GENETIC IMPROVEMENT OF ROOT YIELD AND NUTRITIONAL QUALITY OF CASSAVA (Manihot esculenta Crantz) IN SIERRA LEONE By ISATA KAMANDA (10496574) 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 (CBAS) UNIVERSITY OF GHANA LEGON DECEMBER, 2017 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. .................................................. Isata Kamanda (Student) Prof. Essie. T. Blay (Supervisor) Prof. Isaac. K. Asante (Supervisor) Dr. Agyemang. Danquah (Supervisor) Dr. Beatrice. E. Ifie (Supervisor) Dr. James. B. A Whyte (Supervisor) Dr. Peter. Kulakow (Supervisor) ii ABSTRACT Cassava storage root is a major staple. However, the tuber is poor in nutrients especially micronutrients. The study was conducted to improve the nutritional status of cassava with farmers’ preferred traits in Sierra Leone. The specific objectives were i) to assess adoption challenges, perception and preferences for provitamin A cassava among cassava value chain actors in Sierra Leone. ii) to estimate genetic diversity within provitamin A cassava germplasm using morphological, molecular tools and i-check device for establishing a collection in Sierra Leone. iii) to determine performance and stability of total carotene content and dry matter of selected provitamin A cassava accessions across environments.vi) to characterize F1 progenies for total carotenoid, iron and zinc and protein content using biochemical tool. v) to estimate the combining ability of 12 cassava parents and their F1 progenies for mealiness, dry matter, number of roots and fresh root yield. A participatory rural appraisal (PRA) was conducted in Bombali, Kailahun and Moyamba districts, to identify farmers’ and consumers’ adoption challenges, perceptions and preferences for provitamin. High production cost, low yield, scarcity of planting materials, high cost of fertilizers and agro-inputs, drudgery in peeling and processing and limited access to micro finance loan schemes, were identified as major challenges for provitamin A cassava adoption. The respondents show willingness to accept and adopt provitamin A cassava due to its perceived nutritional quality. A total of 188 cassava accessions cultivated in the southern part of Sierra Leone were assessed using molecular tools and the i-check device. The Cassava accessions were grouped into eight distinct clusters based on the morphological data while they grouped into nine distinct clusters based on the molecular analysis. A significant positive correlation was found between the morphological and molecular data sets (r = 0.104; p < 0.034) but the correlation was rather weak. Thirty provitamin A accessions with higher total iii carotenoid contents were selected to form a collection. The collection evaluated in 3 environments for the GGE biplot analyses for dry matter content (DMC) and total carotenoid content (TCC) showed significant variation among the Genotypes, Environments and their interaction. Genotypes TR-1182 and TR-1313 had the highest performance for DMC and TCC. Njala was identified as an ideal environment for selecting superior genotype for total carotenoid content and Pendembu as ideal for dry matter content. The performance of 868 F progenies 1 (obtained from five crosses involving eight genetically diverse parents) was evaluated for selection of varieties with increased level of micro nutrients. F1 progeny 13 and 33 from cross IITA-TMS-IBA 120004 x IITA-TMS-IBA 120003 recorded the highest (28.0 µg-1) and lowest (6.0 µg -1) values for total carotenoid content with a grand genotypic mean of 14.7 µg -1. F1 progeny 41 and 12 from cross IITA-TMS-IBA 088693x IITA-TMS-IBA 088747 recorded the highest (8.1%) and lowest (4.2%) crude protein content with a grand percentage mean of 5.4%. F1 progeny from cross IITA-TMS-IBA 96/1165 x IITA –TMS-IBA 011368 recorded the iron content with a ranged 45.0 ppm to 59.2 ppm and harvested progeny with the grand mean 12.6 ppm. F1 progeny from cross MM96/81791 x IITA-TMS-IBA 088747 had the highest zinc concentration ranging from 4.5 ppm to 17.7 ppm with a grand mean 8.5 ppm. Micro nutrients analysis on F1 progenies revealed that there is variation for quality traits in cassava. In addition, a 12 x12 diallel study revealed highly significant differences for dry matter, number of roots and fresh root yield at P < 0.01, and 0.05 respectively. General combining ability (GCA) variance was significant for fresh root yield while estimates of specific combining ability (SCA) variance was highly significant for dry matter and number of storage roots. The significant difference observed for most of the traits for the parents and their F1 progeny revealed that genetic diversity exists among the germplasm and Progenies from crosses IBA120004 x IBA961165, IBA 961165 iv x I088693 and IBA120004 x IBA961165 were the best combiners for number of storage roots, mealiness and fresh root yield. v DEDICATION To the glory of God and to my beloved father (Hon) Dr. B.M. Kamanda. My late mother and grandmother of blessed memories Madam Aminata Sheriff and Haja Mrs. Ada Sheriff vi ACKNOWLEDGEMENTS I owe my life to God and His son Jesus Christ for protecting me throughout my study. I am also indebted to our Lord and Savior Jesus Christ for His strange acts and works in my life. Unto Him alone be all the glory, as in the center of this all it’s only Him that I see. I wish to express my sincere thanks to my supervisors: Prof. E.T. Blay, Prof. I.K. Asante, Dr. Agyemang Danquah, Dr. Jim Whyte Dr. Beatrice. E. Ifie and Dr. Peter Kulakow for their patience, diligent guidance, constructive criticism and supervision. I am also grateful to Prof K. Ofori for the suggestions and supports as they were invaluable to the successful completion of my thesis. My gratitude goes to my sponsors. Thank you, West Africa Agricultural Productivity Program, Sierra Leone Agricultural Research Institute, International Institute of Tropical Agriculture, IDRC and West Africa Centre Crop Improvement for sponsoring my study. I want to render my sincere appreciation to the Director General, SLARI. Dr. Joseph Kargbo for creating an enabling environment for the completion of my research thesis. I appreciate the effort of my Center Director, Dr. Abdul Conteh. Director. Njala Agricultural Research Center for his assistance. I am indebted to IITA for providing me with the field and laboratory facilities for my research and their guidance. I appreciate Dr. Alfred Dixon, Dr. Elizabeth Parkes Dr. Robert Asiedu, Prof Michael Abberton, Dr. Maziya Dixon, Mr. Zomana Bamba, Dr. Ishmail Rabbi, Dr. Joseph Sherman-Kamara, Mrs. vii Mary Badejo and Mr. Peter Iluebbey for their benevolence, assistance, constructive critism, and supervision. I want to recognize the staff of Cassava breeding and Food science laboratory in IITA, Ibadan and Ubiaja stations, for their technical support. I wish to express my deepest heartfelt thanks to Mr. Moshood Bakare, Mr. Richard Olayiwola, Mr. Kinsley Akuwa, Mr. Bello and Mr. Adetoro for assisting me with data analysis and constructive editing. I am also grateful to the entire staff of the training unit at IITA. I am indebted to my loving parents. Hon. Dr. and Mrs. B.M. Kamanda for their kind words of encouragements, support and prayers. Special thanks go to my husband, (Abdul) for his, patience and understanding during my absence. I am also grateful to my family and siblings, Pastor. Charles. W. Dorkenoo, Mr. and Mrs. Cummings-John, Mr. and Mrs. Dowe Leigh, Mr. and Mrs. Jusufu Kamanda, Miss. Hawa Kamanda, Miss. Francylida Cummings John, Mr. Augustine Abu, Mrs. Martha Keita and Dr. Carol Markwei for your support and prayers. I sincerely appreciate the staff of Njala and Rokupru Agricultural Research Centers. Mr. Michael Benya, Mr. Sayo Sesay, Mr. Keiwoma Yilla, Mr. Suffian Manasaray, Mr. Osman Nabay, Mr. Emmanuel Henkeley and Mr. Ernest. G. Kamara for their support. viii I wish to express my profound gratitude to the drivers and technicians at the Njala Agricultural Research Center. Mr. Frank Kpandenyange, Mr. John. Lissa, Mr. Mustapha Kamara, Mr. Mamoud Sowe, Mrs Mary Dumboi and Mrs Massah Lissa for helping me out with my field and laboratory activities. I am indebted to Richmond Kyei for his tremendous support since my enrollment into the WACCI programme. Last but not the least to my colleagues in cohort 7, Sanatu, Kanfany, Olumide, Dien, Isatu, Kumba, Bathe, Chika, Kayondo, Solaymane, Lydia, and Ijeoma. It was good meeting you all. ix TABLE OF CONTENTS DECLARATION ............................................................................................................................ ii ABSTRACT ................................................................................................................................... iii DEDICATION ............................................................................................................................... vi ACKNOWLEDGEMENTS .......................................................................................................... vii TABLE OF CONTENTS ................................................................................................................ x LIST OF TABLES ...................................................................................................................... xvii LIST OF FIGURES ...................................................................................................................... xx LIST OF ABBREVIATIONS ..................................................................................................... xxii CHAPTER ONE ............................................................................................................................. 1 1.0 GENERAL INTRODUCTION ............................................................................................. 1 2.0 REVIEW OF LITERATURE ............................................................................................... 5 2.1 Importance of Cassava ...................................................................................................... 5 2.2 Cassava production in Sierra Leone ................................................................................. 5 2.3 Cassava production constraints in Sierra Leone ............................................................... 6 2.4 Agricultural origin and spread of cassava in Africa ......................................................... 7 2.4.1 Taxonomy of cassava ................................................................................................. 8 2.4.2 Root system ................................................................................................................ 9 2.4.3 Variability in colour and provitamin A of cassava storage root .............................. 10 2.5 Cassava growth and development ................................................................................... 12 2.6 Breeding for high beta-carotene content, iron, zinc and protein ............................... 13 2.6.1 Participatory plant breeding .................................................................................. 16 2.6.2 Progress in cassava improvement ............................................................................ 16 x 2.7 Genetic diversity in cassava ............................................................................................ 17 2.8 Morphological and Molecular Markers in cassava ................................................... 19 2.8.1 Morphological markers ..................................................................................... 19 2.8.2 Molecular markers ................................................................................................ 19 2.8.3 Single nucleotide polymorphism (SNP) .............................................................. 21 2.8.4 Morphological and molecular characterization of cassava ................................... 22 2.9 Combining Ability in cassava breeding ................................................................... 22 2.10 Genotype, environment and genotype by environment Interaction .......................... 24 CHAPTER THREE ...................................................................................................................... 27 3.0 ADOPTION CHALLENGES, PERCEPTION AND PREFERENCES FOR PROVITAMIN A CASSAVA AMONG THE CASSAVA VALUE CHAIN ACTORS IN SIERRA LEONE ...................................................................................................................... 27 3.1 Introduction ..................................................................................................................... 27 3.2. Materials and methods .............................................................................................. 29 3.2.1. Study area.......................................................................................................... 29 3.2.2 Sampling procedure .......................................................................................... 30 3.2.4 Data Collection ................................................................................................. 33 3.2.5 Data Analysis ........................................................................................................ 33 3.3 Results ....................................................................................................................... 34 3.3.1 Socio economic characteristics of respondents........................................................ 34 3.3.2 Average yearly farm size per household for cassava production ............................ 35 3.3.3 Cassava cultivation within 5 years in the study areas across the surveyed districts 36 3.3.4 Respondents’ average income in the study areas ..................................................... 37 xi 3.3.5 Adoption challenges of producers for provitamin A cassava .................................. 38 3.3.6 Farmer’s perception for provitamin A cassava across the survey districts .............. 39 3.3.7 Traders perception on cassava root trading across the surveyed districts ............... 42 3.3.8 Provitamin A cassava root trading in the study areas .............................................. 44 3.3.9. Processors perception and preference for cassava roots processing and awareness across the different districts ............................................................................. 45 3.3.10 Perception and preference for cassava roots processing/trading and awareness across the different districts .............................................................................................. 46 3.3.11 Perception and preference of consumers on consumption and awareness of provitamin-A cassava roots across the surveyed districts................................................. 47 3.3.12 Perception and preferences of consumers for Gari consumption across the three districts among value chain actors .................................................................................... 49 3.3.13 Respondents’ perception on the strengths, weakness, opportunities and threat for three districts among the cassava value chain actors ........................................................ 51 3.3.14 SWOT ranking for cassava processors across the three districts ........................... 52 3.3.15 SWOT ranking for cassava traders across the three districts ................................. 54 3.4 Discussion ....................................................................................................................... 56 3.5 Conclusion ................................................................................................................ 59 CHAPTER FOUR ......................................................................................................................... 60 4.0 DIVERSITY STUDIES OF PROVITAMIN A CASSAVA (Manihot esculenta Crantz) IN SIERRA LEONE. ..................................................................................................................... 60 4.1 Introduction ..................................................................................................................... 60 4.2 Materials and Methods .................................................................................................... 61 xii 4.2.1 Land preparation ............................................................................................... 61 4.2.2 Germplasm sources ........................................................................................... 62 4.2.3 Experimental Design ......................................................................................... 62 4.2.4 Molecular Characterization ............................................................................... 64 4.2.5 Data Collection ........................................................................................................ 66 4.2.6 Total carotene determination ................................................................................ 67 4.2.7 Data Analysis .................................................................................................... 68 4.3 Results ....................................................................................................................... 69 4.3.1 Descriptive statistics for the 188 cassava accessions ........................................ 69 4.3.2 Multiple Regression of yield on agro-morphological traits .............................. 71 4.3.3. Correlations among morphological traits of 188 cassava accessions ............... 71 4.3.4. Principal component analysis of yield and yield related traits of 188 cassava accessions .......................................................................................................................... 72 4.3.5. Cluster analysis of the accessions based on Ward’s minimum variance and total carotenoid content ............................................................................................................. 73 4.3.6 Clusters mean and standard deviation of provitamin A cassava accessions ..... 77 4.3.7 Summary statistics of genetic variation of the accessions using SNP markers 77 4.3.8. Cluster groupings of the 188 cassava accessions based on SNP markers. ....... 78 4.3.9 Thirty provitamin-A cassava accessions with varying levels of total carotenoid, yield and dry matter content based on morphological and molecular clustering Analyses. ........................................................................................................................................... 81 4.4 Discussion ....................................................................................................................... 82 4.5 Conclusion ................................................................................................................ 85 xiii CHAPTER FIVE .......................................................................................................................... 86 5.0 GENOTYPE BY ENVIRONMENT INTERACTION ANALYSIS OF PROVITAMIN A CASSAVA IN SIERRA LEONE ............................................................................................. 86 5.1 Introduction ..................................................................................................................... 86 5.2 Materials and Methods .................................................................................................... 88 5.2.1 Experimental materials ......................................................................................... 88 5.2.2 Experimental sites and design ............................................................................... 88 5.2.3 Planting .................................................................................................................... 88 5.2.3 Data collection ...................................................................................................... 90 5.2.4 Data Analysis ........................................................................................................ 91 5.3. Results ....................................................................................................................... 91 5.3.1. Analysis of variance for multilocational trials across 3 locations for dry matter and total carotenoid contents. ........................................................................................... 91 5.3.2 Performance of 30 provitamin A cassava evaluated across 3 test environments for dry matter content and total carotenoid content ................................................................ 92 5.3.3 Polygon view of GGE biplot for dry matter content (%)...................................... 94 5.3.4 Mean performance and stability of genotypes for dry matter content (%) across the 3 environments ............................................................................................................ 96 5.3.5 Discriminativeness vs representativeness for dry matter content (%)across the 3 environments ..................................................................................................................... 98 5.3.6 Winning Genotype and Mega Environment GGE biplot for Total Carotenoid Content ............................................................................................................................ 100 xiv 5.3.7 Mean Performance and Stability for Total Carotenoid Content across three Environments. ................................................................................................................. 102 5.3.8 Discriminativeness vs Representativeness for Total Carotene Content across three Environments. ................................................................................................................. 104 5.4 Discussion ............................................................................................................... 106 5.5 Conclusion .................................................................................................................... 108 CHAPTER SIX ........................................................................................................................... 110 6.0 PERFORMANCE OF F1 CROSSES FOR TOTAL CAROTENOID, PROTEIN, IRON AND ZINC CONTENTS OF STORAGE ROOTS ................................................................ 110 6.1 Introduction ................................................................................................................... 110 6.2 Materials and Methods ............................................................................................ 112 6.2.1. Development of F1 progenies.......................................................................... 112 6.3 Results ..................................................................................................................... 116 6.3.2 Descriptive analysis of F1 progenies ............................................................... 116 6.3.3 Variation in total carotenoid content in roots of F1 progeny .......................... 117 6.3.4 Mean protein content (%) in roots of F1 progeny .......................................... 119 6.3.5 Mean Iron content (ppm) in storage roots of F1 Progeny .............................. 120 6.3.6 Mean zinc content (ppm) in storage roots of F1 progeny ............................... 123 6.4. Discussion ............................................................................................................... 124 6.5 Conclusion .............................................................................................................. 126 CHAPTER SEVEN .................................................................................................................... 127 7.0. GENETIC STUDIES ON MEALINESS, DRY MATTER, ROOT NUMBER AND FRESH ROOT YIELD, IN CASSAVA (Manihot esculenta Crantz) .................................... 127 xv 7.1. Introduction .................................................................................................................. 127 7.2. Materials and Methods ............................................................................................ 130 7.2 1 Establishment of crossing block ......................................................................... 130 7.2.2 Seedling Nursery evaluated at Foya crop site ................................................. 132 7.2.3 Data Collection ............................................................................................... 133 7.2.4 Data analysis ................................................................................................... 134 7.3. Results .......................................................................................................................... 136 7.3.1 Mean performances of crosses for number of storage roots per plant and fresh root yield. 136 7.3.2 Mean performances of crosses for mealiness and dry matter content. ............... 138 7.3.3 Estimates of general combining effect (GCA) and specific combining effect (SCA) 140 7.3.4 Estimates of general combining ability (GCA) effects ................................... 140 7.3.5 Estimates of Specific combining ability (SCA) effects .................................. 141 7.4. Discussion ............................................................................................................... 145 7.5 Conclusion .................................................................................................................... 148 CHAPTER EIGHT ..................................................................................................................... 150 8.0 GENERAL CONCLUSIONS AND RECOMMENDATIONS.................................. 150 8.1 GENERAL CONCLUSIONS ................................................................................. 150 8.2 RECOMMENDATIONS ........................................................................................ 154 REFERENCES ........................................................................................................................... 156 APPENDIX ................................................................................................................................. 175 xvi LIST OF TABLES Table 3.1 Summary of systemic sampling methods, for the selection of respondents in the study areas…………………………….………….…………………………………………………… 32 Table 3.2 Socio economic characteristics of respondents across the three districts ….……….........................................................................................................................35 Table 3.3 Varietal preferences for provitamin-A cassava across the surveyed districts…….........................................................................................................................39 Table 3.4 Farmer’s perception of provitamin-A cassava across the surveyed districts…….................................................................................................................................41 Table 3.5 Cassava root trading across the three districts…............................................................................................................................44 Table 3.6 Cassava roots consumption and awareness across the surveyed districts………………………………………………………………………………..............….48 Table 3.7 Gari consumption across the three districts………………………………..........