaken using isozymes (Lu & Pickersgill, 1993; Stalker et al., 1994), Restriction Fragment Length Polymorphisms (RFLPs) (Kochert et al., 1991; Paik-Ro et al., 1992; Gimenes et al., 2002; Burow et al., 2009); Randomly Amplified Polymorphic DNA (RAPDS) (Lanham et al., 1992; Hilu & Stalker, 1995); in situ hybridization (Raina & Mukai, 1999; Seijo et al., 2004; Custódio et al., 2013) and Simple Sequence Repeats (SSRs) (Hopkins et al., 1998; Nagy et al., 2012). Most of these studies have involved species in section Arachis because of their relatedness with A. hypogaea. Using isozymes, Lu & Pickersgill. (1993) 11 University of Ghana http://ugspace.ug.edu.gh differentiated the only B genome species described (A. batizocoi) from those of A genome and grouped species with A genome according to their life cycle (annual / perennial). Milla et al. (2005) confirmed using AFLP markers the differences between A and B genome species and figured out the close genetic relationship between B and D genomes. 2.7. Genetic diversity of Arachis Most of the variability in the genus was observed in South America, the centre of origin of groundnut. The genetic diversity of the genus is classified into four gene pools according to Singh & Simpson. (1994). The primary genepool consisting of A. hypogaea and its wild form A. monticola; both are AABB genome structure; the secondary genepool consisting of diploid species from section Arachis that are cross compatible with A. hypogaea; they are either A, B, D, F or K genome; the tertiary genepool consisting of species of section Procumbentes that are weakly cross-compatible with A. hypogaea and the quaternary genepool consisting of the remaining wild Arachis species classified into seven other sections. According to the distribution of the two subspecies of Arachis (A. hypogaea ssp. hypogaea and A. hypogaea ssp.fastigiata) and their respective botanical varieties described, seven secondary centers of diversity have been reported (Gregory & Gregory, 1976). The first is the Guarani region (between Paraná basin in Paraguay and southern Brazil), the second in Goias and Minas Gerais regions in Brazil, the third in the Rondônia region and the northeastern Mato Grosso in Brazil, the fourth in Bolivia (eastern slope of the Andes), the fifth from the Peruvian region to the west coast, the sixth in the northeast of Brazil and the seventh at the Ecuadorian region. West Africa and Southeast Asia were described as tertiary centres of diversity. The variability observed in these latter centres results from the combined effects of natural hybridization and selection. 12 University of Ghana http://ugspace.ug.edu.gh In total contrast to what have been observed in terms of phenotypic diversity, little variation has been reported at molecular level in cultivated groundnut (Gregory & Gregory, 1976). In fact several studies reported low levels of polymorphism even between the most distantly related cultivars in cultivated groundnut whatever the marker types used (Halward et al., 1991; Kochert et al., 1996; Gimenes et al., 2002; Tseng et al., 2016). The narrow diversity at the molecular level has been attributed to three major causes. Firstly the recent monophyletic origin of cultivated groundnut greatly reduced the genetic diversity in the gene pools. In fact, studies involving molecular markers suggested that A. hypogaea evolved from a natural single hybridization between two wild species, A. duranensis and A. ipaensis reported to be the progenitors of the cultivated form (Kochert et al., 1996). Thus cultivated groundnut has a single origin (Kochert et al., 1996; Bechara et al., 2010). Figure 2.1: Genetic bottlenecks on crop plants during domestication and plant breeding Different geometric figures represent allelic variations of genes originally found in wild species, but gradually lost through domestication and breeding. Such lost alleles can be recovered only by going back to the wild ancestors of our crop species: from (Tanksley & McCouch, 1997). Secondly, hybridization of the two genomes paved the way to polyploidization, an event that has led to the separation of the cultivated tetraploid from the rest of the diploid species (Hopkins et al., 1999). Barriers to gene flow from related diploid species to domesticated groundnut due to ploidy differences combined with the self-pollinating nature of the crop have blocked genetic 13 University of Ghana http://ugspace.ug.edu.gh exchange between cultivated groundnut and the wild species. The third factor that contributed to reduce the genetic base of cultivated groundnut is the domestication process. Tanksley & McCouch. (1997) clearly illustrated diagrammatically the reduction of the allelic diversity imposed on crop plants during domestication and plant breeding (Figure 2.1). 2.8. Targeted traits for groundnut improvement The targeted traits for improvement are country dependent and are determined by the level of productivity achieved and consumers’ and industry preferences. If productivity is no longer a challenge in developed countries where yields can reach 4 tons per ha, it remains a critical objective in developing countries. Yield and related traits are the most widely targeted traits of groundnut breeding programmes worldwide (Janila et al., 2013; 2016). The most important groundnut traits that contribute to yield are: number of pods per plant, shelling outturn, pod and kernel size and 100-seed weight. Recently, traits associated with yield without direct correlation such as ease in shelling, ease in harvesting (peg strength), reticulation, beak and constriction of pod are also important considerations (Janila et al., 2013). With the use of exotic germplasm to broaden the cultivated genepool, growth habit (erect, spread) is becoming an important breeding trait. In the Sahel context where groundnut plant is also used for animals feeding, the development of varieties with high haulm yield becomes important. In areas where groundnut is grown exclusively under rainfall, the crop cycle should match with the length of the rainy season. This is particularly important with the climate change and the shortening of the rainy season in the Sahel. Thus early maturity is an important trait in breeding programmes mostly in the semi-arid regions. Resistance to a specific groundnut insect pest and or disease is a target trait to reduce yield losses. 14 University of Ghana http://ugspace.ug.edu.gh 2.9. Potential of wild species in groundnut improvement Cultivated groundnut breeding has suffered from the lack of genetic diversity. Conversely, groundnut wild relatives have high diversity (Moretzsohn et al., 2004) and possess desirable alleles that can be used to improve resistance for several biotic and abiotic traits as well as quality traits such as protein and oil content (Nautiyal et al., 2008; Rao et al., 2003 ). While the exploitation of exotic germplasm is now becoming a routine in some crops notably tomato, rice, wheat and barley (Concibido et al., 2003; Feuillet et al., 2008; Bernardo, 2009; Jordan et al., 2011), it remains difficult for other crops like groundnut due mostly to cross incompatibility. Nevertheless, efforts have been deployed by the groundnut scientific community to ease the use of the wild species in breeding programmes. The first wide hybridizations between wild groundnuts with A. hypogaea L were in 1940 by Gregory and Krapovickas (Simpson, 2001). Their first attempts were unsuccessful, but with continued effort, their success rate improved over the years (Gregory & Gregory, 1979). Hammons. (1970) and Simpson & Smith. (1975) released respectively ‘Spancross’ and ‘Tamnut 74’ the first groundnut cultivars derived from interspecific hybridizations. Concerning the hybridization strategies, Simpson. (2001) described two major pathways namely the hexaploid and the amphidiploid. In the hexaploid way, the tetraploid cultivated species is crossed with a wild diploid. This results in a sterile triploid hybrid which, after chromosome doubling gives fertile hexaploid plants. The 4x ploidy level is recovered through successive backcross with the cultivated parent. North Carolina State University and ICRISAT have had success with this pathway (Stalker et al., 1979). However, high sterility, instability of the hybrids and the possibility of loss of chromosomes or chromosomal fragments are the main disadvantages of this method. In the amphidiploid way, two diploid wild species are crossed to 15 University of Ghana http://ugspace.ug.edu.gh produce a diploid hybrid. To obtain the tetraploid species, the obtained F1 interspecific hybrids are treated with colchicine-mediated chromosome doubling. The synthetic tetraploid is then crossed with the tetraploid cultivated parent. This technique has the advantage of directly generating individuals with the same level of ploidy as the cultivated species. The amphidiploid method is actually the pathway mostly used in transferring genes from wild species (Simpson, 1991; Simpson & Starr, 2001; Fonceka et al., 2012). A nematode resistance variety ‘NemaTAM’ was developed using this technique (Simpson et al., 2003). Table 2.1: Trait-related QTL/genes in groundnut characterized using wild relatives Trait/s Population Strategy Genotyping/Ma QTL/Gene Reference pping Root-knot nematodes SNP markers (Leal-Bertioli resistance, drought- RILs (wild × wild) and integrated QTL et al., 2016) related traits and consensus map agronomic/domestication from Shirasawa traits et al. 2013 Water availability, 87 BC3F1 and 55 BC2F2 flowering precocity, seed [cultivated × wild SSR QTL (Fonceka et and pod number, length amphidiploid (A. ipaensis al., 2012) and size, and pod × A. duranensis)] maturity BC4F2 population RAPD Resistance ( Burow et al., Root-knot nematode (cultivated × wilds) associated to 1996) resistance markers Plant growth habit, CSSLs [wild synthetic (A. SSR QTL (Fonceka et height of the main stem, ipaensis × A. duranensis) al., 2012) plant spread × cultivated] (Leal-Bertioli Late leaf spot resistance F2 (A. duranensis × A. SSR, AFLP QTL et al., 2009) stenosperma) Research teams in Brazil, USA and India are currently developing several artificial amphidiploids using different combination of A and B genome species. Nowadays, tremendous progress have been accomplished in using wild species in breeding programmes, thanks to the advance of genomic and biotechnological tools. These tools enable tackling differences in 16 University of Ghana http://ugspace.ug.edu.gh chromosome numbers, identifying wild chromosome segments that contribute significant improvement of traits of agronomical interest as well as dealing with linkage drag and deleterious alleles associated with wild x cultivated crosses. Table 2.1 gives an update of the interspecific mapping populations developed in groundnut and the target traits for which QTL have been identified. 2.10. Genomic resources in groundnut applicable in marker assisted selection 2.10.1 Molecular markers Isozymes were among the earliest genetic markers used in groundnut to study the genetic diversity of cultivated groundnut and its wild diploid relatives. Little variation was found among Arachis hypogaea (Grieshammer & Wynne, 1990). Three out of the 25 isozymes used in their study were polymorphic among 71 lines. However, isozyme polymorphisms were reported to be frequent among wild species compared to cultivated groundnut (Lanham et al., 1992; Lu & Pickersgill, 1993; Stalker et al., 1994). Isozymes have been used in other crops but, as in groundnut, low polymorphism have been generated (Weeden, 1989). The development of genetic markers evolved from the more complex and less informative protein-based molecular marker (isozyme) to the simplest and most informative DNA markers namely randomly amplified polymorphic DNA (RAPD), restriction fragment-length polymorphism (RFLP), amplified fragment-length polymorphism (AFLP), sequence-characterized amplified region (SCAR), simple sequence repeat (SSR) and single nucleotide polymorphism (SNP). RFLP have been used in diversity studies in groundnut and the results were the same as other markers: low polymorphism in Arachis hypogaea (Halward et al., 1991). RAPDs are PCR-based markers and are easy and rapid to perform. The disadvantage of RAPD markers is that only dominant markers are detected. However, they have been used in various studies. SSRs are based on PCR primers 17 University of Ghana http://ugspace.ug.edu.gh designed to flank a set of tandem repeats of a short DNA sequence (Liang et al., 2009). SSRs are highly polymorphic, multiallelic, codominant, abundant in the genome, simple to analyse, and transferable between species (He et al., 2006). Currently, more than 6000 SSRs have been developed and published for groundnut in PeanutBase (http://www.peanutbase.org). Until now, SSRs are being used for genetics and breeding studies in cultivated groundnut. With the genomes sequenced of the wild progenitors of groundnut (Bertioli et al., 2016) SNP markers are being widely used. Significant effort has been made at ICRISAT in collaboration with the University of Georgia in developing SNPs markers (Pandey et al., 2017). Pandey and collaborators developed a high density genotyping ‘Axiom_Arachis’ array with 58 K SNPs. This new technology will contribute to accelerate the process of molecular breeding in cultivated groundnut. Moreover, Bertioli et al. (2016) published the sequences of the genome of the progenitors of cultivated groundnut (A. duranensis and A. ipaensis). This opens new perspectives for the more efficient use of groundnut genomic resources and constitutes a starting point for the whole groundnut genome sequencing which has been initiated by the Groundnut Genome Project (PGP) in collaboration with BGI-Shenzhen (China). 2.10.2 Genetic linkage map A major application of molecular markers is the construction of linkage maps required for QTL mapping and marker-assisted breeding. Linkage maps were constructed using diploid species (either AA or BB genome) as well as tetraploid species with different marker types. Using RFLP, Halward et al. (1993) developed the first linkage map in groundnut. This map was constructed with an F2 population derived from the hybridization between A. stenosperma and A. cardenasii both AA genome species. The first SSR based linkage map in Arachis consisted of 170 markers mapped on 11 LGs (Moretzsohn et al., 2005). A linkage map was also constructed 18 University of Ghana http://ugspace.ug.edu.gh with an F2 population which involved BB genome species A. ipaensis and A. magna (Moretzsohn et al., 2009). The same year, Foncéka et al. (2009) published a SSR-based interspecific tetraploid map using a BC1F1 population [“Fleur11” × (A. ipaënsis × A. duranensis)4x], and 298 loci were mapped in 21 LGs. Recently, Nguepjop et al. (2016) constructed a genetic map of 357 loci using an F2 population [“Fleur11” × (A. ipaënsis × A. batizocoi)4x] where the authors reported homeologous recombination between the A and the K genomes. Cultivated groundnut was also used to construct linkage maps. Two intra-specific maps were first published (Varshney et al., 2009; Hong et al., 2010). These two maps were developed using SSR markers. A linkage map spanning a total distance of 1,446.7 cM with 1685 markers including 1621 SNPs and 64 SSRs was recently published by Zhou et al. (2014). Two consensus map based on 11 mapping populations and 3 recombinant inbred line populations were constructed (Gautami et al., 2012; Shirasawa et al., 2013). 2.11. Exploitation of groundnut wild allelic potential for varietal creation 2.11.1 Interests of integrating molecular breeding approaches For harnessing the beneficial alleles from wild species, particular types of mapping population such as AB-QTL, CSSLs, and NILs have been developed for different crops (Concibido et al., 2003; Feuillet et al., 2008; Bernardo, 2009; Jordan et al., 2011). Figure 2.2 gives a comparison of the different types of population. The development of such interspecific populations has been made possible by the use of molecular markers. DNA markers come as a tool not only to speed up the process of developing new varieties but to increase the efficiency too. The major benefits of integrating molecular markers in breeding approaches are to: 19 University of Ghana http://ugspace.ug.edu.gh Figure 2.2: Comparison of the methods for creating primary and advanced bi-parental mapping populations As shown are the number of backcrosses required and the genotypes of the lines obtained by each of the following methods recombinant inbred lines (RILs), backcross inbred lines (BILs), chromosomes segment substitution lines (CSSLs) and near isogenic lines (NILs). Also shown are Karyotypes of the three CSSLs illustrate how chromosome 1 of the donor can be introgressed into the recurrent parent genetic background (Shakiba & Eizenga, 2014). ü Handle complex traits that may not be amenable by conventional breeding' methods due to their polygenic nature and the large G x E interaction. ü Transfer desirable traits from wild species into improved varieties while getting rid of the unfavourable traits. 20 University of Ghana http://ugspace.ug.edu.gh ü Introgress part or the entire genome of wild specie into cultivated background which enable tapping alleles from wild species and varieties development. ü Select plants on the basis of their genotype and thus optimize time, resources and the cost of phenotyping. ü Avoid linkage drag and limit the failure of phenotype screen, a recurrent problem with field evaluation. ü Pyramid genes or QTL through molecular breeding. Marker-assisted backcrossing (MABC) is extensively used in molecular breeding to improve existing varieties and to pyramid few genes / QTL. It is also used to develop near-isogenic lines (NILs) or chromosome segment substitution lines (CSSLs). The most important limitation of MABC is its inability to deal with several traits simultaneously because of the difficulty of managing large population size (Ribaut & Hoisington, 1998; Varshney et al., 2012). For such traits, marker-assisted recurrent selection (MARS) has been suggested. MARS can target simultaneously several minor and major QTL (Ribaut & Ragot, 2007). Actually in groundnut breeding, MABC is used for improving simple traits such as oil quality traits and resistance to biotic constraints while for more complex traits like drought tolerance where 100 main and epistatic effects QTL were mapped, MARS was proposed (Ravi et al., 2011). 2.11.2 The concept of advanced Backcross (AB) methodology According to Tanksley & Nelson. (1996), the less expected impact of marker-based QTL analysis on the development of varieties may be attributed to two factors: (1) QTL discovery and variety development are separate processes; (2) Most breeding-related QTL studies are targeted toward manipulating quantitative trait variation within elite germplasm. Advanced Backcross QTL (AB-QTL) population was first described by Tanksley et al. (1996) in tomato for 21 University of Ghana http://ugspace.ug.edu.gh overcoming these two factors. It refers to a methodology that allows mapping and introgressing QTL from an un-adapted donor into the background of an elite variety. The creation of an AB mapping population follows generally the steps described by Tanksley et al. (1996): a wild specie is used as the donor parent and is crossed with the recurrent parent, usually an elite cultivar. The hybrid is then crossed with the recurrent parent in subsequent crosses. To avoid the cytoplasmic male sterility, in most cases the donor parent is a male and the recurrent parent is the female. The first generation of backcross results from a cross between the F1 hybrid (s) and the recurrent parent. The resulting BC1 may be backcrossed again with the recurrent parent to generate a BC2 population. Since the publication of Tanksley et al. (1996), the AB-QTL methodology has been used in various crops to tap the variability that exists in crop wild relatives (Huang et al., 2003; Korff et al., 2006). The advantages of the AB populations are that one may develop AB-QTL population (Tanksley et al., 1996), BIL population (Keurentjes et al., 2007), CSSL population (Fonceka et al., 2012) or NIL population (Monforte & Tanksley, 2000). In the process of developing AB-QTL populations, lines with unwanted genes from the donor parents are often removed from the population after phenotypic and genotypic evaluation. The selection in favour of lines with desirable alleles and the genetic background of the recurrent parent can skew the allele’s distribution toward the recurrent parent. Therefore, in more advanced backcrossing generations for example after the BC3 generation, the power of the statistical analysis to detect QTL decreases. Since sequential backcrossing in AB-QTL removes epistatic interactions, the chance of detecting QTL with epistatic interactions among alleles from the donor parent decreases due to the overall lower frequency of donor alleles. Nevertheless the ability to detect additive, dominant, partially dominant or over dominant QTL increases (Tanksley & Nelson, 1996) and 22 University of Ghana http://ugspace.ug.edu.gh because of the homozygosity status of the lines developed, multi-location trials may be done on interesting individual(s) for varietal release. Research to broaden the gene pool of cultivated groundnut using its wild relatives has been undertaken in Senegal resulting in the development of two AB-QTL populations (unpublished data). Moreover a pyramiding programme to increase seed size and biomass is ongoing using the CSSLs lines developed (Fonceka et al., 2012). Such populations would be useful in creating new varieties with desired preferences. 