GENETIC ANALYSIS FOR PANICLE ARCHITECTURE AND GRAIN YIELD IN SORGHUM [Sorghum bicolor (L.) Moench] IN MALI By DRAMANE SAKO (10293984) 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 SCHOOL OF AGRICULTURE COLLEGE OF AGRICULTURE AND CONSUMER SCIENCES UNIVERSITY OF GHANA LEGON DECEMBRE, 2013 University of Ghana http://ugspace.ug.edu.gh i 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. .................................................. DRAMANE SAKO Student .................................................. PROFESSOR ERIC.Y. DANQUAH Supervisor PROFESSOR VERNON GRACEN Supervisor .................................................. PROFESSOR SAMUEL K OFFEI Supervisor ............ ...................................... DR. NIABA TEME Supervisor .................. ................................ DR. JEAN FRANÇOIS RAMI Supervisor University of Ghana http://ugspace.ug.edu.gh ii GENERAL ABSTRACT Sorghum panicle architecture is a complex trait involving growth, elongation and branching pattern. The main objective of this research was to determine the genetic control of sorghum panicle architecture and its involvement in grain yield. Molecular markers associated with statistical analyses were used to identify the genomic regions or QTLs controlling quantitative traits. A population composed of 401 F4 families was derived from a cross between two contrasted parents for panicle traits (Tiandougou and Lata-3). Phenotyping of sorghum panicle architecture was based on two groups of variables: architectural geometry traits related to the length and the diameter and architectural topology traits that relate to the number, density and position. Forty eight sorghum panicle architecture traits and grain yield components were studied in three sowing dates in two years at the research station of Sotuba, Bamako, Mali. A linkage map was constructed using 228 SNP. A total length of the genetic map was 1362.3 cM with average distance between markers of 6.01 cM. In this study, 164 QTLs including 53 major QTLs were detected for the different traits. These putative QTL explained 1.91 to 45.64% of the phenotypic variation observed for each trait. Many of the major QTLs were consistent across the sowing dates. Key regions of the genome in the population Tiandougou/Lata-3 affecting panicle growth, elongation and branching may be utilized to improve agronomic performance through MARS approach. Sorghum grain yield was positively influenced by the following panicle architecture traits: primary branches number and length, rachis basal diameter and the number of nodes per panicle. These traits can be used for indirect selection for yield potential improvement in conventional or molecular breeding methods. QTLs and candidate genes involved in inflorescence architecture in related grasses were co- localized. These associations may be investigated in sorghum to decipher the function and pattern of these candidate genes. University of Ghana http://ugspace.ug.edu.gh iii ACKNOWLEDGEMENTS I am grateful to WACCI, University of Ghana Legon and IER for accepting me and for the quality of training. I will always be grateful to Prof. Eric Danquah and Prof. Samuel Offei for taking time from their busy schedule to visit me twice during a field work and help me with useful guidance for the completion of this thesis. Special thanks to Prof. GRACEN for his supervision and guidance during the thesis write up. My eternal gratitude to Dr. Niaba Teme my in country supervisor for his professional advice, ideas, encouragements, and invaluable inputs in this thesis. I am grateful to CIRAD for genotyping of the population and fruitful exchange with