…….50 Table 3.8 Kendall SWOT ranking for cassava producers across three districts…………...............................................................................................................51 Table 3.9 Kendall SWOT ranking for cassava processors across the three districts………...................................................................................................................53 Table 3.10 Kendall SWOT ranking for cassava traders across the three districts……...………..55 Table 4.1 Germplasm/accession and their pedigree…………………………………...………...63 Table 4.2 Parameters evaluated at 1, 3, 6 and 9 MAP…….........………………………………..66 Table 4.3 Descriptive statistics of some morpho-agronomic traits of 188 cassava accessions…………………………………………………………………………………….…..69 Table 4.4 Multiple regression coefficients of yield on some agro-morphological traits of 188 cassava accessions……………………………………………………………………...………. 71 Table 4.5 Correlations among morphological traits of 188 cassava accessions……….………...72 Table 4.6 Principal component analysis of yield and yield related traits of 188 cassava accessions………………………………………………………………………………………...73 Table 4.7 188 cassava accessions grouped into 8 clusters based on total carotenoid content……………………………………………………………………………………………76 xvii Table 4.8 Number of accession/clusters, mean and standard deviation for provitamin A content……………………………….……………………………………………………..….....77 Table 4.9 Cluster groupings of the 188 cassava accessions based on SNP markers……....….....80 Table 4.10 Thirty provitamin-A cassava accessions with varying levels of total carotenoid, yield and dry matter content…………………………………………………………………...............82 Table 5.1 Climatic data of the three experimental sites…............................................................90 Table 5.2 Analysis of variance dry matter and total carotenoid contents…………………..….. 92 Table 5.3 Mean performance of 30 provitamin A cassava accessions evaluated across three environments………………………………………………………………………………... …..93 Table 6.1 Pedigree and characteristics of breeding lines…………………………………….... 112 Table 6.2 Number of F1 seeds generated, planted and (germinated) per family ……..……………………………………………………………………………………………113 Table 6.3 Descriptive statistics for nutritional quality of F1 progenies of root quality……….....................................................................................................................117 Table 6.4 Variation total carotenoid content of F1 progeny using the color chart and i-check device …………………………………………………………………………………………………..118 Table 6.5 Mean percent protein of F1 progeny (evaluated in seedling nursey trial in Njala) …………………………………………………………………………………………...……...119 Table 6.6 Mean iron concentration of F1 progeny..……………............…………………….…121 Table 6.7 Mean zinc concentration of F1 progeny………………………..………………….....124 Table 7.1 Twelve cassava genotypes used as parents for F1 progenies, their pedigree and important traits…...……………………………..………………………………………………131 Table 7.2 Seeds obtained from 12 x 12 diallel cross of cassava genotypes and number of seeds produced………………...……………………………………………………………………....132 Table 7.3 Phenotypic traits used in the characterization of cassava F1 progenies….……….....133 Table 7.4 Mean performances of crosses for number of roots per plant and fresh root yield………………………………………………………………………………...…………...137 Table 7.5 Mean performances of crosses for mealiness and dry matter content………….….…………………………………………………………………………....139 xviii Table 7.6 Mean squares of 12 x 12 diallel of measured traits in Njala………………………………………………………………………..………...…………140 Table 7.7 Estimate of General combining Ability of 12 cassava genotypes evaluated in Njala …………………………………………………………………………………………………..141 Table 7.8a Estimate of SCA for no of roots and fresh root yield of 66 diallel crosses evaluated in Njala……………………………………………...……………………………………….…….143 Table 7.8b Estimate of SCA for dry matter content and mealiness of 66 diallel crosses evaluated in Njala………………….....…………………………………………………………………... 144 xix LIST OF FIGURES Figure 3.1 Map of Sierra Leone showing surveyed regions………..……………………...….....30 Figure 3.2 Average yearly farm size per household for cassava production……………….........36 Figure 3.3 Cassava cultivation within 5 years in the study areas……………………………......37 Figure 3.4 Respondents’ average income in the study areas……………………………….........38 Figure 3.5 Provitamin A cassava root trading in the study areas………………………….……..44 Figure 3.6 Percentage of respondents processing different cassava products in the study area ……………………………………………………………………………………………………45 Figure 3.7 Respondents interest in provitamin A cassava gari production/trading in the study areas ..……………………………………………………………………………………………46 Figure 4.1 Dendrogram of 188 cassava accessions based on morpho-agronomic traits using Ward’s minimum Variance………………….…………………………………………………...75 Figure 4.2 Dendrogram of 188 cassava accessions based on SNP Markers…………...………..79 Figure 5.1 Map of Sierra Leone; showing trial site ………….……..…………………………..89 Figure 5.2 GGE biplot showing the best genotype for each of the 3 mega environments for dry matter content (%)………………………………..…………..…………………………………..95 Figure 5.3 Stability of dry matter content (%) of 30 Provitamin A cassava genotypes evaluated across three environments ……………………………………………………………………….97 Figure 5.4 Discriminativeness and representativeness of the three environments for dry matter content (%) determination of 30 Provitamin A cassava genotypes ……………….………….....99 Figure 5.5 GGE biplot for total carotenoid content showing the best genotype among the 30 Provitamin A cassava genotypes in each Mega Environment…………………………..…..….101 xx Figure 5.6 Stability for total carotene content of 30 Provitamin- A cassava genotypes evaluated across three Environments. …………………………..………………….…..………………....103 Figure 5.7 Discriminativeness vs representativeness for total carotenoid content of the 30 accessions in the 3 environments…………………..……………………………….……..……105 xxi LIST OF ABBREVIATIONS ACMV African Cassava Mosaic Virus AFLP Amplified Fragment Length Polymorphism ANOVA Analysis of Variance CBB Cassava Bacterial Blight CBSD Cassava Brown Streak Disease CIAT International Centre for Tropical Agriculture CMD Cassava Mosaic Disease CMGs Cassava Mosaic Gemini viruses DMC Dry Matter Content DNA Deoxyribonucleic Acid ESTs Expressed Sequence Tags FAO Food and Agriculture Organization FSRY Fresh Storage Root Yield GCA General Combining Ability GDP Gross Domestic Product GEI Genotype and Environment Interaction IFAD International Fund for Agricultural Development IITA International Institute of Tropical Agriculture MET Multi Environmental Trials NARC Njala Agricultural Research Center NGOs Non-Governmental Organizations PCA Principal Component Analysis xxii QTL Quantitative Trait Loci PIC Polymorphic Information Content PRA Participatory Rural Appraisal RAPD Random Amplified Polymorphic DNA RFLP Restriction Fragment Length Polymorphism PVAC Provitamin A Cassava SAS Statistical Analysis System SCA Specific Combining Ability SCARs Sequence Characterized Amplified regions SGD Specific Gravity Determination SLARI Sierra Leone Agricultural Research Institute SSR Simple Sequence Repeats SNP Single Nucleotide Polymorphism SWOT Strengths, Weaknesses, Opportunities and Threats TCC Total Carotenoid Content UNICEF United Nations Children's Fund (formerly United Nations International Children's Emergency Fund) WAAPP West Africa Agricultural Productivity Program WACCI West Africa Centre Crop Improvement WHO World Health Organization YFC Yellow Flesh Cassava xxiii CHAPTER ONE 1.0 GENERAL INTRODUCTION Cassava (Manihot esculenta Crantz) is an important food crop and source of calories for more than 900 million people in the tropics and sub tropics (FAO, 2014; Nassar, 2006). It is a staple crop for more than 300 million persons in sub-Saharan Africa (Ihemere et al., 2011). The storage root of cassava is the most important source of dietary calories in the tropics after maize and rice. Cassava plays a central role in food and economic security for small-holder farmers and holds an unrealized potential as a cash and food crop as it widely grown and well-adapted into the farming system. The crop is grown by most smallholder farmers due to its ability to yield better than other staple food crops under conditions of extended drought and poor soils (Esuma et al., 2016, Ceballos et al., 2011 and El Sharkawy, 2007. It is a perennial crop native to South America and was among the first crops to be domesticated. It is believed to have been brought to Africa in the 17th century (Okigbo, 1980; Allem, 2002). Cassava cultivation is widespread in Africa, spanning across 40 countries. Generally, cassava serves five purposes: famine reserve crop, rural food staple, cash crop for urban consumption, industrial raw material, and foreign exchange earner (Nweke et al., 2004). In Sierra Leone cassava is the second major staple crop after rice. It has the potential of contributing to the achievement of the millennium development goal of poverty reduction as well as meeting requirement of the pillar II in the poverty reduction strategy paper in Sierra Leone (SLARI Strategic plan, 2012). 1 The agricultural sector in Sierra Leone employs more than 70% of the rural population and contributes 50% of the country’s Gross Domestic Product (GDP) (SLARI Strategic plan, 2012). All parts of the cassava plant are used either as food, animal feed or an industrial feedstock. However, the storage root portion of the plant is mostly used. Notably, the crop is deficient in essential micronutrients (vitamin A, iron, zinc and protein) hence presents a major health problem of nutritional insecurity in communities that heavily rely on cassava (Rice et al., 2004; Gichuki et al., 2010; and Esuma et al., 2016). Over seven billion people are afflicted with micronutrient malnutrition and the numbers are increasing (Mason and Garcia, 1993 and Welch et al., 1997). In sub-Saharan Africa, micronutrient malnutrition is a major public health problem (WHO, 2002). This type of malnutrition results primarily from use of diets deficient in essential vitamins and minerals like iron and zinc. Known as hidden hunger, micronutrient malnutrition can exist even when poor people have enough to eat, but lacks essentials nutrients (FAO, 2003). Diets poor in micronutrients cause illness, blindness, premature death, reduced productivity and impaired mental development. (UNICEF, 2004). The importance of micronutrients cannot be overemphasized, as micronutrient deficiencies remain a huge problem among young children and women in Africa and particularly in Sierra Leone. Micronutrients are essential for the normal functioning of the immune system, growth and development, maintenance of epithelial cellular integrity and for reproduction (Chavez et al., 2007). In sub-Saharan Africa micronutrient deficiencies, are estimated to cause economic losses in productivity of more than U.S. $2.3 billion (UNICEF, 2004 and Nganga et al., 2010). Although there have been efforts to alleviate 2 micronutrient malnutrition, the health problem remains highly prevalent because of inadequate or poor diet intake, which has triggered a call for concern. Different strategies have been adopted in Sierra Leone to reduce micro nutrient deficiencies; these include dietary diversification, food fortification, and supplementation. These strategies have not always proven to be very successful remedies. Irrespective of this, the promising and sustainable strategy to address the micronutrient deficiencies is through food biofortification in Sierra Leone. Biofortification is the process of breeding vitamins and nutrients into food crops, either by using natural breeding techniques or biofortification taking advantage of the genetic variability available in related species and varieties having specific nutrient traits that can be used for breeding (Welch and Graham, 2005; Chávez et al., 2005 and Ceballos et al., 2013). One approach towards the development, use and conservation of biofortified foods involves the screening of the existing plant genetic resources for germplasm with the required nutrient resources and desired quality traits (Harvest plus, 2002). Screening efficiencies can be enhanced using molecular tools to identify genotypes having desirable traits (Njoku, 2012) without the confounding effect of the environment, reduce size of breeding populations, length of time and evaluation cost (FAO, 2004) required for breeding the desired traits. Studies to evaluate the food quality, nutritional and genetic diversity of the cassava germplasm to ameliorate hunger and malnutrition have not previously been carried out in Sierra Leone. Characterization and evaluation of accessions within the cassava germplasm resources in Sierra Leone as well as screening for genetic variation in the germplasm is important in devising 3 optimum management strategies for sustainable utilization and conservation of the resource. As a starting point in breeding for higher micronutrient contents in cassava, with farmers preferred traits like dry matter and fresh root yield, the germplasm should be screened for useful quality traits that may be used or introduced into the breeding program for the development of breeding populations. Regrettably, there is no evidence of adequate variability in protein, iron and zinc content of roots and leaves of cassava varieties in Sierra Leone which can serve as basis for improvement in the nutrient status of cassava Thus, the main thrust of this study seeks to identify and select local and exotic cassava accession in Sierra Leone to establish a collection and develop superior cassava germplasm with improved yield and nutritional qualities. This formed the basis for the specific objectives which were; 1) To assess adoption challenges, perception and preferences for provitamin A cassava among cassava value chain actors in Sierra Leone. 2) To estimate genetic diversity within provitamin A cassava germplasm using morphological, molecular tools and iCheck device. 3) To evaluate the levels and stability of total carotene content and dry matter of selected provitamin A cassava accessions across environments. 4) To characterize F1 progenies for total carotenoid, iron, zinc and protein contents. 5) To estimate the combining ability of 12 cassava parents and their F1 progenies for mealiness, dry matter, number of roots and fresh root yield. 4 CHAPTER TWO 2.0 REVIEW OF LITERATURE 2.1 Importance of Cassava Cassava is a major source of income for most farming households in Africa, creating youth employment opportunities and contributing to poverty alleviation (Njoku, 2012). It is consumed by over 600 million people in Africa, Asia and Latin America (Okoro, 2016). Among all staple crops in sub-Saharan Africa, cassava has been at the forefront as a 'crisis crop' since it can be left in the ground for well over one year and harvested when food shortages arise. The crop ranks third as source of calories in the tropics after maize and rice. According to FAO (2014), Africa contributed 145.77 million tonnes to global cassava production, while Asia and Latin America recorded 89.03 and 30.64 million tonnes, respectively. Although Nigeria was the highest cassava producer between 2010 and 2014 using the largest land area, its average yields were relatively low among the top ten producing countries in Africa. Production in Ghana increased whilst that of Democratic Republic of Congo decreased during the period. Malawi recorded the highest average yields during the period followed by Ghana and Cameroon. 2.2 Cassava production in Sierra Leone Cassava gained its importance after the civil war in Sierra Leone and has become a cash crop that generates income for many households and contributes positively to poverty alleviation. (Fomba et al, 2011). FAO (2014) reports that the average yield of cassava in Sierra Leone from 1990 to 2002 was below 2.0t/ha, followed by an increase to 7.9t/ha in 2004 and to 15.0t/ha in 5 2013. This tremendous progress in cassava productivity represents a positive relationship between yield and production which have grown to meet the rising demands for staple food and industrial applications. Cassava is the second most important staple after rice in Serra Leone. It constitutes an important portion of the diets of rural communities’ and in competition with other traditional staples. As a food crop, cassava fits well into the farming systems of the smallholder farmers in Sierra Leone because it is available all year around, thereby improving household welfare and livelihood. Compared to grains, cassava is more tolerant to low soil fertility and more resistant to drought, pests and diseases. It is largely cultivated by smallholder traditional subsistence farmers who make use of rain fed farming systems characterized by shifting cultivation practices, manual land preparation and inter-cropping systems (Jalloh and Dahniya, 1994). Cassava is grown in all regions of Sierra Leone (Eastern, Southern and Northern parts) with the northern region being the leading producer. The main cultivation period of cassava is from May through to April during the onset of rains. Since cassava is drought tolerant, there has not been any form of irrigation practices with its cultivation in Serra Leone. Varieties are grouped based on their texture, colour, taste, and mealiness (cooking ability). The widely-grown cassava varieties are the white root, which are extremely low in provitamin A. This has posed a nutritional challenge or alarming health issues in communities whose diets heavily rely on cassava. 2.3 Cassava production constraints in Sierra Leone Despite the growing importance of cassava as a nutritional security and income generating crop for subsistence farmers in Sierra Leone, as well as its potential to support the national economic development, the crop’s production output is constrained by a wide range of factors, some of 6 which include: the use of local unimproved low yielding cultivars with low nutritional value and lack of well adapted varieties. This indicates that farmers need better varieties. Furthermore, poor agronomic management practices, shortening fallow periods, declining soil fertility, poor access to market, cassava pests and diseases have limited cassava production in Sierra Leone. The development and yield stability of cassava rely on the quality of planting materials. The use of healthy plant materials is a very important factor in the attainment of good yields. Conversely, cuttings with low vigour which are infested/infected by pests and diseases limit cassava production. The long cropping cycle of cassava, harsh and unpredictable climatic conditions, lack of access to credit facilities and farm inputs also constrained cassava production in Sierra Leone (Spencer, 1997). Poor Government policies and strategies, poor feeding and care practices, limited generation, and dissemination of technologies and preservation of Indigenous Technical Knowledge (ITK), has also resulted in low cycles in cassava production which has engendered chronic malnutrition within communities reliant on cassava and has triggered concerns in Sierra Leone (SLARI Strategic Plan, 2012). Improved cultivars enhanced with high beta carotene, iron, zinc, and protein should be made available to such communities to address the problem of malnutrition in Sierra Leone and regions in sub-Saharan Africa where cassava is consumed as a major staple. 2.4 Agricultural origin and spread of cassava in Africa Historical and scientific evidence supports cassavas to have its origin as South America as the species is an ancient starchy root crop in the region (Allem, 2002). The progenitor of cassava (M. esculenta ssp. Flabellifolia) has been in existence and is adapted to forest and savanna ecozones of the Amazon basin. 7 The crop was domesticated around 4000 BC and has evolved as a food crop from the second and third millennium BC (Allem, 2002; Nassar and Ortiz, 2008). The domestication process initiated natural selection for cultivars with traits such as root size, growth habit and ability of clonal propagation through stem cuttings (Jennings, 1976 and Njoku, 2012). The Portuguese explorers introduced cassava to Africa during the 16th and 17th centuries through their trade within the West African coasts. Africans then spread cassava cultivation and is now found in almost all parts of tropical Africa. Early cultivation started in Fernando Po in the Gulf of Benin and around the Congo River in the 16th century and did not spread through West Africa until the 20th century (Hillocks, 2002; Njoku, 2012). Currently, Africa leads in the production of cassava with Nigeria as the largest producer in the world. Other top producing countries include Brazil, Thailand and Democratic Republic of Congo (FAO, 2010). Realizing the importance of cassava, the International Institute of Tropical Agriculture (IITA) established with its headquarters in Ibadan, Nigeria in 1967 under the guidance of the Consultative Group on International Agricultural Research (CGIAR) was mandated to oversee the development of the crop across Africa 2.4.1 Taxonomy of cassava Cassava is placed in the Fructicosae section of the genus Manihot, which is a member of the Euphorbiaceous (Jennings and Iglesias, 2002: Parkes, 2011). About 98 species have already been identified in the Manihot genus (Rogers and Appan, 1973; Akuwa, 2016). The Fruticose section is made up of growing shrubs adapted to savanna, grassland or desert and is considered less primitive than Arboreae section of Euphorbiaceae, which are made up of tree species (Jennings and Iglesias, 2002). All the species of Manihot have 4x = 2n = 36 chromosomes and could be 8 regarded as polyploids with x = 9, n = 18. Although cassava is normally considered as a polyploid species (El-Sharkawy, 2003), 18 small and similar pairs of associated homologous chromosomes, or bivalents were reported during analyses conducted using diakinesis and metaphase I approach (Jennings and Iglesias, 2002 and Wang et al., 2011). Cassava therefore, can be regarded as a functional diploid crop (Jennings, 1976; De Carvalho and Guerra, 2002; Nassar and Ortiz, 2008). Cassava classification is based on its qualitative characteristics such as leaf shape and size, plant height, stem colour, petiole length and colour, inflorescence and flower colour, storage root shape and colour, earliness, and content of cyanogenic glycosides (Nassar and Ortiz, 2006). Cultivars are classified as bitter and sweet based on cyanogenic glycoside. Onwueme (1987), confirmed against using the level of glycosides as a distinguishing characteristic for cultivars since the exact level of glycosides in a cultivar varies considerably according to the climatic conditions under which the plant is cultivated. Sweet cultivars are reported to have a short growing period with their storage roots maturing early (Nassar and Ortiz, 2006; Amenorpe et al., 2007). It has been reported that the emergence of spontaneous species as well as other Manihot species have been required to occur naturally in Africa and Brazil (Nassar, 1994). 2.4.2 Root system The root system of cassava is made up of feeder roots and storage roots. When propagated from stem cuttings it’s developed adventitious root at the base of the cuttings within the first three weeks (Ekanayake et al., 1997). The root system of cassava affects its establishment in the field. The adventitious roots subsequently develop into a fibrous root system, which absorb water and nutrients from the soil. The length of the fibrous root system reaches 200 cm or more in length (IITA, 1990). Formation of the storage roots begins about eight weeks after planting. Between 9 three to ten fibrous roots start to bulk and become storage roots (Alves, 2002). The fibrous roots that become storage roots lose their ability to absorb water and nutrients considerably. Cassava propagated from true botanical seeds develops the typical tap root system of dicot species within 30-60 days. Increase in roots diameter produces storage roots. Storage roots develop from the activity of cambium and starch accumulation (Alves, 2002). The shelf life of cassava storage roots is very short as compared to other major root crops (Ghosh et al., 1988). Rapid post-harvest deterioration processes occur within 24-72 hours after harvest (Wheatley et al., 1985; Wheatley and Chuzel, 1993). 2.4.3 Variability in colour and provitamin A of cassava storage root Most breeding populations of cassava have white storage roots, although some yellow root landraces have attracted a lot of attention lately. Yellow pigmented cassava root is typically known to be cultivated in a limited way in Colombia, Philippines, Jamaica and some African countries (Oduro,1981), with some yellow landraces also identified in Amazonia in Brazil (Ferreira et al., 2008; Nassar et al., 2009). Wide variation exists in root pigmentation within the global yellow root germplasm. This varies from pale yellow through orange to pink (Nassar et al., 2007). This variation in root pigmentation results from the wide variation in carotenoid contents within the global yellow cassava germplasm. It is important to note that, previous studies have shown that yellow root cassava varieties tend to have low dry matter content (Vimala et al., 2008; Akinwale et al., 2010 and Njoku, 2012) which is associated with poor cooking quality (Vimala et al., 2008). Yellow root cassava has high levels of provitamin A carotenoid and its consumption has been perceived as a sustainable approach for addressing Vitamin A deficiencies. In cassava, intensity of yellow pigment in roots 10 of some genotypes is strongly associated with β-carotene (Sanchez et al., 2006). Enhanced content of β-carotene (provitamin A) in yellow flesh cassava (Chavez et al., 2007 and Sanchez et al., 2006) provides potential opportunity to effectively address vitamin A malnutrition through availability of provitamin A cassava varieties where the crop is largely consumed (Makokha and Tunze, 2005; Nassar and Ortiz, 2010, Esuma et al., 2016). It is interesting to note that global efforts towards breeding cassava for high β-carotene content, is a recent development with slow progress registered towards development and deployment of nutrient rich varieties to farmers (Hershey, 2012; Welch and Graham, 2004). Several attempts have been made to elevate the carotene levels of yellow cassava through the introgression of genes from existing genetic resources (Akuwa, 2016). Carotenoids are a family of C40 isoprenoid pigments including approximately 600 identified structures in higher plants. The accumulation of intermediary carotenoids and their stable natural isomers (Z-iso) varies in accordance with plant species and plant organ types. In higher plants, carotenoids play the role of providing distinct yellow, orange, or red colors to certain organs, such as flowers, fruits, roots and tubers. Beta-carotene is probably the most well-known of the carotenoids; a phyto-nutrient family that represents the most widespread groups of naturally occurring pigments. ß-carotene known as provitamin A is one of approximately 50 carotenoids compounds which can be converted into retinol, an active form of vitamin A, in the body. β- carotene is not synthesized in animals but only in plants and micro-organisms. Carotenoids in plants are derived from the general isoprenoid biosynthesis pathway that takes place in chloroplasts of photosynthetic tissues and in chromoplasts of fruits and flowers. Synthesis of β- carotene is accomplished by insertion of only the missing genes necessary to complete or complement the biosynthetic pathway. Cervantes-Flores (2006) also confirmed the observation 11 of mRNA of genes in the carotenoid pathway in both orange mutant tissues and the 14- unpigmented wild-type tissues in a study conducted in Brassica olearacea. The molecular mechanism responsible for the increasing and variable levels of beta-carotene synthesis in the yellow flesh cassava (YFC) varieties compared with other differently colored varieties is still unknown. 