23 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE 3.0 ASSESSMENT OF FARMERS’ PRODUCTION CONSTRAINTS AND VARIETAL TRAIT PREFERENCES OF GROUNDNUT IN SENEGAL 3.1 Introduction Despite its leading position in terms of production and acreage in the country, groundnut production in Senegal (669329 tons in 2014) is low compared to China (6 million tons), India (5.6 million tons) and Nigeria 1 to 1.5 million tons (FAOSTAT, 2015). Many efforts have been made since 1966 by the National breeding programme in developing varieties that are adapted to Senegalese environments. However, the average national yield is still low and is around 1t/ha on farmers’ fields, which is below the yielding potential of the varieties that have been released. A number of questions related to the adoption of the varieties that have been released include; i) the awareness of farmers of the existence of newly developed varieties that are more adapted to climatic variations; and ii) the characteristics of such varieties in terms of targeted traits preferences of farmers. These questions may help to have a better understanding of the problem and could help the breeder to be more relevant in his/her breeding objectives. Plant breeders, to improve the level of adoption of newly developed varieties, should work closely with farmers through participatory plant breeding. This is where farmers’ knowledge and preferences are integrated during the cultivar development (Bidinger, 1998). Oppong-Sekyere et al. (2015) conducted a Participatory Rural Appraisal (PRA) in Upper East, Upper West and Northern Ghana to identify the major characteristics of groundnut production. They found that drought was the major constraint in Northern Ghana. Egbadzor et al. (2013) assessed cowpea production constraints using matrix scoring, pair-wise ranking and individual interviews in Ghana. Nowadays, such studies are widely conducted by plant breeders to facilitate adoption of newly bred cultivars by farmers (Ceccarelli et al., 2007). In fact, this approach has been 24 University of Ghana http://ugspace.ug.edu.gh demonstrated to enhance adoption and may help plant breeders to identify farmers’ production constraints and preferred traits (Mulatu & Belete, 2001). This approach is also more effective in terms of cost and transferring the most suitable varieties and packages to farmers (Mangione et al., 2006). While such studies have been conducted on groundnut in other countries in Africa (Ntare et al., 2007; Monyo et al., 2012; Oppong-Sekyere et al., 2015), limited information on the main groundnut production constraints and farmers’ preferred traits is available in Senegal where groundnut is one of the most important crops. Most surveys made on groundnut sector in Senegal are on the economic aspect of the crop. Assessment of constraints and farmers’ preferred traits for a specific crop does not have a global application. According to Omanya et al. (2007) plant breeders should develop, based on survey studies, cultivars adapted to a specific location rather than aiming at a wide range of environments. Knowledge of the groundnut production constraints and farmers’ preferred traits would guide the breeding programme to set breeding objectives to meet farmers’ needs. The objectives of this study were to: • identify farmers’ groundnut production constraints; and • identify varietal trait preferences in the groundnut basin of Senegal 25 University of Ghana http://ugspace.ug.edu.gh 3.2 Methodology 3.2.1 Study area The study was conducted between March and April 2015 in two agroecological zones in the groundnut basin in Senegal. These included three representative groundnut growing regions namely Kaolack, Kaffrine and Tambacounda (Figure 3.1). These areas were chosen because they produce groundnut in large quantities and on large areas (ANSD, 2015). There is only one rainy season from July to October and the rainfall amount ranges from 600 to 1200 mm per year. One department was selected in each region. Within each rural commune, three villages were selected providing a total of 9 study sites. Figure 3.1: Location of the study sites on the Senegal map 26 University of Ghana http://ugspace.ug.edu.gh 3.2.2 Sampling procedure The sampling procedure was done with the assistance of the regional agriculture extension officers. A multi-stage sampling procedure was used to select the three most important growing regions (Kaolack, Kaffrine and Tambacounda) and then the biggest producer department in each region. At the department level, three villages were chosen. In Kaffrine and Tambacounda, the chosen villages are in two rural communes and in Kaolack the three villages are in three different rural communes. Seven rural communes were involved in the survey. Nine (9) FGDs comprising of ten to seventeen farmers were organized (one in each village). Local extension staff and sub- district managers at the respective localities facilitated the PRA process by mobilizing farmers to participate in the FGD. In addition within each village, 10 farmers were randomly selected providing a total of 90 producers for the semi-structured interviews. 3.2.3 Data collection Focus group discussions were standardized using a checklist of topics to be covered. The topics discussed included: farm characteristics and farmers’ agronomic practices, farmers’ varietal preferences, marketing of their produce and the perceived constraints affecting the production. The group discussions were followed by formal interviews using semi-structured questionnaires to explore more specific issues. During the survey pictures and drawings were frequently used to illustrate difficult points. Ten important traits including pod and haulm yield, morphometric and morphological descriptors for pod, seed and growth habit were used to assess farmers’ preferred plant characteristics for improved groundnut varieties. Prioritization of the constraints was based on frequency of occurrence and severity. These constraints to groundnut production were identified and listed during the focus group discussions. They were further ranked by each interviewed farmer. 27 University of Ghana http://ugspace.ug.edu.gh 3.2.4 Data analysis Data were coded and entered in Microsoft Excel. The analysis was performed using Statistical Package for the Social Scientists (SPSS version 19). Data were analysed per rural commune. Variables were subjected to descriptive statistics, cross tabulation, one-sample t (two-sided) and chi-square tests. Simple scoring and ranking techniques were used to identify farmers’ production constraints and preferences. Standard errors were used to separate means where necessary. 28 University of Ghana http://ugspace.ug.edu.gh 3.3 Results 3.3.1 Demographic and socio-economic characteristics Of the total interviewed farmers, 88.9% were males while 11.1% were females Table 3.1. For 72.2% of the respondents farming is their only source of income while 27.8% had additional income-generating activities. Most of the interviewed farmers 65.6% and 16.7% have been respectively in Koranic school and local language schools. A small percentage (8.9%) was educated up to primary level, 5.6% secondary level and 3.3% without any education. The most representative range of age in the study areas was between 35 and 60. There was significant difference (p= .004) in terms of age in the selected rural communes. The proportion of the active population involved in groundnut production was 71.1%. About 63% did not have any form of support or training and only 36% of the farmers belonged to farmer association. 3.3.2 Groundnut importance and production level Although the selected rural communes were predominantly groundnut producing areas, other crops like cereals played a key role in the cropping systems. A ranking of the importance of groundnut compared to other crops was done on the basis of the cultivated areas allocated to each crop and the production levels. The average land holding varied from a minimum of 1 to a maximum of 30 ha and significant difference of the importance of the crops grown was observed between locations. In almost all rural communes groundnut occupied the most important piece of land and was ranked number one with an average mean rank of 1.3. 29 University of Ghana http://ugspace.ug.edu.gh Table 3.1: Demographic and socio-economic characteristics of the respondents across locations Rural commune Chi-x2 Variables Modalities G Kaye Kahi Kathiot Kaymor Koussanar Maka Porokhane Total % P-Value Gender 0.239 female 0 3 0 3 3 1 0 10 11 male 10 17 10 7 17 9 10 80 89 Age 0.004 <15 0 0 0 0 1 0 0 1 1 15-35 0 2 3 1 0 5 2 13 14 35-60 3 10 5 7 16 2 8 51 57 >60 7 8 2 2 3 3 0 25 28 Education 0.066 koranic school 9 13 7 9 7 5 9 59 66 local language 0 6 2 1 5 1 0 15 17 none 1 1 0 0 1 0 0 3 3 primary 0 0 1 0 5 2 0 8 9 secondary 0 0 0 0 2 2 1 5 6 MOA 0.006 no 10 11 6 7 7 4 9 54 60 yes 0 9 4 3 13 6 1 36 40 yes 5 16 7 7 16 6 8 65 72 Support 0.115 no 10 12 7 6 11 8 9 63 70 yes 0 8 3 4 9 2 1 27 30 Training 0.009 no 10 16 8 10 11 9 10 74 82 yes 0 4 2 0 9 1 0 16 18 MOA: Member of association 30 University of Ghana http://ugspace.ug.edu.gh Groundnut was followed by pearl millet or corn or sorghum depending on the locality. In the central part of the groundnut basin (Kaolack, Kaffrine), pearl millet was the second most important crop. Elsewhere in the eastern part of the country (Tambacounda region), corn or sorghum was ranked the second most important crop. The comparison of the mean yield revealed significant differences among localities (p=.042) (Table 3.2). Table 3.2: Groundnut pod yield estimated as a function of area and NPK fertilizer applied Guin Kaye Kahi Kathiot Kaymor Koussanar Maka Porokh Total F Yield 1500.25 960.19 967.40 1200.00 1110.79 1289.44 912.50 1112.01 .042 (kg/ha) NPK 116.00 53.50 85.00 45.00 90.20 72.50 28.00 70.43 .05 (kg/ha) Area (ha) 4.90 5.81 8.90 5.70 5.61 5.56 12.78 6.76 The average groundnut yield was estimated at 1112 kg/ha. The highest (1500.25 kg/ha) was recorded in Guinth Kaye in the groundnut basin while the lowest value 912.5 was recorded in Porokhane (Table 3.2). Both localities are in the same region (Kaolack). This result showed differences within the same region. Similar trend was observed for the application of fertilizer. On average, 116 kg/ha of NPK (6 20 10) fertilizer was recorded at Guinth Kaye while 28 kg/ha was recorded at Porokhane. The higher the quantity of NPK applied the higher the yield was. The regression analysis showed that the amount of NPK applied was one of the factors determining the yield (P =.001) and also the area allocated to groundnut did not have any impact on the yield obtained (Table 3.3). 31 University of Ghana http://ugspace.ug.edu.gh Table 3.3: Regression analysis between estimated yield, rate of fertilizer and acreage Model Mean Square F Sig. R2 Regression 962984.662 5.474 .006 Residual 175906.986 Total dose 3.2 96 .00 1 0.1 27 area -.859 .393 3.3.3 Varieties grown and their provenance Six varieties were cultivated in the surveyed zones as shown Figure 3.2. Some varieties were more popular than others depending on the locality. Out of the six varieties, ‘73-33’ was the most cultivated and was ranked number one in almost all rural communes except in Guinth Kaye, Porokhane and Kaymor. The variety ‘28-206’ was predominantly grown in Guinth Kaye and Porokhane while ‘GH119-20’ was the most cultivated in Kamor. N of farmers 25 20 15 10 5 0 GH119_20 73-33 fleur 11 55-437 28-206 unknown Figure 3.2: Cultivated groundnut varieties by community A total of 60 respondents representing 66.67% of the farmers cultivated 73-33. Fleur11 was grown in few communes (Kahi & Kathiot). Two percent of the respondents did not have any idea 32 University of Ghana http://ugspace.ug.edu.gh about the type of variety they grew. Most of the farmers around 83.33% used saved seeds from the previous harvest. Twenty percent of the respondents had previously obtained government- subsidized seeds or procured them from the local market. In Koussanar, farmers belonged to a farmers’ organization and made seed’ multiplications for the next growing season. 3.3.4 Crop management Ninety nine percent of groundnut farmers did not plough their fields. Land preparation consisted mainly of cleaning the field manually using home-made craft tools. The overall results from both focus group discussions and formal surveys demonstrated that farmers sowed groundnut seeds at the onset of the rains in humid soil. All known Senegalese’ types of soil (clayey, sandy-clay and sandy) were used to cultivate groundnut with, however, a predominance of sandy soil (42%). Thirty six percent of the farmers cultivated groundnut on sandy-clay soils. The most common row spacing used was 50 cm (51% of the total respondents). However, 30% of farmers used 40 cm row spacing. While the difference in row spacing did not vary significantly (F= .61), significant differences (F < .05) between rural communes was observed in the distance between plants in a row. Mainly two seed-sowing discs were used 30 cm and 20 cm. In Koussanar and Makacolibatang (Tambacounda region), 30 cm spacing between plants in a row was used while in Kahi, Guinth Kaye and the remaining rural communes 30 and 20 cm were used. Approximately 94% of the farmers treated their seeds against insect pests and other pathogens using fungicides before sowing. Around 66% of farmers used mineral fertilizer (NPK) at different doses. Groundnut was mainly cultivated in rotation with different crops mainly cereals in all localities (98%). In the surveyed zones of Tambacounda, it was cultivated in rotation with maize or sorghum while in Kaffrine and Kaolack pearl millet was the main crop used in rotation. 33 University of Ghana http://ugspace.ug.edu.gh Weeding was done three times in a cropping season and there was no chemical weed control in all study zones. 3.3.5 Farmers trait preferences for groundnut variety Except for tolerance to biotic stresses (P=.007) where highly significant difference was observed among communities, all other preferences ranked by farmers were non-significant (Table 3.4). Most of the farmers identified haulm yield as the second most important character after pod yield, except in Koussanar where good pod filling was ranked number one. The same locality (Koussanar) identified erect plant habit as an important groundnut criterion for variety adoption. Farmers in Kahi, Kathio, Kaymor and Makacolibatang preferred varieties that had more pod weight than those which were light. Farmers preferred intermediate season varieties to very short season varieties. Almost all farmers in the study areas did not want extra-early maturing varieties (less than 80 days) as they were prone to additional rainfalls and animal’ damage. Marketability was ranked among farmer-preferred traits. Table 3.4: Farmers preferences for groundnut traits in different communities Rural communes Total Traits Guinth Kahi Kathiot Kaymor Koussanar Maka Porokhane Mean Rank F Pod yield 1.00 1.00 1.00 1.00 1.19 1.00 1.00 1.04 1 0.783 biotic toleran 5.00 3.67 1.67 3.33 3.25 4.00 3.31 8 0.531 drought toler 2.75 3.00 3.50 4.00 3.38 2.33 3.00 3.08 7 0.007*** Short cycle 4.00 3.00 3.13 1.00 3.33 2.83 3.06 6 0.413 Haulm yield 2.00 2.19 2.43 2.17 2.36 2.00 2.00 2.19 2 0.756 Pod weight 2.00 2.00 2.00 2.50 2.20 3 0.521 Oil content 3.75 4.00 4.00 3.50 3.67 . 3.00 3.80 9 0.919 Erect plant . . . . 2.00 . 3.00 2.25 4 0.276 Good pod fillin 3.00 2.56 2.00 3.00 1.00 2.50 3.50 2.65 5 0.827 Good taste 4.00 5.00 . 4.00 5.00 . . 4.60 10 0.764 *Characteristic with smallest mean rank is perceived to be the most important Tolebiotic: tolerance to biotic, toledrought: tolerance to drought 34 University of Ghana http://ugspace.ug.edu.gh During group discussions, farmers unanimously mentioned that a variety with long and large pod and seed characteristics commanded better acceptability in local markets when sold and produced higher yield. Farmers correlated big pod and seed size with high yielding and marketability. In general, farmers were not concerned much about storability and animal’ feed quality in groundnut haulm and ranked them low. Farmers argued that they had not seen groundnut variety with resistance to storage pests nor with special haulm qualities as feed for animals. 3.3.6 Groundnut production constraints The top five constraints to production across the commune were availability of seed, land degradation, insects in storage, land availability and drought (Table 3.5). Table 3.5: Relative ranking of major groundnut production constraints by community constraints Rural communes Total Guint h Kahi Kathiot Kaymor Koussanar Maka Porokhane Mean F Rank Land degradation 2.31 2.33 2.2 2.06 1 1 4.66 2 0.47 2 Insects at storage 4.01 2 1 3 2.40 1.56 4 2.04 0.06 3 Leaf diseases 3.89 2.78 1 1 2 3.21 3.41 2.56 0.41 7 Drought 4 2.67 3 3.67 1.75 1.75 4 2.36 0.02 5 Land availability 2.17 1 3 2 3.71 4.18 3 2.13 0.60 4 Seeds availability 2.82 1.63 2 1.6 1.67 2.23 2 1.91 0.14 1 Commercialization 3.44 2 2.5 3 1.75 2 2.45 2.50 0.02 6 Lack of tools 5 3 2 2.40 1.67 2.56 2 2.58 0.047 8 Cattle damage 5.07 4.22 4.34 3 3.50 4 5 3.33 0.667 9 *Characteristic with the smallest value is perceived to be the most important Marketing and foliar diseases were ranked sixth and seventh respectively. These constraints were common to groundnut growing areas but the order mostly varied from one commune to another. Insects mainly Caryedon serratus was identified as an important constraint to groundnut storage 35 University of Ghana http://ugspace.ug.edu.gh and ranked as main constraint in Kathiot with foliar diseases. Lack of agricultural tools was ranked in Koussanar, Kaymor and Kathiot as an important constraint. Significant differences were observed in the ranking of drought, commercialization and lack of tools. 36 University of Ghana http://ugspace.ug.edu.gh 3.4 Discussion A small number of women (11%) were represented in this study. This reflects the limited access they have to land. The major explanation of this situation is socio-cultural. In Senegal, land’ acquisition is by inheritance, transmitted from father to son. As a result, few women have their own land. Diatta. (2017) reported the low involvement of women in sorghum and pearl millet cultivation due to their limited access to land. They found that women were more involved in vegetable cultivation which required meticulous techniques. Groundnut was the most important crop grown by farmers in the areas of the study. Production for the market was the major objective followed by animal feeding (production of biomass). These results suggested that the importance of groundnut in farmers’ livelihood is strongly associated with its use as a source of income. Besides groundnut, farmers cultivated mainly in rotation a wide range of crops such as pearl millet, sorghum, maize, rice, fonio, cowpea etc. The option of growing several crops is a strategy to improve resilience (Lin, 2011). According to Traoré et al. (2015) it may prevent farmers from focusing sufficiently on adequate production of one particular crop. Moreover, it allows farmers to have a staple food in addition to cash crop like groundnut. Crop rotation plays an important role in groundnut management and farmers have understood this. In most studied sites, pearl millet was the main crop cultivated in rotation with groundnut. It is known that crop rotation is beneficial for restoring soil fertility and for preventing crop pests and soil borne diseases. This may partly explain why farmers did not apply pesticides or insecticides in their fields. These products are beyond the reach of farmers who also think that such products degrade the soil. The average yield estimated (1.1 t/ha) in the studied sites was not so different from the national average of 0.9 t/ha reported by Sylla. (2016). However, variability was noted among sites regarding yield. This variability was related to the quantity of mineral 37 University of Ghana http://ugspace.ug.edu.gh fertilizer applied as shown by the results. In general, as the dose of NPK applied increased, the yield increased in most of the study sites. But exceptions were observed for example at Koussanar where the yield obtained with fertilizer did not reach the yield in Makacolibantang whereas higher quantities of NPK were applied. These results suggested that besides fertilizer application there were other factors that determined the yield obtained. This difference may due to adequate rainfall, the type of soil or different constraints that farmers faced in groundnut cultivation. Difficulties faced by farmers in groundnut cultivation could be classified in two groups: the first group included the most important constraints ranked by priority by farmers namely seed availability, land degradation, storage pests, land availability and commercialization. The consistency in ranking of these constraints over the communities suggested that these were major constraints affecting groundnut production in the groundnut basin. The first group of constraints involved the whole value chain and could not be addressed by research only except for insects during storage. One of the major insects that plague groundnut production is Caryedon serratus. This insect is a coleopteran which can cause within four months up to 80% of production losses in storage in Senegal (Diome et al., 2013). Entomologists have made huge efforts in determining the mechanisms involved in the infestation process and how to minimize them. They have discovered that the Caryedon species that infest groundnut pods are genetically close to those that infest the pods of a widely distributed legume in most areas of Senegal, Ngigis (Piliostigma reticulatum). Noting that the infestation is almost absent when this wild legume is absent, recommendations have been formulated to reduce the incidence of the Caryedon species on groundnut (Sembene et al., 2006, 2008). 