senior scientists. I am greatly indebted to Dr. Jean François RAMI, Dr. Baptiste GUITTON and Dr. Michel VAKSMANN for their useful contribution in training, data analysis and assisting and advising me during thesis writing for all the valuable inputs. I am particularly thankful to Dr. Evelyne Costes for her help. My sincere thanks to Dr. Aboubacar Touré and his family for making my stay in Ghana enjoyable. I am very grateful to Dr. Dounanke Coulibaly, Dr. Niamoye Yaro, Dr. Youssouf M Diarra and Dr. Ousmane Niangaly for their help. I am thankful to Dr. Mamourou Djourté, Dr. Mamoutou Kouressy and Dr. Mamadou Coulibaly for their help and advice. I am thankful to all my colleagues at the Biotechnology lab, Agro-climatology, Sorghum program and all CRRA of Kayes and Sotuba for sharing their experiences with me, assisting me in field work and data collection. I thank Yacouba Dembelé, Sekouba Sanogo for their invaluable help in the success of the field work. I would like to express sincere appreciation to my uncle Fotigui Sako and my sister Korotoumou Sako for their advice and assistance. Finally, I would like to express my most sincere gratitude to the Generation Challenge Program for providing the funding that allowed me to complete my graduate studies at WACCI, University of Ghana. Dr. Ndeye Ndack Diop and Dr. Theresa Fulton, thank you for your capacity building efforts that facilitated the completion of my thesis. University of Ghana http://ugspace.ug.edu.gh iv DEDICATION This thesis is dedicated to the families of SAKO at Moribabougou, Bougouni and Bamako Hamdallaye. My Mothers Coumba DIALLO, Haby COULIBALY and Fanta MARIKO, my wife, Diarah GOITA and children, Fanta, Mamadou, Minata, Coumba and Fotigui. University of Ghana http://ugspace.ug.edu.gh v TABLE OF CONTENTS DECLARATION I GENERAL ABSTRACT II ACKNOWLEDGEMENTS III DEDICATION IV TABLE OF CONTENTS V LIST OF FIGURES IX LIST OF TABLES XI LIST OF ABBREVIATIONS XII CHAPTER ONE 1 GENERAL INTRODUCTION 1 CHAPTER TWO 5 2.0. LITERATURE REVIEW 5 2.1. Sorghum 5 2.1.1. Economic importance 5 2.1.2. Different types of sorghum and varietal assortment 6 2.1.3. Types of panicles in sorghum 8 2.1.4. Sorghum genome and relationship with members of the Poaceae family 9 2.2. Flowering in sorghum 10 2.2.1. Floral biology: inflorescence biology and morphology 10 2.2.2. Flower emergence: from floral induction to flower development 12 2.2.2.1. Floral induction 12 2.2.2.2. Floral initiation 13 2.2.2.3. Floral differentiation 13 University of Ghana http://ugspace.ug.edu.gh vi 2.3. Panicle architecture 14 2.3.1. Principles of plant architecture: growth, elongation, branching 15 2.3.1.1. Growth 16 2.3.1.2. Elongation 16 2.3.1.3. Branching 17 2.3.2. Application of plant architecture principles on sorghum panicles 17 2.4. Genetic control of panicle development 18 2.4.1. Genes involved in panicle development 18 2.4.1.1. Inflorescence meristem development 18 2.4.1.2. Spikelet pair and branch meristem development 19 2.4.1.3. Spikelet meristem development 19 2.4.2. Hormonal control of inflorescence development 20 2.5. Genetic maps and QTL mapping in sorghum 20 2.5.1. Molecular markers 20 Sorghum genetic maps 21 2.5.3. Principle of QTL mapping 23 2.5.4. Review of QTL in sorghum 24 2.5.5. QTL mapping for inflorescence traits in cereals 25 2.5.5.1. Sorghum 25 2.5.5.2. Rice 26 2.5.5.3. Maize 27 2.5.6. Candidate genes 27 CHAPTER THREE 29 3.0. PHENOTYPING FOR SORGHUM PANICLE ARCHITECTURE AND YIELD IN TIANDOUGOU/LATA-3 F4 POPULATION 29 3.1. Introduction 29 3.2. Materials and Methods 31 3.2.1. Plant Materials 32 3.2.2. Experimental Design 33 3.2.3. Data collection 33 3.2.4. Data analysis 36 