2.5 Cassava growth and development Studies on the growth and development of cassava are not so widespread (Connor et al., 1981; Keating et al., 1982 and Parkes, 2011). Seeds are known to take about 16 days to germinate. However, by carefully filing the sides of the seed coats at the radicle end and by temperature management, germination can be accelerated. Ellis et al., (1982) reported required temperature for cassava seed germination when temperature exceeded 24°C and the best rates occurred at 30 to 35°C. Fourteen days’ dry heat treatment at 60°C is beneficial for newly harvested seeds. Seeds should be stored at 50°C and at 60% relative humidity (IITA, 1978), as they tend to lose viability rapidly during a year storage at ambient temperatures (Kawano, 1978). In cassava, the initial growth phase after seed germination lasts about six weeks after which auxiliary shoots and adventitious roots regenerate. Photosynthesis commences as soon as the first leaves appear. Storage roots developments starts with the initiation of secondary thickening of the adventitious roots, a process observed as early as three weeks after planting (WAP), (IITA,1990). Onwueme (1978) and Vine (1979) reported that the thin root accomplishes the initial penetration through the soil, and the increase in girth or growth occurs afterwards. Soil physical properties and texture which affect storage root yield (Ntawuruhunga, 2000). Cassava storage root differs arbitrarily thereby distinguishing one from another when their thickness surpasses 0.5cm, which is generally reached between one to four months after planting 12 (Veltkamp, 1986). Storage root bulking in cassava is affected by assimilate supply to the roots which is determined by shoot growth and hormonal changes (Williams, 1972). The number, shape, size and angle at which storage roots penetrate the soil, the colour of the outer cork, and internal tissues vary greatly among varieties. Usually there are three to ten storage roots per plant. Roots are cylindrical, 15 to 100cm long, 3 to 15cm in diameter, but could be occasionally branched. Storage root size is a function of root length and root diameter. Root size is genetically controlled in cassava (Dixon et al., 1994a), and this depends on the extent of cell division and starch accumulation processes (Ekanayake et al., 1997). Significant correlation observed between root diameter and yield (Okogbenin and Fregene, 2002) is an indication that diameter can be used to select for high storage root yield in cassava. The final yield is associated with the storage root number, diameter and size (Williams, 1972; Simwambana, 1988; Njoku, 2012). 2.6 Breeding for high beta-carotene content, iron, zinc and protein Improvements in cassava yield and other traits are generally not geared towards the highest possible under favorable conditions but rather obtaining stable yields and durable traits under marginal conditions where cassava is grown at present and is likely to expand in future. (Cock, 1985; El-Sharkawy, 2003). Outstanding yield is achieved first by selecting plants with good genetic structure which maximizes performance and secondly expressing potential resistance or tolerance to factors which limit yield (Ellis et al., 1982). Heterozygosity is the main hybrid vigor requirement for the genetic structure of new varieties and is a major focus of breeding programmes (Nassar et al., 2004). 13 Efforts have been undertaken over the years to determine the genetic potential for increasing the concentrations of bio-available Fe, Zn and provitamin A carotenoids in edible portions of several staple crops including rice, wheat, maize, beans and cassava (Graham and Welch, 1996). Conventional cassava breeding and selection pipeline begins with artificial or open pollinated crossing to obtain botanical seeds, followed with the establishment of the seeds in the seedling nursery and then screening and advancing to different breeding stages; the clonal evaluation trials, preliminary yield trials, advanced yield trials, uniform yield trials, and finally the multi- locational testing and on farm testing towards release (Jennings and Iglesias, 2002; Kawano, 2003). This process lasts for a period of eight to ten years before a variety is released. During the past years, efforts have been made to circumvent the number of years it takes for conventional breeding of cassava varieties. Very recently, modern molecular tools are being employed to shorten the breeding cycle. Iglesias et al. (1997) reported an increased level of carotene contents in roots from 632 clones and found a broad distribution of concentration from less than 0.1 to 2.4mg/100g of fresh roots. Chavez et al., (2005) evaluated 2457 clones and found that carotene content in storage root ranged from 0.102 to 1.040 mg/100g. Jos et al., (1990) screened 654 cassava stocks of indigenous and exotic origin for carotene, and found 21 clones with beta-carotene content ranging from 65I. µ /100g to 670 I. µ /100g. Recent progress has also been reported in the development of quick, inexpensive methods for screening for micronutrients with the use of molecular marker-assisted selection (Wong et al., 2004). The process accelerates the introgression of increased micronutrients from exotic sources into locally adapted, elite varieties. However, this depends on bioavailability of the type of micronutrients and willingness of farmers 14 to adopt such varieties. Variability in carotene content among accessions of national germplasm collections have been reported in India (Moorthy et al., 1990), in Brazil (Ortega-Flores, 1991), in Uganda (Esuma et al., 2016), in Nigeria (Njoku,2012) and in Ghana including iron and zinc (Baafi et al., 2016a). Jos et al. (1990), they however, demonstrated the potential to rapidly increase carotene content in cassava roots through cycles of recurrent selection. They increased the concentration from 4.2 mg/kg of fresh roots in the base population to 14 mg/kg after two cycles of selection and recombination. Hierarchies of screening procedures have been exploited for selection for better nutritional quality traits in cassava. Dixon et al. (2000) reported a significant positive correlation between iron and zinc in cassava and concluded that the potential exists for developing cassava clones with higher levels in both iron and zinc. The broad genetic base combined with recurrent selection has been reported as the most appropriate procedure for improving base population (Bryne, 1984; Fregene et al., 2006; Akuwa, 2016 and Baafi et al., 2016a). A study of 600 cassava clones revealed a zinc concentration of 2.6 and 37 mg/kg with an average of 7.5 mg/kg (Chavez et al., 2005) in storage roots. To supply the minimum daily zinc requirement for individuals consuming between 500 and 1000 kg of cassava would require varieties with at least six-fold zinc levels in the edible parts (Sayre et al., 2011). Through a transgenic approach, Ricachenevsky et al., (2013) used the endosperm-specific overexpression of MTP1 proposed for zinc biofortification in rice to increase zinc concentration in the storage root of cassava. 15 2.6.1 Participatory plant breeding Plant breeding exploits existing variability, gene manipulation and recombination of new genes into plant forms for human uses. Participatory plant breeding is the set of approaches that apply in situations where consumers are involved in the process of developing varieties that address their specific needs for different varietal traits. Early plant breeding was developed essentially as an art when farmers, who were though not educated, applied intuitive knowledge to skillfully and carefully select and retain plants with the most desirable features for crop improvement (Parkes, 2011). This informal breeding during the years have made farmers skillful in selection of varieties with their preferred traits. Most scientists now sourced farmers’ indigenous knowledge on the cultivation of many crops (Francis, 1990; Dapaah et al., 2003; Manu Aduening et al., 2006) as it has been observed and documented that technologies that have been developed with little or no farmers’ participation has gained little or no attention by farmers (Nweke, 2004; Manu-Aduening et al., 2005; Zacarias et al., 2004). This often occurs because farmers’ perception priorities have been different from that of breeders (Manu-Aduening et al., 2005). Kizito et al., (2007) and Parkes, 2011) indicated that farmers used some stable morphological traits like height at first apical branching, petiole colour and culinary attributes such as taste to differentiate and name varieties. 2.6.2 Progress in cassava improvement Hahn et al. (1980), IITA (1982, 1993) and Ceballos et al. (2006a) have reported on yield increases in cassava through genetic improvement. However, despite the proven record of increasing in cassava improvement, there are still many challenges limiting the improvement of the crop. Lawson (1988) reported that cassava genotypes find optimum physiological expression 16 of their genetic potential within narrow ranges of biophysical conditions. Cock (1987) and El- Sharkawy (2003) also reported that only few cassava cultivars attain stability across multiple agro ecologies. A study of growth in cassava concluded that stability in productivity in cassava depends on several factors acting synergistically. These factors include abiotic factors (soils, temperature, photoperiod and latitude), biotic factors (diseases, pests and nematodes) and management practices (Allem and Hahn, 1991). Carter et al. (1992) concluded that 19% of cassava available in Africa is found in mid-altitudes where trends in socio-economic and physical environment favor cassava production. This has provoked more interest in cassava production increment within this ecology since earlier research was focused on the lower altitudes of the tropics where cassava finds its most suitable growth environments (IITA, 1993; FAO, 1996). The variation in plant adaptation is linked with environmental factors that influence differential yield performance of genotypes (Cooper and Hammer, 1996). Understanding the factor that influence the environment is therefore a critical requirement for improving efficiency and effectiveness of a plant breeding programme. 2.7 Genetic diversity in cassava Genetic diversity in plants depends on various evolutionary processes, which include mutation, hybridization, migration and polyploidy (Colombo et al., 2000). Studies have assessed and characterized the genetic diversity in cassava gene pools with the aim of aiding genetic resource conservation into breeding programs. One of the first attempts using molecular markers on a global scale looked at genetic diversity, for potential heterotic groupings in cassava (Fregene et al., 2003; Ferguson et al., 2011). Jennings (1963) suggested that a high genetic diversity of 17 cassava genotypes resulted from introduction of cassava genotypes by immigrants, followed by natural hybridization in the fields. Seeds from spontaneous crosses that occur in field establishes a field seed bank which eventually germinate and produce volunteer seedlings from which farmers make selection and add to their existing cultivars (Elias et al., 2001; Kizito et al., 2007; Manu-Aduening et al., 2005; Mkumbira et al., 2003; Nassar, 2006; Peroni et al., 2007 and Pujol et al., 2002; 2007). The seed bank positively influences genetic diversity (Pujol et al., 2007) with patterns that suggest that the incorporation of volunteer seedlings accounts and supports the increases in intra-varietal genetic diversity (Elias et al., 2000 and Pujol et al., 2007). Fregene et al. (2000) and Tumuhimbise (2013) indicate that the genetic diversity in the East African cassava is threatened by the adaptation to pests and diseases, agronomic practices, and post-harvest uses. Thompson. (2013) however, observed a moderate to high genetic diversity among the cassava accessions evaluated in Ghana. Asante and Offei, (2003) suggested that even though the genetic diversity in Manihot spp is high in Ghana, a very low diversity is observed geographically in regions and is associated with the exchange of planting materials between farmers and selection for similar desired traits. It is believed that centuries of farmers’ methods of selection accounted for the wide range of genetic diversity observed in crops (Jennings and Iglesias, 2002). Nassar (2004), however, indicates that the wide genetic diversity in cassava is due to natural hybridization between the wild Manihot spp and cultivated cassava. One of the first attempts at using molecular markers on a global scale looked at genetic diversity for potential heterotic groupings in cassava (Fregene et al., 2003; Ferguson et al., 2011) 18 Until 2014, only few relatively small-scale genetic diversity assessments of cassava in southern, eastern and central Africa had been exploited using a range of molecular tools (Benesi, 2005; Fregene et al., 2000; Kizito et al., 2005 and Zacarias et al., 2004). 2.8 Morphological and Molecular Markers in cassava 2.8.1 Morphological markers Classification of cultivars and the study of their taxonomic status involving the use of measured qualitative and quantitative parameters of plants can be achieved using different plant breeding tools (Rogers and Appan, 1973; Parkes, 2011). The screening for cultivar identification traditionally depend on botanical traits (Stegemann, 1984). Breeders and geneticists have exploited qualitative characteristics such as leaf and flower attributes to study segregation of genes. Agro morphological traits are determined by more than one gene and are subject to varying degrees of environmental modifications and interactions. 2.8.2 Molecular markers Molecular markers include isozyme and protein, biochemical and DNA markers. A molecular marker could be defined as DNA regions showing sequence polymorphism in different individuals in species (Liu, 1998). DNA fingerprinting is an approach which has been widely accepted and exploited to achieve inter and intra organism differentiation at genotypic species and subspecies levels and other levels (McClean et al., 1994). The strategies have been used for cultivar identification, designed to detect the presence of specific DNA sequences or combination of sequences that uniquely identify the cultivar. Cultivar identification can be successfully attained more accurately through DNA fingerprinting data. The most closely related 19 cultivars are usually distinguished with the DNA fingerprinting methods (Beckman and Soller, 1986). The advantage of DNA fingerprinting over qualitative markers is the dominance of most DNA nucleus, absence of climatological and effects of tissue (Beeching et al., 1993). The application of DNA fingerprinting could be very important in the cultivar and species identification, and could help to create more efficient breeding programmes through the detection of genetic linkages between traits of interest. (Lin et al., 1993). Advances in molecular biology have unveiled the potential of DNA approaches, markers for genetic improvement and identification of cultivar of food crops. Several DNA based markers have been developed for measuring similarity that reveal polymorphism at DNA level in agricultural crops (Kumar et al., 2000). These markers have been proven to be an important tool in genetic variation assessment within and between populations and the elucidation of genetic relationships among cultivars (Lee, 1995; Karp et al., 1996). DNA markers confirmed greater variation than isozymes. DNA composition is also consistent between tissues and this is not affected by environmental changes (Beeching et al., 1993). DNA markers have been extensively used for the development of detailed genetic and physical chromosome maps in a variety of organisms. Molecular markers have been found very useful in conventional breeding for carrying out indirect selection in Quantitative Traits Loci (QTLs). In addition to these major applications, DNA markers can also be used in plant systems for germplasm characterization, genetic diagnosis, characterization of transformants, study of whole genome organization and phylogenetic analysis (Rafalski and Tingey, 1993). Although every marker system has some advantage and disadvantages, the choice of any marker system is dictated by the intended application, convenience, cost and time. 20 2.8.3 Single nucleotide polymorphism (SNP) Single nucleotide polymorphisms (SNPs) are DNA variation in sequence that occur when a single nucleotide (A, T, C and G) in the genome sequence is altered. SNPs can occur in coding (genes) and non-coding regions (introns) of the genome but have been shown to be more commonly situated in introns regions (Rafalski, 2002; Tangphatsornruang, 2008). SNPs are classified into non-coding SNPs, coding SNPs, exonic SNPs, cDNA SNPs and candidate SNPs (Kahl et al., 2004) with regards to their location. SNP markers could be defined as small insertions and deletions (indels) that represent the most available form of naturally occurring genetic variation within some populations. High density SNPs identification would dramatically facilitate progress in cassava genomics and breeding. SNPs are biallelic in nature (two alleles at a locus) meaning they are individually less informative than Simple Sequence Repeats (SSRs), which are generally multi-allelic meaning many alleles are found at a single locus (Ferguson et al., 2011). This drawback is complimented by the abundance and suitability of SNPs for ultra- high throughput genotyping strategies (Rafalski, 2002). SNP markers are extremely important in crop improvement. They are made up of a wide range of applications such as genetic diversity and phylogenetic studies, association mapping, marker diversity increase, QTL mapping, high- throughput MAS and evolutionary biology (Olsen, 2004; Lopez et al., 2005; Njoku, 2012). In cassava, there are limited information on discovery and use of SNP markers. In a study to monitor genetic diversity in cassava, 26 SNPs were identified and characterized using direct sequencing of diverse cassava varieties and an estimated frequency of one SNP every 121 nucleotide was observed (Kawuki et al., 2009). From this study nucleotide diversity varies from 7.8 x 10-4 to 8.6 x 10-3 and individual SNPs had lower polymorphic information content (PIC) values than haplotype based SNPs. SNPs have been used in cassava to understand the genome of 21 cassava origin and its phylogenetic relationship with wild relatives (Schaal et al., 1997; Olsen, 2004). SNP markers are more abundant in the genome, more efficient and cost-effective than SSR genotyping (Thomspon, 2014). 2.8.4 Morphological and molecular characterization of cassava Mathura et al., (1986) confirmed that phenotypic variance in cassava was higher than genotypic variance for traits of agronomic importance, such as weight of storage roots. However, all qualitative descriptors revolve around the three important parts of the cassava plant which can be classified into (a) the leaf characteristics (b) the stem characteristics and (c) the root characteristics (Alves, 2002; Fukuda et al., 2010 and Akuwa, 2016). The Manihot gene pool ranges from a great variety of wild species to numerous domesticated species with very specific unique qualities. Past and present research work has used markers toward molecular linkage maps development to provide a better structural definition of the cassava genome (Fregene et al., 1997 and Rabbi et al., 2014) 2.9 Combining Ability in cassava breeding Combining ability may be determined at two levels. GCA-General combining ability (GCA) and SCA-Specific combining ability (SCA). GCA measures the average performance of a specific inbred in a series of hybrid combinations, while SCA denotes the performance combination of specific inbred in a cross combination (Sprague and Tatium 1942). The importance of general combining ability (GCA) effects is to measure the comparative performance of parents in relation to one another and additive gene action (Byrne, 1984, Falconer and Mackay, 1996). Consequently, for genetic improvement of characters for a new population compared with the parental base population, SCA effects which measures the performance of distribution of crosses 22 in relation to GCA effects of parental combination becomes very essential (Singh, 2001). GCA and SCA effects show both the magnitude and direction of the genetic effects. According to Rajendran (1989), it is necessary to measure the combining abilities of parents before they are exploited in a hybridization program. This is because, selection of parents based on their direct performance may not always be dependable due to the type of gene action involved for expression of the trait and diverse genetic structure of the parents. The importance of mating designs in breeding is a two-way approach; firstly, to collate information for breeding population that can be exploited for selection and secondly, to develop outstanding varieties (Acquaah, 2009). In plant breeding, mating designs and arrangements are exploited by breeders to generate potential varieties. The kinds of mating techniques to use and its arrangements depend upon: i. Predominant type of pollination (self or cross): ii. Type of crossing used (artificial or natural) iii. Type of pollen dissemination (wind or insect) iv. Unique features, such as cytoplasmic or genetic sterility v. Purpose of project and vi. Size of population required. Some commonly used mating designs include; top cross, polycross, North Carolina (i, ii and iii), diallel (1, 2, 3 & 4) and triallels. The information generated from any of these mating designs support scientists to determine appropriate breeding strategy, as well as to evaluate the improvement that can be expected for a given selection intensity (Akuwa, 2016). A complete diallel mating design refers to the one that allows the parents to be crossed in all possible combinations (Schlegel, 2010). This mating design scheme is required to successfully accomplish Hardy–Weinberg equilibrium in a population (Acquaah, 2012; Nduwumuremyi et 23 al., 2013). Different breeding approaches that give information regarding the choice of parents and elucidate the nature and magnitude of various types of gene action involved in the expression of quantitative traits are now being used. They include diallel analysis, line x tester analysis and recurrent selection as well as other procedures currently used in cassava breeding followed by phenotypic selection, as suggested by Kawano (1980) and Akuwa (2016). 2.10 Genotype, environment and genotype by environment Interaction A phenotype (P), is known as the characteristic that is observed, depends on a combination of its genetic constitution, called the genotype (G), and the environment (E) and a component attributed to the interaction between genetic and ecological components (G×E). Since genes are expressed in an environment, the degree of expression of a heritable trait is impacted by its environment. This is usually expressed as: Phenotype = Genotype + Environment + G×E (Falconer and Mackay, 1996; Sleper and Poehlman, 2006; Tumuhimbise, 2014). The equation as stated below for phenotypic expression, it accommodates any variation observed in the phenotype due to variation in the factors resulting in the genotype. The relationship can be described as: VP = VG + VE + VGxE. Where: Where VP = Phenotypic variation VG = Genotypic variation Due VE = Variation because of the environment And VG×E = variation due to genotype by environment interaction effects. Acquaah (2009) explained genetic or heritable variation as the variation that can be attributed to genes that encode for some traits and can be expressed or seen from one traits to the other. Genotypic variation is separated into two components, namely additive and non-additive 24 components (Falconer and Mackay, 1996; Sleper and Poehlman, 2006). Additive variation is because of the cumulative effect of alleles on all gene loci influencing a trait, and is of most value in a crop improvement programme (Falconer and Mackay, 1996). Non-additive variation is separated into dominance variation, caused by the interaction of specific alleles at a gene locus, and epistatic variation, caused by the interaction among gene loci (Falconer and Mackay, 1996). The non-additive is variation normally given little attention since only the additive component of genetic variation determines the heritability of economic traits (Falconer and Mackay, 1996; Sleper and Poehlman, 2006). Genetic or heritable variation in nature originates from mutation, gene recombination, modifications in chromosome numbers, and structure and migration (Falconer and Mackay, 1996). The three phenomena namely mutation, gene recombination, and migration are manipulated more and more extensively to generate genetic variation for their breeding programmes (Acquaah, 2009). Environmental variation is associated with ecological conditions of the site where the crops are grown (Annicchiarico and Perenzin, 1994). Some of these conditions, such as plant to plant competition and population density can be regulated by employing effective agronomic practices, while others, such as biotic and abiotic stresses, are truly uncontrollable. Environmental variation in crops is normally difficult to control because it is non-heritable. For example, when individuals from a clonal population (identical genotypes) are evaluated in a trial site, the plants will exhibit differences in the expression for some traits because of non- uniformity in the environment. Therefore, an understanding for the magnitude and nature of variation and diversity present in the gene pool for the traits of interest is of greatest importance (Akuwa, 2016). Cassava has been subjected to considerable genotype by environment interaction (GEI) studies (Kvitschal et al., 2006; Ssemakula and Dixon, 2007; Lebot, 2009; Thompson, 2013 25 and Esuma et al., 2016). Quantitative trait is especially influenced by environmental effects. Differences in genotypic values may increase or decrease from one environment to another which might cause genotypes to rank differently between environments. Investigation on different cassava genotypes evaluated in contrasting environments have proven that Fresh Storage Root Yield (FSRY) trait is subject to strong GEI (Ssemakula and Dixon, 2007; Aina et al., 2009 and Esuma et al., 2016). Tan and Mak (1995) reported that GEI effects were significant for FSRY, commercial storage root number, harvest index, starch and cyanide content. Buerno (1986) revealed important genotype x location and genotype x location x-year interaction for FSRY when testing so many genotypes in the humid tropics of Brazil. To detect GEI cassava breeders evaluate advanced breeding lines in several environments. The concept of GEI enables plant breeders to determine the genotype among several genotypes that is high yielding and stable in multi-environmental trials (MET) i.e., across different environments (locations, years, growing seasons etc.). Breeding focuses on GEI as a typical requirement for to select the outstanding genotype for a target population of environments (Akinwale et al., 2011 Maroya et al., 2012). Studies reported by (Akinwale et al., (2011), Tumuhimbise et al., (2014) and Agyeman et al., (2015) have indicated a very good variation in fresh root yield across multiple environmental conditions. Ssemakula and Dixon (2007) noted a low influence of GEI on carotenoid content in cassava roots at harvest, based on the analysis of 28 genotypes screened across five multiple environments over two growing seasons. 