38 University of Ghana http://ugspace.ug.edu.gh The second group included drought and leaf diseases. Senegal is located in the Sahelian zone and is subjected to different scenarios of drought. As a result production of many crops is threatened yearly by the inter annual and across location rainfall variability. Drought is characterized by an unequal distribution of rainfall in both time and space (Upadhyaya, 2005). Results showed significant differences in drought ranking between the communities. In the localities of the eastern part of the groundnut basin (Tambacounda region) where the annual average rainfall is 800 mm, drought is perceived to be more important than in the central zone where the annual rainfall is between 500 - 600 mm. This paradoxical situation could be explained by the type of cultivated variety and the distribution of the rains. In the Eastern zone where intermediate and long cycle varieties are grown, two drought periods are generally experienced. One is at seedling stage immediately after the first rainfall events; drought at seedling stage is manageable because of the plasticity of the crop (Reddy et al., 2005). The second occurs during the pod filling stage. Varieties 73-33 and 28-206 are long cycle varieties that are mostly cultivated in the eastern zone while in the central areas a mixture of short and intermediate cycle varieties was grown (55-437, Fleur11 with 90 days). Seed availability and drought were identified as the major challenging constraints in Mali (Ndjeunga et al., 2010; Ntare et al., 2012). Non-availability of seed was also identified as a constraint in Karnataka State in India (Varman et al., 2012). Oppong-Sekyere et al. (2015) reported drought as a major constraint to groundnut in northern Ghana. In Mozambique, drought was the main constraint to cowpea production (Chiulele et al., 2011). The years 2002 and 2011 were the most challenging years in groundnut cultivation in Senegal due to low and erratic rainfall patterns. The impact of such events is not only extremely damaging, but it can also be lasting. Development of varieties better adapted to cope with the impact of drought is required as drought is unpredictable and erratic. In Senegal, Clavel et al. (2006, 2008) worked 39 University of Ghana http://ugspace.ug.edu.gh to identify drought tolerant germplasm and breeding lines but they experienced difficulties due to the limited genetic resources available in Arachis hypogaea to handle complex traits like drought. In this context, expanding the genetic base of cultivated groundnut could play an important role in finding new sources of variability. In Kathiot and Kaymor (Nioro zone), leaf diseases were the major constraints that affected groundnut production while in the other localities, leaf diseases were not perceived by farmers as important. Environmental conditions such as weather, susceptibility of the host plants related to the type of varieties grown could explain such disease occurrence. However, good agricultural practices may help in the management of such diseases. Groundnut is cultivated in rotation with pearl millet or with other cereal species mostly as it was the case in most of the studied zones. For farmers, crop rotation is a strategy which can restore soil fertility, avoid severe foliar diseases and prevent soil borne diseases. Farmers did not apply any fungicides for leaf diseases. These results indicate that breeding for improved drought tolerance and leaf diseases resistance should constitute major breeding objectives to ensure high yield and varietal acceptance by farmers in specific zones. Farmers indicated pod and haulm yield as the most important criteria for selecting groundnut varieties. High pod yield was the most highly ranked criterion in all surveyed sites; followed by biomass yield. Farmers gave importance to seed and pod characteristics. Farmers correlated pod characteristics to pod yield. However, scientific studies have shown that there was no correlation between pod yield and pod and seed size characteristics (Fonceka et al., 2012). Another important consideration for farmers is the attractiveness and the business side of large pods and seeds. The fact that groundnut is a cash crop could justify why farmers paid much attention to pod and seed size. Similar consideration was given to big seeds and pods by farmers in Mali. In a 40 University of Ghana http://ugspace.ug.edu.gh study conducted in Mali by Sanogo. (2017), pod yield, pod and seed size played an important role in farmers’ choices in addition to pod yield. Oil content and good taste received little attention. This result could be different if the survey had taken into account the industrial processors and end users. The agreement between farmers and breeders in some of the criteria used for selecting groundnut varieties suggests the need for collaborative work to improve the efficiency of selection. 41 University of Ghana http://ugspace.ug.edu.gh 3.5 Conclusion The most important constraints across sites were seed availability, land degradation, insects at storage and land availability. Drought was the fifth constraint with differences between sites. Marketing and foliar diseases were important constraints to farmers. Insects mainly Caryedon serratus was identified as an important constraint to groundnut storage and was ranked as main constraint in Kathiot with leaf diseases. Pod and haulm yield were the most preferred traits in groundnut variety. Large pod and seed were also important traits in groundnut variety. 42 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR 4.0 ASSESSMENT OF AGRONOMIC POTENTIAL OF AN ADVANCED BACKCROSS QTL POPULATION AND IDENTIFICATION OF PROMISING BREEDING LINES 4.1 Introduction Exploiting the genetic potential of wild species has been of great interest in several plants mostly for crops with narrow genetic base (Tanksley & McCouch, 1997; Zamir, 2001; Zhang et al., 2017). Cultivated groundnut has a narrow genetic base (Kochert et al., 1991) and wild groundnut species have been reported to be rich in alleles that can be used to enrich the cultigen genetic background. For a long time, wild groundnut species have been used as donors to transfer useful characters, mainly biotic resistance traits, in cultivated varieties (Mallikarjuna et al., 2004). Because of the polyploidy differences between the tetraploid Arachis hypogaea and the diploid species of the section Arachis, two methods have been mainly used to move genes from wild species to cultivated species: the tri-hexaploïd technique and the amphidiploid technique (Simpson, 2001; Fávero et al., 2006). The success of the first method has been limited due to the high rate of sterility and the instability of the hybrids. The latter method has the advantage of generating tetraploid hybrids with the same level of ploidy as the cultivated species. This method has been used to transfer nematode resistance genes from A. cardenasii into A. hypogaea (Simpson & Starr, 2001; Simpson et al., 2003). In terms of resistance to biotic stresses, wild species have been greatly exploited. However, for traits related to yield, wild species have not been of much interest in groundnut improvement. Nevertheless, in crops such as tomato, barley and rice, many studies have demonstrated that genomic regions from wild species can be used to improve complex traits including yield (Gur & Zamir, 2004; Fu et al., 2010; Nevo & Chen, 2010). The under exploitation of wild species to improve complex traits may be partly attributed to their inferior phenotypic value associated often with undesirable characters. 43 University of Ghana http://ugspace.ug.edu.gh In fact the wild species’ alleles have no agronomical interest in a wild genetic background. However what would be of interest in terms of yield components is a small proportion of the wild genome in a cultivated genetic background. Fonceka et al. (2012) developed an interspecific genetic material using a synthetic allotetraploid created from the progenitors of the cultivated groundnut (Arachis duranensis and Arachis ipaensis). They found a valuable benefit of using wild species in improving agronomic traits in groundnut. In the present work, a different synthetic allotetraploid donor ISATR52B was used to: • develop an interspecific mapping population, • assess the agronomic potential of the introgressed groundnut lines; and • identify superior and stable genotypes for agronomic traits from the introgressed lines. 44 University of Ghana http://ugspace.ug.edu.gh 4.2 Materials and Methods 4.2.1 Population development A synthetic allotetraploid ISATR52B derived from a cross between [A. duranensis (AA genome) X A. valida (BB genome)] 4x was developed and provided by ICRISAT. The allotetraploid was then used to develop a population of 200 BC2F4:6 progenies at CERAAS (Figure 4.1). The synthetic wild allotetraploid ISATR52B was used as the male donor and Fleur11 a cultivated Spanish type variety of A. hypogaea used as the recurrent female parent. Fleur11 is popular moderate yielding variety introduced from China in 1990. The advanced backcross population BC2F4:6 was developed using two successive generations of backcrossing of selected F1 individuals to Fleur11 followed by four generations of self-fertilization (from BC2F1 to BC2F4) using single seed descent (SSD). Then seed multiplication was done in bulk up to BC2F4:6. Crosses were performed under greenhouse at the Centre d’Etudes Regional pour l’Amelioration de l’Adaptation a la Secheresse (CERAAS, Senegal) between 2012 and 2015. The Fleur11 was used as female parent to produce the F1, BC1F1 and BC2F1 generations. The process of the crossing was as follows: Fleur11 plants used as female were manually emasculated the afternoon of the day before each crossing step. The pollinations were performed early in the morning the following day between 7 am and 10 am by capping the pollen sacs of the flowers of the male parent on the pistils of the female parent. Flowers that had not been pollinated were removed manually. F1 hybrids were selected based on phenotypic appearance. For the BC1 and BC2, microsatellite markers were applied to select the true BC1 and BC2 individuals. Eight SSRs were used to select the BC1 and 20 markers for the BC2. Figure 4.1 gives a summary of the crossing scheme and the number of selected plants advanced at each generation. 45 University of Ghana http://ugspace.ug.edu.gh Figure 4.1: Description of the development of the advanced backcross QTL (AB-QTL) population 46 University of Ghana http://ugspace.ug.edu.gh 4.2.2 Study sites and field design The 200 lines along with the recurrent parental line Fleur11 were evaluated in five environments. Two water regimes were imposed in 2015, well-watered (WW) and water limited (WS) (pod filling), during the off season (October - January) at ISRA’s station of Nioro du rip (13°43’N; 15°46’W) located in South Groundnut basin zone. The two experimental conditions received the same quantity of water from sowing up to 65 days after sowing (DAS). A total of 30 mm of water per week divided in two irrigations /week was applied. From 66 DAS, irrigations were stopped in the WS until harvest (53 days of stress). The WW received in total 331 mm and the WS received 157 mm of water. The lines were also evaluated in 2016 during the rainy season (July - October) in three locations: Nioro du rip, Darou located both in south- eastern part of the Groundnut basin and the third location was ISRA’s station of Sinthiou Maleme (13°49’N; 13°55’ W) (Figure 4.2). Figure 4.2: Location of the study sites 47 University of Ghana http://ugspace.ug.edu.gh Each experiment was arranged in an augmented incomplete block design made of 25 blocks of 8 families with 5 checks including Fleur11. The block size was 13 entries. The plot size was 1 row of 3 m length with 10 plants, 30 cm between plants and 50 cm inter-row spacing. Standard field management for groundnut was applied in all experiments. 4.2.3 Data collection Thirteen quantitative and three qualitative traits were recorded in one or more environments. Except for 50% days to flowering, the plant growth habit and the main stem height, all traits were recorded after harvest. The mean phenotypic value of each trait was expressed on a per plant basis by averaging the values of a minimum of 6 to a maximum of 10 plants. Days to flowering (Dflo): § Number of days from sowing to the first flower appearance date on 50% of the plants in a row Plant architecture: § The plant height (PH) was measured from the base of cotyledonary axil up to the terminal bud. Groundnut growth habit (GH) is usually recorded on a 1–6 scale based on the IBPGR of ICRISAT (1992) descriptors where, 1 = procumbent 1, 2 = procumbent 2, 3 = decumbent 1, 4 = decumbent 2, 5 = decumbent 3, and 6 = erect. However, this method yields qualitative data that may not be suitable for QTL mapping. Therefore, in this study, the method applied was that one described by Fonceka et al. (2012) where GH trait was represented by the ratio of the length of the creeping part of a given lateral branch to the total length of the lateral branch. This technique generates quantitative data. No data transformation was performed. 48 University of Ghana http://ugspace.ug.edu.gh Pod and seed morphology: § Pod length (PL) and width (PW), seed length (SL) and width (SW) were evaluated on 30 pods and 30 seeds respectively using a caliper with a digital display. These traits were measured in all environments. Data on pod beak and constriction were recorded in one site on 30 pods by using a 0-9 scale according to the descriptors of groundnut. For Pod constriction, only pods with at least two seeds were considered. The data obtained were transformed into frequencies by multiplying the number of pods of a given class by the value of the class (0, 3, 5, 7 or 9). The sum of the frequencies divided by the total number of pods (30) yielded an average that was used for QTL analysis. Yield components: § The yield components were determined based on the whole plant, the pod and haulm dry mass. The total plant biomass was first weighed to determine the total biomass per plant (TPW). Then pods were removed and weighed to determine the total pod weight per plant (PPW). The haulm weight per plant (HW) was calculated as the difference between the total biomass and the total pod weight. One hundred (100) pods (HPW) were randomly sampled and weighed. The 100 pods were then shelled and the percentage at maturity (PMAT) was determined by counting mature pods characterized by a dark black internal pericarp colour. All seeds contained in the 100 pods were weighed and mature seeds were separated from immature seeds and counted. The weight of 100 seeds (HSW) was calculated as the weight of mature seeds divided by the number of mature seeds multiplied by 100. 4.2.4 Statistical analysis Data collected at Darou in 2016 were removed from the analysis due to the high number of BC2F4:6 families with missing data. Data normality was checked with the Shapiro test for normality. Basic statistical analyses (mean and standard deviation) were calculated for each trait. The combination 49 University of Ghana http://ugspace.ug.edu.gh location and year was considered as an environment. Since the environment x genotype (GxE) interaction was significant, data from each environment were subjected to independent analysis of variance (ANOVA) using the general linear model (GLM) procedure implemented in SAS for the analysis of data issued from augmented block designs. The macros of Parsad et al. (2011) were used to estimate the block effects within the experiments, the environments, the genotypes and the genotype-by-environment interactions. The linear effect model for such ANOVA was as follow: ! = $ + & + ' (&) + * + * + & + , Where Y is the trait, $ is the overall population mean, E is the environmental effect, B (E) the block within the environmental effect, G is genotype effect, G x E is the environment x genotype interaction, and , is the experimental error. All effects were treated as fixed. Pearson correlation coefficients were computed for combined localities using PROC CORR procedure of SAS 9.4 (SAS Institute Inc, 2013). Broad-sense heritability was estimated in each location based on the formula: -. = ((/01 − /03)/5)/ (((/01 − /03)/5) + (/03/5)) Where, MSG and MSE are the genetic and residual mean squares and r the number of block in the experiment. Statistical analyses were done in SAS/STAT version 9.4 (SAS Institute Inc 2013). A composite selection index called index of agronomic stability (IAS) was used to select the best AB families based on pod and seed size traits in each environment. IAS combined four variables namely hundred pod weight (HPW), hundred seed weight (HSW), seed length (SL) and seed width (SW). These four traits were selected on the basis of the PRA results reported in this research and correlation patterns. IAS was computed according to Trouche et al. (2011): 6789 = :; ∗ =>; − ?; ÷ A; ; 50 University of Ghana http://ugspace.ug.edu.gh where xij is the phenotypic value of the line i for trait j, mj the population mean performance and sj the standard deviation of all lines for trait j, aj is the relative weighting of trait j in the index, where j varied from 1 to 3 with: 1 = HPW; 2 = SL; 2 = SW; 3 = HSW. Stability analysis was performed using the pod and biomass yield data of the selected genotypes. Genotype main effects plus genotype x environment interaction (GGE) biplot analysis were computed in GenStat (12th edition) to evaluate the yield stability following the model described by Yan et al. (2007); Alwala et al. (2010); Rao et al. (2011): !9B − !B = CDE9DFBD + C.E B.FB. + G9B Yij is the mean yield of genotype i in environment j, Yj is the mean yield across all genotypes in environment j, λ1 and λ2 are the values for principal component PC1 and PC2, ξi1 and ξj2 are the PC1 and PC2 scores for genotype i, ηj1and ηj2 are the PC1 and PC2 scores for environment j and Σij is the residual of the model associated with the genotype i in environment j. 51 University of Ghana http://ugspace.ug.edu.gh 4.3 Results 4.3.1 Traits variation and correlations Except for plant growth habit (GH), all other traits showed continuous normal or almost-normal distribution in all studied environments (Figure 4.3). Plant growth habit (GH) showed a wide range of morphologies, ranging from completely prostrate to totally erect. Except for growth habit (GH), pod beak (PB) and pod constriction (PC), the mean population values for all other quantitative traits tended to be skewed towards the phenotypic value of the recurrent Fleur11 parent (Table 4.1). Figure 4.3: Distribution and variation of studied traits among the BC2F4:6 progenies Excluding the date to flowering (Dflo) trait that displayed little variation ranging from 17 to 22 days, all traits studied showed extreme phenotypes indicating transgressive segregation for these 52 University of Ghana http://ugspace.ug.edu.gh traits. The analysis of variance was highly significant P<0.01 for all traits in all four environments except in Nioro and Sinthiou 2016 where TPW and HW did not display significant differences among tested genotypes (Table 4.1). The trend of the estimated broad sense heritability values (H2) was quite similar in the studied sites for characters such as pod and seed size. PL, PW, SL, SW and PB displayed high broad sense heritability ranging from 0.6 to 0.9 in all environments except in Nioro where SL showed a low value of 0.3. A similar pattern was observed for yield related traits HPW and HSW in all sites with H2 ranging from 0.5 to 0.8. For yield traits such as TPW, PPW and HW, moderate to low heritability estimates were detected respectively in the off-season and rainy season environments. Table 4.1: Summary statistics of traits in the studied environments Symbol Well- Water-watered stressed Nioro16 Sinthiou16 F H2 F H2 F H2 F H2 Dflo 2.1** 0.5 1.5** 0.3 5.8** 0.8 _ _ PH 3.2** 0.7 3.3** 0.7 1.7** 0.4 1.5** 0.3 GH <001 0.9 <001 _ _ _ _ _ PB 3.3** 0.7 _ _ _ _ _ _ PC 2.2** 0.5 _ _ _ _ _ _ PL 7.9** 0.9 4.9** 0.8 13.8** 0.9 9.1** 0.9 PW 3.0** 0.7 2.5** 0.6 2.5** 0.6 3.3** 0.7 SL 5.0** 0.8 6.7** 0.9 1.4** 0.3 9.1** 0.9 SW 6.2** 0.8 2.6** 0.6 5.9** 0.8 3.7** 0.7 TPW 3.2** 0.7 1.9** 0.5 1.3ns 0.2 1.3ns 0.2 PPW 3.0** 0.7 2.8** 0.6 1.4** 0.3 2.0** 0.5 HW 4.7** 0.8 2.7** 0.6 1.3ns 0.2 1.3ns 0.2 HPW 5.0** 0.8 4.1** 0.8 4.1** 0.8 2.3** 0.6 HSW 2.1** 0.5 6.6** 0.8 3.8** 0.7 2.5** 0.6 PMAT 2.0** 0.5 2.6** 0.6 3.7** 0.7 1.8** 0.4 Dflo: flowering, PH: plant height, GH: growth habit, PB: pod beak, PC: pod constriction, PL: pod length, PW: pod width, SL: seed length, SW: seed width, TPW: total plant weight, PPW: plant pod weight, HW: haulm weight, HPW: hundred pod weight, HSW: hundred seed weight, PMAT: Percentage maturity. 