3.2.5. Correlation, Principal Component Analysis and ANOVA for panicle traits 37 3.2.6. Calculation of Broad sense heritability 37 3.2.7. Construction of linear model and BLUP extraction 38 3.2.8. Analysis on BLUP 38 University of Ghana http://ugspace.ug.edu.gh vii 3.2.9. Contribution of panicle traits to panicle grain yield 38 3.3. Results 39 3.3.2. Experimental design data analysis 39 3.3.3. Performance of the F4 families and their parents 40 3.3.4. Analysis of Variance (ANOVA) 41 3.3.5. Broad sense heritability estimates 43 3.3.6. Principal Component Analysis 45 3.3.7. Correlation among traits 48 3.3.8. Transgressive segregation 49 3.3.9. Estimation of grain yield per panicle using panicle architecture traits 53 3.3.10. Branch length, grain number and rachis diameter at different positions on panicle 53 3.4. Discussion 55 3.5. Conclusion 59 CHAPTER FOUR 60 4.0. QUANTITATIVE TRAIT LOCUS (QTL) ANALYSIS OF PANICLE ARCHITECTURE IN THE TIANDOUGOU/ LATA-3 BREEDING POPULATION 60 4.1. Introduction 60 4.2. Material and methods 61 4.2.1. Plant material 61 4.2.2. Leaf sampling and DNA extraction 61 4.2.3. Genotyping 62 4.2.4. Phenotyping 62 4.2.5. Construction of Genetic maps 62 4.2.6. QTL Detection 63 4.2.7. Digenic epistasis detection 64 4.2.8. QTLs overlapping confidence interval across studies 64 4.3. Results 65 4.3.1. Summary of QTLs detected in Tiandougou/Lata-3 F4 families population 65 4.3.2. QTLs detected for the most important sorghum panicle architecture traits and grain yield in Tiandougou/Lata-3 F4 families using Simple Interval Mapping followed by refine QTLs 66 4.3.2.1. Panicle grain yield 66 4.3.2.2. Number of primary branches per panicle 67 4.3.2.3. Number of internodes per panicle 69 4.3.2.4. Rachis base diameter 70 University of Ghana http://ugspace.ug.edu.gh viii 4.3.2.5. Panicle length 71 4.3.2.6. Average number of grains per primary branch 72 4.3.3. Consistent QTLs detected in Tiandougou/Lata-3 F4 population across the three sowing dates in the combined data (BLUP) 73 4.3.4. Consistent major QTLs cluster 77 4.3.5. Pleiotropic effects 79 4.3.6. Clustered heatmap analysis 82 4.3.6. Epistasis effect of QTLs detected in F4 families derived from the cross between Tiandougou and Lata-3 84 4.3.7. Identification of QTLs with overlapping confidence interval for sorghum panicle architecture traits across reported studies 88 4.4. Discussion 90 4.5.Conclusion 94 CHAPTER FIVE 96 5.0. IDENTIFICATION OF CANDIDATE GENES INVOLVED IN SORGHUM PANICLE ARCHITECTURE IN MAIZE AND RICE 96 5.1. Introduction 96 5.2. Methods 98 5.3. Results 100 5.3.1. Candidate genes related to meristem regulation during inflorescence development 100 5.3.2. Candidate genes related to specification of the spikelet meristem identity 104 5.3.3. Candidate genes related to plant growth and development 106 5.3.4. Phylogenetic analysis 108 5.4. Conclusion 110 CHAPTER SIX 111 6.0. GENERAL DISCUSSION 111 6.1. Phenotyping for sorghum panicle architecture and yield in Tiandougou/Lata-3 F4 population 111 6.2. QTLs analysis of panicle architecture in the Tiandougou/Lata-3 breeding population 114 University of Ghana http://ugspace.ug.edu.gh ix 6.3. Identification of candidate genes involved in sorghum panicle architecture in maize and rice 116 GENERAL CONCLUSIONS AND RECOMMENDATIONS 118 RECOMMENDATIONS 119 REFERENCES 120 APPENDIXES 138 Appendix 4.1: Genotyping Material and method at CIRAD lab 138 Appendix 4.2: QTLs map for sorghum panicle architecture traits in F4 families from Tiandougou/Lata-3 141 Appendix 4.3: Quantitative Trait Loci (QTLs) detected in Tiandougou/Lata-3 F4 families population 146 LIST OF FIGURES Figure 2.1: Head types of cultivated sorghum. Type I consists for wild races and is considerably more diffuse than type 2 [Original picture from Harlan and de wet, (1972)] 7 Figure 2.2: Sorghum races A = bicolor, B = durra, C = caudatum, D = guinea and E = kafir (Picture adapted from Harlan and de wet, 1972). 