26 CHAPTER THREE 3.0 ADOPTION CHALLENGES, PERCEPTION AND PREFERENCES FOR PROVITAMIN A CASSAVA AMONG THE CASSAVA VALUE CHAIN ACTORS IN SIERRA LEONE 3.1 Introduction Cultivated varieties of cassava in Sierra Leone either have white or cream/light yellow roots and high dry matter content but are deficient in vitamin A. The development of biofortified provitamin A cassava by IITA and research partners is a strategy to address the deficiency of vitamin A of white cassava root varieties. Biofortified yellow root cassava has β-carotene, a dietary precursor of vitamin A, which is known to be responsible for the yellow to orange colour of flesh storage roots (Rodriguez Amaya and Kimura, 2004; Njoku, 2012). Vitamin A is important for immune competence and good vision, as well as, cellular differentiation, growth, and reproduction. In cassava, β-carotene is the most abundant carotenoid which can be efficiently converted to vitamin A, E, ascorbic acid, enzymes and proteins; belongs to the biological antioxidant network converting highly reactive radicals (•OH) and free fatty peroxy radicals to less active species, thereby protecting the body against oxidative damage to cells (Packer, 1992). In human nutritional studies, vitamin A activity is expressed as retinol equivalent, and 6 μg of all-trans-β-carotene has the biological (vitamin A) activity of 1 μg retinol (Okaka et al., 2001). The dietary allowance (RDA) of vitamin A for adults (men and women) and children (4 to 9 years) are 0.75 and 0.3 to 0.4 mg/day retinol activity equivalents (REjm A)/day, respectively (Okaka et al.,2001; Njoku, 2012). These recommended dietary requirements are not adequately supplied in diets, especially in children, pregnant women and the under privileged in most developing countries including Sierra Leone. 27 In a cassava biofortification breeding programme for vitamin A, iron, zinc and protein, the following factors must be considered: farmers’ and consumers’ acceptability of the new product and bioavailability of the vitamin and nutrients. In Sierra Leone, adoption rates for new or improved technologies have been rather low; this is because during the early stages of the breeding process, farmers and consumers are not involved. However, it has been recommended that to increase the acceptability and adoption rate for biofortified provitamin A cassava cultivars, farmers and consumers should be included in formulating research objectives and in the selection of varieties through participatory methods. Participatory rural appraisal (PRA) is a technique that brings all stakeholders together with the aim of addressing a challenge and narrowing the communication gap between scientists and farmers. PRA forms the basis for stakeholders to work as a team to understand the target environment, identify the constraints and together formulate research agenda with plant breeders. This allows greater participation of farmers in the research process, leading to more effective and efficient information gathering and quick adoption of new research technologies. Adebo (2000) pointed out that, participatory approach is one of the ways that farmers’ perceptions can be captured and accounted for as hard data; the people’s resourcefulness and creativity can be challenged and captured and breeders may demonstrate respect for the insight and knowledge of the farmers in Sierra Leone. The present study aimed to solicit information from end-users in Sierra Leone regarding adoption challenges, perception and preferences for provitamin A cassava. This design allows the research team to source qualitative and quantitative data from study participants. This design supports a variety of analytical techniques including econometric and non-econometric analyses. 28 The objective was to i) assess perception and preferences for provitamin A yellow cassava roots and its gari among cassava value chain actors in Southern, Eastern and Northern region in Sierra Leone. 3.2. Materials and methods 3.2.1. Study area The study was conducted in Sierra Leone a country located on the West Coast of Africa which lies between latitudes 6o55’ and 10o00’N and longitudes 10o16’ and 13o18’W, and covers a total area of 72 000 km2. The climate is a monsoon humid tropical type with two distinct seasons: rainy season from May to October and dry season from November to April. Provisional results of a 2015 population census showed a total population of 7,075,641 people in Sierra Leone as opposed to 2004 of 4.5 million populations. Sierra Leone is divided into four regions: The Eastern, Western, Northern and Southern Regions. Each district is divided into districts of unequal sizes. Each district is divided into chiefdoms, which in turn are divided into districts and enumeration areas (EAs). This study focused on three districts of three regions and fifteen chiefdoms in Sierra Leone. The three regions eastern, northern and southern were purposively selected based on the production, processing and distribution patterns of cassava or cassava products within the country. Each region is divided into districts of unequal sizes. The three districts namely Kaliahum, Bombali and Moyamba districts were randomly sampled from the eastern, northern and southern regions, respectively. (Figure 3.1). Five chiefdoms from each district were purposively selected based on the level of cassava production, processing and utilization. 29 Figure 3.1 : Map of Sierra Leone showing surveyed regions 3.2.2 Sampling procedure The mixed method research design as “qual quant” (qualitative and quantitative) approaches was exploited for this study. This approach is an integrated research paradigm that combines various schools of philosophy, such as positivism and realism, within the research design. Mixed-method design is normally used when researchers are interested in gaining a rich and deeper understanding of a research problem. A mixed-method research design was applied to the study .Mixed methods research is a methodology for conducting research that involves 30 collecting, analyzing and integrating quantitative (e.g., experiments, surveys) and qualitative (e.g., focus groups, interviews) research data from study participants to deepen understanding of a research problem. The qualitative methods exploited in the study include: focus group discussions and observations, while survey and documentary review were applied to gather quantitative information. 3.2.3 Sample size determination Sampling refers to the procedure in which a sample is selected from an individual or a group of people of certain kind for research purpose. In sampling, the population is divided into a number of parts called sampling units. For this study, a multistage sampling was used to arrive at 450 value chain actors that were then distributed among the cassava value chain actors, (producers, processors, traders and consumers). The ‘value chain’ actors are stakeholders who played important role in the structure and performance of the cassava value chain. Anderson’s (2007) sampling method was adopted for determining sample sizes using the formula sampling size: 𝑧2𝑝𝑞 𝑛 = = 1 𝑑2 Where n = the sample size, z =1.96, p = proportion of population (the proportion of cassava producers, processors, traders and consumers in the three regions of Sierra Leone), q = a weighting variable computed as 1-p and d = the margin of error. Since the three regions in the country are known for agricultural production as a main livelihood activity, proportion of the population was estimated as 0.5 considering that the exact proportion 31 of the population was unknown. To ease identification among the cassava value chain actors in the sampled population, 8% was used as the margin of error for producers and consumers, while 11.32% was used as margin of error for traders and processors in the study. The sample size for producer was computed using Anderson’s (2007) formula below: 𝑧2𝑝𝑞 𝑛 = = 1 𝑑2 1.962(0.5)(0.5) 𝑛 = = 150 0.08002 Since producers and consumers were households, sample size of producers was the same as that of consumers. A total of 150 producers and 150 consumers were sampled for all the three regions with each having 50 producers and 50 consumers. Traders and processors were not household. The sample size for traders and processor were determined by Anderson’s formula as follows: 𝑧2𝑝𝑞 𝑛 = = 1 𝑑2 1.962(0.5)(0.5) 𝑛 = = 75 0.113162 The sample size for processors and traders in all the three regions was 75 with each region having 25 traders and processors each (Table 3.1) Table 3.1 Summary of systematic sampling methods used for the selection of respondents in the study areas. Stage Category Sampling method Sample size Stage: 1 3 regions (East, North and South) Purposive sampling technique 450 actors Stage: 2 3 districts (Kaliahum, Bombali and Simple random techniques 150 actors per district Moyamba) Stage 3 15 chiefdoms Simple random techniques 30 actors Stage 4 Chiefdom level Producers Systematic sampling 10 Producers Consumers Systematic sampling 10 Consumers Traders Simple random techniques 5 Traders Processors Simple random techniques 5 Processors Note: Actors consist of consumers, producers, processors and traders 32 Community based information sharing, awareness creation and sensitization programs on the benefits of Provitamin A cassava in addressing malnutrition were undertaken across different districts before initiating the PRA study in the selected chiefdoms. 3.2.4 Data Collection The following data were collected. The primary data set on provitamin A cassava was sourced from producer, processors, traders and consumers of cassava using pre-tested structured questionnaires for household interviews and semi- structured questionnaires for the focus group discussion. The household data set was collected through an electronic data capturing device using CSPro software, while, the focus group was exploited using charts and notepads. A total of 450 questionnaires were administered to respondents consisting 150 producers, 150 consumers, 75 processors and 75 traders. Qualitative and quantitative data sets were collected through focus group discussions with 4 groups of persons: youths, adult male and adult female and a pulled gender group. Efforts were made to capture gender issues within each category that consisted of 15 persons. Sixty persons were interviewed in total per chiefdom through group discussions. Secondary data were sourced from relevant journals, textbooks, internet, other related research projects, and extensive review of relevant literature on cassava value chain in Sierra Leone. 3.2.5 Data Analysis Data collected were analyzed using the SAS 9.3, SPSS and Microsoft Office Excel 2010. Descriptive statistics such as frequency distribution tables, arithmetic means, standard deviation and percentage, were used to describe the socio-economic characteristics. SWOT variables of the various actors in the study area were ranked with the use of Kendall’s coefficient of concordance. 33 3.3 Results 3.3.1 Socio economic characteristics of respondents Results in Table 3.2 show the socio-economic characteristics of the respondents. 50% of farmers interviewed in Bombali district were adults between the ages of 36 to 60, while 38% and 46% of the farmers contacted in Moyamba and Kailahun, districts respectively, were youth below 36 years. The proportion of males was higher than female. Males constituted 86% in Bombali, 98% in Kailahun and 64% in Moyamba. The results also show those respondents with no education in Moyamba where the highest with 48% followed by Bombali 44% and Kailahun 38% and almost all respondents were married; 88% in Bombali, 88% in Kailahun and 76% in Moyamba. In Bombali and Kailahum districts, 90% and 52% of the respondents, respectively, did not belong to any organization. However, 56% in Moyamba district belonged to different organizations such as Agricultural Business Centers and Farmers Based Organizations. Although 60%, 54% and 57% of respondents in Bombali, Kailahum and Moyamba districts respectively revealed that labour exchange formed the major benefit derived from belonging to organizations, an additional 40% of respondents in Bombali revealed that they also benefit from loan facilities within their affiliated organizations. 34 Table 3.2: Socio economic characteristics of respondents across the three districts Districts Socio economic Bombali Kailahun Moyamba characteristics (N = 50) (N = 50) (N = 50) Freq % Freq % Freq % Age group Adults 36 - 60 years 25 50 19 3 8 2 3 4 6 Aged Above 60 years 4 8 4 8 2 4 Youths below 36 years 21 42 27 54 25 50 Total 50 100 50 100 50 100 Gender Female 7 14 1 2 1 8 36 Male 43 86 49 98 32 64 Total 50 100 50 100 50 100 Educational level None 22 44 19 3 8 2 4 4 8 Koranic 4 8 2 4 2 4 Primary 15 30 7 14 9 18 Junior secondary school 6 12 14 28 9 18 Senior secondary school 2 4 6 12 5 10 Tertiary 1 2 2 4 1 2 Total 50 100 50 100 50 100 Marital status Married 44 8 8 44 8 8 38 76 Single 3 6 5 10 7 14 Widow/widower 2 4 0 0 3 6 Divorced 1 2 0 0 0 0 Separated 0 0 1 2 2 4 Total 50 100 50 100 50 100 Membership in an organization Yes 5 10 24 48 2 8 5 6 No 45 90 26 52 22 44 Total 50 100 50 100 50 100 If yes, key benefits derived from organization Labour exchange 3 60 13 54 1 6 5 7 Loan facilities 2 40 2 8 6 21 Input 0 0 4 17 1 4 Marketing 0 0 2 8 4 14 Information sharing 0 0 3 13 1 4 Total 5 100 24 100 28 100 3.3.2 Average yearly farm size per household for cassava production It was observed that 50%, 70% and 70% of the interviewees in Bombali, Kailahun and Moyamba districts, respectively, cultivate between 1 to 3 acres for cassava yearly (Figure 3.2). 35 Figure 3.2 Average yearly farm size per household for cassava production 3.3.3 Cassava cultivation within 5 years in the study areas across the surveyed districts Figure 3.3 show that all the respondents in Bombali, Kailahun and Moyamba districts have cultivated cassava for the past 5 years. 36 100 100 100 100 90 80 70 60 Yes 50 No 40 30 20 10 0 0 0 0 Bombali Kailahun Moyamba Figure 3.3: Cassava cultivation within 5 years in the study areas 3.3.4 Respondents’ average income in the study areas Most respondents earn less than Leones (SLL) 500,000 which is equivalent to US 67.67, annually (Figure 3.4) 37 percentages (%) 90 80 70 60 50 Less than 500,000 40 500,000 - 1,000,000 Above 1,000,000 30 20 10 0 Bombali Kailahun Moyamba Figure 3.4: Respondents’ average income in the study areas 3.3.5 Adoption challenges of producers for provitamin A cassava Results in Table 3.3 indicate the importance attached to the different factors that determines the adoption of cassava cultivars by respondents in the three districts. Cultivar selections made by farmers across the surveyed areas Bombali, Kailahun and Moyamba were based on traits of preference ranked in order of their importance. The first four traits of utmost interest to the 38 percentages (%) farmers in Bombali district were; high yielding, early maturing, edible (mealiness) and root size. However, while in Kailahun the order of importance for the traits were high yielding, early maturing, root size and petiole colour. In the case of Moyamba, the four most important traits were high yield, early maturing, root size and mealiness. In all the three districts, high yielding and early maturing were the key traits for selection in that order. Table 3.3: Varietal preferences of producers for provitamin A cassava across the surveyed districts Districts Bombali (N = 50) Kailahun (N = 50) Moyamba (N = 50) Cassava traits Percent Rank Percent Rank Percent Rank High Yield 92 1 98 1 58 1 Early maturing 72 2 94 2 48 2 Root size 52 4 46 3 42 3 Mealiness 70 3 4 7 38 4 Skin colour 4 6 34 5 24 6 Petiole colour 2 7 44 4 12 8 Plant size 2 7 4 7 28 5 Branching pattern 6 5 0 9 24 6 Root colour 0 9 8 6 16 8 Figures in brackets denote traits of preference ranked in order of importance for variety adoption by producers in each district; 1 = highest importance, 9 = lowest importance 3.3.6 Farmer’s perception for provitamin A cassava across the survey districts Results in Table 3.4 shows that 24% of respondents in Bombali, 6% in Kaliahum and 4% in Moyamba are aware of provitamin A cassava. However, 17%, 67% and 50% respondents in Bombali, Kaliahum and Moyamba are aware of yellow flesh cassava varieties. Also 66%, 33% and 50% of the respondent classified the variety SLICASS 11 (cream colour) as yellow fleshed. The study also reveal that only respondents in Bombali 67% have cultivated SLICASS 11 and 39 they source the planting material from NGOs. The presence and involvement of Village Hope, an NGO, in distributing planting materials and processing cassava in Bombali district coupled with the robust activities of SLARI’s extension office has contributed greatly in influencing the availability and accessibility of the yellow flesh cassava across the district. The percentage of interviewees in Bombali, Kailahun and Moyamba districts willing to adopt provitamin A cassava varieties once available are 98%, 96% and 96%, respectively. Overall, more than 95% of the respondents across the three districts indicated willingness to adopt new varieties of provitamin A cassava when available as they get to appreciate and perceive their nutritional quality. 40 Table 3.4: Farmer’s perception of provitamin A cassava across the surveyed districts Districts Bombali Kailahun Moyamba (N = 50) (N = 50) (N = 50) Farmer’s perception Freq % Freq % Freq % Heard about yellow flesh cassava Yes 12 2 4 3 6 2 4 No 38 76 47 94 48 96 Total 50 100 50 100 50 100 Name of yellow flesh cassava variety SLICASS 11 8 6 6 1 33 1 50 Yellow flesh 2 17 2 67 1 50 Don't know 2 17 0 0 0 0 Total 12 100 3 100 2 100 Have you planted yellow flesh cassava Yes 8 67 0 0 0 0 No 4 33 3 100 2 100 Total 12 100 3 100 2 100 Source of planting of yellow flesh cassava variety MAFFS 1 13 0 0 0 0 NGOs 7 88 0 0 0 0 Total 8 100 0 0 0 0 Means of sourcing the yellow flesh cassava variety Purchasing 1 13 0 0 0 0 Gift 7 88 0 0 0 0 Total 8 100 0 0 0 0 Did you plant yellow flesh cassava last season Yes 7 8 8 0 0 0 0 No 1 13 0 0 0 0 Total 8 100 0 0 0 0 Proportion land area planted with yellow flesh 0 - 25% 4 5 7 0 0 0 0 26 - 50% 2 29 0 0 0 0 Above 75% 1 14 0 0 0 0 Total 7 100 0 0 0 0 Purpose for growing yellow flesh in the HH Processing to other cassava products 7 100 0 0 0 0 Boil and eat 0 0 0 0 0 0 Total 7 100 0 0 0 0 Willingness to grow yellow cassava variety Yes 42 9 8 48 9 6 48 96 No 1 2 2 4 2 4 Total 43 100 50 100 50 100 41 3.3.7 Traders perception on cassava root trading across the surveyed districts Almost all the traders in Bombali district sell creamy or white (non-provitamin A) cassava roots while 8% in Kailahun and 52% (Table 3.5) in Moyamba sell white roots. No yellow cassava root was sold in the study area. All respondents in Bombali and 50% and 53% of the respondents in Kailahun and Moyamba districts respectively, claimed that they source cassava roots from the periodic markets. One hundred percent and 50% of suppliers of cassava roots, in Bombali and Kailahun districts respectively, obtain their cassava roots from other traders, whereas, 47% in Moyamba obtained theirs directly from farmer’s field. Hundred percent of respondents in Bombali and Kailahun reported that they sell cassava roots at the daily market, while less than 50% of respondents in Moyamba district sell their cassava roots through road side markets. Consumers are the most preferred customers of cassava roots across the three surveyed districts with 100% scores each in both Bombali and Kailahun and 67% in Moyamba. 42 Table 3.5: Traders perception on cassava root trading across the three districts Districts Bombali (N Kailahun (N Moyamba (N Cassava root trading = 8) = 26) = 29) Freq % Freq % Freq % Sell cassava roots Yes 7 88 2 8 15 52 No 1 12 24 92 14 48 Total 8 100 26 100 29 100 Flesh colour of cassava root sold White 0 1 00 1 5 0 11 73 Cream 7 0 1 50 4 27 Yellow 0 0 0 0 0 0 Total 7 100 2 100 15 100 Source of cassava roots Own production 0 0 0 0 7 4 7 Periodic market 7 100 1 50 8 53 Daily market 0 0 1 50 0 0 Road side market 0 0 0 0 0 0 Total 7 100 2 100 15 100 Suppliers of cassava roots Farmers 0 0 0 0 10 6 7 Wholesalers 2 29 0 0 5 33 Other traders 5 71 2 100 0 0 Total 7 100 2 100 15 100 Selling location of cassava roots Periodic market 0 0 0 0 5 33 Daily market 7 100 2 100 3 20 Road side market 0 0 0 0 7 44 Total 7 100 2 100 15 100 Preferred customers of cassava roots Consumers 7 1 00 2 1 00 10 67 Wholesalers 0 0 0 0 0 0 Other traders 0 0 0 0 5 33 Total 7 100 2 100 15 100 43 3.3.8 Provitamin A cassava root trading in the study areas There’s no evidence of trading or selling provitamin A cassava roots in the three districts (Figure 3.5). However, there is strong willingness among farmers in all the districts to sell provitamin A cassava roots when made available. 100 100 100 87 90 80 70 60 50 40 30 20 10 0 0 0 0 Bombali Kailahun Moyambo HS% WLS% HS - Have sold yellow flesh cassava root, WLS – Would like to sell yellow flesh cassava root Figure 3.5: Provitamin A cassava root trading in the study areas 44 Percentages(%) 3.3.9. Processors perception and preference for cassava roots processing and awareness across the different districts Survey of respondents processing different cassava products in the study areas shows that Gari is the most important processed product in all the 3 districts, followed by Fufu in Moyamba and Kailahun and Tho in Bombali district (Fig 3.6). Starch processing had moderate importance in all 3 districts and accounted for around 30% of processing in Moyamba and Bombali and about 22% in Kailahun. Ninety-six percent of respondents in Moyamba, 100% (Bombali) and 78% (Kailahun) of the interviewees indicated willingness and preference for provitamin A gari when it is made available. 100.0 100.0 100.0 100.0 94.4 80.0 75.0 63.6 60.0 45.5 40.0 33.3 27.3 22.2 22.2 16.7 20.0 9.1 4.2 5.6 0.0 Moyamba Bombali Kailahun HS - Have sold yellow flesh gari, WLS – Would like to sell yellow flesh gari Figure 3.6: Percentage of respondents processing different cassava products in the study area 45 Percentages (%) Gari Fufu Tho Flour Starch Gari Fufu Tho Flour Starch Gari Fufu Tho Flour Starch 3.3.10 Perception and preference for cassava roots processing/trading and awareness across the different districts Figure 3.7 shows that 100% of respondents across all three districts indicated interest in trading in provitamin A cassava gari. 100 100 100 100 90 80 70 60 50 40 30 20 10 0 0 0 0 Bombali Kailahun Moyamba HS% WLS% HS - Have sold yellow flesh gari, WLS – Would like to sell yellow flesh gari Figure 3.7 Respondents interest in provitamin A cassava gari production/trading in the study areas 46 3.3.11 Perception and preference of consumers on consumption and awareness of provitamin-A cassava roots across the surveyed districts Cassava roots consumption and awareness rates are presented in (Table 3.6). Over 90% of respondents across the three surveyed districts eat cassava. The widely consumed variety is the white flesh root with 97% Bombali, 100% Kailahun and 96% Moyamba of the respondents consuming it. There is very little knowledge on provitamin A cassava across the three surveyed districts with 14%, 6% and 6%, indicating awareness levels in Bombali, Kailahun and Moyamba respectively. Of the few respondents who indicated awareness of yellow cassava, only 50%, 33% and 0% consume the yellow cassava in Bombali, Kailahun and Moyamba respectively. However, all the respondents (100%) indicated interest in provitamin A cassava, and showed willingness to accept and introduce provitamin A cassava into family meals mainly because of its nutritional value. Poundability was least considered by all responded as a criterion for introducing provitamin A cassava to their families. 47 Table 3.6: Cassava roots consumption and awareness of consumers across the surveyed districts Cassava root consumption Districts Bombali Kailahun Moyamba (N = 49) (N = 50) (N = 54) Freq % Freq % Freq % Eat cassava roots Yes 48 97 50 100 52 96 No 1 3 0 0 2 4 Total 49 100 50 100 54 100 Roots flesh color White 44 91 47 94 51 98 Cream 0 0 3 6 1 2 Yellow 4 8 0 0 0 0 Total 48 100 50 100 52 100 Aware of YF cassava root Yes 7 14 3 6 3 6 No 41 86 47 94 49 94 Total 44 100 50 100 52 100 If yes, do you consume yellow flesh cassava root Yes 3 50 1 33 0 0 No 3 50 2 67 3 100 Total 6 100 3 100 3 100 If yes.do you prefer yellow flesh cassava roots Yes 3 100 1 100 0 0 No 0 0 0 0 0 0 Total 3 100 1 100 0 0 If yes, why? Nutritional values 3 100 1 100 0 100 Willingness to introduce YF cassava to family? Yes 41 100 49 100 54 100 No 0 0 0 0 0 0 Total 49 100 50 100 54 100 Reasons for willingness to introduce YF cassava to family Nutritional value 27 56 37 74 31 57 Taste 8 16 10 20 16 29 Flavor 1 2 3 6 3 6 Accessibility 12 24 0 0 3 6 Poundability 1 2 0 0 1 2 Total 49 100 50 100 54 100 Do you use cassava to wean infants? Yes 5 10 21 42 21 39 No 44 90 29 58 33 61 Total 49 100 50 100 54 100 Flesh colour of cassava root used White 5 100 21 100 20 95 Cream 0 0 0 0 1 5 Yellow 0 0 0 0 0 0 Total 5 100 21 100 21 100 Willing to use YF cassava root as weaning food? Yes 49 100 50 100 54 100 No 0 0 0 0 0 0 Total 49 100 50 100 54 100 48 Table 3.6: Cassava roots consumption and awareness of consumers across the surveyed districts (cont’d) Districts Bombali (N = 49) Bombali B o mbali Cassava root consumption (N = 49) (N = 49) Freq % Freq % Freq % Price difference between YF cassava root and other roots Yes 0 0 22 44 12 22 No 6 12 28 56 5 9 Don't know 43 88 0 0 37 69 Total 49 100 50 100 54 100 Advantage of YF cassava roots Nutritional value 47 95 20 40 32 59 Good taste 2 4 16 32 7 13 Fine colour 0 0 14 28 15 28 Total 49 100 50 100 54 100 Disadvantage of YF cassava root Bitter taste 0 0 4 8 0 0 Not available 49 100 12 24 23 43 No awareness 0 0 34 68 31 57 Total 49 100 50 100 54 100 3.3.12 Perception and preferences of consumers for Gari consumption across the three districts among value chain actors Table 3.7 shows that the respondents in Bombali 47%, Kailahun 50% and Moyamba 54% districts consumed gari. Virtually all the gari eaten in the 3 districts were white gari constituting 100%, 96% and 98% of the gari consumed in Bombali, Kailahun and Moyamba respectively. All the consumers interviewed had no knowledge of provitamin A gari, but all of them were willing to adopt it. Above 90% of the respondents prefer provitamin A gari, mainly because of its nutritional value. Respondents in Bombali 4%, Kailahun 26% and Moyamba 17% use white gari as weaning food. Over 90% of consumers interviewed across the three districts were willing to introduce provitamin A gari as weaning food. 49 Table 3.7 Gari consumption across the three districts Districts Bombali Kailahun Moyamba Gari consumption (N = 49) (N = 50) (N = 54) Freq % Freq % Freq % Eat gari Yes 23 4 7 25 5 0 29 54 No 26 53 25 50 25 46 Total 49 100 50 100 54 100 Color of gari White 49 100 48 96 53 98 Cream 0 0 2 4 1 2 Yellow 0 0 0 0 0 0 Total 49 100 50 100 54 100 Aware of YF gari Yes 0 0 0 0 0 0 No 49 100 50 100 54 100 Total 49 100 50 100 54 100 Adopt YF gari Yes 45 92 49 9 8 49 91 No 4 8 1 2 5 9 Total 49 100 50 100 54 100 Use gari to wean infants Yes 2 4 13 2 6 9 17 No 47 96 37 74 45 83 Total 49 100 50 100 54 100 If yes, colour of gari White 2 100 9 6 9 7 7 8 Cream 0 0 4 31 2 22 Total 2 100 13 100 9 100 Willing to use YF gari as weaning food Yes 45 9 2 47 9 4 50 93 No 4 8 3 6 4 7 Total 49 100 50 100 54 100 Advantage of YF gari Nutritional value 49 100 22 44 36 67 Good taste 0 0 15 30 16 30 Fine colour 0 0 3 6 2 3 Don't know 0 0. 