53 University of Ghana http://ugspace.ug.edu.gh Significant correlations were found between most of the traits (Table 4.2). The values in bold were highly significant at P = 0.001. Pod and seed morphology related traits namely PL, PW, SL, and SW showed significant positive correlations with HPW and HSW. PL, PW, SL and SW displayed also significant positive correlations between them. The correlation values ranged from 0.45 to 0.65. High positive correlations were observed among pod and seed morphometric traits. PMAT was not correlated to any of the studied traits except for PPW (0.36). The highest correlation coefficients were obtained between TPW and HW (0.93), HPW and PL (0.65), HSW and SW (0.58) and between PPW and HPW (0.55). HPW was highly correlated with PPW (0.55). The main stem height (PH) was positively correlated with almost all traits except for PMAT and negatively correlated with GH. Plant growth habit (GH) was negatively correlated with all studied traits. The days to flowering (Dflo) was not correlated to any of the studied traits. 54 University of Ghana http://ugspace.ug.edu.gh Table 4.2: Correlation between studied traits in one site Traits Dflo PH GH TPW PPW HW HPW PMAT HSW PL PW SL SW Dflo 1 PH -0.26 1 0.0001 GH -0.00589 -0.22789 1 0.934 0.0012 TPW 0.07174 0.46782 -0.0931 1 0.3127 <.0001 0.1898 PPW -0.16632 0.1892 -0.03521 0.2859 1 0.0192 0.0076 0.6224 <.0001 HW 0.12367 0.43526 -0.08195 0.9375 -0.06531 1 0.0826 <.0001 0.251 <.0001 0.3606 HPW -0.2099 0.40292 -0.29652 0.3051 0.55287 0.13026 1 0.0029 <.0001 <.0001 <.0001 <.0001 0.0674 PMAT -0.11057 0.07267 0.02787 0.0164 0.36612 -0.10885 0.2213 1 0.1191 0.3065 0.6953 0.8174 <.0001 0.1269 0.0016 HSW -0.17989 0.40729 -0.21968 0.2009 0.15773 0.1602 0.5346 0.05976 1 0.011 <.0001 0.0018 0.0044 0.0269 0.0245 <.0001 0.4018 PL -0.11685 0.29179 -0.17233 0.2638 0.41467 0.13366 0.6505 0.12179 0.4509 1 0.1003 <.0001 0.0149 0.0002 <.0001 0.0611 <.0001 0.0866 <.0001 PW -0.16886 0.29979 -0.08719 0.2072 0.1412 0.18422 0.5481 -0.00005 0.4707 0.5679 1 0.0168 <.0001 0.2196 0.0032 0.0472 0.0094 <.0001 0.9995 <.0001 <.0001 SL -0.13421 0.32244 -0.19221 0.1595 0.15666 0.12264 0.5124 -0.09721 0.5329 0.6945 0.5002 1 0.0588 <.0001 0.0065 0.0244 0.0279 0.086 <.0001 0.172 <.0001 <.0001 <.0001 SW -0.21845 0.41327 -0.16499 0.2663 0.24198 0.17743 0.5842 0.02304 0.5885 0.6134 0.5511 0.8101 1 0.0019 <.0001 0.0196 0.0001 0.0006 0.0124 <.0001 0.7461 <.0001 <.0001 <.0001 <.0001 Values in bold are highly significantly correlated at 0.1 55 University of Ghana http://ugspace.ug.edu.gh 4.3.2 Performance of the population in the studied environments The characterization of the 200 progenies showed significant variation among sites for plant pod weight (PPW) and for plant biomass weight or haulm weight (BW=HW) (Figure 4.4). For HW, the population performed better at Sinthiou and in the well-watered condition than in the other environments. For PPW, the population showed best performance at Nioro 2016 and in the well- watered condition. On the other hand traits such as pod and seed size were more stable and displayed little variation between tested environments. Within environment, the population displayed important variation except for seed width in Sinthiou 2016. Figure 4.4: Traits variation among studied sites Traits: BW or HW (biomass weight or haulm weight), PPW (plant pod weight), SW (seed width), SL (seed length), PW (pod width), PL (pod length); Sites: N16 (Nioro 2016), S16 (Sinthiou 2016), WW (well-watered), WS (water stressed). 56 University of Ghana http://ugspace.ug.edu.gh 4.3.3 Selection of the promising lines based on IAS Using the selection index IAS, the top 50 genotypes were selected in each of the four environments (Table 4.1 in appendix). In the well-watered environment, 106 genotypes exhibited an IAS score between 0.45 and 13.09. These families outperformed the local recurrent parent Fleur11 which scored 0.43. The water limited environment ranked second with 88 lines that outperformed the recurrent parent Fleur11 for HSW, pod and seed size traits, followed by Sinthiou (49 genotypes) and Nioro (44 genotypes). Considering all four environments, 11 BC2F4:6 genotypes listed in Table 4.3 were constantly present and exhibited IAS scores varying between 3.83 and 15.38. 4.3.4 Pod yield performance and stability of the selected genotypes The top 11 genotypes selected for HPW, HSW, and pod and seed sizes plus Fleur11 were subjected to GxE biplot analysis in order to identify the best performing genotype for a specific environment and to identify the more stable genotype across environments based on pod yield. The GGE biplot explained 86.67% of the total variation for pod yield with PC1 and PC2 accounting for 56.41% and 30.26%, respectively (Figure 4.5). The lines B7_32_8_1 (243), B7_23_10_3 (66) and Fleur11 (0) appeared as the highest yielding genotypes as they are present on the far right-hand side of the biplot towards the pointing arrow of the AEC abscissa (Figure 4.5). Moreover, these three lines were also stable as they are positioned very near to the AEC abscissa. B7_28_14_1 (189) and Fleur11 (0) were the most stable lines across environments as these two lines had a very low PC2 score near zero. 57 University of Ghana http://ugspace.ug.edu.gh Table 4.3: Top selected genotypes across environments based on IAS indices Genotypes IAS WW WS N16 S16 B12_14_2_2 5.66 14.40 15.13 6.58 B12_14_2_5 11.48 5.37 8.04 13.09 B7_20_3_1 8.39 11.79 12.17 7.60 B7_21_2_8 7.69 3.83 7.09 6.31 B7_23_10_3 11.18 12.37 15.38 7.61 B7_23_10_8 6.67 4.26 9.10 5.94 B7_27_5_7 11.36 5.73 4.63 8.76 B7_28_14_1 6.96 7.11 9.05 7.65 B7_30_6_5 6.97 6.64 5.61 9.74 B7_32_7_5 11.15 12.27 7.83 13.11 B7_32_8_1 9.36 11.21 8.43 14.46 IAS: Index of Agronomic Stability; environments: WW, WS, N16, S16 B7_23_10_8 (70), B7_21_2_8 (34) and B7_30_6_5 (217) were regarded as consistently the lowest yield genotypes as they are on the far left-hand side of Figure 4.4. The B7_30_6_5 (217) was the more unstable line as it is away from AEC abscissa. 58 University of Ghana http://ugspace.ug.edu.gh Figure 4.5: Ranking and stability of selected genotypes based on mean pod yield AEC: average environment coordinate; the single arrowed horizontal line represents the AEC abscissa and the double arrowed vertical line represents the AEC ordinate. The direction of AEC arrowhead indicates increasing yield. The vertical projection on the AEC indicates stability. Greater the projection, higher is the instability. The “which won where” pattern (Figure 4.6) displayed five sectors based on genotypes performance for pod yield. Sinthiou 2016 (S16) and Nioro 2016 (N16) were clustered together in one sector and well-watered (WW) and water stressed (WS) clustered together in a separate sector. The best performer for a given site is located on the vertex of each corresponding sector at which the site (s) is represented. Hence, B7_32_8_1 (243) was the best performer at S16 and N16. 59 University of Ghana http://ugspace.ug.edu.gh Figure 4.6: GGE biplot for the “which won where” pattern of genotypes and environments 60 University of Ghana http://ugspace.ug.edu.gh 4.4 Discussion Although the stress imposed in 2015 was not sufficient to discriminate the genotypes for drought stress tolerance in terms of pod yield probably because of the late application of the stress and the weather conditions, this study showed significant genetic differences for various traits among the 200 genotypes of the interspecific BC2F4:6 population. The higher variations were observed for TPW, PPW and HW traits in the different environments which displayed low broad sense heritability in Nioro16 and Sinthiou16 and moderate heritability under water stress and well- watered environments. However, high heritability for pod and biomass yield was reported in groundnut (Songsri et al., 2008). This result reflects the complex and polygenic nature of these traits and the large part of the environmental component on the expression of these characters. This is part of the reasons why breeding for such traits is difficult and time consuming because small genetic gain can be achieved (Tester & Langridge, 2010; Sinclair, 2011). An alternate way is to breed for surrogate traits. HPW and HSW have been reported to be positively correlated to pod yield in groundnut (Gomes & Lopes, 2005; Luo et al., 2017). High and moderate positive correlations between HPW and PPW (0.74) and between HSW and PPW (0.34) were observed, indicating that such traits may be used as alternate traits to improve pod yield. Pod and seed' sizes were found to be highly correlated with HPW and HSW (Gomes & Lopes, 2005; Luo et al., 2017); the values ranged in the present study from 0.51 to 0.76. Additionally, pod and seed size displayed high heritability and were less influenced by the environments. These results suggest that important genetic gains can be achieved in a rapid manner for pod and seed size and therefore indirectly on pod and seed weight. Earlier studies have demonstrated that improvement in seed size, seed weight, and number of pods per plant largely contributes to enhanced pod yield (Rathnakumar et al., 2012). 61 University of Ghana http://ugspace.ug.edu.gh When several traits are being simultaneously improved, a selection index has been suggested as a good approach (Lin, 1978; Bänziger & Lafitte, 1997; Yan & Frégeau-Reid, 2008). In the present study, a two step selection was carried out to select genotypes with desired traits based on the economic weights given to each trait using selection index and subsequently, the top performer and stable genotypes were identified. It is worth noting that the weights attributed to each of the traits were somewhat subjective and depended on the breeding objectives. Siddique et al. (2006) reported a path analysis approach for improving groundnut yield. For both strategies the main objective was to perform indirect selection with secondary traits to improve more complex traits. Safari et al. (2013) reported that selection index for increasing oil yield, grain pod ratio and 100 grain weight of groundnut genotypes would be between 5 and 62% more efficient than direct selection, depending on the trait and selection method. GGE biplot analysis applied in this study helped to identify the best performing and most stable genotypes B7_32_8_1 and B7_23_10_3. It was used to select five high oil content and stable genotypes (ICGV 05155, ICGV 06049, ICGV 06041, ICGV 06420 and ICGV 03043) promoted to multilocation trials for their release for cultivation and use as parents in breeding programmes (Janila et al., 2016). Both lines B7_32_8_1 and B7_23_10_3 were erect plants. However, B7_32_8_1 exhibited deep pod constriction value (6.69) while B7_23_10_3 exhibited pronounced pod beak (6.60). These are some unwanted traits associated with the use of wild species in groundnut. Nevertheless, the identification of molecular markers linked to such traits may help to counteract such effects. The rainy season environments (S16 and N16) clustered together and were different from off season environments (Well-watered and Water-limited). These findings were reported earlier by Janila et al. (2016) and may be attributed to the regular moisture available through irrigations in post rainy seasons. 62 University of Ghana http://ugspace.ug.edu.gh 4.5 Conclusion An advanced backcross QTL population was developed. The important genetic variation and the transgressive segregation observed attest of the potential of the wild donor to enlarge the genetic base of the cultigen. The variation and low heritability observed for pod and biomass yield across environments confirm the complex nature for such traits and the need to use secondary traits to improve them. Hundred pod weight, hundred seed weight, pod and seed size were less influenced by the environment and exhibited high heritability and were correlated to pod yield. The use of the selection index (IAS) allowed the identification the best fifty genotypes in each environment among which eleven were constantly present in all environments. Among the 11 genotypes selected, B7_32_8_1 and B7_23_10_3 were identified as the best performing genotype for pod yield. 63 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE 5.0 QUANTITATIVE TRAIT LOCI (QTL) MAPPING FOR POD YIELD RELATED TRAITS AND USE IN GROUNDNUT 5.1 Introduction Cultivated groundnut is a useful legume widely grown in most tropical and subtropical areas of the world. One of the most important groundnut breeding objectives in developing countries is to increase pods and biomass yields because of the double use of this crop in agro-pastoral systems. Such traits are complex and polygenic and are therefore difficult to improve using direct selection. However the use of yield surrogate traits seems to have wide application in groundnut breeding schemes. Yield related traits such as the height of main stem (HMS), pod and seed characteristics may contribute to yield in groundnut (Holbrook & Stalker, 2003; Shirasawa et al., 2012). Indices like 100 pod weight (HPW), weight of 100 seed (HSW) quantified by pod and seed size and percentage of maturity are important and related to yield of groundnut. Such indices have showed high correlation with pod and seed size (Shirasawa et al., 2012). Marker-assisted selection (MAS) has been described as an efficient way to improve complex traits. The use of molecular markers appears to be a powerful tool in accelerating and increasing selection efficiency especially in the case of traits with low heritability. The construction of a genetic map and the identification of major genes or regions of the genome (QTL delimited by markers) involved in the variation of the traits of interest are prerequisites for the application of MAS (Francia et al., 2005). To identify genomic regions underlying important agronomic traits, quantitative trait locus (QTL) analysis has been widely conducted. This method allows the identification of tightly linked markers to the QTL that can be further deployed in marker- assisted breeding. 64 University of Ghana http://ugspace.ug.edu.gh While MAS application has been a routine procedure for various crops (Buerstmayr et al., 2002; Zhang et al., 2003; Fan et al., 2006; Foiada et al., 2015; Jiao et al., 2015), in cultivated groundnut, because of the low genetic polymorphism (Moretzsohn et al., 2004), QTL mapping has been slow in the past. In recent years, progress has been achieved in the creation of new sources of genetic variation using wild groundnut species and in the generation of new genomic resources. Many QTL for various characters have been mapped either in cultivated groundnut and/or in interspecific genetic material. Successes in MAS’ application have been reported on oil quality, rust, nematode and late leaf spot improvement (Khedikar et al., 2010; Chu et al., 2011; Shoba et al., 2012; Varshney et al., 2014). To date there is no documented report on the deployment of MAS for yield related traits in groundnut. Nevertheless, an important number of QTL for yield components have been detected and studies are still in progress (Selvaraj et al., 2009; Fonceka et al., 2012; Shirasawa et al., 2012; Jiang et al., 2014; Pandey et al., 2014; Faye et al., 2015; Huang et al., 2015; Chen et al., 2016; Chen et al., 2017). Among these QTL studies, only two reported QTL analysis were conducted on interspecific genetic material using either the advanced backcross QTL (AB-QTL) QTL approach (AB-QTL) (Fonceka et al., 2012) or QTL mapped using chromosome segment substitution lines (CSSLs) (Fonceka et al., 2012). About half of the QTL positive effects were associated with alleles of the wild parent. Seventeen QTL for pod and seed size were mapped into 7 linkage groups (LGs) and explained 8 to 23.7% of the observed phenotypic variation in the AB-QTL genetic material. Some of these QTL related to biomass, pod and seed characteristics were also mapped in the CSSLs thus allowing to validate some of them and to identify interesting chromosome segment lines bearing major and constant QTL associated with these traits. Such interesting material (CSSLs) have been used in tomato (Lecomte et al., 2004; 65 University of Ghana http://ugspace.ug.edu.gh Vidavski et al., 2008; Hanson et al., 2016) and in rice (Ando et al., 2008; Wang et al., 2012) to pyramid useful QTL. Pyramiding QTL has been conducted in groundnut for nematode resistance and high oleic acid traits (Chu et al., 2011) but to date there are no reports on its deployment for yield related traits. The objectives of the present work were to: • identify QTL associated with yield related traits in groundnut using an advanced backcross (AB) population, • estimate the introgression rate of the selected genotypes; and • pyramid identified and validated fragments from wild relatives bearing QTL for increased pod and seed sizes. 66 University of Ghana http://ugspace.ug.edu.gh 5.2 Materials and Methods 5.2.1 QTL mapping 5.2.1.1 Population development and phenotypic characterization The plant material used for the QTL study consisted of the advanced backcross QTL (AB-QTL) population of 200 BC2F4:6 progenies described in the previous section. Field evaluation was described in the previous section (chapter four). 5.2.1.2 Map construction and molecular analysis The genetic linkage map was constructed using the BC1 population of 150 progenies derived from the cross between the cultivated variety Fleur11 and the synthetic allotetraploid ISATGR52B ((A. valida (BB) x A. duranensis (AA)) 4x) hereafter called AvAd. The synthetic allotetraploid AvAd combines the AA genome of A. duranensis (2n =2x =20), a close wild relative and one of the most probable ancestors of A. hypogaea, and the BB genome of A. valida (2n =2x =20), a wild relative taxa that was reported to pair with the B genome of the cultivated species during meiosis. AvAd was used as male donor parent and Fleur11 was as female and recurrent parent. Genomic DNA of both parents and BC1 progenies were extracted from young leaves according to the MATAB protocol as described by Fonceka et al. (2009). Polymorphism between parents was assessed using 602 primer pairs of SSR, mainly selected from a previous study (Nguepjop et al., 2016). Out of the 602 SSRs 330 polymorphic markers were used to genotype the parents and the 150 progenies according to PCR amplification as described by Nguepjop et al. (2016). PCR products were separated by electrophoresis run at a constant 95 W for 1 - 2h in a DNA- Sequencer (LI-COR 4300 DNA Analyzer, Lincoln, USA). Segregation for parental alleles was examined visually in the gels and data were manually scored using the type codes employed for 67 University of Ghana http://ugspace.ug.edu.gh BC1 population in Jelly 2.017b application (Rami, unpublished). For the BC1 population, each polymorphic allele was scored as a locus and then tested for goodness-of-fit to the expected segregation ratio of 1:1 using the chi-square test (P<0.05). On the basis of the genetic map produced in BC1, 128 polymorphic SSR markers were selected in order to cover the genetic map with a distance of about 10 cM between adjacent markers. The genotyping of the BC2F4 generation was done using a LI-COR 4200 dual-dye DNA analyzer (LI- COR Bio-sciences, Lincoln, Nebraska USA). Segregation for parental alleles was examined visually in the gels and data were manually scored using the type codes employed for BC2F4 population in Jelly 2.017b application (Rami, unpublished) where A= AA homozygous for the cultivated parent, B= BB homozygous for the alleles of the wild parent and H= AB heterozygous. 