8 Figure 2.3: Orthologous gene families between sorghum, Arabidopsis, Rice and poplar. [Original picture from Paterson et al. (2009)]. The numbers of gene families (clusters) and the total numbers of clustered genes are indicated for each species and species intersection. 10 Figure 2.4: Gradual primordial differentiation into sorghum inflorescence. [Original picture from Singh et al. (1997)] 11 Figure 2.5: Scanning electron micrographs of Sorghum bicolor apices showing the transition from vegetative to floral state. 1a and 1b are vegetative; leaf primordial are evident at the arrow. 1c through 1g are floral, with floral primordial evident at the arrow heads. 1a through 1g are all printed at same magnification, with magnification bar shown in 1e. 1a through 1d boxed areas on left enlarged 2x on right (Original picture from Verbeke et al., 1990). 14 University of Ghana http://ugspace.ug.edu.gh x Figure 2.6: Components of sorghum panicle architecture traits investigated (modified, Witt-Hmon et al., 2013). 15 Figure 3.1: Methodological approach for measuring panicle architecture traits 34 Figure 3.2: Broad sense heritability estimates for panicle architecture traits and grain yield 43 Figure 3.3: Gradient in broad sense heritability for length of panicle traits 45 Figure 3.4: Contribution of sorghum panicle architecture traits to the dimension 1 and 2 46 Figure 3.5: Contribution of sorghum panicle architecture traits to the dimension 2 and 3 47 Figure 3.6: Distribution of four panicle traits showing transgressive segregation in the first sowing date (SB1), 2011, Sotuba, Mali 50 Figure 3.7: Distribution of four panicle traits showing transgressive segregation in the second sowing date (SB2), 2011, Sotuba, Mali 51 Figure 3.8: Distribution of four panicle traits showing transgressive segregation in the first sowing date (SB1’), 2012, Sotuba, Mali 52 Figure 3.9: Primary and secondary branches length; grain number and rachis diameter at second, median and penultimate nodes 54 Figure 4.1: QTLs detected for panicle grain yield in Tiandougou/Lata-3 F4 families population 67 Figure 4.2: QTLs detected for number of primary branches per panicle in Tiandougou/Lata-3 F4 families population 68 Figure 4.3: QTLs detected for number of internodes per panicle in Tiandougou/Lata-3 F4 families population 69 Figure 4.4: QTLs detected for rachis base diameter in Tiandougou/Lata-3 F4 families population 70 Figure 4.5: QTLs detected for panicle length in Tiandougou/Lata-3 F4 families population 72 University of Ghana http://ugspace.ug.edu.gh xi Figure 4.6: QTLs detected for average number of grains per primary branch in Tiandougou/Lata-3 F4 families population 73 Figure 4.7: Major consistent QTL cluster on SBI-03 77 Figure 4.8: Major consistent QTLs cluster on SBI-06 78 Figure 4.9: QTLs Clustered heat map for sorghum panicle architecture traits in F4 families derived from bi-parental population 83 Figure 4.10: Digenic epistasis detected for number of grains on primary branch at median node (NG_MN) using QTL IciMapping. 