10 20 0 0 Total 49 100 50 100 54 100 Disadvantage of YF gari Not available 49 100 19 3 8 13 2 4 No awareness 0 0 31 62 41 76 Total 49 100 50 100 54 100 50 3.3.13 Respondents’ perception on the strengths, weakness, opportunities and threat for three districts among the cassava value chain actors Table 3.8 shows the producers’ strengths, weaknesses, opportunities and threats (SWOT) across the three surveyed districts. Access to land availability was ranked as the most important strength, limited access to credit finance for weakness, availability of improved varieties for opportunities and lack of external funding in Bombali and high cost of agricultural machinery for Kailahum and Moyamba for threats across the surveyed districts. Table 3.8 Kendall SWOT ranking for cassava producers across three districts Bombali Kailahun Moyamba SWOT Mean Rank Mean Rank Mean Rank Strengths Agricultural land 1.5 1 1.04 1 2.00 1 Labour 2.02 2 1.98 2 2.74 2 Improved planting material 2.88 3 3.04 3 4.02 3 Finance and credit 4.92 4 5.16 5 4.82 5 Member of FBO 5.5 5 5.58 6 4.82 5 Experience and knowledge 5.5 5 6.06 7 5.28 7 Have processing facilities 5.68 7 5.14 4 4.32 4 P value <0.0001 <0.0001 <0.0001 Kendall’s W 0.71 0.83 0.31 Weaknesses Limited access finance and credit facilities 2.04 1 1.22 1 2.45 1 High transportation cost of tubers 2.26 2 2.48 2 2.74 2 Limited access to market 3.08 3 3.14 3 3.09 3 Lack agricultural machines and equipment 3.5 4 3.82 4 3.4 5 Low agricultural productivity 4.12 5 4.34 5 3.32 4 P value <0.0001 <0.0001 <0.064 Kendall’s W 0.30 0.59 0.01 Opportunities Availability of improved varieties 1.82 1 1.08 1 2.38 1 Strong support from government and 2.9 2 3.08 3 3.82 5 NGOs Training on improved agronomic practices 3.02 3 3.76 4 3.3 4 Availability of markets 3.62 4 4.08 5 2.76 3 Availability of processing centres 3.64 5 3.00 2 2.74 2 P value <0.0001 <0.0001 <0.0001 Kendall’s W 0.22 0.54 0.13 51 Table 3.8: Kendall SWOT ranking for cassava producers across three districts (cont’d) Bombali Kailahun Moyamba SWOT Mean Rank Mean Mean Rank Threats Lack of external funding 2.00 1 4.00 5 3.02 5 High transport fares 2.92 2 3.62 4 3.14 3 High market competition 2.86 3 2.84 3 3.18 4 High interest rates on loan 3.46 4 2.40 2 2.94 2 High cost of agricultural machinery 3.76 5 2.14 1 2.72 1 P value <0.0001 <0.0001 <0.611 Kendall’s W 0.18 0.25 0.01 3.3.14 SWOT ranking for cassava processors across the three districts Table 3.9 shows processors’ strengths, weaknesses, opportunities and threats across the three surveyed districts. Knowledge and labour availability were ranked first and second for strength across the surveyed districts. Market and lack of finance and credit were observed as weaknesses across the three districts. Improved varieties and improved processing technologies were ranked as the most important opportunities in Bombali while for Kailahun and Moyamba improved varieties and strong linkages constituted the most important opportunities. High cost of processing technologies and high cost of labour were ranked as threats. 52 Table 3.9: Kendall SWOT ranking for cassava processors across the three districts Bombali Kailahun Moyamba SWOT Mean Rank Mean Rank Mean Rank Strengths Knowledge 2.18 2 1.22 1 1.50 1 Labour 2.00 1 2.61 2 2.19 2 Equipment 4.64 6 3.28 2 3.73 3 Finance and credit 4.14 5 4.83 6 4.69 6 Market and storage 4.00 3 4.50 4 4.65 5 Linkage with farmers 4.05 4 4.56 5 4.25 4 P value <0.001 <0.0001 <0.0001 Kendall’s W 0.36 0.57 0.52 Weaknesses Market 3.00 2 1.67 1 2.04 1 Finance and credit 1.45 1 2.31 2 2.38 2 Processing equipment 3.00 2 2.81 3 3.21 3 Training 4.09 3 3.67 4 4.19 4 High transport fares 4.18 4 4.75 5 4.12 5 Low production 5.27 5 5.81 6 5.06 6 P value <0.0001 <0.0001 <0.0001 Kendall’s W 0.49 0.70 0.39 Opportunities Improve varieties 2.85 2 2.03 1 2.23 1 Strong linkages 5.30 6 2.92 2 3.02 2 High demand for gari 3.35 4 3.47 4 3.60 3 Provision of training 4.15 5 4.17 5 4.44 6 Improved processing technologies 2.25 1 3.28 3 3.92 5 Processing centres with modern equipment 3.10 3 5.14 6 3.79 4 P value <0.001 <0.0001 <0.001 Kendall’s W 0.04 0.35 0.18 Threats High labour cost 2.45 2 2.06 1 2.19 2 High cost of processing technologies 1.64 1 2.97 2 2.17 1 High interest rates on loan 4.64 5 3.94 4 3.52 3 Market diversity and competition 4.05 4 4.08 5 3.71 4 Inadequate supply of raw materials 2.73 3 3.69 3 4.56 5 Theft 5.50 6 4.25 6 4.85 6 P value <0.0001 <0.002 <0.0001 53 3.3.15 SWOT ranking for cassava traders across the three districts Table 3.10 reveals traders’ strengths, weaknesses, opportunities and threat across the three survey districts. High demands for product, storage facilities and membership in traders’ organization were ranked as the most important strengths in Bombali, Kailahun and Moyamba, respectively. Lack of access to finance and credit were the major weaknesses across the three districts (Bombali, Kailahun and Moyamba). Availability of financial institutions and availability of produce/products were ranked as the greatest opportunity for Bombali followed by Kailahun and Moyamba districts, while high transportation cost and no external funding were ranked as the most important threat across the three surveyed districts. 54 Table 3.10: Kendall SWOT ranking for cassava traders across the three districts Bombali Kailahun Moyamba SWOT Mean Rank Mean Rank Mean Rank Strengths 4.00 5 2.50 2 3.41 5 Access to market 2.51 2 2.21 1 2.81 2 Have storage facilities 3.43 3 3.21 3 2.74 1 Membership in trader's organisation 3.51 4 3.82 5 3.30 4 High demand of product 1.60 1 3.41 4 2.90 3 P-value 0.23 0.001 0.351 Kendall's W 0.353 0.178 0.04 Weaknesses Lack of finance and credit 1.51 1 2 .40 1 2 .30 1 Lack of storage facilities 3.40 3 2.61 2 3.10 3 Lack of market facilities 2.62 2 2.82 3 2.31 2 Low availability of cassava products 3.63 4 3.71 5 3.82 5 High market dues 3.91 5 3.61 4 3.61 4 P-value 0.019 0.004 0.001 Kendall's W 0.369 0.342 0.192 Opportunities Cassava product always available 2 .31 2 1 .90 1 1.90 1 High demand for cassava products 2.30 2 20 2 2.21 2 Availability of financial institutions 2.00 1 3.10 4 2.80 2 Availability of markets 3.52 4 3.42 3 3.10 4 P-value 0.275 < 0.001 0.002 Kendall's W 0.086 0.255 0.176 Threats High taxation 3 .00 3 3.01 3 2.40 2 High transportation cost 1.51 1 2.40 2 2.11 1 No external funding 1.61 2 1.61 1 2.40 2 Market diversity and competition 3.94 4 3.23 3 3.01 4 P-value < 0.001 < 0.001 0.089 Kendall's W 0.781 0.255 0.078 55 3.4 Discussion The PRA approaches used in this study yielded meaningful results as cassava value chain actors willingly shared their experiences and knowledge. Youth below 36 years in the surveyed areas constituted more than 50 percent of the cassava farmers interviewed in Moyamba, Kailahun and Bombali. This implies that the youth were actively involved in cassava cultivation in Moyamba and Kailahun. However, in Bombali the adult population was highest. Age is said to be a primary latent characteristic in adoption decisions. This agrees with Okoye (2004), who reported that young people adopt innovations faster than old people. Nwaru (2004) reported that the ability of a farmer to break risk is innovative and starts from 30 years, but decreases with age. Eighty-six percent, 98% and 64% of respondents in Bombali, Kailahun and Moyamba were male indicating the predominance of males in cassava farming. The very high percentage of men compared to women in cassava farming in the surveyed districts could possibly be due to lack of mechanization facilities, labour intensiveness and task demands which may discourage women from undertaking cassava cultivation and production independently (SLARI Cassava Value Chain Report unpublished). This was contrary to findings of Adebayor and Salahu (2007), Oyegbami et al. (2010) and Thompson (2013) who reported of higher percentage of female (women) in cassava cultivation and production. Cassava farming is a gender-friendly occupation where men and women play active roles. Cassava cultivation activities which range from brushing, burning, ploughing and harvesting were mainly undertaken by men, while women participated in activities like planting, weeding and transportation of harvested storage roots. Although women spend more time in agricultural activities, unfortunately, most women have less access to information technology (FAO, 1988). 56 Education has been reported to be very important as it helps to refine a person’s perceptions of issues and help him/her to make reasonable decisions based on available information. Low level of formal education of about two-third of the farmers in all three districts contributed to the slowdown of the adoption of released technologies (improved varieties). Ajibefun and Aderinola (2004) observed that education facilitates adoption of new varieties; hence the low level of education among farmers in all the three districts sampled could influence the selection and adoption of introduced improved yellow root cassava varieties and provitamin A gari. Creating awareness and promoting their nutritional value and benefits in a dissemination program will accelerate adoption. Eighty-eight percent of respondents in both Bombali and Kailahun and 76% in Moyamba which makes up the farming population in the three districts were legally married implying that the land on which cassava is being produced is owned by family, rented by family or secured as a family property. The average land holdings in the surveyed districts ranged between 0.5 to 6 hectares. This implies that only a small parcel of land is allocated to cassava production and that subsistence agriculture predominates. This does not provide enough motivation to try innovations. Ochola, (2006) found that adoption of new technologies may be affected by the land sizes used by the farmers for agricultural purposes. Studies on adoption have shown that farm size positively correlates with the adoption of an introduced technology (FAO, 2002; Njoku et al., 2012). Forty percent of the farmers in Moyamba district, became members of an organization for them to have access to loan opportunities while farmers in the other two districts (Kailahun and Bombali) are members of organization mainly because of labour exchange. This shows that access to loan and labour are the most important resource required for cassava production in the surveyed area as majority of the farmers are poorly resourced. The survey conducted confirms that most of the cassava 57 produced in all the districts is processed into gari for sale locally and for family consumption. Majority of the farmers grow variety SLICASS 4 (non-provitamin A) because of its high percentage dry matter content. Dry matter content is a quality trait responsible for dry texture in cassava products, and this is preferred by most farmers especially in gari processing. There is a great demand for gari and a well-structured market outlet (Gbangbatok) where traders purchase bags of gari to sell as retailers. White root varieties are largely consumed in all the surveyed districts because of the low awareness of provitamin A varieties. SLICASS 11 is cream coloured which does not qualify it to be called a biofortified variety, as biofortified cassava variety is a cassava which should have been developed through hybridization with an increased level of total carotenoid content (provitamin A). Its nutritional status would have been increased. In the present study, respondents showed a high level of willingness to adopt and promote the use of provitamin-A or yellow cassava roots and products, due to its perceived nutritional benefits. This agrees with the results of Nkonya and Featherstone (2001), who found that varieties with farmers’ preferred traits were easily adopted. Farmers’ personal experience influenced decisions on what varieties they grew. There is no available market outlet where yellow cassava or provitamin A storage roots/products can be obtained across the three surveyed districts. The major constraints identified in this study including high cost of transportation impeding the growth and expansion of cassava production in the districts surveyed. These finding agrees with Parkes, (2011) who reported that high cost of transportation negatively affects cassava production. 58 3.5 Conclusion Farmers appreciate new technologies that have an added advantage over their current existing technologies. There is high prospect for this new technology (provitamin A cassava) to be adopted in Sierra Leone as revealed in this study. Consumers’ preference for provitamin A cassava root is linked to the crop’s nutritional benefits as informed through the awareness and sensitization campaigns conducted before the data collection period across the districts. Provitamin A cassava has been proven to improve nutrition especially among the vulnerable group (children less than five years, pregnant and lactating women). All consumers interviewed had no prior knowledge about provitamin A gari, but were willing to adopt it. Above 90% of the respondents preferred provitamin A gari, after the sensitization and awareness campaign during the survey about its nutritional value. This underscores the importance of creating awareness of nutritional value of provitamin A cassava genotypes/accessions in Sierra Leone. Cassava value- chain actors’ (Producers, consumers, Traders and Processors) preferences to further improve production and adoption of provitamin A cassava is understood. The study confirms that Moyamba district should serves as a hub for gari processing and marketing in Sierra Leone. A reliable database on provitamin-A cassava has been developed from this study. 59 CHAPTER FOUR 4.0 DIVERSITY STUDIES OF PROVITAMIN A CASSAVA (Manihot esculenta Crantz) IN SIERRA LEONE. 4.1 Introduction Genetic diversity provides species with the ability to adapt to changing environments. Several studies have been reported on the use of morphological descriptors to determine the genetic diversity among cassava genotypes (Rimoldi et al., 2010 and Asare et al., 2011 and Thompson, 2013). Recent advances in molecular biology techniques have led to the development of important tools for genetic studies in several plant species. The accuracy in accession characterization may therefore, be enhanced/achieved with the use of molecular markers associated with morphological traits. A lot of research has been undertaken in plant genetic diversity using molecular markers including DNA (Ferreira et al., 2008 and Rimoldi et al., 2010), such as amplified fragment length polymorphism (Benesi et al., 2010), simple sequence repeats (Alves et al., 2011; Parkes, 2009; Oliveira et al., 2012 and Costa et al., 2013) and single nucleotide polymorphism (Kizito et al., 2005; Tangphatsornruang et al., 2008; Ferguson et al., 2011; Thompson, 2013 and Rabbi, 2015). With recent advances in high throughput genotyping technologies, single nucleotide polymorphic markers (SNPs) are increasingly becoming markers of preference for plant genetic studies and breeding. SNPs are the most common type of genetic variation among species, involving just a change in a single nucleotide. Many Expressed Sequence Tags (ESTs) have been exploited to explain and detect SNPs in maize (Zea mays L.) (Ching et al., 2002) and Soybean (Glycine max L. Merr.) (Zhu et al., 2003). Lopez et al. (2005) and Rabbi et al., (2014; 2015) have also reported on SNPs detection from ESTs in cassava. 60 Cassava being an outbreeding and highly heterogeneous crop, possesses an extreme level of phenotypic plasticity and thereby, lacks the potential for unified classification system for cultivars (Kawano et al., 1978). Consequently, characterization of agronomic traits becomes a challenge (Carvalho and Schaal, 2010). To conduct diversity studies on cassava germplasm in Sierra Leone, there is need to augment the existing collection with cassava germplasm collection. This engenders the need for assessing the existing collection to identify gaps that need to be filled. A collection was exploited to quantify the diversity of provitamin A cassava germplasm in Sierra Leone. One hundred and eighty-three provitamin A cassava accessions and five released varieties selected from clonal trials established at Taiama in 2014 (Personal communication with Dr. J.B.A. Whyte). The objectives of the present study were to: i) characterize the 183 provitamin A cassava germplasm and 5 released varieties using; a) morphological traits b) total-carotenoid content and c) SNP markers (ii) develop a collection for conservation. 4.2 Materials and Methods 4.2.1 Land preparation The land preparation was undertaken to control early weeds emergence and to ensure better crop establishment. The trial was conducted without supplemental irrigation and was weeded regularly until canopy closure. 61 4.2.2 Germplasm sources A total of 183 provitamin A cassava accessions with varying yellow colour were selected from the Sierra Leone’s germplasm development program (Table 4.1). Selections were made based on their performance in terms of storage root yields, dry matter content, and pest and disease tolerance/resistance, plant architecture, nutritional quality and flowering ability. Cocoa, SLICASS 4, SLICASS 6, SLICASS 7 and SLICASS 11 (cream fleshed) varieties were used as checks. 4.2.3 Experimental Design The trial was laid out in an Alpha lattice design with two replications at the Njala Agricultural Research Institute (NARC), Foya crop site, Njala, representing the transitional rain forest agro climatic zone (Van Vuure et al., 1972; Odell et al., 1974). Each replication had four blocks with 47 entries per block. The blocks were separated by 1m and 2m alleys between and within blocks to reduce intra and inter block plant competition respectively. Each entry was grown on 10m row ridge at a spacing of 1m x 1m between and within ridges, respectively. Cassava cuttings of 20-25 cm length were obtained from healthy stem cuttings and horizontally planted. The established trial was evaluated for one cropping season (2015-2016). 62 Table 4.1 Germplasm/accession and their pedigree Accession Accession Accession Accession Pedigree Pedigree Pedigree Pedigree Name Name Name Name TR 1563 IBA 082708 TR 0334 IBA 070675 TR 1389 IBA 083724 TR 1361 IBA 070557 TR 1337 IBA 011368 TR 1610 IBA 30572 TR 1259 IBA 070738 TR 0189 IBA 990313 TR 0421 IBA 051652 TR 0631 MM 090564 TR 1182 SM 3374 TR 1269 IBA 070593 TR 1207 SM 3374 TR 1233 SM 3374 TR 1543 IBA 102429 TR 1533 IBA 102429 TR 0267 IBA 961439 TR 0998 SM 3666 TR 0975 GM 3594 - 12 TR 1762 IBA 070749 TR 0626 MM 050626 TR 1744 IBA 070749 TR 1155 IBA 101438 TR 0015 IBA051740 TR 0431 IBA 011735 TR 1153 IBA 101438 TR 1404 IBA 083724 TR 0018 IBA051740 TR 0085 IBA 050311 TR 0886 IBA 102480 TR 1202 SM 3374 TR 1073 IBA 100224 TR 1295 IBA 011412 TR 0446 IBA 070620 TR 0955 IBA 101645 TR 0890 IBA 102480 TR 1627 TMEB 693 TR 0974 GM 3594 TR 0520 IBA 071313 TR 0316 IBA 050099 TR 0224 IBA 000351 TR 1565 IBA 082708 TR 1208 SM 3374 TR 1199 SM 3374 TR 1578 BA 011371 TR 0785 IBA 011206 TR 0843 SM 3444 TR 1144 IBA 100198 TR 0222 IBA 020134 TR 1569 IBA 082708 TR 1113 IBA 982101 TR 0982 GM 3594 - TR 1755 IBA 070749 TR 0713 SM 3434 TR 0893 IBA 102480 TR 1244 IBA 070738 TR 0854 KIBAHA TR 0423 IBA 071393 TR 1316 IBA 070520 TR 1279 IBA 070593 TR 1051 IBA 961089A TR 0887 IBA 102480 TR 0693 IBA 102286 TR 1008 SM 3666 TR 0261 IBA 961439 TR 1785 IBA 980505 TR 1593 IBA 30572 TR 0861 KIBAHA TR 1201 SM 3374 TR 0025 Z 960012 TR 1598 IBA 30572 TR 0983 GM 3594 TR 0894 IBA 102710 TR 1374 IBA 070557 TR 0282 IBA 050303 TR 1031 IBA 100403 TR 0232 BA 010169 TR 1562 IBA 082708 TR 1350 IBA 083849 TR 0683 IBA 102286 TR 1302 IBA 070520 TR 1236 IBA 082708 TR 0957 IBA 101645 TR 0772 IBA I011086 TR 1128 IBA 100198 TR 0838 SM 3444 TR 1422 IBA 102612 TR 1229 SM 3374 TR 1808 IBA 070539 TR 0688 IBA 102286 TR 0932 IBA 070337 TR 0118 IBA 970219 TR 1556 IBA 082708 TR 1480 IBA 980581 TR 1689 TMEB 2026 TR 0840 SM 3444 63 Table 4.1 Germplasm/accession and their pedigree (con’t) Accession Accession Accession Accession Pedigree Pedigree Pedigree Pedigree Name Name Name Name TR 0172 IBA 010732 TR 0937 IBA 101040 TR 1349 IBA 083849 TR 0396 IBA 070525 TR 0382 BA 011404 TR 0743 IBA 101094 TR 0927 IBA 070337 TR 1788 IBA 980505 TR 0384 IBA 011404 TR 1540 IBA 102429 TR 0810 IBA 011206 TR 0485 IBA 051654 TR 1688 TME B2026 TR 0747 IBA 100252 TR 0718 IBA 100449 TR 1152 IBA 101438 TR 1437 IBA 102612 TR 1348 IBA 083849 TR 0907 SM 3434 TR 0990 SM 3666 TR 0696 IBA 102286 TR 1438 IBA 102612 TR 0335 IBA 030007 TR 1004 SM 3664 TR 0033 IBA 050327 TR 1477 IBA 980581 TR 1327 IBA 070520 TR 0679 KALESO TR 1034 SM 3444- 2 TR 1243 IBA 070738 TR 1666 IBA 070703 TR 1515 IBA 102429 TR 0700 IBA 102286 TR 0807 BA 011206 TR 1748 IBA 070749 TR 1735 IBA 070749 TR 1463 IBA 980581 TR 0707 SM 3434 TR 0856 KIBAHA TR 1448 IBA 980581 TR 0365 IBA 011663 TR 1007 SM 3666 TR 1359 IBA 070557 TR 1322 IBA 070520 TR 1620 TMEB 693 TR 0299 IBA 051625 TR 0744 IBA 101094 TR 0399 IBA 070525 TR 0289 IBA 961632 TR 1289 IBA 011412 TR 0881 IBA 102480 TR 1525 IBA 102429 TR 1603 IBA 30572 TR 0851 SM 3444 TR 1405 IBA 083724 TR 1753 IBA 070749 - TR 1505 IBA 102429 TR 0295 IBA 051625 TR 0385 IBA 961551 TR 1501 IBA 102429 TR 1849 TME B778 - TR 1590 IBA 30572 TR 1223 SM 3374 TR 0019 IBA961039 TR 0031 IBA 050311 TR 0918 IBA 101803 TR 0868 KIBAHA TR 0296 IBA 051625 TR 0319 IBA 050099 TR 1133 IBA 100198 TR 1313 IBA 070520 TR 1502 IBA 102429 TR 1198 SM 3374 TR 1331 IBA 070520 TR 0480 BA 051654 Cocoa Local Cultivar TR 1256 IBA 070738 TR 0461 IBA 051654 TR 1266 IBA 070738 SLICASS 4 Released Variety TR 1557 IBA 082708 TR 1419 IBA 102612 TR 1071 IBA 100649 SLICASS 6 Released Variety TR 0535 IBA 020091 TR 0368 IBA 011663 TR 0703 SM 3434 SLICASS 7 Released Variety TR 1360 IBA 070557 TR 1527 IBA 102429 TR 0560 MM 980747 SLICASS 11 Released Variety 4.2.4 Molecular Characterization The Dellaporta method of DNA extraction (Dellaporta et al., 1983) was adopted at the Institute of Tropical Agriculture (IITA), Nigeria. Genomic DNA was extracted from young fresh leaves of the 188 cassava genotypes. 200 mg of young leaves tissue were harvested into a genogrinder tube with two steel balls inside. The tubes together with the leaves were freeze-dried inside a container of liquid nitrogen and later genogrinded at 5,000 rpm for 2min. 400 μl of DNA 64 extraction buffer was added and incubated at 65oC for 20min with 2min intermittent mixing. After incubation, 250μl of chilled 5M potassium acetate was added and iced for 20min. The 188 samples were removed from ice and centrifuged at 10,000 rpm for 10 min. The supernatant solution was transferred into a fresh tube, 450 μl cold isopropanol was added and mixed by inversion four to five times. The content stayed at -20oC for 5 min and later centrifuged at 10,000 rpm for 10 min. The retained pellets were allowed to dry for about 30 min at room temperature and later dissolved in 300 μl sterile distilled water. 300 μl of chloroform-isoamyalcohol was added, mixed and centrifuged at 10,000 rpm for 10 min. The supernatant solution was carefully transferred into another tube and 25 μl of 3M sodium acetate and 500 μl of ethanol were added. The tube was mixed by inversion and kept at -20oC for 1 hr. The samples were centrifuged at 10,000 rpm for 10 min. DNA pellets were washed twice in 70% solution and air dried at room temperature for 30 min. The DNA pellet in each genotype sample was re-dissolved in 250 μl sterile distilled water and kept in -20oC. After preparation of the samples plate and ID, 2μl of the DNA samples were loaded into the Nanodrop on a nucleic acid file for quality assessment. For genotyping-by-sequencing library preparation. the ApekI restriction enzyme (recognition site: G|CWCG) that produces less variable distributions of read depth and therefore a larger number of scorable SNPs in cassava (Hamblin and Rabbi, 2014) was used. Two 96-plex GBS libraries were constructed as described by Elshire et al., (2011) and sequenced at the Institute of Genomic Diversity at Cornell University using the Illumina HiSeq2500. Raw read sequences were processed through cassava GBS production pipelines developed using TASSEL 5.0 V2. The GBS-derived SNPs were further filtered using the TASSEL software (Bradbury et al., 2007) to retain only polymorphic SNPs. Initially filtered for minor allele frequency (MAF<0.05), the generated 5,634 SNPs were processed under the Next Generation Cassava project. The resulting 65 SNP dataset was used for the diversity analysis study among the 188 cassava accessions already phenotyped and analyzed. Results from both the phenotype and genotype analyses were compared to check the correspondence between the two. 4.2.5 Data Collection Data collection was undertaken at 1, 3, 6 and 9 months after planting (MAP), on the parameters listed below using the IITA cassava descriptor (Fakuda et al., 2010): Table 4.2 Parameters evaluated at 1, 3, 6 and 9 MAP Traits Parameter Leaf color Color of stem epidermis Number of leaf lobes Color of stem cortex Length of leaf lobe Growth habit of stem Width of leaf lobe Prominence of foliar scars Lobe margin Leaf retention Pubescence of apical leaves Level of branching Color of apical leaves Height at 1st branching Orientation of petiole Height at 2nd branching Petiole color Height at 3rd branching Leaf area Color of end branches of adult plant Length of stipule Percentage sprout Stipule margin African Cassava Mosaic Disease Stem color Cassava Green Mite Stem diameter base Cassava Anthracnose Disease Stem diameter-1foot below Harvesting was done at 12 MAP (August – September). The following parameters were taken at harvest: number of marketable roots (no), number of non-marketable roots (no), total number of storage roots (no), roots weight/tuber (kg), inner skin color, and outer skin color, ease of peel, root shape, marketable weight (kg), and non- marketable weight (kg). Dry matter content 66 expressed as a percentage was determined by selecting three representative storage roots. These were bulked, washed, peeled and sliced using knives. Slices were randomly selected and weighed to obtain a 100g fresh mass sample per genotype before being dried for 48hours in an oven at 80°C. The dried samples were then re-weighed to obtain the dry mass. 4.2.6 Total carotene determination The 188 provitamin-A cassava accessions were screened at harvest using a scale of 1 – 6 in the color chat .75 harvested provitamin-A accessions were selected from the field screening based on color chat. Five storage roots samples were taken to the nutritional laboratory for analysis. The fresh roots were washed with tap water to eliminate sand and other substances that could serve as contaminants. The roots were then peeled and placed on a chopping board, and divided into 4 equal longitudinal parts. The two opposite parts were pulled together. A portion was used for the iCheck analysis to determine total carotenoid content levels whilst the remaining was used for dry matter determination. The selected portion for the iCheck analyses were chopped into cubes mixed together and divided into 4 equal parts. The two opposite parts were again pulled together, mixed and divided as described above until a composite sample was obtained. The samples were processed further for iCheck analysis as follows: Five grams of each sample was weighed into a medium mortar and ground with 20ml of distilled water as the solvent. Each resultant solution was poured into a well labeled falcon tube. The falcon tubes were shaken vigorously to obtain homogenous slurry and 0.4ml of the suspension was injected into the iCheck reagent vials. The vials were shaken vigorously for 10sec and allowed to stand for 5min. The suspension inside the vial separated into two distinct phases (a clear upper phase and a turbid lower phase). Reading was done using the iCheck device. 