5.2.1.3 Genetic map and QTL statistical analysis The Fleur11xAvAd map was computed using Mapdisto software (Lorieux, 2012). Markers were grouped into LGs applying the following MapDisto parameters: LOD of 3 - 7, maximum recombination frequency of 0.3 and Kosambi mapping function. Marker order within LG was estimated using the “order” and “ripple” commands. LGs were drawn using SpiderMap 1.6 (Rami, unpublished). Chi-square test was used to test possible segregation distortion for each marker by comparing for each locus the frequency of the genotypic classes obtained with that expected in a BC2F4 segregating population. QTL identification was performed with R/qtl package (version 1.41 - 6) (Broman & Sen, 2009). The trait mean values from each individual environment were used. The association between trait and marker data was calculated by single- marker regression (SMR) and the locations of the detected QTL were estimated using simple interval mapping (SIM) (Haley & Knott, 1992). The confidence interval estimates of the QTL 68 University of Ghana http://ugspace.ug.edu.gh location were obtained using the 1.5-LOD support interval method (Broman & Sen, 2009). The critical LOD threshold value at a type I error rate of 5% was determined to indicate a significant QTL effect for each trait using 1000 permutations. The percentage of the phenotypic variation (PV%) explained by a QTL was estimated using an additive multi-QTL model which involved all detected QTL for a given trait using the function “fitqtl” implemented in R/qtl (Broman & Sen, 2009). For a QTL to be declared, two conditions were set: 1) the threshold value should be significant that is, higher than the threshold at 5% and 2) the number of individuals with BB genotypes at that locus should be ≥ to 5. QTL were declared major if the phenotypic variance explained was higher than 10% and minor when it accounted less than 10% (Collard et al., 2005). QTL for different traits were considered to be co-located when their positions with highest LOD scores (peak) were located in the same marker intervals. The contribution of wild alleles was determined by computing the additive effect of the QTL following the formula: ! = (%&& − %(()/+ Where ,-- and ,.. represent the phenotypic values of the BB and AA individuals respectively. The QTL were illustrated graphically on the linkage groups using Spidermap software version 1.5.7b (Rami, unpublished). The percentage of introgression of the wild genome into the genetic background of the selected genotypes was estimated using GGT graphical genotypes software version 2.0. The genotype of the selected line B7_32_8_1 was visualized graphically using the GGT software. 69 University of Ghana http://ugspace.ug.edu.gh 5.2.2 Pyramiding of QTL using CSSLs 5.2.2.1 Material Six chromosome segment substitution lines (CSSLs) developed from a previous study (Fonceka et al., 2012) were used. Each of the CSSL carries a particular chromosome fragment derived from the wild donor that governs the variation of a particular character. Differences between the CSSLs reside in the length of the wild target chromosomic segment that has been introgressed and the location of the insert fragment. Details of the selected lines involved in the crosses and the molecular markers (SSRs) used to follow the introgression are given in Table 5.1. Table 5.1: Details of the CCSLs involved in the development of the pyramiding populations Lines Traits Segment Linkage % Crosses size (cM) group donor genome 12CS_037 seed length and 25.5 B5 4.0% 12CS_031*12CS_037 biomass 12CS_031 Hundred seed 89.0 A9 6.5% 12CS_031*12CS_039 weight and biomass 12CS_069 hundred seed 9.5 B6 1.2% 12CS_069*12CS_037 weight 12CS_039 pod and seed 15.9 A8 2.3% 12CS_037*12CS_039 length 12CS_028 yield stability 42.3 A9 2.7% 12CS_069*12CS_028 and seed length 12CS_120 yield stability 31.3 A1 1.7% 12CS_028*12CS_120 70 University of Ghana http://ugspace.ug.edu.gh 5.2.2.2 DNA extraction from seeds A modified MATAB protocol used for DNA extraction from leaves was applied to extract DNA from seeds. The different steps involved in the extraction are described in (Figure 5.1): The end of the seed (1/3) about 100 mg from the opposite side of the embryo was cut and put in 2 ml Eppendorf tube. The rest of the seed was kept in a 15ml conical tube and stored in the fridge at 4 °C. The tubes containing the 100 mg of the seed and the rest of the seed was correctly identified. The same process was applied to the total number of seeds to be extracted. Figure 5.1: Description of the way to cut the seed for DNA extraction Briefly, the 100mg of seed were ground in a mortar and pestle and dissolved in 750µL of MATAB buffer at 74°C. The samples were incubated 20 minutes at 74°C and cooled for 5 minutes at room temperature. A volume of 750µL of CIA (24:1) was added in each sample and all samples were shaken gently until homogenization before centrifugation at 12000 rpm for 20 71 University of Ghana http://ugspace.ug.edu.gh minutes. 300 µl of supernatant was harvested and 300 µl of phenol-chloroform added. Homogenize and centrifuge. After centrifugation, 200 µl of supernatant was transferred to a tube of 1.5 µl. Then 200ul of cold isopropanol was added. This was homogenized gently and then centrifuge for 20 minutes. Pellets were washed with 500 µL of 70% ethanol; air dried and dissolved in 100 µL of TE 1X. 5.2.2.3 F1 control and selection of the F2 lines For the quality control of the F1 plants of each of the crosses, one SSR marker mapped in a wild introgression segment from one parent was used. Heterozygous individuals were selected and the corresponding stored seeds were sown to obtain the F2 seeds. Among the F2 seeds, desired events which mean F2 individuals homozygous at both loci for the two corresponding fragments from both parents were selected. For that two SSRs each located on the wild fragment associated with the QTL of each of the parents involved in the cross were used. This is like a foreground selection. A background selection was not needed because the parents differed almost exclusively on the introgressed wild chromosome segment meaning that except the introgressed fragment, their genetic background remains similar. PCR amplifications were performed as described in Nguepjop et al. (2016). 72 University of Ghana http://ugspace.ug.edu.gh 5.3 Results 5.3.1 Marker segregation and polymorphism 5.3.1.1 Linkage map A set of 250 polymorphic primers providing a strong amplification was selected to construct the Fleur11xAvAd map. A total of 341 loci were obtained out of which 79 (23.2%) deviated from the expected Mendelian segregation ratio of 1:1 (P<0.05). Except for one locus (PM032), all distorted loci were skewed toward the cultivated allele suggesting an occurrence of gametic or unintentional selection toward the recurrent parent. In total, 330 loci were mapped onto 20 LGs that spanned a cumulative length of 1792 cM with an average marker density of 6.4 cM. LGs were classified into 10 pairs according to homeologous loci (Figure 5.2). In A-genome, 172 loci were distributed in 10 LGs with a mean number of 17.2 (6 - 24) markers per LG. Each LG spanned an average length of 96.4 cM (45 - 111.2 cM) with inter-marker spacing of 6.1 cM. In B-genome, 158 loci were mapped in 10 LGs with a mean number of 16 markers per LG (9 - 23). Each LG spanned 88.6 cM on average (59.4 - 137 cM) with inter-marker spacing of 6.4 cM. The number of bridge markers within each pair of homoeologous LG ranged from 2 (LG1) to 11 (LG5) with an average of 8 (Figure 5.2). A good collinearity was found among homeologous pairs with the exception of the pairs 1 and 6. Likewise, a quadrivalent relationship was found between the pairs 7 and 8 thought to be bridges loci between the LGs B7 and A8 (Figure 5.2). A total of 29 (11.9%) primer pairs detected duplicated loci within the genome and the proportion of duplicated loci was slightly higher on the B-genome than the A (54% versus 46%). Most duplicated loci were clustered onto LGs and some duplicated blocks were found in the LGs B10, B2, B7, A6, B6, A1 and A10. 73 University of Ghana http://ugspace.ug.edu.gh Figure 5.2: Genetic map derived from Fleur11 x AvAd population LGs were grouped into homoeologous pairs according to homeologous loci among the A and B genomes connected by blue dashed lines. The LGs deriving from the A genome, named from A1 to A10 and the B genome named from B1 to B10. The map distances in Kosambi map units (cM) of each LG are shown on the left, and the loci marker names are on the right. Duplicated loci are identified by the number 1, 2 or 3 after the suffix A or B. Distorted Loci are identified by asterisk after the locus name. The colour and number of asterisk indicate the direction and the intensity of the segregation distortion respectively. Blue: markers skewed toward the alleles of the cultivated parent. Red: markers skewed toward the alleles of the wild parent. 74 University of Ghana http://ugspace.ug.edu.gh Map location of distorted markers was also analysed. Except for 2 loci, distorted loci (78) were clustered across LGs where 21 (26.9%) were mapped onto A genome (LGs A4, A7 and A9) while 57 (73.1%) were mapped onto B genome (LGs B3, B4, B6, B7 and B10). 5.3.1.2 QTL detection The framework map of 250 SSR markers was used for QTL mapping. Out of the 128 markers selected for the genotyping, 26 SSRs were removed due to bad genotyping quality. Hundred and two markers that amplified 119 loci were used for the QTL analysis. LOD significance thresholds estimated by 1000 permutation tests for each trait ranged from a minimum of 2.49 to a maximum of 9.93 (Table 5.2). At least one QTL was detected for each of the 10 traits analysed except for pod beak. A total of 38 QTL were mapped on 17 LGs and mostly on LGs belonging to the B genome with a percentage of 63.2 (Figure 5.3). A summary of QTL is provided in Table 5.2. Days to flowering § Four QTL for Dflo were mapped on 4 LGs (A4, A8, B1 and B10). For the QTL on LGs A4 and B10, the flowering precocity was associated with the alleles of the wild parent. Together they explained 17.09% of the phenotypic variance. The detected QTL on LG B10 represents alone 12% of the observed phenotypic variance and consequently can be considered as a major QTL. 75 University of Ghana http://ugspace.ug.edu.gh Table 5.2: Summary of the detected QTL Trait LG Conf1 Peak Conf2 R2 Add Lod Increase parent SL_ww A1 36.0 55.0 67.0 1.9 0.1 4.7 ISATR52B SW_ww A1 36.0 54.0 62.0 2.2 0.1 4.6 ISATR52B PH_ww A10 79.5 86.6 93.0 0.8 -1.4 4.1 Fleur11 PC_ww A2 44.6 85.8 85.8 3.0 1.1 7.9 ISATR52B SL_WS A2 14.0 42.8 98.0 1.4 0.1 4.6 ISATR52B PH_WS A3 -3.5 2.5 12.5 0.7 -1.7 4.3 Fleur11 PH_ww A3 -0.5 4.6 11.5 3.8 -1.5 5.0 Fleur11 Dflo_N16 A4 92.0 99.0 108.0 5.1 -1.2 2.5 Fleur11 HSW_S16 A4 98.0 106.6 112.0 1.0 1.8 4.8 ISATR52B HPW_S16 A5 18.6 31.1 46.6 1.1 -16.1 3.4 Fleur11 HPW_WS A5 13.6 24.6 44.6 0.6 -2.7 4.7 Fleur11 PW_S16 A6 36.0 51.9 83.0 2.2 0.3 4.2 ISATR52B SW_N16 A7 0.0 25.0 57.0 1.3 -0.2 4.1 Fleur11 Dflo_WS A8 0.0 7.0 20.0 6.4 0.4 3.4 ISATR52B Dflo_N16 B1 0.0 40.0 53.0 5.4 0.4 2.6 ISATR52B PC_ww B1 75.0 77.0 77.0 0.3 1.0 4.3 ISATR52B PH_ww B1 38.0 40.0 57.0 2.2 0.4 3.7 ISATR52B Dflo_ww B10 33.0 36.2 40.0 12.0 -0.2 5.6 Fleur11 PC_ww B2 23.0 68.0 70.0 1.3 1.3 3.5 ISATR52B HPW_WS B3 94.0 130.5 135.0 1.6 -5.5 4.2 Fleur11 HPW_ww B3 91.0 112.0 134.0 1.7 -6.2 3.9 Fleur11 HSW_WS B3 93.0 114.5 132.0 2.2 -3.0 5.0 Fleur11 HSW_ww B3 112.0 119.0 130.0 1.6 -3.1 6.9 Fleur11 PC_ww B3 26.5 63.5 137.0 0.6 1.4 4.3 ISATR52B PW_S16 B3 109.0 119.5 132.0 1.7 -0.1 4.7 Fleur11 PW_WS B3 111.0 121.0 129.0 0.3 -0.7 4.6 Fleur11 PW_ww B3 105.0 120.5 129.0 1.3 -0.5 5.2 Fleur11 SL_WS B3 114.0 121.5 129.0 0.3 -0.4 9.9 Fleur11 SL_ww B3 96.0 108.5 117.0 1.4 0.2 4.6 ISATR52B SW_N16 B3 110.0 121.5 131.0 0.7 -0.2 3.6 Fleur11 SW_S16 B3 110.0 129.0 134.0 1.7 -0.1 3.8 Fleur11 SW_WS B3 114.0 115.5 119.0 1.5 -0.3 5.4 Fleur11 SW_ww B3 96.0 111.0 121.0 1.4 -0.1 4.8 Fleur11 HPW_WS B4 18 28.0 52.0 0.8 -2.5 3.5 Fleur11 GH_WS B5 0 4.0 13.0 1.0 1.9 8.6 ISATR52B GH_ww B5 0 6.0 18.0 0.8 0.3 4.8 ISATR52B PL_WS B8 29.1 29.1 49.5 0.8 0.4 4.4 ISATR52B PL_ww B8 44.5 49.6 54.5 2.7 -0.3 4.7 Fleur11 Dflo: days to flowering, PC: pod constriction, SW: seed width, SL: seed length, PW: pod width, PL: pod length, GH: growth habit, HPW: 100 pod weight, HSW: 100 seed weight, PH: plant height, WW, WS, S16 and N16: environments Plant architecture § Four QTL located on LGs A3, A10 and B1 were detected for PH. Two of the QTL PH_ww and PH_ws were consistent across the two environments. Both QTL were located on the same LG A3 and almost at the same position with an R2 of 4.5%. Increase in plant height phenotypic value was associated with the cultivated parent Fleur11 alleles 76 University of Ghana http://ugspace.ug.edu.gh at QTL on LGs A3 and A10. On LG B1 the PH increase was associated with alleles from the synthetic amphidiploid donor parent. Two QTL were associated with plant growth habit (GH) and both were located on the same LG B5. At the mapped QTL, the spreading growth habit was linked to the wild parental alleles. A total of six QTL involved in the plant architecture was detected. Yield related traits § Hundred pod weight (HPW) and hundred seed weight (HSW) are the most direct yield related traits in groundnut. A total of 5 QTL for HPW have been identified in three out of the four environments with phenotypic variation (PVE) ranging from 0.6 to 1.7%. Two QTL were detected on LG A5, two on B3 and one QTL on LG B4. The increase in HPW at the identified QTL was all associated with the alleles of the cultivated parent Fleur11. QTL for HPW on LGs A5 and B3 were consistent across two environments. Three QTL for HSW have been mapped on LGs A4 and B3. The two QTL detected on LG B3 colocalized with QTL for HPW and at those QTL, Fleur11’ alleles contributed positively to increase HSW. But at the QTL identified on LG A4, increase in hundred seed weight was associated with the amphidiploid alleles. Pod and seed morphology § A total of 10 QTL were detected for traits related to pod morphology: four QTL for pod constriction (PC) spread over 4 different linkage groups (A2, B1, B2 and B3). These QTL were detected in only one environment and individually explained 0.3% to 3% of the phenotypic variance (Table 5.2). At all detected QTL the amphidiploid wild alleles tended to confer deep constriction. 77 University of Ghana http://ugspace.ug.edu.gh 78 University of Ghana http://ugspace.ug.edu.gh Figure 5.3: Location on the genetic map of detected QTL Each QTL is represented by a triangle located at QTL peak and indicating the sign of the additive effect (upward: positive effect from the wild parent, downward: positive effect from the cultivated parent), and by a box representing the confidence interval. The size of triangle is proportional to the part of phenotypic variance explained by the QTL (R2). QTL identified for the same trait have same colour. The name of the QTL is the trait followed by the location where the QTL was detected. 79 University of Ghana http://ugspace.ug.edu.gh Two QTL for PC were located on homeologous regions on LGs A2 and B2 and at both loci the wild alleles conferred deep constriction. Six QTL were identified for pod size: two for PL localized on LG B8 and four for PW localized on LG A6 and B3. QTL for PL on LG B8 were common to two environments and their positions on the LG overlapped. But the two QTL were in repulsion meaning that phenotypic contribution of the cultivated and wild parents differed. At all QTL (3 in total) identified for PW on LG B3, increase in pod width was associated with the cultivated parent Fleur11. However at the QTL on LG A6, the wild parent brought the increase in pod width. The QTL analysis for seed morphology revealed four QTL for SL and six QTL for SW. The QTL mapped on LGs A1, A2, A7 and B3 taken individually, explained 0.3% to 2.2% of the observed phenotypic variance. Increase in seed length was associated with the synthetic amphidiploid alleles at all detected QTL except for one of the two QTL (SL_ws) where the positive alleles were associated with Fleur11. The QTL detected on LGs A1 and A2 were specific to one environment. Four out of the six QTL identified for SW were mapped across the four locations on LG B3 at proximal positions with the favourable alleles coming from Fleur11. The only QTL identified for SW with the positive effect brought by the wild parent was SW_N16 specific to an environment localized on LG A1. A co-localization of QTL for SL and SW for the same environment was found on LG A1 and at both QTL, the alleles of the synthetic donor parent conferred superior seed size. Most of the detected QTL for pod and seed size (PW, SW and SL) were distributed on LG B3 and co-localized with QTL for yield related traits (HPW and HSW) in a region between 93.4 and 130.1cM. 80 University of Ghana http://ugspace.ug.edu.gh 5.3.1.3 Wild alleles’ introgression rate and examination of the genotype of the selected line From the genotyping data, the quantity of wild donor alleles present in the genome of the top 11 genotypes selected from chapter four was estimated (Table 5.3). Table 5.3: Percentage of introgression of the wild donor into the selected genotypes Genotypes AA AB BB NA B12_14_2_2 86.9 11 1.8 0.4 B12_14_2_5 96.6 2.4 1 0 B7_20_3_1 89.4 3.4 7.3 0 B7_21_2_8 97.1 0 2.9 0 B7_23_10_3 97.6 0 2.4 0 B7_23_10_8 89 5.9 4.6 0.5 B7_27_5_7 89.3 4 7.2 0.5 B7_28_14_1 95.6 2.8 1.6 0 B7_30_6_5 83.6 12.4 1.6 2.4 B7_32_7_5 88.8 1.3 9.5 0.3 B7_32_8_1 82.1 1.7 14.5 1.7 AA: cultivated allele, AB: heterozygous, BB: wild allele, NA: missing data The range of the introgressed genome from the synthetic amphidiploid varied from 1% (B12_14_2_5) to 14.5% (B7_32_8_1). The best genotype identified B7_32_8_1 had the highest introgression rate of 14.5% of BB which would correspond roughly to the rate of introgression expected in a BC2 generation. The graphical visualization of the genome of the line B7_32_8_1 (Figure 5.4) shows introgressions of the wild chromosomic segments on LGs A1, B1, A2, A3, B3, A6, A9 and B9. 81 University of Ghana http://ugspace.ug.edu.gh Figure 5.4: Graphical representation of the genotype of the introgressed LGs of B7_32_1_8. QTL with positive effect associated with the wild alleles have been located on LGs A1 (1), B1 (11), A2 (2), B3 (13) and A6 (6) close to the positions of introgressed regions of the selected genotype. On LGs A1 (1) and A6, QTL associated with pod and seed characteristics (SL, SW and PW) were mapped. QTL associated with an increase in seed length and seed width co- localized on LG A1 between 35 and 56 cM. On LG A6, QTL for pod width was located between 31.86 and 70.12 cM. On LG B1, QTL associated with the shortening of the flowering time was located between 11.08 and 41.63 in a position comprised within the introgressed region of the line B7_32_8_1 (Figure 5.4). QTL associated with deep constriction have also been mapped on LGs A2 and B3 and at these QTL, the effect was associated with the wild alleles. These regions corresponded also to the introgressed regions on LGs A2 and B3 of the line B7_32_8_1. 82 University of Ghana http://ugspace.ug.edu.gh 5.3.2 Selection of the F2 individuals For each of the cross, five F2 individuals that accumulated both target fragments each deriving from one parent involved in the cross were selected based on genotyping data. Three of the five selected F2 seeds from the cross 12CS_37 x 12CS_39 named 037039_30, 037039_75 and 037039_78 are shown (Figure 5.5). For each F2, the remaining part of the seed kept in the fridge was sown to produce the F3 generation. The number of F3 seeds harvested from each F2 plant varied from 12 to 30. A picture of the F3 pods from the cross between 12CS_37 and 12CS_39 is given in Figure 5.5. The F3 families will be evaluated in different locations in 2018. The effect of the pyramiding process will be evaluated and their additive effect estimated. Cross 12CS_37*12CS_39; desired event: homozygous at TC6E01 and homozygous at TC1E05 Figure 5.5: Selection of the desired F2 individuals based on marker genotype. The F2 progenies are distributed from lane 1 to 96. Each primer pair amplified two bands. One band segregated for each primer. The green and red arrows indicate F2 individuals that are fixed at the locus TC1E05 on LGA8 bearing the fragment from 12CS_39 and are fixed at TC6E01 on LGB5 bearing the fragment from 12CS_37 respectively. 