67 4.2.7 Data Analysis Data sets from these trials were subjected to selected statistical packages for analysis. Analytical procedures comprised the following: descriptive statistics using XLSTAT (2010) and MINITAB 15 programs, multiple regressions analysis to identify variables that best discriminate among the classes, Principal Component Analysis (PCA) to examine the structure of the correlations between the variables. Correlation matrixes were used to visualize associations among variables and parameters (Daulfrey, 1976). Cluster analyses were performed to group observations together using the method of Ward’s minimum variance distance with SAS 9.4. A dendrogram was plotted from the computed similarity values to show the relationship among the accessions. The accessions were grouped based on the varying levels of total carotenoid content. Basic diversity indices for the population of the 188 cassava accessions were calculated using Power marker (Liu & Muse, 2005) and GenAlex version 6.41 (Peakall and Smouse, 2006). The Power maker software was used to generate the following statistics: number of alleles per locus, major allele frequency, observed heterozygosity (Ho), expected heterozygosity (He) and polymorphic information content (PIC) (Bostein and White, 1980). PIC values were calculated with the equation: PIC=1-ΣP2i－Σ 2P2i Where: ΣP2i = sum of each squared ith haplotype frequency. A Mantel matrix test (Mantel, 1967) was carried out to compare the extent of agreement between dendrograms derived from morphological and molecular data using the distance matrices. The pairwise genetic distance (identity-by-state, IBS) matrix was calculated among all individuals using PLINK (Purcell et al., 2007). A Ward’s minimum variance hierarchical cluster 68 dendrogram was built from the IBS matrix using the analyses of phylogenetic and evolution (ape) package in R. 4.3 Results 4.3.1 Descriptive statistics for the 188 cassava accessions Significant differences were observed among the 188 accessions for all the measured traits. Severity scores for African Cassava Mosaic Disease, Cassava Bacteria Blight and Cassava Green Mite variably ranged from 1 to 4 in the studied population. Yield per hectare ranged from 0.2 to 42.5t/ha while dry matter content ranged from 4.0 to 44.5% (Table 4.3) Table 4.3 Descriptive statistics of some morpho-agronomic traits of 188 cassava accessions Descriptive statistics Time of data Trait collection Standard (MAP) Minimum Maximum Mean Deviation Sprouting (%) 1 6.5 10 9.56 0.6 ACMD Incidence (%) 1, 3, 6 and 9 0 4.25 0.08 0.42 ACMD Severity (score) 1, 3, 6 and 9 0.75 2 1.04 0.13 CAD Incidence (%) 1, 3, 6 and 9 0 2.75 0.11 0.41 CAD Severity (score) 1, 3, 6 and 9 0.5 2.75 1.05 0.23 CBB Incidence (%) 1, 3, 6 and 9 0 4 0.41 0.6 CBB Severity (score) 1, 3, 6 and 9 0.5 4.5 1.15 0.34 Mealybug incidence (%) 9 0 9 3.22 2.17 Mealybug severity (score) 9 1 6.5 2.54 0.84 CGM Incidence (%) 9 2 8 5.27 1.66 CGM Severity (score) 9 2 9 3.31 0.72 Colour of apical lobe (score) 3 3 9 6.8 1.61 Colour of apical lobe (score) 9 0 9 6.71 1.74 Plant height (cm) 6 65.5 284.5 155.69 26.12 Height of branching (cm) 6 37 196.5 85.83 29.38 Stem diameter base (cm) 6 1.07 3.94 1.51 0.26 Stem diameter (mid height) (cm) 6 1.03 2.25 1.53 0.2 Leaf area (cm2) 6 10.24 73.93 34.13 11.04 Leaf retention (score) 6 1.75 4.5 2.87 0.5 Shape of central leaflet (score) 6 1.75 6.25 3.13 0.94 69 Table 4.3 Descriptive statistics of some morpho-agronomic traits of 188 cassava accessions (con’t) Descriptive statistics Time of data Trait collection Standard (MAP) Minimum Maximum Mean Deviation Petiole colour (score) 6 0.5 7 1.94 1.48 Petiole colour (score) 9 1 8 3.2 1.54 Leaf colour (score) 6 1.5 5 3.69 0.87 Leaf colour (score) 9 3 6 3.94 0.77 Colour of leave vein (score) 6 3 18.75 3.85 1.73 Petiole length (cm) 6 3 32.95 14.79 6.09 Orientation of petiole (score) 6 0.5 7 2.55 1.13 Number of leaf lobes (no) 6 3.75 8 6.18 0.89 Length of leaf lobe (cm) 6 3.13 15.15 11.15 1.61 Width of leaf lobe (cm) 6 1.08 7.05 3.05 0.81 Lobe margin (score) 6 1.5 8 4.38 1.87 Length of stipules (cm) 9 1 4 2.97 0.22 Stipule margin (score) 9 1 5 1.31 0.59 Prominence of foliar scars colour (score) 9 3 6 4.93 0.39 Stem colour (score) 6 4 8 6.47 0.79 Colour of stem exterior (score) 9 1 7 2.55 0.71 Colour of stem epidermis (score) 9 4 8.5 6.52 1.1 Colour of end branches of adult plants 9 1 32.5 4.62 2.47 (score) Stem colour (score) 6 4 8 6.47 0.79 Color of stem exterior (score) 9 1 7 2.55 0.71 Color of stem epidermis (score) 9 4 8.5 6.52 1.1 Color of end branches of adult plants 9 1 32.5 4.62 2.47 (score) Mean number of storage root (no) 12 7.5 88 44.83 14.21 Yield (t/ha) 12 0.24 42.5 12.09 5.69 Mean weight per storage root (kg) 12 0.09 28 0.47 2.62 Dry matter content (%) 12 4 44.5 29.56 6 Root size (score) 12 2 7 4.93 1.07 Root shape (score) 12 1 5 2.76 0.62 Outer root colour (score) 12 1 4 3.4 0.72 Inner root colour (score) 12 1 3 1.9 0.36 Pulp colour (score) 12 1 3 2.01 0.19 Ease of peeling (score) 12 2 7 2.83 0.53 Biomass (kg) 12 2.5 13.5 9.99 1.91 Field carotene (score) 12 1 4.5 3.34 0.72 Total carotene level (µg/g fresh root 12 3.6 13.7 7.6 2.74 weight) MAP= Month after Planting, ACMD=African cassava mosaic Disease, CAD= Cassava Anthracnose Disease, CBB-Cassava Bacterial Blight, CGM-Cassava Green Mite 70 4.3.2 Multiple Regression of yield on agro-morphological traits Marketable (MWET) and non-marketable weights (NMWET) were the most discriminating characters among the accessions with a very high b value of 1.43 (Table 4.4). Table 4.4 Multiple regression coefficients of yield on some agro-morphological traits of 188 cassava accessions Traits b± Se ACMDI (1MAP) 0.003±0.001 ACMDI (3MAP) 0.000±0.001 LA (9MAP) -0.000±0.000 MROT 0.000±0.000 NMROT 0.000±0.000 TSR 0.000±0.000 MWET 1.429±0.000** NMWET 1.43±0.000** TWET 0.000±0.000 WSROT 0.002±0.003 DM -0.000±0.000 **Significant at 1%, ACMDI=African cassava mosaic disease incidence, LA=Leaf area, MROT=Marketable roots, NMROT=Non-marketable roots, TSR= Total no of storage roots, MWET=marketable weight, NMWET=Non- marketable weight, WSROT=Storage root weight, DM=dry matter 4.3.3. Correlations among morphological traits of 188 cassava accessions Table 4.5 shows 49 Pearson’s correlation coefficients among some measured traits. Yield was positively associated with storage root weight (r=0.51**) and root size (r=0.45**). Correlation among yield and all other traits were not significant. Significant correlation coefficients were however observed between INCOL with BIOMASS (0.32*) and WSROT with BIOMASS (0.48*). 71 Table 4.5 Correlations among morphological traits of 188 cassava accessions Variables YLD WSROT DMC RZ RS OCOL INCOL PCOL EPEEL WSROT 0.51** DMC 0.12 0.09 RZ 0.45** 0.16 0.04 RS 0.00 0.04 -0.01 -0.19 OCOL -0.03 0.08 0.11 -0.07 -0.00 INCOL -0.06 0.24 -0.06 -0.15 0.04 0.05 PCOL 0.14 0.11 0.01 0.20 0.08 0.02 -0.3 EPEEL 0.06 0.04 0.01 0.18 0.05 -0.02 0.09 -0.05 BIOMASS -0.08 0.48** -0.12 0.02 -0.08 -0.05 0.32* 0.01 0.05 YLD=Yield; WSROT=Storage root weight; DM=Dry matter; RZ=Root size; RS= Root shape; OCOL=Outer color; INCOL=Inner color; PCOL= Petiole color; EPEEL= Ease of peel; Biomass, **=significant at1% level 4.3.4. Principal component analysis of yield and yield related traits of 188 cassava accessions The first four PCs together accounted for 51.50 % of the total phenotypic variation among the 188 accessions (Table 4.6). PC1 axis had an eigenvalue of 1.92 and acounted for 17.5%, of the total variation whereas PC2, PC3 and PC4 axes had eigenvalues of 1.43, 1.20 and 1.11, and acounted for 13.00%, 10.54% and 10.10% of the total variation respectively. Storage root weight, root size and yield had postive loadings on PC1, whereas root shape had negative loading. Dry matter content and root shape had positive loadings on PC2 , while RCOL had a negative loading. Storage root weight, root size, root shape, outer skin colour and pulp colour had positive loadings in PC3. PC4 axis contained negative loadings for dry matter, root size and outer color with positive loadings for root shape, pulp color and ease of peel. 72 Table 4.6 Principal component analysis of yield and yield related traits of 188 cassava accessions Principal Component Analysis Variables PC1 PC2 PC3 PC4 YLD 0.6 0.03 0.06 0.03 WSROT 0.46 -0.02 0.36 -0.06 DMC 0.19 0.31 0.09 -0.43 RZ 0.50 -0.17 0.32 -0.45 RS -0.50 0.14 0.48 0.37 OCOL 0.03 0.04 0.49 -0.53 PCOL -0.12 -0.12 0.38 0.35 EPEEL 0.23 -0.11 0.20 0.43 BIOMASS 0.15 -0.26 -0.03 0.04 RCOL -0.06 -0.62 0.07 0.02 Eigenvalue 1.92 1.43 1.20 1.11 Variance % 17.50 13.00 10.90 10.10 Cumulative % 17.50 30.5 41.40 51.50 YLD=Yield; WSROT=Storage root weight; DMC=Dry Matter; matter content RZ=Root size; RS= Root shape; OCOL=Outer color; INCOL=Inner color; PCOL= Petiole color; EPEEL= Ease of Peel; Biomass; RCOL= Root color 4.3.5. Cluster analysis of the accessions based on Ward’s minimum variance and total carotenoid content Figure 4.1 revealed that the semi partial R – squared derived from the cluster analysis ranged from 0.00 to 0.15. At 0.04 eight clusters were observed suggesting a moderate to high diversity among the accession. Table 4.7 showed the 188 accessions grouped into eight clusters. The number of accessions per cluster varied from 1 in Cluster C to 56 in cluster D. Clusters A, B, E, F and G had 2, 43, 51, 5 and 28 accessions respectively. Only 1 out of the initial 2 accessions in Cluster H was identified as a carotenoid accession whereas, 11, 34, 27, 2 and 2 accessions in clusters B, D, E, G and H had higher levels of total carotenoid content. Cluster F did not have any accession with measurable total carotenoid content. The one accession in cluster C could be considered as an outlier or a unique accession. Thirty provitamin A accessions with higher levels of total carotenoid were selected from five out of the eight clusters. These included 3 from Cluster B, 13 from cluster D, 9 from E, 4 from G and 1 from H. Accessions TR-0399, TR-0707, 73 TR-0222, TR-1337, TR-1569, TR-1313, TR-0998, TR-1755, and TR-1557 had higher total carotenoid content levels of 11.1, 10.9, 13.1, 11.8, 10.3, 11.7, 13.7, 10.7, and 11.2 mg/100g-1 . 74 Figure 4.1 Dendrogram of 188 cassava accessions based on morpho-agronomic traits using Ward’s minimum Variance 75 Table 4.7 188 cassava accessions grouped into 8 clusters based on total carotenoid content C=A C=B C=B C=C C=D C=D C=E C=E C=F C=G C=H TR 1259 TR 0382 TR 0703 TR 0700 TR 0982 TR 1289 TR 0868 TR 0957 TR 0480 TR 0384 Cocoa TR 1808 TR 1361 TR 1133 TR 0693 TR 1501 TR 0881 TR 0520 TR 1533 TR 1688 TR 0843 TR 0296 SLICASS 11 TR 1244 TR 1525 TR 1279 TR 0983 TR 0890 TR 0399 TR 1034 TR 0679 TR 0893 TR 1557 TR 1515 TR 1540 TR 1788 TR 0955 TR 0707 TR 0025 TR 1556 TR 1152 TR 0232 SLICASS 4 TR 1603 TR 0485 TR 1144 TR 1437 TR 0019 TR 1243 TR 1182 TR 1256 TR 1359 TR 0990 TR 0368 TR 1313 TR 1419 TR 1543 TR 0446 TR 0854 TR 0261 TR 0267 TR 1071 TR 1448 TR 0743 TR 1610 TR 1348 TR 0927 TR 0431 TR 0718 TR 0907 TR 1735 TR 1762 TR 1051 TR 0316 TR 1128 TR 0886 TR 1198 TR 0031 TR 0535 TR 1327 TR 1748 TR 1207 TR 0807 TR 1236 TR 0299 TR 0626 TR 1374 TR 0688 TR 1004 TR 0840 TR 1438 TR 1755 TR 1593 TR 0937 TR 0918 TR 1480 TR 1199 TR 0772 TR 0015 TR 0785 TR 1785 TR 0319 TR 0365 TR 0018 TR 0861 TR 0975 TR 1229 TR 1463 TR 0747 TR 1269 TR 1527 TR 1502 TR 0385 TR 1562 TR 1155 TR 1223 TR 0172 TR 1316 TR 1569 TR 0998 TR 1565 TR 0334 TR 0894 TR 0838 TR 1350 TR 0118 TR 0713 TR 1073 TR 0189 TR 1007 TR 1590 TR 0696 TR 0289 TR 0887 TR 0396 TR 1266 TR 1113 TR 0295 TR 1322 TR 0932 TR 1295 TR 1578 TR 1849 TR 1689 TR 1477 TR 1505 TR 1331 TR 1598 TR 0974 TR 0851 TR 0744 TR 0033 TR 1620 TR 0085 TR 0631 TR 1666 TR 0421 TR 1208 TR 1201 TR 1405 TR 1233 TR 1389 TR 0560 TR 1744 TR 1337 TR 1349 TR 1404 SLICASS 7 TR 1031 TR 1008 TR 1753 TR 0282 TR 1360 TR 1202 TR 0683 TR 1153 TR 0222 TR 0423 TR 0810 TR 1627 TR 1302 TR 0224 TR 0335 TR 0856 TR 1422 TR 1563 TR 0461 SLICASS 6 76 4.3.6 Clusters mean and standard deviation of provitamin A cassava accessions Results in Table 4.8 shows the different cluster groups for provitamin A containing cassava accessions with their means and standard deviations. Cluster D has the highest number of total carotenoid containing accessions with a mean and standard deviation of (7.3±2.9), while two clusters G and H had the lowest number of total carotenoid accessions with a mean and standard deviation of 10±00 and 9.3± 2.6, respectively. Clusters A, C and F did not have carotenoid cassava accessions. Table 4.8 Number of accession/clusters, mean and standard deviation for provitamin A content Cluster Name Accessions TCC Mean and Standard Deviation D 34 7.3±2.9 E 27 8.1±2.6 B 11 7.2±2.3 G 2 10±00 H 2 9.3±2.6 4.3.7 Summary statistics of genetic variation of the accessions using SNP markers Summary statistics for number of alleles observed, expected heterozygosity and polymorphic information content are presented in Appendix 2. An average of 2 alleles per locus was observed from the analysis. The expected heterozygosity was lowest for TR 1233 (0.15) and SLICASS 6 (0.15) and highest in TR 1525 (0.23) with a mean of 0.36. The observed heterozygosity per individual observation ranged from 0.30 (TR 1233) to 0.47 (TR 1525) with a mean of 0.38. The mean of observed heterozygosity was higher than the expected heterozygosity. This substantiates the heterozygous nature of most of the accessions and the fact that cassava is largely cross-pollinated. However, the major allele frequency of all the markers used in the observations was generally below 0.95, indicating that they were all polymorphic. PIC values 77 ranged from 0.11 in TR 1233 to 0.18 in TR 1199 and TR 1525. The mean PIC is 0.28. The higher the PIC value the more informative is the marker. 4.3.8. Cluster groupings of the 188 cassava accessions based on SNP markers. The 188 accessions were grouped into 9 clusters based on the 5,643 SNP markers (Figure 4.2). Clusters A, B, C, D, E, had 21, 7, 10, 16 and 11 accessions, while cluster F, G, H and I consisted of 50, 44, 12 and 7 accessions. Clusters A, B, D, E, F, G, H and I had 5, 3, 7, 6, 20, 23, 5 and 3 accessions with varying levels of total carotenoid content. Cluster C did not have any carotenoid accessions. A comparison of the two dendrogram based on Mantel matrix test showed a significant positive but weak correlation between the morphological and molecular data sets (r = 0.104;P<0.034) 78 C D E B F A I G H Figure 4.2 Dendrogram of 188 cassava accessions based on SNP Markers 79 Table 4.9 Cluster groupings of the 188 cassava accessions based on SNP markers Cluster A Cluster B Cluster C Cluster D Cluster E Cluster F Cluster F Cluster G Cluster G Cluster H Cluster I TR 0018 TR 0679 TR 0918 TR 1755 TR 0299 TR 0626 TR 0172 TR 1389 TR 0267 TR 0907 TR 1269 TR 0222 TR 0282 TR 0772 TR 1337 TR 0718 TR 0118 TR 0840 TR 1244 TR 1051 TR 0485 TR 0693 TR 1788 SLICASS 4 TR 1590 TR 0421 TR 1808 TR 1229 TR 1259 SLICASS 6 TR 1502 TR 1201 TR 1202 TR 1155 TR 1505 TR 1620 SLICASS 7 TR 1661 TR 0683 TR 1525 TR 1327 TR 1031 TR 1610 TR 0851 TR 1556 TR 1534 TR 0295 TR 0886 TR 0713 TR 1256 TR 0937 TR 0932 TR 1349 TR 1313 TR 1182 TR 3168 TR 1735 TR 1753 TR 0261 01/1635 TR 1128 TR 0431 TR 0856 TR 1289 TR 1223 TR 1849 TR 0894 TR 0747 TR 0703 TR 1198 TR 0335 TR 0399 TR 0461 TR 0224 TR 0957 TR 1266 TR 1477 TR 0890 TR 0085 TR 0744 TR 0975 TR 1593 TR 0365 TR 1322 TR 0560 TR 0990 TR 1296 TR 0033 TR 0868 TR 1688 TR 1302 TR 0535 TR 0289 TR 1569 TR 0688 TR 0881 TR 1331 TR 0189 TR 0743 TR 0927 TR 1152 TR 0025 TR 1359 TR 1557 TR 0843 TR 1243 TR 1034 TR 1236 TR 1562 TR 0700 Cocoa TR 1603 TR 1071 TR 0019 TR 1689 TR 0998 TR 1073 TR 1437 TR 0983 TR 0446 TR 1527 TR 1361 TR 1762 TR 1374 TR 1533 TR 0015 TR 1627 TR 1563 TR 0520 TR 0232 TR 1360 TR 1008 TR 0893 TR 1540 TR 1004 TR 1422 TR 0631 TR 1348 TR 0838 TR 1785 TR 1279 TR 1405 TR 0031 TR 0319 TR 1133 TR 1598 TR 0480 TR 1113 TR 0316 TR 1007 TR 1565 TR 1438 TR 0974 TR 0955 TR 0907 TR 1208 TR 1350 TR 1748 TR 0795 TR 0334 TR 1578 TR 1480 TR 1144 TR 0887 TR 1199 TR 0810 TR 0296 TR 1744 TR 0696 TR 0382 TR 0861 TR 0384 TR 0747 TR 1405 TR 1448 TR 0423 TR 1543 80 4.3.9 Thirty provitamin-A cassava accessions with varying levels of total carotenoid, yield and dry matter content based on morphological and molecular clustering Analyses. Thirty provitamin A accessions with higher levels of total carotenoid were selected from five out of the eight and eight out of the nine clusters derived from morphological and molecular (SNPs) clustering analyses with some showing appreciable high dry matter content and yield for the multilocational testing across the three environments (Table 4.10). Although TR-1208, TR-1152 and TR-0713 were selected and formed part of the carotenoid accessions, they had the lowest levels of total carotenoid content of 8.9, 9.9 and 8.2 respectively. 81 4.10 Thirty provitamin-A cassava accessions with varying levels of total carotenoid, yield and dry matter content Total Phenotypic Genotypic Dry matter Accession Yield carotenoid cluster name cluster name content content (µg-1) TR 0747 B B 4.3 29.5 10.9 TR 0365 B F 2.3 25.5 7 TR 0560 B G 7.5 25.5 9.7 TR 1208 D G 7.5 39.5 8.9 TR 0461 D F 2 23 11.5 TR 1337 D D 14.6 25.5 11.8 TR 1569 D G 21.8 26.5 10.3 TR 0683 D F 5 28.5 10.2 TR 1198 D D 7 28.5 10.8 TR 1313 D H 11 35 11.7 TR 0696 D F 6.5 12.5 11.1 TR 1322 D G 13 29.5 9.9 TR 1350 D G 8 29.5 9 TR 0907 D H 6 31.5 9.1 TR 1557 D A 10.6 18 11.2 TR 1152 D G 4.8 33 8.08 TR 0232 E G 22.8 27 9.9 TR 1279 E G 6.3 35.6 9.1 TR 0031 E A 6.9 29.5 10.3 TR 0222 E A 7.8 37 13.1 TR 0998 E H 2.8 38.1 13.7 TR 1755 E D 5.3 24 10.7 TR 1182 E I 10.8 24 10.4 TR 1753 E B 16.8 35 8.6 TR 0713 E E 7.5 28 8.2 TR 0423 G F 6.5 25.5 8.7 TR 0384 G F 5.5 27 10.6 TR 1327 G G 4.5 21 11.1 TR 0399 G F 11.8 25.5 11.1 4.4 Discussion Significant variation observed among the economically important traits such as African cassava mosaic disease, yield and dry matter content (DMC) among the 188 accessions studied offers a prospect for progress in cassava breeding program in Sierra Leone. The marketable tuber weight which is positively associated with yield and commonly preferred by farmers to increase their income and enhance their livelihood showed the highest variability which can be used for yield improvement. Similarly, the non-significant very low negative correlation (-0.06) between dry 82 matter content and inner color (INCOL) in the current study does appear to be contrary to findings from Latin American germplasm evaluated at CIAT (Esuma et al., 2016). It is worth noting that combined selection for both total carotenoid content (INCOL) and dry matter in Latin America has been underway much longer than in Africa and probably explains why the Latin American yellow cassava also have high dry matter content. Thus, such negative correlations could be lost during the several cycles of recombination (Ceballos et al., 2013 and Esuma et al., 2016). Diversity studies of cassava germplasm has been widely undertaken worldwide (Bolanos, 2001; Chavez et al., 2005; Morillo et al., 2009; Fregene, 2007; Parkes, 2011; Njoku, 2012 and Thompson, 2013) with little or no attention in Sierra Leone. In the present study, descriptive analysis of the cassava accessions based on selected traits showed the existence of high variability among the accessions. These findings were in confirmation with the findings of Carvalho and Schaal (2001) who reported a high degree of variability among 94 cassava accessions of Brazilian origin. Raghu et al. (2007) in a similar study also identified a high level of diversity among 58 cassava accessions based on 29 morphological traits. Lyimo et al. (2012) reported significant variability among 39 cassava accessions of Tanzanian origin using 14 morphological traits. Thompson (2013) observed a moderate to high diversity among 150 accessions using 25 morphological traits in Ghana. In a similar study, Raghu et al. (2011) mentioned that 24 morphological traits out of 28, contributed to the total variation observed. In the present study, clustering based on similarity index of both qualitative and quantitative traits grouped the 188 cassava accessions into 8 and 9 distinct clusters based on morphological and molecular analyses respectively. In a similar study, Carvalho and Schaal (2001) identified 22 distinct clusters using 94 cassava accessions. Raghu et al. (2007) also identified six distinct groups using 58 accessions. Since morphological traits are influenced by the environment, 83 molecular markers which are not influenced or controlled by the environment are preferable in genetic diversity studies (Kaemmer et al., 1992; Gepts, 1993; Njoku, 2012 and Thompson, 2013). The study by Kawuki et al. (2009) was the first published report where SNPs were used for diversity studies in cassava. They identified, characterized some SNP markers and assessed their utilization in cassava diversity assessment. The present study seems to be first reported case in Sierra Leone where SNP markers have been exploited in cassava diversity study on provitamin A cassava accessions. Using the 5,634 SNP markers, 95% of them were polymorphic. The informativeness of a genetic marker is measured by the polymorphic information content (PIC). The Mean PIC value observed for this study was 0.28. Kawuki et al. (2009) reported a PIC value of 0.29 in 74 cassava accessions using 26 SNP while Thompson (2013) also reported PIC value of 0.29 using 150 cassava accessions. PIC values for SNP markers are generally low as observed in genetic diversity studies in other crops. For instance, Yang et al. (2011) reported PIC value of 0.34 in maize genotypes using 884 SNP markers. Although morphological and SNP data grouped the accessions into eight and nine distinct clusters respectively, some similarities were observed. Related accessions were grouped in the same cluster while unrelated accessions were grouped in separate clusters. Accessions TR 1337, TR 1198, TR 0747, TR 0713 and TR 1327 were both morphologically and genetically similar. The significant positive but low correlation (r = 0.104; p < 0.034) observed between the two dendrograms revealed by the Mantel matrix test could be attributed to the relatedness of the accessions within the studied population. This could explain why the morphological and molecular analysis showed similar accessions between the two clusters. 84 4.5 Conclusion The present morphological and molecular assessment studies showed that provitamin A cassava accessions in Sierra Leone have moderate to high diversity based on total carotenoid content, based on morphological and molecular assessment studies. The results obtained will serve as a guide and basis of germplasm management and improvement for total carotenoid content, yield and African cassava mosaic disease resistance. The diversity of the provitamin A cassava accessions was sufficient to enable the creation of a collection of 30 provitamin A cassava accession with diverse genetic background from the different cluster groups. 85 CHAPTER FIVE 5.0 GENOTYPE BY ENVIRONMENT INTERACTION ANALYSIS OF PROVITAMIN A CASSAVA IN SIERRA LEONE 5.1 Introduction Plant breeders have established that the expression of quantitative traits such as crop yield is controlled not only by the genetic make-up of the variety but also the environment (Carpena et al., 1982). Dixon and Nukenine (2000) defined genotype by environment interaction (GEI) as the change in a cultivar’s behavior over environments, from differential response of the cultivar, to various edaphic, climatic and biotic factors. Assessment of GEI effects for a given trait is therefore useful in understanding varietal stability (Acquaah, 2012). A significant GEI limits the usefulness of superior genotypes. Breeders address the GEI challenge by evaluating genotypes in multiple environments to select better adapted genotypes with high and stable performance in different environments (Fakuda et al., 2002, Nassar and Ortiz, 2006, Esuma et al., 2016). The GGE biplot is a decision-making tool that allows identification of stable and good performing genotypes in test environments towards subsequent release (Farshadfar et al., 2013 and Rao et al., 2011). GGE biplot is a data visualization tool, which graphically displays GEI in a two-way table (Yan, 2001). GGE biplot is an efficient tool for mega –environment analysis (e.g. which won where), thereby specific genotypes can be recommended to a special mega –environments (Yan and Tinker, 2006); genotype screening (the mean performance and stability), and environmental assessment provides the power to discriminate among genotypes in preferred environments). GGE biplot analysis is increasingly being used in GEI data analysis in agriculture (Yan, 2001; 86 Crossa et al., 2010; Yan and Hunt ,2002; Yan and Tinker, 2005; Samonte et al., 2005; Dehghani et al., 2006; Yan and Tinker, 2006). Winning genotypes and mega-environments in a polygon-view of a GGE biplot is the most effective approach in identifying winning genotypes and mega environments. A polygon is drawn by joining the genotypes that are located farthest from the biplot origin, while all other genotypes are contained in the polygon. Mean performance and stability of genotypes can be visualized on a GGE biplot by drawing an average environment coordinate (AEC) on genotype- focused biplot. The AEC is decomposed into two axes, which are perpendicular to each other, the abscissa and the ordinate. Evaluation of test locations is done by defining three parameters, namely: the ability to discriminate between genotypes (discrimination ability), the ability to represent the target environment (representativeness) and the biplot distance from an ideal location (desirability index) (Xu et al., 2013). Provitamin A cassava genotypes have featured so distinctly in biofortification because they have an increasing level of micronutrients, such as carotenoids (Iglesias et al., 1997; Chávez et al., 2005; Ssemakula et al., 2007 and Esuma et al., 2016). However, the acceptance of micronutrient biofortified genotypes largely depends on their agronomic qualities eg, including, dry matter content, fresh root yield, resistance to major pests and diseases, and the stability of these traits over time and space. Though cassava is widely adapted to a variety of environmental conditions, usually the adaptability of most white flesh varieties is narrow with large GEI effects (Dixon et al, 1994b; Dixon and Nukenine, 1997). Due to poor social infrastructure and high poverty levels, food fortification and supplementation have been less impacting (Boy et al., 2009, Mayer et al., 2008, Thompson and Amoroso, 2011). The national cassava breeding programme in Sierra Leone has initiated new strategies geared towards developing and advancing provitamin A 87 cassava that incorporates farmer preferred traits, especially high dry matter content and fresh storage root yield. It is envisioned that this initiative will culminate in deployment of provitamin A cassava varieties for purposes of improving the nutritional status of populations that depend upon cassava as a major staple. The objectives of this study were to i) determine the GEI among selected provitamin A accessions. ii) identify and select stable genotypes with high carotenoid levels and dry matter. 