83 University of Ghana http://ugspace.ug.edu.gh 5.4 Discussion In the present study, a BC1 mapping population was used to construct a linkage map containing 330 loci that spanned a total length of 1792 cM with an average marker density of 6.4 cM. This is the first SSR-based map reported on AB-genome tetraploid Arachis involving Arachis valida as a BB genome donor. When comparing the produced map to other published AB tetraploid involving wild species maps, the length and density of the produced map was relatively similar to that of Fonceka et al. (2009) which covered 1843.7 cM with a maker density of 6.1 cM between adjacent markers. The map from this present study was, however, smaller than the consensus map which integrated marker information of 16 linkage maps (Shirasawa et al., 2013), smaller than the map produced by Burow et al. (2001) using RFLP. The rate of distorted marker was similar to the one published by Burow et al. (2001) (24.2% versus 25%). However, this rate is higher compared to the one (10.7%) reported by Fonceka et al. (2009). Segregation distortion is a common feature exhibited by most linkage maps in most plants. Genetically, segregation distortion has been attributed to differences in chromosomes rearrangements, DNA content, abortion of male or female gametes, or the selective fertilization of particular gametic genotypes (Liang et al., 2006). In this particular case, the distorted alleles were skewed toward the cultivated parent. High level of segregation distortion were reported in rice mostly in interspecific crosses (Xu et al., 1997). Therefore, the relatively high proportion of distorted markers may likely imply a huge variability in the parental genotypes as it is often the case in interspecific crosses. The nearly good collinearity found among homeologous pairs suggested relatively conserved genomic sequences between A and B genomes. Groundnut wild species have long been used as useful sources of disease resistance genes. Introgression of disease resistance into cultivated species has been successful and varieties have 84 University of Ghana http://ugspace.ug.edu.gh been released and cultivated ‘Coan’ (Simpson & Starr, 2001), ‘NemaTAM’ (Simpson et al., 2003), ‘Tifguard’ (Holbrook et al., 2008) and Webb (Simpson et al., 2013). However, groundnut wild species have had limited application to decipher more complex traits such as yield surrogate traits due to their low phenotypic performance. In this study, an AB-QTL approach was utilized to evaluate the genetic potential of wild species to improve agronomic traits in cultivated groundnut. Thirty eight (38) QTL for 11 traits were mapped. For more than half of the detected QTL, the positive effects were associated with the wild alleles. There were 10 QTL involving HSW, pod and seed size (PW, PL, SL and WS) for which the wild parent contributed positively to increase the phenotypic value of the trait. These traits clustered on LGs A1, B8 and largely on LG B3. Moreover these traits showed high positive correlations suggesting pleiotropic effects or a number of linked genes responsible for the observed variations. However, for most of the QTL mapped on LG B3, the positive effects were associated with the cultivated parent alleles. In relation to the findings of Huang et al. (2015) study, one could believe that LG B3 may harbour genes related to pod and seed traits which can be deployed in marker-assisted selection even though most QTL mapped on LG B3 in the present study, have positive effects associated with the cultivated parent. Furthermore, Zhang et al. (2017) found a region of 6.89 Mb on chromosome B3 harbouring 107 selective genes one of which codes for Gibberellin-related protein and three genes encoding flavonoid biogenesis and regulation and five nucleotide-binding site leucine-rich repeat NBLRR- encoding genes. As a conclusion, this region may be involved in different traits like plant resistance, plant type related traits and seed related traits. An unanswered question is: is it the same region that is involved in the variation of most of the detected QTL in the present study? Further research such as fine mapping may be 85 University of Ghana http://ugspace.ug.edu.gh conducted to investigate the genes responsible for the QTL on LG B3 even if in the present study, increase of the traits is associated with the cultivated parent. For all identified QTL for PC and GH, the wild alleles tended to increase the value of the trait: deep pod’ constriction and spread growth habit. This result demonstrated in a way the undesirable effects associated with the use of wild species in breeding programmes. However, the detection of QTL and markers associated with these traits could be used to counter-select such regions during introgression processes. While QTL that affected pod and seed size appeared to cluster together, those affecting pod morphology seemed to be dispersed across the genome. Fonceka et al. (2012) highlighted a similar observation and attributed it to some extent to a process related to domestication. A salient result which came out from the present study was the differences in QTL distribution between A and B genomes and the non-consistency of QTL between homeologous LGs while good collinearity between sub-genomes was reported in the present study and from others (Feng et al., 2012; Shirasawa et al., 2013; Nguepjop et al., 2016). Except for PC for which 2 QTL were mapped on LGs A2 and B2, QTL for all remaining traits were identified on non homeologous LGs. Fonceka et al. (2012) reported 96% of QTL cases mapped in non-homeologous regions. This fact could be explained partly by the segregation of alleles in one genome versus the other. No QTL has been identified on LGs A9, B6 and B7 probably due to the lack of wild chromosome introgressions on these particular LGs as a result of strong disequilibrium within the population and/or lack of polymorphism genes located in the cited LGs associated with the variations of studied characters. The examination of the genotype of the B7_32_8_1 line showed a correspondence between the localization of some of the QTL associated with seed and pod characteristics and the introgression regions of that particular line. This result indicates probably 86 University of Ghana http://ugspace.ug.edu.gh an accumulation of useful QTL associated with the wild alleles which makes this particular line to be of interest. In addition, QTL linked to deep pod constriction have also been mapped on genomic regions corresponding to wild introgression regions in B7_32_8_1. This fact could explain the deep pod constriction that is observed on B7_32_8_1. By comparing QTL detected in the present work to other QTL identified for the same trait in other studies, new QTL were mapped. QTL for PH were located on LGs A3, A4, A7, B4, B7, B8, and B10 (Fonceka et al., 2012; Huang et al., 2015). In addition to the QTL identified on A3, two new QTL have been mapped on A10 and B1 groups in this study; and increase in plant height on LG B1 was associated with the wild alleles. They explained moderate PV. This result confirms the findings reported on the polygenetic control of PH (Chen et al., 2017). For pod and seed related traits, previous studies reported QTL on A2, A3, A5, A7, A8, A9, A10, B2, B3, B5, B6, and B9 LGs (Fonceka et al., 2012; Shirasawa et al., 2012; Jiang et al., 2014; Pandey et al., 2014; Huang et al., 2015; Chen et al., 2016; Chen et al., 2016; Huang et al., 2016). Recently, QTL for pod and seed related traits have also been mapped on B1, B4, B7 and B8 (Chen et al., 2017). In this study, QTL for these traits have been mapped on A1, A4 and A6. For all the three newly mapped QTL, increase in the phenotypic value of the trait was associated with the alleles of the wild parent. The distribution of QTL for yield related traits throughout the genome confirms the complex genetic architecture of yield. Moreover, the fact that no QTL has been mapped for pod and haulm yield attest of the complex nature of those traits. 87 University of Ghana http://ugspace.ug.edu.gh 5.5 Conclusion From the study 38 QTL were mapped among which 5 were considered as new. For four out of the five newly mapped QTL, increase of the phenotypic value was associated with the alleles of the wild parent. For about half of the detected QTL, increase of the phenotypic value of the trait was associated with the alleles from the wild relatives. Most of the identified QTL for seed related traits were co-localized on LG B3 and for most of them increase of the trait was brought by the cultivated parent. QTL for traits associated with the negative effects from the wild species were also mapped and markers linked to them identified. This result will be useful to counter- select such non desirable traits during introgression processes. The success of the DNA extraction performed from seeds has permitted selection of desired F2 lines namely 037039_30, 037039_75 and 037039_78 bearing the two wild fragments in the pyramiding process. 88 University of Ghana http://ugspace.ug.edu.gh CHAPTER SIX 6.0 GENERAL CONCLUSIONS AND RECOMMENDATIONS 6.1 General conclusions The Senegalese groundnut sector is one of the most fruitful and important sectors in the country. With the liberalization of the export market and the overgrowing population, producers have more challenges: Produce more on smaller surfaces. Although Senegal is among the top five African producers, yield has declined over the years. The increase in the production is attributed more to cultivation of larger acreage. Many studies have reported the opinions, ideas and participation of farmers regarding the economic aspect but to the best of my knowledge none related to agronomic traits. In this study, a participatory rural appraisal was conducted to assess farmers’ preferences and production constraints and see how the information obtained could help in related researches. The findings imply that during the selection process for developing new varieties attention should be paid to high pod yield, high biomass yield, big pod and seed’ size as well. After the commercialization, drought is the most important abiotic constraint highlighted by farmers. One of the most important limitations that researchers and breeders are facing in groundnut improvement is the low genetic variation available in the cultivated groundnut. In the present work a mapping based population was developed using wild groundnut species. The material developed displayed large phenotypic variation which attests of the potential of the wild species in widening the genetic base of the cultivated groundnut. The variation and low heritability observed for pod and biomass’ yield across environments confirm the complex nature for these traits and the need to use related traits to improve more complex traits. The use of the selection index (IAS) permitted the identification of the best fifty genotypes in each environment. The 89 University of Ghana http://ugspace.ug.edu.gh phenotypic values of the selected genotypes were higher than those of the cultivated parent Fleur11. Among the eleven lines, B7_32_8_1 was identified as the best performing genotype. The genotyping of the BC1 population using 250 polymorphic SSRs allowed production of a genetic map of 330 loci on 20 LGs spanning a cumulative length of 1792 cM with an average marker density of 6.4 cM. LGs were classified into 10 pairs according to homeologous loci. In A-genome, 172 loci were distributed in 10 LGs with a mean number of 17.2 (6 - 24) markers per LG. Each LG spanned on average length of 96.4 cM (45 - 111.2 cM) with inter-marker spacing of 6.1 cM. In B-genome, 158 loci were mapped in 10 LGs with a mean number of 16 markers per LG (9 - 23). Each LG spanned on average of 88.6 cM (59.4 - 137 cM) with inter-marker spacing of 6.4 cM. The QTL analysis mapped 38 QTL for 15 agronomic traits. For yield traits (TPW, PPW and HW) no QTL were identified. Among the 38 QTL, 5 were considered as new loci. For half of the detected QTL, increase of the phenotype value of the trait was associated with the wild alleles. The shortening of the flowering time QTL was associated with the wild alleles. QTL for traits associated with the negative effects from wild species (deep pod constriction, pronounced pod beak and spread growth habit) were also mapped and markers linked to them identified. This result will be useful to counteract such undesirable effects during introgression processes. The use of the genomic DNA extracted from seed to pyramiding useful fragments was a good approach to save time, space and therefore money. It has allowed selecting efficiently the desired F2 individuals which are homozygous at the target loci. The obtained results demonstrate the positive contribution of the wild species to groundnut improvement and offer novel perspectives for exploiting the unused useful alleles available in groundnut wild species. 90 University of Ghana http://ugspace.ug.edu.gh 6.2 Recommendations • Senegalese’ groundnut breeding programme has to pay more attention to farmers’ priorities and engage farmers through participatory varietal selection for them to identify and select varieties that meet their preferences before release. • Breeding for varieties that are drought tolerant and that carry preferred traits would be welcomed by the farmers. Lastly farmers need to have access to land and to enough seeds of quality. • As wild species are known to be tolerant to several biotic and abiotic stresses, it could be interesting to characterize the developed material for drought and leaf spot diseases to identify sources of resistance which may be further used in breeding programmes. • Something that could be interesting is to look at the quality of the haulm to see whether there is an improvement in terms of quality related to the use of wild species • The identified lines should be subjected to multi-locations and multi-years testing for release. • The QTL analysis may be run with more abundant molecular markers (SNP) to confirm the QTL identified in the present study. • The obtained F3 seeds derived from the pyramiding process should be subjected to multi- location and multi-year evaluations to estimate the effect(s) of the accumulation of the two wild segments on the traits. Their pod yield performances should be evaluated as well and compared to Fleur11. • It could be interesting to look at the number of pods per plant of the selected lines to evaluate the impact of an increase in pod size. 91 University of Ghana http://ugspace.ug.edu.gh REFERENCES Abdou, Y. A., Gregory, W. C., & Coopers, W. E. (1971). Sources and Nature of Resistance to Cercospora arachidicola and Cercosporidium personatum ( Beck & Curtis ) Deighton in Arachis Species. Peanut Science, 6–11. Alwala, S., Kwolek, T., McPherson, M., Pellow, J., & Meyer, D. (2010). A comprehensive comparison between Eberhart and Russell joint regression and GGE biplot analyses to identify stable and high yielding maize hybrids. Field Crops Research, 119(2-3), 225–230. Ando, T., Yamamoto, T., Shimizu, T., Ma, X. F., Shomura, A., Takeuchi, Y., … Yano, M. (2008). Genetic dissection and pyramiding of quantitative traits for panicle architecture by using chromosomal segment substitution lines in rice. Theoretical and Applied Genetics, 116(6), 881–890. ANSD. (2015). Agence Nationale de la Statistique et de la Demographie. Retrieved March 1, 2015, from http://www.ansd.sn/ Bänziger, M., & Lafitte, H. R. (1997). Environments, Efficiency of Secondary Traits for Improving Maize for Low-Nitrogen Target. Crop Science, 37, 1110–1117. Bechara, M. D., Moretzsohn, M. C., Palmieri, D. A., Monteiro, J. P., Bacci, M., Martins, J., … Gimenes, M. A. (2010). Phylogenetic relationships in genus Arachis based on ITS and 5.8S rDNA sequences. BMC Plant Biology, 10, 255. Bernacchi, D., Beck-Bunn, T., Eshed, Y., Lopez, J., Petiard, V., Uhlig, J., … Tanksley, S. (1998). Advanced backcross QTL analysis in tomato. I. Identification of QTLs for traits of agronomic importance from Lycopersicon hirsutum. Theoretical and Applied Genetics, 97(3), 381–397. 92 University of Ghana http://ugspace.ug.edu.gh Bernardo, R. (2009). Genomewide selection for rapid introgression of exotic germplasm in maize. Crop Science, 49(2), 419–425. Bertioli, D. J., Cannon, S. B., Froenicke, L., Huang, G., Farmer, A. D., Cannon, E. K. S., … Chu, Y. (2016). The genome sequences of Arachis duranensis and Arachis ipaensis , the diploid ancestors of cultivated peanut. Nature Genetics, 48(4), 438–446. Bidinger, F. R. (1998). Farmer participation in pearl millet research in Namibia. In Proc. participatory plant improvement. MS Swaminathan Research Foundation (MSSRF)- International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) work (21–30). Birthal, P. S., Rao, P. P., Nigam, S. N., Bantilan, C. S., & Bhagavatulu , S. (2010). Groundnut and soybean economies of Asia: facts; trends and outlook. Patancheru; Andhra Pradesh; India. Broman, K. W., & Sen, S. (2009). A Guide to QTL Mapping with R/qtl, Statistics for Biology and Health. Buerstmayr, H., Lemmens, M., Hartl, L., Doldi, L., Steiner, B., Stierschneider, M., & Ruckenbauer, P. (2002). Molecular mapping of QTLs for Fusarium head blight resistance in spring wheat . I . Resistance to fungal spread ( Type II resistance ). Theorical and Applied Genetics, 104, 84–91. Burks, A W., Shin, D., Cockrell, G., Stanley, J. S., Helm, R. M., & Bannon, G. A. (1997). Mapping and mutational analysis of the IgE-binding epitopes on Ara h 1, a legume vicilin protein and a major allergen in peanut hypersensitivity. European Journal of Biochemistry, 245(2), 334–339. Burow, M. D., Simpson, C. E., Starr, J. L., & Paterson, A. H. (2001). Transmission genetics of 93 University of Ghana http://ugspace.ug.edu.gh chromatin from a synthetic amphidiploid to cultivated peanut (Arachis hypogaea L.): Broadening the gene pool of a monophyletic polyploid species. Genetics, 159, 823–837. Burow, M.D., Simpson, C.E., Faries, M.W., Starr, J.L., & Paterson, A.H. (2009). Molecular biogeographic study of recently described B- and A-genome Arachis species, also providing new insights into the origins of cultivated peanut. Genome, 52 (2), 107–119. Ceccarelli, S., Grando, S., & Baum, M. (2007). Participatory plant breeding in water-limited environments. Experimental. Agriculture, 43, 411–436. Chee, P., Draye, X., Jiang, C.-X., Decanini, L., Delmonte, T. A., Bredhauer, R., … Paterson, A. H. (2005). Molecular dissection of interspecific variation between Gossypium hirsutum and Gossypium barbadense (cotton) by a backcross-self approach : I . Fiber elongation. Theoretical and Applied Genetics, 111, 757–763. Chen, W., Jiao, Y., Cheng, L., Huang, L., Liao, B., Tang, M., … Jiang, H. (2016). Quantitative trait locus analysis for pod- and kernel-related traits in the cultivated peanut (Arachis hypogaea L .). BMC Genetics, 1–9. Chen, X., Li, H., Pandey, M. K., Yang, Q., Wang, X., Garg, V., & Li, H. (2016). Draft genome of the peanut A-genome progenitor (Arachis duranensis) provides insights into geocarpy , oil biosynthesis , and allergens. In Proceedings of the National Academy of Sciences, United States of America, 113, 6785 – 6790. Chen, Y., Ren, X., Zheng, Y., Zhou, X., Huang, L., Yan, L., … Jiang, H. (2017). Genetic mapping of yield traits using RIL population derived from Fuchuan Dahuasheng and ICG6375 of peanut (Arachis hypogaea L .). Molecular Breeding, 37(2), 1–14. Chiulele, R. M., Mwangi, G., Tongoona, P., Ndeve, A. D., & Ehlers, J. D. (2011). Assessment of 94 University of Ghana http://ugspace.ug.edu.gh farmers ’ perceptions and preferences of cowpea in Mozambique. African Crop Science Conference Proceedings, 10 (1993), 311–318. Chu, Y., Wu, C. L., Holbrook, C. C., Tillman, B. L., & Person, G. (2011). Marker-Assisted Selection to Pyramid Nematode Resistance and the High Oleic Trait in Peanut. Plant Genome, 4, 110–117. Clavel, D., Diouf, O., Khalfaoui, J. L., & Braconnier, S. (2006). Genotypes variations in fluorescence parameters among closely related groundnut (Arachis hypogaea L.) lines and their potential for drought screening programs. Field Crops Research, 96(2-3), 296–306. Clavel, D., Sarr, B., Marone, E., & Ortiz, R. (2008). Agronomy for sustainable development. Italian Journal of Agronomy, 3(3), 77–78. Collard, B. C. Y., Jahufer, M. Z. Z., Brouwer, J. B., & Pang, E. C. K. (2005). An introduction to markers , quantitative trait loci (QTL) mapping and marker-assisted selection for crop improvement : The basic concepts. Euphytica, 142, 169–196. Concibido, V., La Vallee, B., Mclaird, P., Pineda, N., Meyer, J., Hummel, L., … Delannay, X. (2003). Introgression of a quantitative trait locus for yield from Glycine soja into commercial soybean cultivars. Theoretical and Applied Genetics, 106(4), 575–582. Custódio, A. R., Seijo, G., & Valls, J. F. M. (2013). Characterization of Brazilian accessions of wild Arachis species of section Arachis (Fabaceae) using heterochromatin detection and fluorescence in situ hybridization (FISH). Genetics and Molecular Biology, 36(3), 364–370. Diatta, C. (2017). Development of sorghum [Sorghum bicolor (L) Moench] for resistance to grain mold in senegal. West Africa Centre for Crop Improvement (WACCI) Legon/ACCRA. 