5.2 Materials and Methods 5.2.1 Experimental materials Genotypes used for the study are the 30 selected provitamin A cassava of high carotenoid content selected from the collection established in chapter 4. 5.2.2 Experimental sites and design Trials were planted for one season beginning in October, 2016 to September 2017, at three locations in three different agroecological zones namely; Kambia (north savannah grassland), Njala (Transitional Rain Forest) and Pendembu (Forest Zone). The experiment and climatic information are showed in Figure 5.1 and Table 5.1 respectively. The trials were laid out in a randomized complete block design using three rows of 8m per plot with three replications. 5.2.3 Planting Planting was done at a spacing of 1 × 1 m, giving a density of 10,000 plants ha−1. To increase the chances of sprouting and uniform plant establishment, all stakes used for planting were taken 88 from the middle portions of mature stems. Adjacent plots were separated by 2 m alleys. Weeding was done as necessary. Figure 5.1 Map of Sierra Leone; showing trial sites 89 Table 5.1 Climatic data of the three experimental sites Agro-climatic Regions Characteristics Kambia Njala Pendembu Mean temperature. (OC) 28.2 28.5 28.6 Rainfall (mm) 2280 2730 2660 Altitude (m) 150 – 300 150 - 300 300 – 600 Lophira savannah, savannah woodland, Savannah woodland, Dominant vegetation mixed tree savannah, montane grassland, and Forest and forest re-growth. upland grassland, and forest re-growth. forest re- growth. Drainage depressions, Plateau with undulating Plateau with undulating Dominant land form undulating plains, high lying plains, and plains, rolling plains, and low plateau, and rolling hills. hills. hills. Average length of growing 255 ± 10 270 – 300 314 ± 9 period (days) 5.2.3 Data collection Data were collected at various crop growth stages on morphological parameters using the IITA Cassava descriptor by Fakuda et al., (2010). At harvest (11 months after planting), the following data were taken, number of marketable roots (MKR), non-marketable roots (NMKR) marketable weight (MKW), nonmarketable weight (NMKW), total carotenoid content (TCC), fresh root yield (FRYD) and dry matter content (DMC). Data were taken on eight plants from the inner rows of each experimental plot. Samples were taken from five roots out of the eight harvested plants to measure dry matter content (DMC) as total carotenoid content (TCC). 90 5.2.4 Data Analysis Data were subjected to statistical analyses. GGE Biplot analysis was used to determine GGE interaction and stability performance (Yan and Hunt, 2002). The GGE biplot model equation is as follows; Yij − yj = λl ξil ηjl + λ2 ξi2 ηj2 + εij where yij = is the mean yield of genotype, i = in environment j, yj = is the mean of genotypes in environment j, λl and λ2 = are the eigenvalues for PC1 and PC2, ξil and ξi2 = are the scores of genotype i, ηjl and ηj2 = are the scores for environment j, εij= is the residual term related to the mean of genotype i in environment j. GGE biplots (version 4.1) was used for graphical analysis to identify genotypes with broad or specific adaptation to target environments. 5.3. Results 5.3.1. Analysis of variance for multilocational trials across 3 locations for dry matter and total carotenoid contents. The results of the analysis of variance are shown in Table 5.2. Differences in DMC and TCC were not significant among the genotypes. However, significant environmental differences were observed among the three sites for both traits. Genotype x environment interaction was significant for only DMC. 91 Table 5.2 Analysis of variance dry matter and total carotenoid contents Source Df Dry matter content Total carotenoid content Genotype (G) 29 15.09 3.04 Environment (E) 2 767.08** 52.21** G*E 58 45.45* 3.3 Error 161 32.49 3.2 5.3.2 Performance of 30 provitamin A cassava evaluated across 3 test environments for dry matter content and total carotenoid content Results show significant differences among genotypes across the 3 environments for TCC but not for DMC (Table 5.3). Generally dry matter content varied between 35% and 54% across the three sites. Mean dry matter content above 50% were recorded for 13 genotypes across the 3 sites. TR 0384 (54.39%), TR 0232(53.26%) and TR 0998(42.63%) had the highest dry matter content in Pendembu, Njala and Kambia, respectively. Although TR 0998 had the highest dry matter in Kambia, it had very low values in Njala and Pendembu. Total carotenoid content for the genotypes for all the 3 sites varied between 5 and 13 µg -1 of fresh weight. TR 1313 had the highest level of carotenoid content with a mean of 10.6 µg -1 of fresh weight across 3 locations followed by TR 0683 and TR 1152 with 10.28 µg -1 of fresh weight and 10.20 µg/g of fresh weight, respectively. The accessions with the lowest levels of total carotenoid in each location were TR 0423 (Pendembu), TR 1337 (Njala) and TR 1279 (Kambia). Accession with highest total carotenoid content were recorded for TR 0998 (Pendembu), TR 0222 (Njala) and TR 0907 (Kambia) with total carotenoid content of 13.02, 11.34 and 11.24 µg -1 of fresh weight, respectively. 92 Table 5.3 Mean performance of 30 provitamin A cassava accessions evaluated across three environments Entry Dry matter content Total carotenoid content (%) (µg-1) No Genotype Kambia Njala Pendembu Mean Kambia Njala Pendembu Mean 1 TR 0747 43.98 42.20 44.51 43.56 10.65 9.61 10.07 10.11 2 TR 0365 42.73 44.22 45.84 44.26 7.69 10.79 11.60 10.03 3 TR 0560 39.39 48.44 48.20 45.34 9.95 7.85 9.03 8.94 4 TR 1208 45.79 51.51 44.47 47.26 9.54 10.37 10.13 10.01 5 TR 0461 40.23 46.67 45.19 44.03 7.07 9.72 10.64 9.14 6 TR 1337 39.55 50.24 47.01 45.60 9.76 6.59 10.13 8.83 7 TR 1569 43.39 47.81 42.87 44.69 7.66 8.97 10.62 9.08 8 TR 0683 40.74 45.93 48.91 45.19 9.04 10.37 11.44 10.28 9 TR 1198 47.55 45.78 46.59 46.64 5.77 8.22 10.79 8.26 10 TR 1313 39.39 45.55 51.10 45.35 9.47 9.98 12.40 10.62 11 TR 0696 41.21 47.84 44.14 44.40 8.53 9.24 10.36 9.38 12 TR 1322 41.50 45.98 45.03 44.17 5.82 9.44 11.17 8.81 13 TR 1350 44.84 41.47 44.05 43.45 7.83 10.43 10.48 9.58 14 TR 0843 38.04 44.29 51.17 44.50 8.23 8.66 10.17 9.02 15 TR 0907 43.16 44.99 51.15 46.43 11.24 9.50 9.16 9.97 16 TR 1557 42.35 44.54 48.12 45.00 7.83 7.63 9.54 8.33 17 TR 1152 39.94 46.38 42.59 42.97 10.02 9.93 10.64 10.20 18 TR 0232 43.72 53.26 44.11 47.03 8.27 9.26 9.34 8.96 19 TR 1279 39.14 48.34 50.80 46.09 4.85 9.04 11.15 8.35 20 TR 0031 38.72 49.10 45.72 44.51 10.07 8.23 11.00 9.77 21 TR 0222 37.15 47.27 46.74 43.72 7.93 11.34 9.11 9.46 22 TR 0998 46.48 39.61 41.80 42.63 9.00 8.03 13.02 10.02 23 TR 1755 38.45 52.73 46.55 45.91 8.82 8.64 11.09 9.52 24 TR 1182 34.93 53.03 46.25 44.74 9.06 10.52 9.30 9.63 25 TR 1753 41.15 50.07 49.53 46.92 9.57 7.91 10.37 9.28 26 TR 0713 39.98 50.93 45.08 45.33 8.41 8.29 10.31 9.00 27 TR 0423 39.50 41.05 48.48 43.01 8.62 9.39 7.39 8.47 28 TR 0384 46.32 39.12 54.39 46.61 9.06 8.88 9.58 9.17 29 TR 1327 43.50 41.68 44.89 43.36 8.27 10.40 10.33 9.67 30 TR 0399 44.58 39.68 51.69 45.32 4.90 8.41 9.11 7.47 Mean 41.58 46.32 46.90 44.93 8.46 9.19 10.31 9.32 SE 3.82 5.1 3.81 2.70 0.28 0.19 0.04 1.00 LSD (5%) 7.69 10.23 7.65 5.30 0.58 0.39 0.09 2.10 CV % 11.3 13.5 9.9 12.70 18.44 11.86 10.67 19.10 93 5.3.3 Polygon view of GGE biplot for dry matter content (%) The PC1 and PC2 axes together explained 87.5% of the total variation observed for dry matter content (Figure 5.2). The polygon view of the GGE biplots revealed that the three locations fell into 3 mega environments. The three mega environments were Pendembu (Pen) Forest zone environment 1, Kambia (Kam) North savanna grassland environment 2 and Njala (Nja) Transitional rain forest environment 3. In Pendembu (Environment 1) the vertex/best performing genotype was TR 0384 (28). Other accessions in the Pen mega environment were TR 0907 (15), TR 0423 (27), TR 1557(16) and TR 0399 (30). TR 0998 (22) was the vertex genotype in mega environment 2 (Kam). Other entries within the Kam mega environment were TR 0747 (1) TR 1327 (29), and TR 1350 (13). TR 1182 (24) was the vertex genotype in mega environment 3 (Nja) which also had these entries TR 0031 (20), TR 0222 (21) and TR 1755 (23). 94 Figure 5.2: GGE biplot showing the best genotype for each of the 3 mega environments for dry matter content (%) Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, PEN-Pendembu, KAM-Kambia, NJA-Njala 95 5.3.4 Mean performance and stability of genotypes for dry matter content (%) across the 3 environments Figure 5.3 shows the stability for dry matter content of the 30 genotypes across the 3 mega environments. Thus, TR 1182 (24) was observed as the genotype with the highest dry matter content across test environments followed by 4 entries TR 1755 (23), TR 0843 (14), TR 1279 (19) and TR 1753 (25) constituted the top. The projection on the AEC ordinate, depending on the length measures the stability of the genotypes across test environments. The shorter the length of the projection, the more stable is the high dry matter content for the associated genotype. Consequently, the most stable entries were TR 0365 (2), TR 1279 (19) and TR 0033 (21). The arrow on the AEC abscissa delineates the ideal genotypes and entry TR 1182 (24) is the closest followed by TR1753 (25) and TR1279 (19) respectively. 96 Figure 5.3: Stability of dry matter content (%) of 30 Provitamin A cassava genotypes evaluated across three environments Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, Pen-Pendembu, Kam-Kambia, Nja-Njala 97 5.3.5 Discriminativeness vs representativeness for dry matter content (%)across the 3 environments Figure 5.4 shows the discriminativeness vs representativeness for dry matter content (%) of the 3 environments. The biplot identified Njala having the longest vector, as the most discriminatory followed by Pendembu while Kambia was the least discriminatory (Figure 5.4). Kambia was identified as the most representative of the test environments. The arrow on the AEC abscissa delineates the ideal environments. Among the test environments Pendembu had its vector closest to the arrow, it was therefore considered the ideal environment for dry matter content production. 98 Figure 5.4: Discriminativeness and representativeness of the three environments for dry matter content (%) determination of 30 Provitamin A cassava genotypes Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, Pen-Pendembu, Kam-Kambia, Nja-Njala 99 5.3.6 Winning Genotype and Mega Environment GGE biplot for Total Carotenoid Content The GGE biplot show the response of the 30 evaluated pro vitamin A cassava genotypes for total carotene content across the three-test environment. The polygon view of biplot (Figure 5.5), showed that all the test environments fell into 3 of the 7 sectors on the biplot and thereby identified the 3 mega environments namely Pendembu (Pen), Njala (Nja) and Kambia (Kam). In the Pendembu environment TR 0365 (2) was the vertex genotype with entries TR1350 (13) and TR 0222 (21) TR 0747 (1) being the other members. The vertex genotype in environment 2 (Njala) was TR 1313 (10) followed by TR 0683 (8), and TR 0998 (22), while TR 0907 (15) was the vertex genotype in Kambia, the third mega environment. 100 Figure 5.5: GGE biplot for total carotenoid content showing the best genotype among the 30 Provitamin A cassava genotypes in each Mega Environment Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, PEN-Pendembu, KAM-Kambia, NJA-Njala 101 5.3.7 Mean Performance and Stability for Total Carotenoid Content across three Environments. Stability of total carotenoid/carotene content across the three environments is shown in Figure 5.6. TR 1313 (10) had the highest total carotenoid content followed by TR 1152 (17) and TR 0683 (8) and the genotype with the lowest total carotenoid content was TR 0399 (30). Other low performers included TR 1198 (9), TR 0423 (27) and TR 1557 (16). Of the 30 Pro-vitamin-A cassava genotypes evaluated for total carotenoid content,15 had values above the mean of the population.TR 1557 (16), TR 1152 (17), TR 1198 (9), TR 0843 (14) and TR 0461 (6) but neither of them were the highest performer for carotenoid content. Therefore, among the top ten performers TR 1313 (10), TR 0683 (8), and TR 0998 (22) were observed as being the most stable genotypes. The arrow on the AEC abscissa delineates the ideal genotypes and entry TR1313 (10) was the closest followed by TR 0683 (8) and TR 1152 (17). 102 Figure 5.6 : Stability for total carotene content of 30 Provitamin- A cassava genotypes evaluated across three Environment. Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, PEN-Pendembu, KAM-Kambia, NJA-Njala 103 5.3.8 Discriminativeness vs Representativeness for Total Carotene Content across three Environments. Figure 5.7 shows the discriminatory ability and representativeness of the test environments for total carotene production. The biplot identified Kambia with the longest vector, as the most discriminating while Pendembu and Njala did not differ much. Njala formed the smallest angle with the AEC abscissa and was therefore identified as the most representative of the test 3 environments. 104 Figure 5.7: Discriminativeness vs representativeness for total carotenoid content of the 30 accessions in the 3 environments Key: 1-TR0747, 2-TR0365, 3-TR0560, 4-TR1208, 5-TR0461, 6-TR1337, 7-TR1569, 8-TR0683, 9-TR1198, 10- TR1313, 11-TR0696, 12-TR1322, 13-TR1350, 14-TR0843, 15-TR0907, 16-TR1557, 17-TR1152, 18-TR0232, 19- TR1279, 20 -TR0031, 21 - TR0222, 22 - TR0998, 23 - TR1755, 24 - TR1182, 25 - TR1753,26-TR0713,27- TR0423,28-TR0384, 29-TR1327, 30-TR0399, PEN-Pendembu, KAM-Kambia, NJA-Njala 105 5.4 Discussion The performance of cassava is subject to strong influence of genotype, environment and genotype*environment interaction (Aina et al., (2007); Aina et al., (2009); Ntawuruhunga et al., (2010). It is a routine practice for plant breeders to evaluate different genotypes in multi- locational trials to be able to compute the stability and performance of a genotype. The significant variability observed in the dry matter content (DMC) among the cassava genotypes presents improvement opportunities for the crop in Sierra Leone. The significance of G x E interactions for the dry matter content (DMC) revealed that genotypes responded differently across the tested environments. An obvious deduction from the low environmental effect on TCC is that selection for the trait can effectively be achieved by evaluating target genotypes in one location.). Genotypic by environment was observed over the different environments as indicated by crossover performances for some of the genotypes. This led to variations in the mean ranks of the genotypes in the different environments (Dixon et al., (2000) and Marlosetti et al., (2009). This implies different adaptation by the different genotypes suggesting the need to identify and select location specific genotypes for different environments The GGE biplot permits identification of outstanding genotypes in each mega-environment “which-won-where” (Yan et al., 2000; Yan, 2001, 2002; Yan and Tinker, 2006. In this study the high percentage of total variation explained by the first and second principal component axes PC1 =57.99 and PC2=29.49, PC1 = 50.62 and PC2= 24.85 for both dry matter content and caroteniod content captured the largest variation while component 2 contributed below 30% of the observed variation. 106 Different genotypes excelled for dry matter and total carotenoid production in the three mega- environments, - Njala, Pendembu and Kambia. Genotypes TR 0365 (2), TR 1313 (10) and TR 0907 (15) were the vertex/best performing for total carotenoid content while TR 0384 (28), TR 0998 (22) and TR 1182 (24) were the best genotypes for dry matter content across the 3 mega environments. However, they can be recommended for specific environments where they performed well. On the other hand, TR 0560 (3), TR 1337 (6), TR 1152 (17), TR1279 (19) and TR 0399 (30), fell into sectors that contained none of the tested locations for total carotenoid content while TR1208 (4), TR 1313 (10), TR 0696 (11), TR 0232 (18), and TR 1279 (19) fell into sectors that did not belong to any of the 3-test environment for dry matter content. They were therefore not the best in any of the 3-test locations and were probably the poorest in some of the locations tested for total carotenoid content and dry matter content respectively (Yan, 2001). Entry TR 1182 (24) and entry TR 1313 (10) were identified as the best yielders while entry TR 0461 and TR 0399 were the lowest performing genotypes for total carotenoid content and dry matter content respectively. Genotypes TR 0560 and TR 0461 showed stability for dry matter predicting that these genotypes possess the ability to prevent substantial fluctuation in dry matter over a range of environmental conditions. TR 1208 (4) and TR 1337 (6) and TR 0461(5) and TR 1350(13) were the most stable genotypes for dry matter content. TR 1313 (10) and TR 1182 (24) were the most ideal for total carotenoid content and dry matter content. These genotypes were the most outstanding and responsive at the mega-environmental levels for total carotenoid content and dry matter content, respectively. From the study, Njala was ideal for selecting superior genotype for total carotenoid content while Pendembu seems to be the ideal environment for dry matter content. This finding agrees with (Olayiwola et al., 2013) who 107 reported ideal genotypes from his studied population. Nonetheless, the best performing genotypes identified in this study could form the material for such genetic improvement through hybridization. DMC in the 30 genotypes studied here was less than that in varieties commonly grown by farmers in Sierra Leone. Typically, in a cassava variety selection scheme, screening involves no less than five candidate genotypes for total carotenoid content and dry matter content that are key drivers of variety adoption (Fakuda et al., 2002, Owusu and Donkor, 2012, Abdoulaye et al., 2014 and Esuma et al., 2016). Therefore, a breeding programme targeting development and advancement of provitamin A rich cassava varieties could exploit on-station trial evaluation or assessment in identifying carotene-rich genotypes that can be subjected to multi-locational evaluations where focus shifts to other traits highly influenced by environmental effects. This strategy would save costs while increasing precision to identify best performers for root yield, carotenoid content and dry matter content. 5.5 Conclusion The environment and genotype x environment interaction for the two studied traits (dry matter and total carotenoid content) of the 30 provitamin A genotypes showed significant differences (P < 0.001). To realize gains in translating investments in cassava bio-fortification research into impact on human nutrition, breeding efforts will need to focus on advancing varieties that combine high levels of both DMC and TCC alongside high-yielding genetic backgrounds. Nonetheless, the best performing genotypes identified in this study could form the material for such genetic improvement through hybridization in Sierra Leone. DMC value in the 30 provitamin A genotypes were less than that in the local cultivar (Cocoa) commonly grown by farmers in Sierra Leone. With respect to each trait studied, genotypes 108 showed differences in performance in specific environments. Genotypes TR 1182 and TR 1313, were the highest performers for dry matter and total carotenoid content. For the dry matter content alone genotypes TR 0683, TR 1152, TR 1208, TR 0365, and TR 1182 were the best. The identified superior genotypes could be used as parents for breeders to improve other accessions for increased carotenoid concentration and dry matter contents in provitamin A cassava varieties in Sierra Leone. 109 CHAPTER SIX 6.0 PERFORMANCE OF F1 CROSSES FOR TOTAL CAROTENOID, PROTEIN, IRON AND ZINC CONTENTS OF STORAGE ROOTS 6.1 Introduction Cassava is primarily a carbohydrate source that is often considered a low-quality food, when compared to most cereals (Nganga, 2010), and of some major traditional staples (Chavez et al., 2004). This is probably attributable to the low vitamins, protein and other nutrients and micronutrients content of its storage roots. Micronutrients which include folic acid, vitamin A, iodine, iron and zinc are required in minute quantities for human health, growth and development (Baafi et al., 2016b). They play crucial roles in metabolism and maintenance of tissue function (Shenkin, 2006 and Baafi et al., 2016b). Micronutrient deficiencies are chronic deprivation of these nutrients and constituted an alarming public health problem adversely affecting one third of the population globally (Darnton-Hill et al., 2005 and Simon et al., 2013). Micronutrient deficiencies have been estimated to cost sub- Saharan African economies more than $2.3 billion in lost productivity (UNICEF, 2004). WHO (2009) reported that almost 100 million preschool children suffer from vitamin A deficiency worldwide. Vitamin A deficiency affects about one- third of preschool children globally, of which 1% develop of these results to night blindness (West, 2002; Howe and Tanumihardjo, 2006; and WHO, 2009), while anemia, caused by inadequate intake of iron, and has affected 1.6 billion people globally (De Benoist et al., 2008; WHO, 2014). Severe and extensive deficiencies also prevail for zinc (Baafi et al., 2016b). These nutrient deficiencies have been of greatest global health concern. 110 According to the WHO (2009), three possible interventions to reduce nutritional deficiencies in humans have been established: (1) improving vitamin A, protein and nutrients availability through a diversified diet, (2) increasing the consumption levels of micronutrients by biofortification of staple foods, and (3) supplementation through vaccines. The use of vaccine is not sustainable and can hardly reach all affected persons especially those in rural communities where cassava is consumed as a major staple. However, sustainable solutions to malnutrition can be developed through linking agriculture, nutrition and health. Cassava can be improved through biofortification to increase its provitamin A, protein and micronutrient contents (HarvestPlus, 2002). This is an impressive step in developing a food systems approach to reducing malnutrition. This approach provides solution to the root causes of micronutrient malnutrition, targeting the marginalized people, involves built-in delivery mechanisms, and is scientifically feasible and cost-effective (HarvestPlus, 2002). As part of an integrated food systems approach, biofortification represents the best means for enabling rural households to improve family health and nutrition in sustainable ways (HarvestPlus, 2002). Cassava is a vegetatively propagated crop however the crop that undergoes natural outcrossing and is amenable to controlled crossing for genetic improvement. Improving vitamin- A, protein, iron and zinc content in cassava cultivars with traits preferred by farmers and consumers through biofortification (Dixon et al., 2005), would go a long way towards preventing vitamin-A, protein, iron and zinc deficiencies in Sierra Leone where cassava is largely consumed as the second major staple after rice. The development of biofortified cassava storage root varieties provides a cheap source of provitamin- A precursors (total carotenoid content), protein and minerals that are essential for the supply of quality food for better nutrition. This will properly address vitamin A and other nutritional deficiencies and ensure food security. 111 Therefore, the objective was to evaluate the levels of total carotenoid content, protein, iron and zinc contents in F1 cassava progeny to identify elite clones and potential sources of germplasm for nutritional improvement in cassava. 6.2 Materials and Methods Table 6.1 Pedigree and characteristics of breeding lines Genotype Sex Pedigree Characteristics IITA-TMS-IBA011368 F 940561 X 940263 Vitamin A, early bulking, high root yield IITA-TMS-IBA 120004 F IBA011224 X IBA011368 Vitamin A, early bulking, high root yield Zinc, high dry matter, resistant to ACMD, high root IITA-TMS-IBA 91/00416 F 91934 X 63397 yield IITA-TMS-IBA088693 F 05/0303 X 06/0657 Protein, resistant to ACMD, high root yield IITA-TMS-UBJ 120003 M Not Available Vitamin A, early bulking, high root yield IITA-TMS-IBA 96/1165 M O88/00210 HS Iron, high dry matter, resistant to ACMD, high root yield IITA-TMS-IBA088747 M FBA-1 X 05/0561 Protein, resistant to ACMD, high root yield Zinc, high dry matter, resistant to ACMD, high root MM96/81791 M Not Available yield 6.2.1. Development of F1 progenies Three breeding lines with high beta-carotene, two with high protein and zinc content and 1 with high iron contents were intercrossed to generate genetic materials (Table 6.1). The crossing block was established at Ubiaja, Nigeria in May 2015. The soil within this location comprises sandy, clay loam. Ubiaja is characterized by a bimodal rainfall pattern, with two distinct rainy seasons and dry seasons of nearly equal length. Ubiaja is an ideal location for cassava breeding program as most cassava genotypes flower at the location. The crossing block trial was laid out at the randomized complete block design in two replications with ten stands per genotypes been planted. 112 Controlled pollination using diallel mating design was undertaken starting three months after planting for period of 2-months according to the standard procedure described by Kawano (1980). Pollination bags were used to enclose flowers about to open to prevent uncontrolled pollination. Pollen was collected in the morning (7 to 8 am) and pollination was done later in the day between (11 am to 2 pm). Pollinated flowers were tagged and labelled. After pollination, the female flowers were covered with the pollination bags for one week. Harvesting and collection of seeds were undertaken two months after pollination. Seeds were sorted, labeled and stored till the time of seedling nursery establishment. Table 6.2 Number of F1 seeds generated, planted and (germinated) per family Female (♀) IBA120004 IBA91/00416 IBA011368 IBA088693 UBJ120003 300 (230) 40 (8) MM96/81791 78 (62) IBA96/1165 2 00 (200) IBA088747 250 (167) NB: Values in brackets are the number of seeds that germinated in each cross combination 6.2.2 Establishment of seedlings F1 progenies Nursery evaluation A total of 868 seeds obtained from parental cross combinations were sown directly into a well- prepared soil following a flotation test in June 2016 for one cropping season (2016-2017) (Table 6.2). At Foya crop site Njala, in Sierra Leone under a transitional rain forest agro climatic zone seeds were sown at a spacing of 0.30 m x 1 m using serpentine arrangement in an RCBD in two replications. An alley of 2m was established between blocks to reduce inter-block plant competition. Cocoa, a local variety was planted around the trial to ensure uniform exposure to African Cassava Mosaic Disease. NPK 20-10-10 fertilizer, was applied eight weeks after planting at the recommended rate of 450 kg/ha. The trial was conducted under rain fed 113 Male (♂) conditions. Field maintenance was done as required. Harvesting was done at 12 MAP per individual cross combination. 