95 University of Ghana http://ugspace.ug.edu.gh Diome, T., Ndong, A., Kébé, K., Thiaw, C., Ndiaye, A., Doumma, A., … Sembène, M. (2013). Effect of agro-ecological zones and contiguous basin crops of groundnut (Arachis hypogaea) on the structuring and genetic diversity of Caryedon serratus (Coleoptera: Chrysomelidae, Bruchinae) in the sub-region of West Africa. Journal of Asia-Pacific Entomology, 16(3), 209–217. Dwivedi, S., & Upadhyaya, H. D. (2008). Enhancing Crop Gene Pools with Beneficial Traits Using Wild Relatives. Plant Breeding Reviews, 30(3), 180–230. Egbadzor, K. F., Yeboah, M., Offei, S. K., Ofori, K., & Danquah, E. Y. (2013). Farmers ’ key production constraints and traits desired in cowpea in Ghana. Journal of Agricultural Extension and Rural Development, 5(January), 14–20. Fan, Z., Robbins, M. D., & Staub, J. E. (2006). Population development by phenotypic selection with subsequent marker-assisted selection for line extraction in cucumber (Cucumis sativus L .). Theoretical and Applied Genetics, 112, 843–855. FAOSTAT. (2015). http://faostat.fao.org/faosat/ (last accessed 20 February 2015). Fávero, A. P., Simpson, C. E., Valls, J. F. M., & Vello, N. A. (2006). Study of the evolution of cultivated peanut through crossability studies among Arachis ipaënsis, A. duranensis, and A. hypogaea. Crop Science, 46(4), 1546–1552. Faye, I., Pandey, MK., Hamidou, F., Rathore, A., Ndoye, O., Vadez, V., & Varshney, RK. (2015) Identification of quantitative trait loci for yield and yield related traits in groundnut (Arachis hypogaea L.) under different water regimes in Niger and Senegal. Euphytica, 206:631–647 Feng, S., Wang, X., Zhang, X., Dang, P. M., Holbrook, C. C., Culbreath, A. K., … Guo, B. 96 University of Ghana http://ugspace.ug.edu.gh (2012). Peanut (Arachis hypogaea) Expressed Sequence Tag Project: Progress and Application. Comparative and Functional Genomics. 373768, 9. Feuillet, C., Langridge, P., & Waugh, R. (2008). Cereal breeding takes a walk on the wild side. Trends in Genetics, 24(1), 24–32. Foiada, F., Westermeier, P., Kessel, B., Ouzunova, M., Wimmer, V., Mayerhofer, W., … Carolin, C. (2015). Improving resistance to the European corn borer : a comprehensive study in elite maize using QTL mapping and genome wide prediction. Theoretical and Applied Genetics, 128(5), 875–891. Foncéka, D., Hodo-Abalo, T., Rivallan, R., Faye, I., Mbaye Ndoye Sall, Ousmane Ndoye, Alessandra P Fávero, D. J. B., Glaszmann, J.-C., … Rami, J.-F. (2009). Genetic mapping of wild introgressions into cultivated peanut: a way toward enlarging the genetic basis of a recent allotetraploid. BMC Plant Biology, 9(1), 103. Fonceka, D., Tossim, H. A., Rivallan, R., Vignes, H., Lacut, E., de Bellis, F., … Rami, J. F. (2012). Construction of Chromosome Segment Substitution Lines in Peanut (Arachis hypogaea L.) Using a Wild Synthetic and QTL Mapping for Plant Morphology. PLoS ONE, 7(11). Fonceka, D., Tossim, H.-A., Rivallan, R., Vignes, H., Faye, I., Ndoye, O., … 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. Fox, A. T., Sasieni, P., du Toit, G., Syed, H., & Lack, G. (2009). Household peanut consumption as a risk factor for the development of peanut allergy. The Journal of Allergy and Clinical Immunology, 123(2), 417–423. 97 University of Ghana http://ugspace.ug.edu.gh Francia, E., Tacconi, G., Crosatti, C., Barabaschi, D., Bulgarelli, Dall'Aglio, E., & Vale, G. (2005). Marker assisted selection in crop plants. Plant Cell, Tissue and Organ Culture, 82, 317–342. Fu, Q., Zhang, P., Tan, L., Zhu, Z., Ma, D., Fu, Y., … Sun, C. (2010). Analysis of QTLs for yield-related traits in Yuanjiang common wild rice (Oryza rufipogon Griff.). Journal of Genetics and Genomics, 37(2), 147–157. Fulton, T. M., Beck-Bunn, T., Emmatty, D., Eshed, Y., Lopez, J., Petiard, V., … Tanksley, S. D. (1997). QTL analysis of an advanced backcross of Lycopersicon peruvianum to the cultivated tomato and comparisons with QTLs found in other wild species. Theoretical and Applied Genetics, 95(5-6), 881–894. Garcia, G. M., Stalker, H. T., Shroeder, E., & Kochert, G. (1996). Identification of RAPD, SCAR, and RFLP markers tightly linked to nematode resistance genes introgressed from Arachis cardenasii into Arachis hypogaea. Genome, 39(5), 836–845. Gautami, B., Pandey, M. K., Vadez, V., Nigam, S. N., Ratnakumar, P., Krishnamurthy, L. ...Varshney, R. K. (2012). Quantitative trait locus analysis and construction of consensus genetic map for drought tolerance traits based on three recombinant inbred line populations in cultivated groundnut (Arachis hypogaea L.). Molecular Breeding, 30:757–772. Gimenes, M.A., Lopes, C.R., Galgaro, M.L., Valls, J.F.M., & Kochert, G. (2002). RFLP analysis of genetic variation in species of section Arachis, genus Arachis (Leguminosae). Euphytica, 123, 421–429. Gomes, R. L. F., & Lopes, Â. C. de A. (2005). Correlations and path analysis in sunflower. Crop Breeding and Applied Biotechnology, 5, 105–112. 98 University of Ghana http://ugspace.ug.edu.gh Grandillo, S., & Tanksley, S. D. (2003). Advanced backcross QTL analysis :Results and perspectives. In R. Tuberosa, R. . Phillips, & M. Gale (Eds.), In the Wake of the Double Helix: From the Green Revolution to the Gene Revolution (115–132). Bologna, Italy. Gregory, M. P., & Gregory, W. C. (1979). Exotic germplasm of Arachis L. interspecific hybrids. Journal of Heredity, 70(3), 185–193. Gregory, W.C. (1976). Arachis hypogaea (Leguminosae- Papilionadae). Evolution of Crop Plants, 134, 151. Grieshammer, U., & Wynne, J. C. (1990). Mendelian and Non-Mendelian Inheritance of Three Isozymes in Peanut (Arachis hypogaea L .)’. Peanut Science, 17(2), 101–105. Gur, A., & Zamir, D. (2004). Unused natural variation can lift yield barriers in plant breeding. PLoS Biology, 2(10), 1610-1615. Haley, C. S., & Knott, S. A. (1992). A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity, 69(4), 315–324. Halward, T. M., Stalker, H. T., Larue, E. A., & 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. Halward, T., Stalker, H. T., & Kochert, G. (1993). Development of an RFLP linkage map in diploid peanut species. Theoretical and Applied Genetics, 87(3), 379–384. Hammons, R. O. (1970). Spancross—a new peanut variety, Reg. no. 3. Crop Science, 10, 459– 460. Hammons, R. O. (1994). The origin and history of the groundnut. In The Groundnut Crop ( 24– 99 University of Ghana http://ugspace.ug.edu.gh 42). Springer Netherlands. Hanson, P., Lu, S. F., Wang, J. F., Chen, W., Kenyon, L., Tan, C. W., … Yang, R. Y. (2016). Conventional and molecular marker-assisted selection and pyramiding of genes for multiple disease resistance in tomato. Scientia Horticulturae, 201, 346–354. He, G., Woullard, F. E., Marong, I., & Guo, B. Z. (2006). Transferability of Soybean SSR Markers in Peanut (Arachis hypogaea L.). Peanut Science, 33, 22–28. Hilu, K. W., & Stalker, H. T. (1995). Genetic relationships between peanut and wild species of Arachis section Arachis (Fabaceae): Evidence from RAPDs. Plant Systematics and Evolution, 198(3), 167–178. Ho, C., Mccouch, R., & Smith, E. (2002). Improvement of hybrid yield by advanced backcross QTL analysis in elite maize. Theoretical and Applied Genetics, 105, 440–448. Holbrook, C. C., Timper, P., Culbreath, A. K., & Kvien, C. K. (2008). Registration of “Tifguard” peanut. Journal of Plant Peanut 2. Holbrook, C., & Stalker, H. (2003). Peanut Breeding and Genetic Resources. Plant Breeding Reviews, 22, 297-355 Hollowell, J. E., Isleib, T. G., Tallury, S. P., Copeland, S. C., & Shew, B. B. (2008). Screening of Virginia-Type Peanut Breeding Lines for Resistance to Cylindrocladium Black Rot and Sclerotinia Blight in the Greenhouse. Peanut Science, 1(35), 18–24. Hong, Y., Chen, X., Liang, X., Liu, H., Zhou, G., Li, S., … Guo, B. (2010). A SSR-based composite genetic linkage map for the cultivated peanut (Arachis hypogaea L.) genome. BMC Plant Biology, 10(1), 17. 100 University of Ghana http://ugspace.ug.edu.gh Hopkins, M. S., Casa, A. M., Wang, T., Mitchell, S. E., Dean, R. E., Kochert, G. D., & Kresovich, S. (1998). Discovery and Characterization of Polymorphic Simple Sequence Repeats (SSRs) in Peanut. Crop Science, 39(4), 1243–1247. Huang, L., He, H., Chen, W., Ren, X., Chen, Y., Zhou, X., … Jiang, H. (2015). Quantitative trait locus analysis of agronomic and quality-related traits in cultivated peanut (Arachis hypogaea L.). Theoretical and Applied Genetics, 128(6), 1103–1315. Huang, L., Ren, X., Wu, B., Li, X., Chen, W., Zhou, X., & Chen, Y. (2016). Development and deployment of a high-density linkage map identified quantitative trait loci for plant height in peanut ( Arachis hypogaea L .). Scientific Report. 6,39478. Huang, X. Q., Cöster, H., Ganal, M. W., & Röder, M. S. (2003). Advanced backcross QTL analysis for the identification of quantitative trait loci alleles from wild relatives of wheat ( Triticum aestivum L.). Theoretical and Applied Genetics. 106(8), 1379–1389. Huang, X. Q., Kempf, H., Ganal, M. W., & Röder, M. S. (2004). Advanced backcross QTL analysis in progenies derived from a cross between a German elite winter wheat variety and a synthetic wheat (Triticum aestivum L.). Theoretical and Applied Genetics, 109, 933–943. Hull, E. ., & Carver, W. (1938). Peanut Improvement. Florida Agricultural Experiment Annual Report, 36–40. Husted, L. (1936). Cytological Studies an the Peanut, Arachis. II Chromosome number, morphology and behavior, and their application to the problem of the origin of the cultivated forms. Cytologia, 7(3), 396–423. Jambunathan, R. (1991). Groundnut Quality Characteristics. In Uses of tropical grain legumes. Patancheru, India: ICRISAT. 267–275 101 University of Ghana http://ugspace.ug.edu.gh Janila, P., Manohar, S. S., Patne, N., Variath, M. T., & Nigam, S. N. (2016). Genotype × Environment interactions for oil content in peanut and stable high-oil-yielding sources. Crop Science, 56(5), 2506–2515. 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 Science, 4(February), 23. Janila, P., Variath, M. T., Pandey, M. K., Desmae, H., & Motagi, B. N. (2016). Genomic Tools in Groundnut Breeding Program : Status and Perspectives. Frontiers in Plant Science, 7(March), 2004–2013. Jiang, H., Huang, L., Ren, X., Chen, Y., Zhou, X., Xia, Y., … Lei, Y. (2014). Diversity characterization and association analysis of agronomic traits in a Chinese peanut ( Arachis hypogaea L .) mini - core collection. Journal of Integrative Plant Biology, 56(2), 159–169. Jiang, R., Manson, J. E., Stampfer, M. J., Liu, S; Willett, W. C., & Hu, F. (2002). Nut and Peanut Butter Consumption and Risk of Type 2 Diabetes in Women. The Journal of the American Medical Association, 288(20), 2554–2560. Jiao, Y., Vuong, T. D., Liu, Y., Li, Z., Noe, J., Robbins, R. T., … Nguyen, H. T. (2015). Identification of quantitative trait loci underlying resistance to southern root-knot and reniform nematodes in soybean accession PI 567516C. Molecular Breeding, 35(6), 1–10. Jordan, D. R., Mace, E. S., Cruickshank, A. W., Hunt, C. H., & Henzell, R. G. (2011). Exploring and exploiting genetic variation from unadapted sorghum germplasm in a breeding program. Crop Science, 51(4), 1444–1457. Keurentjes, J. J. B., Bentsink, L., Alonso-Blanco, C., Hanhart, C. J., Vries, H. B., Effgen, S., … 102 University of Ghana http://ugspace.ug.edu.gh Koornneef, M. (2007). Development of a near-isogenic line population of Arabidopsis thaliana and comparison of mapping power with a recombinant inbred line population. Genetics, 175(2), 891–905. Khedikar, Y. P., Gowda, M. V. C., Sarvamangala, C., Patgar, K. V., Upadhyaya, H. D., & 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, 971–984. Kochert, G., Stalker, H. T., Gimenes, M., & Galgaro, L. C. (1996). RFLP and cytogenetic evidence on the origin and evolution of allotetraploid domesticated peanut, Arachis hypogaea (Leguminosae). American Journal of Botanic, 83, 1282–1291. Kochert, G., Halward, T., Branch, W. D., & Simpson, C. E. (1991). RFLP variability in peanut (Arachis hypogaea L.) cultivars and wild species. Theoretical and Applied Genetics, 81(5), 565–570. Koppelman, S. J., Hefle, S. L., Taylor, S. L., & de Jong, G. a H. (2010). Digestion of peanut allergens Ara h 1, Ara h 2, Ara h 3, and Ara h 6: a comparative in vitro study and partial characterization of digestion-resistant peptide. Molecular Nutrition & Food Research, 54, 1711–1721. Krapovickas, A., & Gregory, W. C. (1994). Taxonomía del género “ Arachis (leguminosae).” Bonplandia, 8, 1–186. Lanham, P. G., Fennell, S., Moss, J. P., & Powell, W. (1992). Detection of polymorphic loci in Arachis germplasm using random amplified polymorphic DNAs. Genome, 35(5), 885–889. Leal-Bertioli, S. C. M., Cavalcante, U., Gouvea, E. G., Ballén-Taborda, C., Shirasawa, K., 103 University of Ghana http://ugspace.ug.edu.gh Guimarães, P. M., … Moretzsohn, M. C. (2015). Identification of QTLs for Rust Resistance in the Peanut Wild Species Arachis magna and the Development of KASP Markers for Marker-Assisted Selection. G3 Genes Genomes Genetics, 5(7), 1403–13. Leal-Bertioli, S. C. M., Farias, M. P., Silva, P. Í. T., Guimarães, P. M., Brasileiro, A. C. M., Bertioli, D. J., & Araujo, A. C. G. (2010). Ultrastructure of the Initial Interaction of Puccinia arachidis and Cercosporidium personatum with Leaves of Arachis hypogaea and Arachis stenosperma. Journal of Phytopathology, 158(11-12), 792–796. Leal-Bertioli, S. C. M., José, A. C. V. F., Alves-Freitas, D. M. T., Moretzsohn, M. C., Guimarães, P. M., Nielen, S., … Bertioli, D. J. (2009). Identification of candidate genome regions controlling disease resistance in Arachis. BMC Plant Biology, 9, 112. Leal-Bertioli, S. C. M., Moretzsohn, M. C., Roberts, P. A., Ballén-Taborda, C., Borba, T. C. O., Valdisser, P. A., … Bertioli, D. J. (2015). Genetic Mapping of Resistance to Meloidogyne arenaria in Arachis stenosperma: A New Source of Nematode Resistance for Peanut. G3 Genes Genomes Genetics, 6(2), 377–90. Lecomte, L., Duff, P., Buret, M., Servin, B., Hospital, F., & Causse, M. (2004). Marker-assisted introgression of five QTLs controlling fruit quality traits into three tomato lines revealed interactions between QTLs and genetic backgrounds. Theoretical and Applied Genetics, 109(3), 658–668. Liang, S. X., Zhen, S. X., Zhen, Z. T., Xian-liang, S., Xue-zhen, S., & Tian-zhen, Z. (2006). Review article Segregation distortion and its effect on genetic mapping in plants. Chinese Journal of Agricultural Biotechnology, 3(3), 163–169. Liang, X., Chen, X., Hong, Y., Liu, H., Zhou, G., Li, S., & Guo, B. (2009). Utility of EST- 104 University of Ghana http://ugspace.ug.edu.gh derived SSR in cultivated peanut (Arachis hypogaea L.) and Arachis wild species. BMC Plant Biology, 9(1), 35. Lin, B. B. (2011). Resilience in Agriculture through Crop Diversification: Adaptive Management for Environmental Change. BioScience, 61(3), 183–193. Lin, C. Y. (1978). Index selection for genetic improvement of quantitative characters. Theoretical and Applied Genetics, 52(2), 49–56. Lorieux, M. (2012). MapDisto : fast and efficient computation of genetic linkage maps. Molecular Breeding, 30, 1231–1235. Lu, J., & Pickersgill, B. (1993). Isozyme variation and species relationships in peanut and its wild relatives (Arachis L. — Leguminosae). Theoretical and Applied Genetics, 85(5), 550– 560. Luo, H., Ren, X., Li, Z., Xu, Z., Li, X., Huang, L., … Jiang, H. (2017). Co-localization of major quantitative trait loci for pod size and weight to a 3.7 cM interval on chromosome A05 in cultivated peanut (Arachis hypogaea L.). BMC Genomics, 18(1), 58. Mallikarjuna, N., Pande, S., Jadhav, D. R., Sastri, D. C., & Rao, J. N. (2004). Introgression of disease resistance genes from Arachis kempff-mercadoi into cultivated groundnut. Plant breeding, 123, 573–576. Mangione, D., Senni, S., Puccioni, M., Grando, S., & Ceccarelli, S. (2006). The cost of participatory barley breeding. Euphytica, 150(3), 289–306. Mendes, A. J. . (1947). Estudos citológicos no gênero Arachis. Bragantia Campinas, 7(11), 257– 267. 105 University of Ghana http://ugspace.ug.edu.gh Metcalfe, D. D. (1996). Food allergy in adults. In D. D. Metcalfe, H. A. Sampson, & R. A. Simon (Eds.), Food allergy: adverse reactions to food and food additives (183–191). Cambridge: Blackwell Science Inc. Milla, S. R., Isleib, T. G., & Stalker, H. T. (2005). Taxonomic relationships among Arachis section. Arachis species as revealed by AFLP markers. Genome, 48(1), 1–11. Monforte, A. J., & Tanksley, S. D. (2000). Development of a set of near isogenic and backcross recombinant inbred lines containing most of the Lycopersicon hirsutum genome in a L. esculentum genetic background: a tool for gene mapping and gene discovery. Genome, 43(5), 803–813. Monyo, E., Chamanga, A., Maideni, F., Mponda, O., Mizambwa, F., & Anane, M. (2012). Enhancing Groundnut Productivity and Production in Eastern and Southern Africa. In I. C. R. I. for the S.-A. Tropics (Ed.), Four Seasons of learning and Engaging Smallholder Farmers-Phase 1. (33–53). India. Moretzsohn, M. C., Barbosa, A. V., Alves-Freitas, D. M., Teixeira, C., Leal-Bertioli, S. C., Guimarães, P. M., … Gimenes, M. A. (2009). A linkage map for the B-genome of Arachis (Fabaceae) and its synteny to the A-genome. BMC Plant Biology, 9, 40. Moretzsohn, M. D. C., Hopkins, M. S., Mitchell, S. E., Kresovich, S., Valls, J. F. M., & Ferreira, M. E. (2004). Genetic diversity of peanut (Arachis hypogaea L.) and its wild relatives based on the analysis of hypervariable regions of the genome. BMC Plant Biology, 4, 11. Moretzsohn, M. C., Leoi, L., Proite, K., Guimaraes, P. M., Leal-Bertioli, S. C., Gimenes, M. A., Martins, W. S., Valls, J.F., Grattapaglia, D., Bertioli, D. J. (2005). A microsatellite-based, gene-rich linkage map for the AA genome of Arachis (Fabaceae). Theoretical and Applied 106 University of Ghana http://ugspace.ug.edu.gh Genetics, 111, 1060–1071. Mulatu, E., & Belete, K. (2001). Participatory varietal selection in lowland sorghum in eastern Ethiopia: impact on adoption and genetic diversity. Experimental Agriculture, 37 (2), 211- 229. Muñoz, N., Liu, A., Kan, L., Li, M.-W., & Lam, H.-M. (2017). Potential Uses of Wild Germplasms of Grain Legumes for Crop Improvement. International Journal of Molecular Sciences, 18(2), 328. Nagy, E. D., Guo, Y., Tang, S., Bowers, J. E., Okashah, R. A., Taylor, C. A., … Knapp, S. J. (2012). A high-density genetic map of Arachis duranensis, a diploid ancestor of cultivated peanut. BMC Genomics, 13(1), 469. Nautiyal, P. C., Rajgopal, K., Zala, P. V., Pujari, D. S., Basu, M., Dhadhal, B. A., & Nandre, B. M. (2008). Evaluation of wild Arachis species for abiotic stress tolerance: I. Thermal stress and leaf water relations. Euphytica, 159(1-2), 43–57. Ndjeunga, J., Ntare, B. ., Abdoulaye, A., Ibro, A., Zarafi, M. ., Cisse, Y., … Micko, I. (2010). Farmer preferences for groundnut traits and varieties in West Africa: Cases of Mali, Niger and Nigeria. International Crops Research Institute for the Semi-Arid Tropics, 27. Nevo, E., & Chen, G. (2010). Drought and salt tolerances in wild relatives for wheat and barley improvement. Plant, Cell & Environment, 33(4), 670–685. Nguepjop, J. R., Tossim, H., Bell, J. M., Rami, J., Sharma, S., Courtois, B., … Sane, D. (2016). Evidence of Genomic Exchanges between Homeologous Chromosomes in a Cross of Peanut with Newly Synthetized Allotetraploid Hybrids. Frontiers in Plant Science, 7(November), 1–12. 107 University of Ghana http://ugspace.ug.edu.gh Nigam, S.N . (2005). yield potential versus yield barrier in groundnut. International Crops Research Institute for the Semi-Arid Tropics, India. Ntare, B., Ndjeunga, J., Waliyar, F., Kodio, O., Echekwu, C. A., Kapran, I., … Sako, K. (2007). Farmer Participatory Evaluation and Dissemination of Improved Groundnut Varieties in West Africa. International Crops Research Institute for the Semi-Arid Tropics, India, 36. Ntare, B., Waliyar, F., Ndjeuga, J., Diarra, A., Amadou, A., Abdoulaye, I., … Adamu, A. (2012). Enhancing Groundnut Productivity and Production in Drought –Prones Areas of West and Central Africa. In Four Seasons of learning and Engaging Smallholder Farmers-Progress of Phase 1 (Abate.,55–80). India: International Crops Research Institute for the Semi-Arid Tropics. Omanya, G. O., Weltzien-rattunde, E., Sogodogo, D., Sanogo, M., Hanssens, N., Guero, Y., & Zangre, R. (2007). Participatory varietal selection with improved pearl millet in West Africa. Experimental Agriculture, 43, 5–19 Oppong-Sekyere, D., Akromah, R., Akpalu, M. M., Ninfaa, A. D., & Nyamah, E. Y. (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. Paik-Ro, O. G., Smith, R. L., & Knauft, D. A. (1992). Restriction fragment length polymorphism evaluation of six peanut species within the Arachis section. Theoretical and Applied Genetics, 84(1), 201–208. Pande, S. and Rao, J. N. (2001). Resistance of Wild Arachis Species to Late Leaf Spot and Rust 108 University of Ghana http://ugspace.ug.edu.gh in Greenhouse Trials. Plant Disease, 85(8), 851–855. Pandey, M. K., Agarwal, G., Kale, S. M., Clevenger, J., Nayak, S. N., Upadhyaya, H. D., … Liang, X. (2017). Development and Evaluation of a High Density Genotyping “ Axiom _ Arachis ” Array with 58 K SNPs for Accelerating Genetics and Breeding in Groundnut. Scientific Reports, 7, 40577. Pandey, M. K., Monyo, E., Ozias-akins, P., & Liang, X. (2012). Advances in Arachis genomics for peanut improvement. Biotechnology Advances, 30(3), 639–651. Pandey, M. K., Upadhyaya, H. D., Rathore, A., Vadez, V., Sheshshayee, M. S., Sriswathi, M., … Varshney, R. K. (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. Parsad, R., Dhandapani, A., & Khandelwal, M. K. (2011). SAS Macro for Analysis of Data from Augmented Block Designs. Peñaloza, A. P. S., & Valls, J. F. M. (2005). Chromosome number and satellited chromosome morphology of eleven species of Arachis (leguminosae). Bonplandia Argentina, 14(1-2), 65–72. Rabjohn, P., Helm, E. M., Stanley, J. S., West, C. M., Sampson, H. A., Burks, A. W., & Bannon, G. A. (1999). Molecular cloning and epitope analysis of the peanut allergen Ara h 3. Journal of Clinical Investigation, 103(4), 535–542. Raina, S. N., & Mukai, Y. (1999). Detection of a variable number of 18S-5.8S-26S and 5S ribosomal DNA loci by fluorescent in situ hybridization in diploid and tetraploid Arachis species. Genome, 42(1), 52–59. 109 University of Ghana http://ugspace.ug.edu.gh Rao, N. K., Reddy, L. J., & Bramel, P. J. (2003). Potential of wild species for genetic enhancement of some semi-arid food crops. Genetic Resources and Crop Evolution, 50(7), 707–721. Rao, P. S., Reddy, P. S., Rathore, A., Reddy, B. V. S., & Panwar, S. (2011). Application GGE biplot and AMMI model to evaluate sweet sorghum {Sorghum bicolor) hybrids for genotype  environment interaction and seasonal adaptation. Indian Journal of Agricultural Sciences, 81(5), 438–444. Rao, V. R.; Murthy, U. R. Botany – morphology and anatomy (1994). In: Smartt, J.(ed.) The Groundnut Crop. London:Chapman & Hall. 43-95. Rathnakumar, A. ., Hariprasanna, K., & Lalwani, H. B. (2012). Genetic improvement in Spanish type peanut, Arachis hypogaea L. varieties in India over the years. Journal of Oilseeds Ressources, 27, 1–7. Ravi, K., Vadez, V., Isobe, S., Mir, R. R., Guo, Y., Nigam, S. N., … Varshney, R. K. (2011). Identification of several small main-effect QTLs and a large number of epistatic QTLs for drought tolerance related traits in groundnut (Arachis hypogaea L.). Theoretical and Applied Genetics, 122 (6), 1119–1132. Reddy, T. Y., Reddy, V. R., & Anbumozhi, V. (2005). Physiological responses of groundnut (Arachis hypogaea L.) to drought stress and its amelioration: A review. Acta Agronomica Hungarica, 51(2), 205–227. Revoredo, C. L., & Fletcher, S. M. (2002). World Peanut Market : An Overview of the Past 30 Years. Research Bulletin, The University of Georgia, (437). Ribaut, J. M., & Hoisington, D. (1998). Marker-assisted selection: new tools and strategies. 110 University of Ghana http://ugspace.ug.edu.gh Trends in Plant Science, 3(6), 236–239. Ribaut, J. M., & Ragot, M. (2007). Marker-assisted selection to improve drought adaptation in maize: The backcross approach, perspectives, limitations, and alternatives. Journal of Experimental Botany, 58(2), 351–360. Ross-Ibarra, J., Morrell, P. L., & Gaut, B. S. (2007). Plant domestication, a unique opportunity to identify the genetic basis of adaptation. Proceedings of the National Academy of Sciences of the United States of America, 104, 8641–8. Safari, P., Honarnejad, R., & Esfahani, M. (2013). Indirect selection for increased oil yield in peanut : comparison selection indices and biplot analysis for simultaneous improvement multiple traits. International Journal of Biosciences, 3(8), 87–96. Sanogo, O. (2017). Early generation selection for drought related traits in groundnut (Arachis hypogaea L.) in Mali. West Africa Centre for Crop Improvement (WACCI) Legon/Accra. Sanou, A., Adam, M., Brocke, K. V., & Trouche, G. (2014). Enquête sur l’adoption et la diffusion des variétés de sorgho issues de la sélection participative dans les régions Centre- Nord et Boucle du Mouhoun : résultats et discussion. Bilan Thématique Programmé : Production agricole et sécurité alimentaire en Afrique de l’Ouest. AMSP; UGCPA-BM, l’INERA et CIRAD, 43. Savage, G. & Keenan, J.I. (1994). The composition and nutritive value of groundnut kernels. In The Groundnut Crop: A scientific basis for improvement. editors: Smartt. Chapman & Hall. London. Seijo, J. G., Graciela, I. L., Fernández, A., Krapovickas, A., Ducasse, D., & Moscone, E. A. (2004). Physical mapping of the 5S and 18S-25S rRNA genes by FISH as evidence that 111 University of Ghana http://ugspace.ug.edu.gh Arachis duranensis and A. ipaensis are the wild diploid progenitors of A. hypogaea (Leguminosae). American Journal of Botany, 91(9), 1294–1303. Selvaraj, M. G., Schubert, A. M., Ayers, J. L., & Baring, M. R. (2009). Identification of QTLs for pod and kernel traits in cultivated peanut by bulked segregant analysis. Electronic Journal of Biotechnology, 12(2). Sembene, M. (2006). The origin of groundnut infestation by the seed beetle Caryedon serratus (Olivier) (Coleoptera: Bruchidae): Results from cytochrome B and ITS1 gene sequences. Journal of Stored Products Research, 42(2), 97–111. Sembene, M., Rasplus, J., Silvain, J., & Delobel, A. (2008). Genetic differentiation in sympatric populations of the groundnut seed beetle Caryedon serratus ( Coleoptera : Chrysomelidae ): new insights from molecular and ecological data. International Journal of Tropical Insect Science, 28(3), 168–177. Septiningsih, E. M., Prasetiyono, J., Lubis, E., Tai, T. H., Tjubaryat, T., Moeljopawiro, S., & McCouch, S. R. (2003). Identification of quantitative trait loci for yield and yield components in an advanced backcross population derived from the Oryza sativa variety IR64 and the wild relative O. rufipogon. Theorical and Applied Genetics, 107, 1419–1432. Shakiba, E., & Eizenga, G. (2014). Unraveling the Secrets of Rice Wild Species. Rice - Germplasm, Genetics and Improvement, 1–58. Shirasawa, K., Bertioli, D. J., Varshney, R. K., Moretzsohn, M. C., Leal-Bertioli, S. C. M., Thudi, M., … Isobe, S. (2013). Integrated consensus map of cultivated peanut and wild relatives reveals structures of the a and b genomes of arachis and divergence of the legume genomes. DNA Research, 20, 173–184. 112 University of Ghana http://ugspace.ug.edu.gh Shirasawa, K., Koilkonda, P., Aoki, K., Hirakawa, H., Tabata, S., Watanabe, M., … Isobe, S. (2012). In silico polymorphism analysis for the development of simple sequence repeat and transposon markers and construction of linkage map in cultivated peanut. BMC Medical Research Methodology, 12(80). Shoba, D., Manivannan, N., Vindhiyavarman, P., & Nigam, S. N. (2012). SSR markers associated for late leaf spot disease resistance by bulked segregant analysis in groundnut (Arachis hypogaea L .). Euphytica, 188, 265–272. Siddique, M., Malik, M.F.A. & Awan, S.I. (2006). Genetic Divergence , Association and Performance Evaluation of Different Genotypes of Mungbean (Vigna radiata). International Journal of Agriculture & Biology, 8(6), 793–795. Simpson, C., & Smith, O. D. (1975). Registration of Tamnut 74 peanut. Crop Science, 15, 603– 604. Simpson, C. E. (1991). Pathways for Introgression of Pest Resistance into Arachis hypogaea L. Peanut Science, 18(1), 22–26. Simpson, C. E., & Starr, J. L. (2001). Registration of “COAN” Peanut. Crop Science, 41, 918. Simpson, C. E., Starr, J. L., Baring, M. R., Burow, M. D., & Cason, J. M. (2013). Registration of “Webb.” Peanut. Journal. Plant Registration, 7, 265–268. Simpson, C. E., Starr, J. L., Church, G. T., Burow, M. D., & Paterson, A. H. (2003). Registration of “NemaTAM” peanut. (Registrations Of Cultivars). Crop Science, 43(4), 1561-1562. Simpson, E. (2001). Use of Wild Arachis Species/Introgression of Genes into A. hypogaea L. Peanut Science, 28(2), 114–116. 113 University of Ghana http://ugspace.ug.edu.gh Sinclair, T. R. (2011). Challenges in breeding for yield increase for drought. Trends in Plant Science, 16(6), 289–293. Singh, A. K., & Simpson, C. E. (1994). Biosystematics and genetic resources. In The Groundnut Crop, 96–137. Netherlands: Springer. Smartt J, S. H. (1982). Speciation and cytogenetics in Arachis. In Peanut Science and Technology, American Peanut Research and Education Society. Pattee HE, 21–49. Yoakum. Smartt, J., & Stalker, H. T. (1982). Speciation and cytogenetics in Arachis. In: Pattee HE, Young CT (eds) Peanut science and technology. American Peanut Research and Education Society, Yoakum/TX, 21- 49. Songsri, P., Jogloy, S., Kesmala, T., Vorasoot, N., Akkasaeng, C., Patanothai, A., & Holbrook, C., C. (2008). Heritability of Drought Resistance Traits and Correlation of Drought Resistance and Agronomic Traits in Peanut. Crop Science, 48(6), 2245–2253. Stalker, H. T. (1997). Peanut (Arachis hypogaea L.). Field Crop Research, 53, 205–217. Stalker, H. T., & Dalmacio, R. D. (1986). Karyotype analysis and relationships among varieties of Arachis hypogagea L. Cytologia, 51(79), 617–629. Stalker, H. T., Phillips, T. D., Murphy, J. P., & Jones, T. M. (1994). Variation of isozyme patterns among Arachis species. Theoretical and Applied Genetics, 87(6), 746–755. Stalker, H. T., & Simpson, C. E. (1995). Germplasm resources in Arachis. In H. E. Pattee & H. T. Stalker (Eds.) . In Advances in peanut science. 14–53. Stillwater: Okla. Stalker, H. T., Wynne, J. C., & Company, M. (1979). Variation hypogaea in progenies diploid of 114 University of Ghana http://ugspace.ug.edu.gh an Arachis species. Euphytica, 28(701), 675–684. Subrahmanyam, P., Moss, J. P., & Rao, V. R. (1983). Resistance to peanut rust in wild Arachis species. Plant Disease, 67 (2), 209-212. Sylla, F. (2016). Oilseeds and Products Annual 2016. USDA Foreign Agricultural Service. Isleib, T.G., Wynne, J.C., & Nigam, S.N. (1994). Groundnut breeding. In The Groundnut Crop, 552–623. Tanksley, S., & Nelson, J. (1996). Advanced backcross QTL analysis: a method for the simultaneous discovery and transfer of valuable QTLs from unadapted germplasm into elite breeding lines. Theoretical and Applied Genetics, 92, 191–203. Tanksley, S. D., Grandillo, S., Fulton, T. M., Zamir, D., Eshed, Y., Petiard, V., … Beck-Bunn, T. (1996). Advanced backcross QTL analysis in a cross between an elite processing line of tomato and its wild relative L. pimpinellifolium. Theoretical and Applied Genetics, 92, 213–224. Tanksley, S. D., & McCouch, S. R. (1997). Seed Banks and Molecular Maps: Unlocking Genetic Potential from the Wild. Science, 277 (5329), 1063–1066. Tester, M., & Langridge, P. (2010). Breeding Technologies to Increase. Science, 327, 818–822. Traoré, V. S., Néya, B. J., Camara, M., Gracen, V., Offei, S. K., & Traoré, O. (2015). Farmers’ Perception and Impact of Rice Yellow Mottle Disease on Rice Yields in Burkina Faso. Agricultural Sciences, 06(09), 943–952. Trouche, G., Lançon, J., Aguirre Acuña, S., Castro Briones, B., & Thomas, G. (2011). 115 University of Ghana http://ugspace.ug.edu.gh Comparing decentralized participatory breeding with on-station conventional sorghum breeding in Nicaragua: I.Agronomic performance. Field Crops Research, 121, 19–28. Tseng, Y. C., Tillman, B. L., Peng, Z., & Wang, J. P. (2016). Identification of major QTLs underlying tomato spotted wilt virus resistance in peanut cultivar Florida-EPTM ‘113’. BMC Genetics, 17:128. Upadhyaya, H. D. (2005). Variability for drought resistance related traits in the mini core collection of peanut. Crop Science, 45(2003), 1432–1440. Valls, J.F.M. (2005). New species of Arachis (Leguminosae) from Brazil, Paraguay and Bolivia. Bonplandia, 14, 35–63. van Heerwaarden, J., Doebley, J., Briggs, W. H., Glaubitz, J. C., Goodman, M. M., de Jesus Sanchez Gonzalez, J., & Ross-Ibarra, J. (2011). Genetic signals of origin, spread, and introgression in a large sample of maize landraces. Proceedings of the National Academy of Sciences, 108(3), 1088–1092. Varman, P. ., Viswanatha, K. ., Nadaf, H. ., Konda, C. ., & Nigam, S. . (2012). Enhancing Groundnut Productivity and Production in India. In Four Seasons of learning and Engaging Smallholder Farmers-Progress of Phase 1 (Abate., pp. 15–32). India: International Crops Research Institute for the Semi-Arid Tropics. Varshney, R. K., Bertioli, D. J., Moretzsohn, M. C., Vadez, V., Krishnamurthy, L., Aruna, R., … Hoisington, D. A. (2009). The first SSR-based genetic linkage map for cultivated groundnut (Arachis hypogaea L.). Theoretical and Applied Genetics, 118(4), 729–739. Varshney, R. K., Pandey, M. K., Janila, P., Nigam, S. N., Sudini, H., Gowda, M. V. C., … Nagesh, P. (2014). Marker assisted introgression of a QTL region to improve rust resistance 116 University of Ghana http://ugspace.ug.edu.gh in three elite and popular varieties of peanut (Arachis hypogaea L.). Theoretical and Applied Genetics, 127(8), 1771–1781. Varshney, R. K., Ribaut, J. M., Buckler, E. S., Tuberosa, R., Rafalski, A. J., & Langridge, P. (2012). Can genomics boost productivity of orphan crops? Nature Biotechnology, 30(12), 1172–1176. Vidavski, F., Czosnek, H., Gazit, S., Levy, D., & Lapidot, M. (2008). Pyramiding of genes conferring resistance to Tomato yellow leaf curl virus from different wild tomato species. Plant Breeding, 127(6), 625–631. Korff, M.V., Wang, H., Léon, J., & Pillen, K. (2006). AB-QTL analysis in spring barley: II. Detection of favourable exotic alleles for agronomic traits introgressed from wild barley (H. vulgare ssp. spontaneum). Theoretical and Applied Genetics, 112(7), 1221–1231. Wang, P., Xing, Y., Li, Z., & Yu, S. (2012). Improving rice yield and quality by QTL pyramiding. Molecular Breeding, 29(4), 903–913. Weeden, N. F. (1989). Applications of isozymes in plant breeding. In J. J (Ed.), Plant Breeding Reviews 6, 11–54. Portland: Timber Press. Witcombe, J. R., Joshi, K. D., Gyawali, S., Musa, A. M., Johansen, C., Virk, D. S., & Sthapit, B. R. (2005). Participatory Plant Breeding Is Better Described As Highly Client-Oriented Plant Breeding. I. Four Indicators of Client-Orientation in Plant Breeding. Bangladesh and International Plant Genetic Resources Institute (IPGRI-APO), 41(10), 299–319. Xu, Y., Zhu, L., Xiao, J., Huang, N., & McCouch, S. R. (1997). Chromosomal regions associated with segregation distortion of molecular markers in F2, backcross, doubled haploid, and recombinant inbred populations in rice (Oryza sativa L.). Molecular and General Genetics, 117 University of Ghana http://ugspace.ug.edu.gh 253(5), 535–545. Yan, W., & Frégeau-Reid, J. (2008). Breeding Line Selection Based on Multiple Traits. Crop Science, 48(2), 417–423. Yan, W., Kang, M. S., Ma, B., Woods, S., & Cornelius, P. L. (2007). GGE biplot vs. AMMI analysis of genotype-by-environment data. Crop Science, 47(2), 643–655. Zamir, D. (2001). Improving plant breeding with exotic genetic libraries. Nature Reviews. Genetics, 2(12), 983–989. Zhang, H., Mittal, N., Leamy, L. J., Barazani, O., & Song, B. H. (2017). Back into the wild— Apply untapped genetic diversity of wild relatives for crop improvement. Evolutionary Applications, 10(1), 5–24. Zhang, T., Yuan, Y., Yu, J., Guo, W., & Kohel, R. J. (2003). Molecular tagging of a major QTL for fiber strength in Upland cotton and its marker-assisted selection. Theoretical and Applied Genetics, 106, 262–268. Zhang, X., Zhang, J., He, X., Wang, Y., Ma, X., & Yin, D. (2017). Genome-Wide Association Study of Major Agronomic Traits Related to Domestication in Peanut. Frontiers in Plant Science, 8(1611). Zhou, X., Xia, Y., Ren, X., Chen, Y., Huang, L., Huang, S., … Jiang, H. (2014). Construction of a SNP-based genetic linkage map in cultivated peanut based on large scale marker development using next-generation double-digest restriction-site-associated DNA sequencing (ddRADseq). BMC Genomics, 15(1), 351. 118 University of Ghana http://ugspace.ug.edu.gh APPENDICES Table 4.1: Selected genotypes based on IAS in each site WW WS genotypes PPW HW IAS genotypes PPW HW IAS B7_25_7_7 9.5 70.5 13.10 B12_14_2_2 12.4 39.2 14.40 B7_26_6_3 39.4 51.1 12.56 B7_32_6_6 18.8 43.5 13.72 B12_14_2_5 18.4 106.6 11.48 B7_23_10_3 20.3 52.5 12.37 B7_30_9_1 13.1 89.6 11.37 B7_32_7_5 17.0 36.3 12.27 B7_27_5_7 14.3 76.1 11.36 B7_20_3_1 17.4 47.7 11.79 B7_23_10_3 26.1 64.7 11.18 B7_32_8_1 22.7 17.2 11.21 B7_32_7_5 27.3 72.9 11.15 B7_27_5_1 11.4 32.1 10.60 B7_32_6_3 20.6 45.6 10.73 B7_29_7_7 16.4 36.3 10.24 B7_21_14_3 18.9 34.3 10.34 B7_22_7_2 19.8 33.8 9.39 B7_26_4_8 26.9 37.6 9.95 B7_30_6_4 31.1 41.4 9.33 B7_32_8_1 25.2 32.4 9.36 B7_24_13_9 18.4 38.5 9.27 B7_27_13_3 11.9 76.7 9.25 B7_33_7_2 17.3 15.1 8.81 B7_23_16_2 20.2 68.3 8.80 B7_30_9_3 18.0 32.8 8.47 B7_20_3_1 16.9 65.3 8.39 B12_14_2_7 21.5 53.9 8.46 B12_11_18_6 9.6 79.1 8.21 B7_23_10_1 9.9 32.4 8.04 B7_30_3_4 25.7 55.8 8.17 B7_26_4_9 5.2 35.9 7.92 B7_27_16_6 20.6 55.4 8.15 B7_26_6_3 13.8 40.8 7.73 B7_21_17_1 39.0 80.6 7.87 B7_22_12_1 22.8 39.5 7.58 B7_24_13_9 25.5 43.5 7.81 B7_30_9_1 8.7 69.9 7.51 B7_21_2_8 13.9 42.9 7.69 B12_11_18_3 15.0 37.5 7.46 B7_21_2_1 23.6 82.9 7.53 B7_26_16_1 24.9 44.3 7.42 B7_24_5_4 21.9 11.8 7.26 B7_25_22_5 19.7 32.3 7.23 B7_33_7_1 11.8 133.6 7.21 B7_28_14_1 15.9 43.6 7.11 B7_30_6_5 24.4 4.8 6.97 B7_30_3_5 14.5 52.3 6.92 B7_28_14_1 20.4 42.6 6.96 B7_27_10_5 11.9 40.2 6.84 B7_25_4_8 27.5 23.4 6.74 B12_11_18_6 12.9 49.4 6.67 B7_23_10_8 12.8 89.7 6.67 B7_30_6_5 25.7 22.0 6.64 B7_32_8_2 34.9 42.6 6.32 B7_30_9_8 3.9 60.8 6.50 B7_33_7_2 27.0 28.0 6.19 B7_21_17_1 29.2 44.4 6.35 B7_21_14_6 15.1 34.3 6.15 B7_33_7_4 8.3 19.9 6.00 B7_32_2_1 16.3 49.9 6.01 B7_26_16_4 26.5 40.7 5.97 B7_24_17_3 30.9 71.9 5.99 B12_14_9_5 18.3 47.0 5.75 B12_11_18_5 24.8 24.7 5.90 B7_27_5_7 18.2 31.2 5.73 B7_32_7_4 34.5 60.1 5.88 B7_32_6_3 10.9 26.0 5.63 B7_20_10_2 8.1 45.1 5.71 B12_14_2_5 12.0 58.3 5.37 B12_14_2_2 11.1 35.1 5.66 B7_27_13_3 10.5 44.7 5.05 119 University of Ghana http://ugspace.ug.edu.gh B7_25_4_6 32.4 54.0 5.26 B7_25_22_6 20.4 36.9 4.85 B7_26_16_4 35.4 65.5 5.20 B7_24_17_3 27.9 42.7 4.85 B12_17_17_6 27.9 66.8 5.15 B7_25_7_6 13.9 64.3 4.85 B7_22_12_1 24.6 54.3 5.13 B7_33_7_1 13.4 50.6 4.84 B7_27_16_3 8.4 27.1 5.10 B12_14_9_1 15.3 34.0 4.79 B7_24_5_7 2.4 56.9 4.89 B7_23_16_3 19.0 20.0 4.74 B7_29_14_4 19.7 59.1 4.86 B7_24_8_1 16.2 41.4 4.67 B7_27_16_7 27.0 70.6 4.78 B7_27_16_9 34.2 29.6 4.60 B7_27_13_10 20.2 61.3 4.74 B7_32_8_3 15.0 52.4 4.52 B12_14_2_7 21.2 114.9 4.73 B12_14_9_4 14.5 30.7 4.27 B7_24_8_8 28.8 51.6 4.71 B7_23_10_8 9.2 47.4 4.26 B7_29_7_7 24.0 91.5 4.67 B7_20_5_5 28.4 25.1 4.16 B7_20_5_9 34.3 86.0 4.66 B12_17_17_3 20.6 28.8 4.12 B7_23_16_3 21.0 45.4 4.51 B7_21_2_8 11.4 43.3 3.83 PPW: plant pod weight; HW: haulm weight Table 4.1: Selected genotypes based on IAS in each site (continued) N16 S16 genotypes PPW HW IAS genotypes PPW HW IAS B7_23_10_3 22.1 36.0 15.38 B7_32_8_1 30.6 74.6 14.46 B12_14_2_2 29.1 67.3 15.13 B7_32_7_5 23.8 57.5 13.11 B7_26_4_8 17.2 12.4 14.97 B12_14_2_5 19.5 63.6 13.09 B7_20_3_1 33.6 47.8 12.17 B7_23_16_4 16.2 53.6 12.26 B7_22_12_1 31.5 65.2 11.32 B7_26_4_8 26.9 59.0 12.04 B7_25_22_6 24.1 33.8 11.14 B7_20_5_5 24.4 47.5 11.41 B7_27_16_7 29.8 32.6 10.87 B7_21_14_3 21.9 61.9 11.17 B7_23_16_4 18.9 40.5 10.30 B7_32_2_3 27.6 64.0 10.98 B7_21_2_1 28.5 50.7 10.23 B7_27_16_3 21.3 68.5 10.11 B7_25_7_7 4.9 25.1 9.98 B12_17_17_6 38.7 64.2 10.03 B12_11_18_6 23.7 31.0 9.88 B7_21_2_1 18.4 58.7 9.82 B7_20_5_5 40.7 30.9 9.77 B7_30_6_5 14.0 30.3 9.74 B7_27_3_1 20.5 21.4 9.61 B7_20_10_10 10.1 52.7 9.56 B12_14_9_4 35.3 63.3 9.23 B7_25_7_7 10.7 30.4 9.46 B7_23_10_8 25.0 35.7 9.10 B7_24_8_8 24.1 52.5 9.42 B7_28_14_1 24.9 33.1 9.05 B7_30_6_6 36.0 83.1 8.98 B7_25_22_8 20.7 21.2 8.92 B7_27_5_7 25.3 43.6 8.76 B7_32_8_1 46.0 31.9 8.43 B7_33_7_2 29.1 51.0 8.43 B7_32_6_6 17.5 31.0 8.21 B7_30_9_3 21.4 60.4 8.42 B7_20_10_10 9.3 58.9 8.21 B7_21_2_6 16.8 45.6 8.37 B12_14_2_5 26.4 55.7 8.04 B7_21_17_1 22.6 50.3 8.33 B12_11_18_3 27.9 13.5 7.97 B12_14_9_4 23.6 49.5 7.95 B7_27_10_5 21.5 26.6 7.90 B7_29_7_7 16.6 45.9 7.77 120 University of Ghana http://ugspace.ug.edu.gh B7_26_16_4 32.2 50.0 7.87 B7_24_17_3 25.2 69.4 7.72 B7_20_10_7 18.9 33.7 7.83 B7_28_14_1 23.9 73.0 7.65 B7_32_7_5 14.9 33.0 7.83 B7_23_10_3 26.0 61.3 7.61 B7_24_5_2 27.1 39.9 7.81 B7_20_3_1 17.9 73.9 7.60 B12_14_2_3 26.7 40.6 7.61 B7_30_3_5 24.3 71.0 7.31 B7_21_2_8 21.9 19.2 7.09 B7_33_7_1 12.1 86.2 7.03 B7_25_19_1 16.9 31.9 7.01 B7_27_10_2 15.2 63.2 6.97 B7_30_9_8 22.0 49.9 6.84 B7_20_10_2 12.4 64.3 6.91 B7_25_22_3 14.3 60.1 6.68 B7_21_14_1 27.5 53.1 6.85 B7_21_14_6 6.2 36.0 6.63 B7_22_7_2 23.4 46.3 6.69 B12_11_12_4 15.1 11.4 6.61 B7_32_8_2 21.0 64.0 6.65 B12_17_17_6 24.6 36.9 6.44 B7_27_3_1 16.3 38.4 6.60 B7_21_14_3 25.8 30.4 6.41 B12_14_2_2 26.6 53.2 6.58 B7_24_8_1 21.6 40.9 6.32 B7_32_8_3 10.3 72.8 6.51 B7_32_8_2 33.6 58.8 6.27 B7_25_4_6 26.6 34.3 6.43 B7_25_4_5 19.4 25.6 6.12 B7_21_2_8 18.0 57.7 6.31 B7_27_16_9 21.8 19.0 6.07 B12_14_2_7 19.4 61.4 6.25 B7_27_13_10 15.5 31.5 5.89 B12_11_18_5 31.3 57.1 6.03 B7_30_6_5 9.0 0.0 5.61 B7_23_10_8 15.1 50.0 5.94 B7_27_5_2 15.1 25.2 4.96 B12_11_18_2 18.8 46.9 5.78 B7_30_6_4 31.2 40.1 4.83 B7_32_7_1 23.6 44.1 5.73 Fleur11 26.9 34.8 4.77 B7_32_6_6 16.1 42.2 5.66 B7_25_7_6 15.1 22.0 4.72 B7_25_13_6 25.5 70.7 5.64 B12_14_2_1 21.6 27.7 4.71 B12_11_18_4 20.5 55.2 5.55 B7_27_5_7 16.7 65.5 4.63 B7_32_7_4 36.2 40.0 5.50 B7_20_5_9 35.8 40.3 4.54 B7_23_16_2 22.9 76.9 5.45 B12_11_18_2 42.7 8.2 4.46 Fleur11 23.5 56.7 5.30 121