6.2.3 Data Collection Percentage fruit sets were determined by dividing total number of fruits collected with total number of flowers pollinated (Table 6.1). Number of seeds generated per cross are presented in Table 6.2. Cassava is highly heterozygous which makes each cross between two different parent plants genetically distinct with large variation in the ensuing F families. Families comprised 1 868F progenies from five crosses. The field was maintained by hand weeding as and when 1 necessary. At harvest, data was collected on a set of traits according to the standard procedure outlined by Fukuda et al. (2010). During harvesting data were collected on the following parameters number of roots per plant, pulp color of root, inner color, outer color and root size. Storage root samples were selected form harvested plant stand with storage roots. Storage root samples with different sizes were randomly selected for determination of total carotenoid, protein, iron and zinc contents. Selected storage root samples with roots from 3 cm or more in diameter, without cracks, insect damage or rotten parts were sent to the food science and nutrition laboratory of Njala Agricultural Research Centre Njala, for sample preparation for total carotenoid content analysis. For protein, iron and zinc analyses, the harvested storage roots samples were shipped to crop utilization laboratory IITA, Ibadan for analysis. About 200g of the dried storage root samples were ground into fine powder using a benzene 750 watts’ electric blender for analysis. Crude 114 percent Protein analysis was done at the crop utilization laboratory IITA, Ibadan, while samples for iron and zinc determination were shipped to the Microchem laboratory, Pretoria, South Africa for analysis. 6.2.4 Total carotenoid determination Total carotenoids content was estimated following IITA standard operating laboratory procedure using iCheck device. Procedure for estimation of total carotenoids in cassava using iCheck device as described in Section 4.2.7 (page 68). 6.2.5. Determination of Crude Protein Procedure for estimating percent crude protein followed the IITA standard operating laboratory procedure in the food science laboratory using Kjeldahl procedure. Two grams of dried milled cassava flour was weighed into a digestion tube, one tablet of selenium catalyst, 5ml of conc 80 % sulphuric acid and 5ml of conc hydrogen peroxide as antifoaming agent were added and placed in a digestion block preheated at 4200C in a fume cupboard. After digestion, the tube was cooled and transferred to the Automated Kjeldahl distillation apparatus (Kjeltec 8400). The digest was diluted with distilled water and equal volume of 40% NaOH was added and thereafter steam distilled into a receiver flask containing 4% boric acid solution, 3 drops of methyl red indicator and bromocresol green. A total of 50ml distillate was collected and titrated with normal hydrochloric acid solution using colorimetric end-point detection for estimation of the total Nitrogen. Samples were analyzed in duplicate. Three random samples were selected from the total batch of samples and re-analyzed to check on the accuracy and reproducibility of the method. The average values were recorded. The Nitrogen content was calculated using the formula 115 (Va – Vb) ∗ Normality of HCL 𝑥 1.4007 𝑥100 % Nitrogen = Sample Weight Where Va = Volume, in mL, of standard HCl required for sample Vb = Volume, in mL, of standard HCl required for blank % Crude Protein (CP) = % Nitrogen x 6.2 6.2.6 Iron and Zinc determination These were carried out at the Microchem laboratory, Pretoria, South Africa. Being a private company, protocols for their analyses were not provided. 6.2.5 Data Analysis Descriptive statistics were computed using XLSTAT (2010) and Genstat.v12. 6.3 Results 6.3.2 Descriptive analysis of F1 progenies Table 6.3 shows the performance of the 667 germinated F progenies evaluated in the seedling 1 nursery trial from five crosses. The summary analysis for the F1 progenies showed differences in mean values of the quality traits measured. The coefficient of variation (CV) ranged from 14.8 – 58.6 %, for the quality traits. 116 Table 6.3 Descriptive statistics for nutritional quality of F1 progenies of root quality Summary Statistics Traits Total Carotenoid F1 progenies assessed Content (µg _1 ) Protein (%) Iron (PPM) Zinc (PPM) Mean 14.7 5.4 12.6 8.5 Minimum 6.0 4.2 4.5 4.5 Maximum 28.0 8.1 59.2 17.7 Standard error of mean 0.5 0.1 0.6 0.5 Coefficient of variation (%) 34.7 14.8 58.6 37.5 6.3.3 Variation in total carotenoid content in roots of F1 progeny The performance on total carotenoid content of the F1 progeny is presented in (Table 6.4). Progenies 13 and 33 from cross IITA-TMS-IBA120004 x IITA-TMS-IBA120003 recorded the highest 28.0 -1 -1 µg and lowest 6.0 µg values of total carotenoid content with a grand mean of 14.7 µg-1. Ten progeny had total carotenoid values higher than the mean. 117 Table 6.4 Variation total carotenoid content of F1 progenies using the color chart and i- check device F1 progeny No F1 progeny No (IITA-TMS- TCC (IITA-TMS- TCC IBA120004 x IITA- (µg-1) IBA120004 x IITA- (µg-1) TMS-IBA120003) TMS-IBA120003) 1 13.5 31 12.7 2 12.6 32 12.7 3 14.9 33 6.0 4 12.3 34 12.8 5 13.8 35 7.5 6 13.0 36 15.0 7 14.9 37 16.9 8 16.9 38 6.6 9 18.2 39 6.4 10 13.6 40 7.4 11 12.5 41 15.7 12 13.3 42 11.8 13 28.0 43 7.6 14 21.2 44 7.1 15 14.2 45 7.3 16 12.9 46 6.2 17 13.3 47 7.1 18 13.0 48 6.7 19 19.5 49 9.1 20 13.6 50 10.4 21 10.9 51 10.0 22 13.1 52 6.4 23 14.2 53 7.9 24 14.8 54 8.6 25 14.1 55 9.1 26 14.4 56 7.3 27 12.7 57 8.8 28 13.2 58 7.1 29 14.4 59 16.4 30 14.8 60 9.4 Grand Mean 14.7 118 6.3.4 Mean protein content (%) in roots of F1 progeny Progeny numbers 41 and 12 from cross IITA-TMS-IBA 088693 x IITA-TMS-IBA 088747 recorded the highest 8.1% and lowest 4.2% percent protein respectively with a grand mean of 5.4% (Table 6.5). Twenty four out of the 58 harvested progeny had levels higher than the mean with values ranging from 5.5 to 8.1%. Table 6.5 Mean percent protein of F1 progeny (evaluated in seedling nursery trial in Njala) Protein Protein F1 progeny No (IITA-TMS- F1 progeny No (IITA- (%) (%) IBA 088693 x IITA-TMS- TMS-IBA 088693 x IITA- IBA 088747) TMS-IBA 088747) 1 6.0 31 4.9 2 4.5 32 8.1 3 7.3 33 6.3 4 4.5 34 4.9 5 5.5 35 5.6 6 6.9 36 4.9 7 5.8 37 4.4 8 4.2 38 4.2 9 5.4 39 5.4 10 4.8 40 6.1 11 4.8 41 4.3 12 5.2 42 4.2 13 6.0 43 5.5 14 5.6 44 4.6 15 5.0 45 6.1 16 5.4 46 5.2 17 5.5 47 6.4 18 6.0 48 5.4 19 5.4 49 4.8 20 6.1 50 5.0 21 4.3 51 5.3 22 4.9 52 4.8 23 6.3 53 5.5 24 6.1 54 5.8 25 4.9 55 4.8 26 5.9 56 5.2 27 6.6 57 5.9 28 4.3 58 5.5 29 5.1 30 5.1 Grand Mean 5.4 119 6.3.5 Mean Iron content (ppm) in storage roots of F1 Progeny Iron concentration of progeny from cross IITA-TMS-IBA 96/1165 x IITA –TMS-IBA 011368 ranged from 45.0 ppm to 59.2 ppm for harvested progeny (Table 6.6). Sixty-one progeny had higher levels than the grand mean (12.6 ppm), of which 22 had levels above 20.0 ppm. 120 Table 6.6 Mean iron concentration of F1 progeny F1 progeny F1 progeny No F1 progeny No (IITA- (IITA- No (IITA- F1 progeny No TMS-IBA TMS-IBA TMS-IBA (IITA-TMS- Concentration Concentration Concentration Concentration 96/1165 x 96/1165 x 96/1165 x IBA 96/1165 x (ppm) (ppm) (ppm) (ppm) IITA – IITA – IITA – IITA –TMS- TMS-IBA TMS-IBA TMS-IBA IBA 011368) 011368) 011368) 011368) 1 9.9 26 8.4 51 32.2 76 13.1 2 7.8 27 7.2 52 18.8 77 11.8 3 4.5 28 16.7 53 17.8 78 7.2 4 12.4 29 11.2 54 17.2 79 14.4 5 4.7 30 12.4 55 7.5 80 15.8 6 20.3 31 18 56 15 81 10.4 7 9 32 25.6 57 7.3 82 15.6 8 21.9 33 13.9 58 8.7 83 11.3 9 11.2 34 17.9 59 14.9 84 8.2 10 11.6 35 9.5 60 7.6 85 4.7 11 16.2 36 23.7 61 9.2 86 10.4 12 5.8 37 5.8 62 10.2 87 9.9 13 9 38 14.7 63 5.9 88 28.9 14 6.9 39 16.8 64 13.9 89 13.2 15 7.6 40 12.9 65 9 90 10 16 10.9 41 12.7 66 7 91 17.6 17 7.2 42 21.4 67 7.8 92 20.4 18 8.8 43 14.6 68 7.5 93 25.1 19 6.4 44 9.7 69 23.7 94 7.3 20 8 45 7.9 70 12.2 95 16.8 21 9.1 46 6.1 71 14.6 96 15.5 22 7.4 47 17.5 72 23.1 97 12.4 23 16.3 48 24.6 73 30.1 98 38.5 24 11.1 49 17.4 74 6.2 99 11.7 25 7.5 50 11.7 75 10.6 100 17.7 101 10.9 117 45 133 7.6 149 17.3 102 14.2 118 16.3 134 8.9 150 10.4 121 Table 6.7 Mean iron concentration of F1 progeny (cont’d) F1 progeny F1 progeny No F1 progeny No (IITA- (IITA- No (IITA- F1 progeny No TMS-IBA TMS-IBA TMS-IBA (IITA-TMS- Concentration Concentration Concentration Concentration 96/1165 x 96/1165 x 96/1165 x IBA 96/1165 x (ppm) (ppm) (ppm) (ppm) IITA – IITA – IITA – IITA –TMS- TMS-IBA TMS-IBA TMS-IBA IBA 011368) 011368) 011368) 011368) 103 7.8 119 24.2 135 7.8 151 9.8 104 6.8 120 29.9 136 10 152 11.2 105 15.6 121 59.2 137 9.1 153 8.1 106 9.2 122 22.2 138 6.9 154 13.1 107 11.6 123 18 139 7.7 155 7.7 108 10.8 124 15.9 140 9.8 156 8.9 109 9.7 125 34.2 141 8.9 157 8.7 110 8.7 126 33.8 142 8.5 158 9.1 111 12.4 127 16 143 9.3 159 6.6 112 7.7 128 7.8 144 17.2 160 10.5 113 11.3 129 18.4 145 9.6 161 8.4 114 20.4 130 30.6 146 7.1 162 9.3 115 22.4 131 13 147 8.6 116 29.8 132 10.9 148 7.2 Grand Mean 12.6 122 6.3.6 Mean zinc content (ppm) in storage roots of F1 progeny The zinc content of F1 progeny from cross MM96/81791 x IITA-TMS-IBA 088747 ranged from 4.5 ppm to 17.7 ppm (Table 6.7). Fourteen progeny recorded higher values than the grand genotypic mean (8.5 ppm) while 20 progeny recorded values that were lower than the grand mean. Table 6.7: Mean zinc concentration of F1 progeny F1 progeny No Concentration (MM96/81791 x IITA- (ppm) TMS-IBA 088747) 1 14.9 2 10.9 3 8.8 4 6.8 5 15.4 6 9.6 7 4.8 8 9.5 9 7.3 10 11.1 11 8.0 12 10.3 13 4.5 14 7.1 15 9.3 16 5.4 17 5.8 18 5.3 19 6.0 20 6.6 21 8.6 22 12.1 23 5.9 24 9.2 25 7.1 26 8.2 27 6.2 28 8.1 29 6.7 30 7.2 31 10.3 32 4.5 33 17.7 34 5.6 Grand Mean 8.5 123 6.4. Discussion Performance of the F1 progenies from different crosses showed a high variation in root nutrient which offers range of genetic variations. Similar findings were reported by Nganga (2010), Micheal et al. (2015) and Akuwa. (2016). Biochemical analysis on the F1 progenies revealed a high variation in the root nutrient quality traits (total carotenoid, protein, iron and zinc contents) for which genotypic selection can be applied. In addition, it is the performance of the individual F1 or clone that cassava breeders are most interested in because, it forms the basis for plant breeders to build efficient breeding programme in a crop like cassava (Tumuhimbise et al., 2014; Micheal et al., 2015; Akuwa, 2016 and Esuma et al., 2016). Total carotenoid content in the F1 progeny ranged from 6.0 µg -1 to 28.0 µg-1 with mean of 14.7µg -1. The observed mean of the total carotenoid content of the F1 progenies was comparable to the mean 14.7 μg g−1 reported for populations developed at CIAT (Ceballos et al., 2013). The high level of total carotenoid content revealed in cassava breeding populations at CIAT is a result of several years of cyclic selection processes exploited to primarily improve the total carotenoid levels in cassava to levels above 15 μg −1. IITA’s cassava breeding program has benefitted from the effective germplasm and seed exchange program with CIAT, thus benefited from clones of carotenoid levels above 15 μg −1. There is a high probability that parents used in hybridization are related to the CIAT germplasm. The mean carotenoid content was higher than findings of Maroya et al. (2012) and Ssemakula and Dixon. (2007), who reported 3.6 μg −1 and 5.0 μg−1 for cassava breeding populations evaluated at IITA and that of Esuma et al. (2016 where a mean of 3.8 μg-1 was obtained for population in Uganda. 124 There were progenies with high levels of crude protein that were above the 2% threshold value (Nganga, 2010). These high levels of crude protein content have also been reported in some landraces and improved varieties of cassava in CIAT (Chavez et al., 2005). The authors suggested that the variation in crude protein content are high genetic in nature which could be exploited for improving the protein content of cassava. In another study in CIAT by Ceballos, (2006b) large differences were observed in crude protein content of cassava roots ranging from 0.95% to 6.42%. He suggested that a considerable proportion of these differences are genetic in nature and therefore are excellent possibilities for exploiting these differences and further increasing them by traditional breeding methods. However, I’m suggesting the use of a different analytical method which is not based on the use of nitrogen since the cassava samples that are analysed in most studies even though dry, still have locked nitrogen in hydrogen cyanide that gives false high levels of protein. Nganga (2010) also reported a similar trend of high levels of percent crude protein ranging from 1.35% to 3.45%. Some of the F1 progeny from the cross combination IITA-TMS-IBA 91/00416 x MM96/81791 had high zinc content (8.4 ppm), but lower than the findings of Nganga (2010) who reported above 64.0 ppm from one of his test sites in Kenya. CIAT (2006) reported that higher levels of zinc positively correlated with the beta carotene conversion to vitamin A when high carotenoid cassava cultivars are consumed. Chavez et al., (2005) have also reported that there is genotypic variation for zinc content in cassava. Also, some of the F1 progeny from cross IITA-TMS-IBA 96/1165 x IITA–TMS-IBA 011368 gave higher levels of iron but less than values observed in cassava collections from Meso America (Chavez et al.,2005), for collections from Nigeria (Dixon et al., 2005) and for Kenyan collections (Nganga 2010). 125 Dixon et al., (2000) reported values ranging from 4 to 19 ppm for iron and 4 to 8 ppm for zinc that were similar to the ranges observed in the present study as they also observed a significant positive correlation between iron and zinc. Baafi (2016a) reported strong positive genotypic association among total carotenoid content, iron and zinc indicating sufficient variability for these traits in sweet potato. The observation from this study showed that F1 progenies developed from this study could also serve as useful parents or source in any biofortification programmes. 6.5 Conclusion Genetic variation existed among the cassava F1 progenies for total carotenoid content, protein, iron and zinc contents. The observed values for total carotenoid, protein, iron and zinc contents in the F1 progenies indicate their potential for improving the nutritive value of cassava in Sierra Leone. Higher levels of micronutrients which are favorable for the human diet were recorded for some progenies evaluated within this population. An important aspect of enhancing micronutrient levels in cassava roots is maintaining a good agronomic background of micro nutrient-rich genotypes (Dixon et al., 2005). 126 CHAPTER SEVEN 7.0. GENETIC STUDIES ON MEALINESS, DRY MATTER, ROOT NUMBER AND FRESH ROOT YIELD, IN CASSAVA (Manihot esculenta Crantz) 7.1. Introduction Breeding for increased mealiness, number of roots, dry matter content and good fresh root yield enhancing the production of cassava to meet consumer acceptance for cassava storage root and its products has become the next challenge in cassava breeding in Sierra Leone. These traits are essential for cassava production as they are the most preferred traits for end-use consumption. Although other breeding mating designs had been successfully used for cassava, they are limited in that, they do not provide information for estimation of specific combining ability (SCA), which is important in the inheritance of key traits such as number of storage roots and fresh root yield (Ceballos et al.,2004, 2015; Crossa et al., 2010). Subsequent shifts in cassava breeding schemes have seen an increased production of full-sib progenies (Ceballos et al., 2012; Nassar and Ortiz, 2006). The selection of suitable parents and good mating designs are keys to successful breeding schemes. The importance of mating designs in cassava breeding cannot be neglected as it provides information on the genetic control of the character under investigation, generates a breeding population to be used as a basis for the selection and development of potential varieties, estimating the genetic gain and provides information for evaluating the parents used in breeding program (Acquaah, 2012).The full-sib crossing schemes employ controlled pollinations, where selected mating designs are used to generate families from specific parental combinations, 127 facilitating genetic studies alongside production of breeding populations with traits of interest (Esuma et al., 2016, Nduwumuremyi et al., 2013). The diallel mating design, particularly, has become popular for cassava breeding simply because it facilitates generation of useful information on genetics of key agronomic traits and allows identification of parents with superior combining ability for developing breeding populations (Kulembeka et al., 2012; Parkes et al., 2011; Tumuhimbise et al., 2014; Zacarias and Labuschagne. 2010 and Akuwa, 2016). It is this genetic information that guides breeders to deploy appropriate strategies for crop improvement (Acquaah, 2012; Nduwumuremyi et al., 2013). Knowledge of general combining ability (GCA) of parental lines is particularly helpful for predicting genetic gains in a breeding program (Falconer and Mackay, 1996). Hayman (1954) elaborated on the procedure for statistical analyses based on diallel data, which partitions total variation into GCA of the parents and SCA of crosses. Combining ability could be described as the relative potentials of an inbred line or a clone, when allowed to mate with another inbred line or clone, to transmit desirable traits or specific trait to the next generation (Chaudhari, 1971). It facilitates the prediction of the behavior of a line when exploited as a progenitor in a hybrid and compliments the selection of superior breeding lines for hybrid combination and for studying the nature of genetic variation). Griffing (1956) reported a method of analyzing combining abilities using the genetic estimates of the parent and hybrid components of diallel analysis through general and specific combining abilities. Falconer and Mackay (1996) described general combining ability (GCA) as the mean performance of the progenitors in all its crosses and it is expressed as a deviation from the mean of all crosses. This average behavior of parents in 128 crosses (GCA), calculates the breeding value of a given genotype because of additive gene effects (Ceballos et al., 2004; Micheal et al., 2015). Specific combining ability (SCA) is described as the deviation of individual crosses from the average performance of parents, because of the presence of dominance effects. Understanding the expression of gene action would be important for formulating breeding techniques to generate and develop desired traits. Therefore, information on combining abilities are required to identify suitable and superior progenitors and genotypes which can be hybridized for the development of elite cultivars and progenies varieties that would ultimately ensure sustained production and productivity by subsistence farmers in Sierra Leone. The cassava breeding programme in Sierra Leone has been evaluating half sib seeds and clones sourced from the cassava breeding unit in IITA. Ibadan. In Cassava breeding, various mating designs and arrangements are exploited by breeder to generate improved cassava genotypes or varieties. The present study is probably the first breeding research undertaken for Sierra Leone. Currently cassava varieties released in Sierra Leone are white or cream fleshed with little amount of high dry matter, mealiness and increased yield. To employ suitable breeding strategies for genetic improvement of economically important traits which are quantitatively inherited, diallel experimental design was used for this study. The objectives of the study were to: i) estimate the combining abilities of twelve cassava genotypes for mealiness, dry matter content, number of roots and fresh root yield. 129 ii) determine the gene action for mealiness, dry matter content, number of roots and fresh root yield, and iii) identify and select superior families for the development of elite clones. 7.2. Materials and Methods 7.2 1 Establishment of crossing block The crossing block was established in May 2015 in Ubiaja, Edo state, Nigeria. The soil within this location comprises sandy clay loam. Ubiaja has a bimodal rainfall pattern, with two distinct rainy seasons and dry seasons of nearly equal length. The peak rainfall occurs between April to May and September to November. Twelve cassava parents (Table 7.1) were selected based on mealiness, dry matter content, pest and disease resistance, plant architecture, flowering ability and fresh root yield from the IITA genetic gain as well as landraces from Ghana. The 12 genotypes (Table 7.1) were crossed in a 12 x 12 diallel mating design without reciprocals and selfs. Controlled pollinations were undertaken three months after planting. Hand weeding was performed as required. Controlled hand pollination was carried out according to the standard procedure described by Kawano (1980) (See Section 6.2.1). Total seeds obtained was 5,382 (Table 7.2) 130 Table 7.1 Twelve cassava genotypes used as parents for F1 progenies, their pedigree and important traits Genotype Pedigree Traits IITA-TMS-IBA 120003 Vitamin A, early bulking, high root yield UCC 20012 (246) High dry matter/Mealy, resistant to ACMD, high root yield IITA-TMS-IBA 693 High dry matter/Mealy, resistant to ACMD, high root yield TMEB 419 High dry matter/Mealy, resistant to ACMD, high root yield TMEB 1 High dry matter/Mealy, resistant to ACMD MM96/8179 Zinc, high dry matter, resistant to ACMD, high root yield IBA011224 X IITA-TMS-IBA 120004 Vitamin A, early bulking, high root yield IBA011368 96/1165 O88/00210 HS Iron, high dry matter, resistant to ACMD, high root yield 91/00416 91934 X 63397 Zinc, high dry matter, resistant to ACMD, high root yield IITA-TMS-IBA088747 FBA-1 X 05/0561 Protein, resistant to ACMD, high root yield IITA-TMS-IBA088693 05/0303 X 06/0657 Protein, resistant to ACMD, high root yield IITA-TMS-IBA011368 940561 X 940263 Vitamin A, early bulking, high root yield IBA - Ibadan; IITA- International Institute of Tropical Agriculture; TMS - Tropical Manihot Selection; NB- Parent1-UBJ120003: Parent2- UCC2001 (246): Parent3- TMEB693:Parent4- TMEB419: Parent5-TMEB1: Parent6- MM961781: Parent7-IBA120004: Parent8- I961165: Parent9- I9100416: Parent10 – I088747: Parent11- I088693: Parent12- I011368 131 Table 7.2: Seeds obtained from 12 x 12 half diallel cross of cassava genotypes and number of seeds produced Female UB120003 UCC2001(246) TMEB693 TMEB419 TMEB 1 MM961871 IBA120004 IBA961165 I9100416 I088747 I088693 I011368 Male UBJ120003 77 71 148 20 46 96 54 75 100 100 UCC2001(246) 13 148 10 29 93 150 95 100 150 151 TMEB693 151 30 76 78 59 44 150 90 75 TMEB419 21 29 75 70 35 98 75 7 TMEB1 75 23 64 38 100 250 86 MM961871 60 114 72 100 88 100 IBA120004 150 56 83 106 100 I961165 67 100 84 150 I9100416 21 45 18 I088747 143 200 I088693 124 I011368 7.2.2 Seedling Nursery evaluated at Foya crop site A total of 40 seeds were randomly selected after germination test from each of the 65 F1 families or cross combination. Seeds were planted at a spacing of 0.30 cm x 1 m in June, 2016 in two replications at Foya crop site Njala, representing the transitional rain forest agro climatic zone in Sierra Leone (Van Vuure et al., 1974; Odell et al., 1974) for one cropping season (2016--2017). The trial was established in a randomize complete block design using serpentine design within the 2 replications in three blocks each. The ‘Cocoa’ cassava variety (an ACMV susceptible variety) was planted around the trial to ensure screening for ACMD. Fertilizer, NPK 20-10-10 was applied 8 weeks after planting at the recommended rate of 450 kg/ha. The blocks were separated by 2m alleys. The trial was conducted without supplementary irrigation, and field maintenance was undertaken throughout the period of evaluation as necessary. 132 7.2.3 Data Collection At harvest data were collected on some yield related (Table 7.3) traits per the standard procedure outlined by Fukuda et al. (2010). Evaluation was done on 40 plants per family. Table 7.3 Phenotypic traits used in the characterization of cassava F1 Progeny Time of Mode of Remarks Traits Procedure evaluation Scoring (phenotypic class) Number of harvested Record the number of roots 12 MAP (at harvest) count Number of stand harvested and roots per family harvested per plot of each clone. counted Number of storage root Record the number of plant stand 12 MAP (at harvest Count Number of roots harvested and harvested counted Number of root per plant The most frequent occurrence was 12 MAP (at harvest) Count 3 = small sized roots, 5= medium observed and recorded. sized roots 7 = large sized roots Fresh storage root weight Roots were placed in a clean 12 MAP (at harvest) Weighing Shoot fresh weight (stem and synthetic bag and weighed using a leaves) spring balance scale set to zero using the empty bag. Shoot weight (kg) Shoots were tied with a twig and 12 MAP (at harvest) Harvested samples weighed using a spring balance Weighing scale set to zero. Mealiness Root were cooked for 25 minutes 12 MAP (at harvest) Hand Harvested samples and a scale of 1,2, and 3 was used screening to screened the cooked samples MAP = Month After Planting Dry matter content (DMC) expressed as a percentage was determined by selecting at least 3 storage roots from a bulk of storage roots of each plant. The roots were washed, peeled and sliced into piece. The sliced of the six fresh samples were weighed to obtain 100grams before drying for 48hours in an oven at 80°C. The dried samples were then reweighed to obtain the dry mass and the DMC was calculated as; DMC (%) = (Oven dried root weight / Fresh root weight) × 100 -1 Fresh root yield (FYLD in t ha ) Fresh root yield was estimated as from fresh storage root weight (FRW) for each family; 133 -1 FYLD (t ha ) = Fresh root weight × 10 000 Mealiness Mealiness is an important trait in breeding for quality assessment, it is a method used in assessing the cooking quality of mealy cassava genotypes. Mealiness was estimated following IITA standard laboratory procedure. Cassava stands with at least three roots were randomly selected from cross combinations.
Samples were prepared by washing each root with clean water to remove all soil. The roots were peeled and washed again and then dried with paper towel. Each root was sliced into four equal parts. The sliced samples were placed into a pot when the temperature of the boiling water attained 100oC. The samples were cooked for 25 min on a gas cooker. A scale of 1 to 3 was used to assess mealiness, where 1 = 10-30% mealy, 2 = 40- 60% mealy, and 3 = 70-100 mealy. 7.2.4 Data analysis Data were subjected to statistical analyses using SAS version 9.3. Analytical tools employed included general analysis of variance and diallel analysis of variance for combining abilities and estimates.
 The general analysis of variance was used to test if the sources of variation had any significant influence on the character. Diallel analysis of variance for combining ability was performed using mean values, following Griffing‘s Model I Method IV (1956).
 134 Yijkl =μ +Ri +Gj+Gk + Sjk +Eijkl Where; Y = is the l-th observation of the i-th replication for the jk-th cross; ijkl μ = is the overall mean; Ri = is the fixed effect of the i-th replication, i=1 to b; G or G = is the random general combining ability (GCA) effect of the j-th female or the k-th j k male ~Normally and Independently Distributed (NID) (0, σ2G), j, k=1 to p and j