85 Figure 5.1: Phyllogenetic tree for candidate genes involved in inflorescence architecture 109 LIST OF TABLES Table 3.1: Data on sorghum panicle architecture variables with code collected in 2011 (SB1 and SB2) and 2012 (SB1’) at Sotuba research station in Mali (modified Segura et al., 2006) 35 Table 3.2: F probability (Fpr) of REML analysis on 401 F4 families across three sowing dates at Sotuba research station, Mali, 2011 and 2012 39 Table 3.3: Mean comparison of the parents and F4 progenies, T test probability of the parents Lata- 3 and Tiandougou vs progenies across the three sowing dates at Sotuba research station, Mali 41 Table 3.4: ANOVA for sorghum panicle traits and grain yield and its components 42 Table 3.5: Correlation among sorghum panicle architecture traits and grain yield per panicle 49 Table 4.1: Summary of significant QTLs detected for sorghum panicle architecture traits and grain yield in Tiandougou / Lata-3 F4 families using Simple Interval Mapping followed by refine QTLs 65 Table 4.2: Topological consistent QTLs detected in Tiandougou/Lata-3 population 74 Table 4.3: Geometrical and grain yield consistent QTLs detected in Tiandougou/Lata-3 population 76 University of Ghana http://ugspace.ug.edu.gh xii Table 4.4: Pleiotropic QTLs detected on SBI-01 and SBI-02 79 Table 4.5: Pleiotropic QTLs detected on SBI-03 and SBI-04 80 Table 4.6: Pleiotropic QTLs detected on SBI-06 and SBI-07 81 Table 4.7: Significant digenic epistasis QTL detected in F4 families derived from Tiandougou x Lata-3 using BLUP data 86 Table 4.8: QTLs with overlapping confidence interval identified from five (5) studies on sorghum panicle architecture traits. 89 Table 5.1: List of candidate genes related to meristem regulation during inflorescence development 103 Table 5.2: List of candidate genes related to specification of spikelet meristem identity 105 Table 5.3: List of candidate genes related to plant growth and development 107 LIST OF ABBREVIATIONS AFLPs: Amplified fragment length polymorphism ANOVA: Analysis of Variance BIP: bi-parental populations BLAST: Basic Local Alignment Search Tool BLUP: Best Linear Unbiased Predictor BM: Branch Meristem CIRAD: Agricultural Research for Development cM: Centimorgan DArTs: Diversity Array Technology DAP: Di-Ammonium Phosphate University of Ghana http://ugspace.ug.edu.gh xiii DNA: Deoxyribo Nucleic Acid FM: Floral Meristem GBS: Genotyping by Sequencing IBPGR: International Board for Plant Genetic Resources ICRISAT: International Crops Research Institute for the Semi-Arid-Tropics IPGRI: International Plant Genetic Resources Institute LOD: Logarithm of Odds MARS: Marker Assisted Recurrent Selection NCBI: National Center for Biotechnology Information PCA: Principal component analysis PCR: Polymerase Chain Reaction QTL: Quantitative Trait Locus REML: Restricted Maximum Likelihood RFLPs: Restriction Fragment Length Polymorphisms RIL: Recombinant inbred line SAM: Shoot Apical Meristem SIM: Simple interval mapping SM: Spikelet Meristem SNPs: Single Nucleotide Polymorphisms SPM: Spikelet Pair Meristem SSRs: Simple Sequence Repeats University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE GENERAL INTRODUCTION Cultivated sorghum [Sorghum bicolor (L.) Moench] is an annual C4 photosynthetic monocot, diploid with haploid chromosome number of 10 (2n =2x =20). The physical size of its genome is 730 Mbp (Paterson et al., 2009). Sorghum crop ranks fifth in production after wheat, maize, rice, and barley and has a predominant contribution towards food and fodder security in the arid and semi-arid regions of the world (Srinivas et al., 2009). Food security for an increasing world population depends largely on the ability of plant breeders to increase sorghum grain yield, which depends on several yield components (Pushpendra et al., 2006). Sorghum grain yield increase is of primary interest in food security for millions of rural families in the arid and semi-arid areas of the world. Grain yield is determined by the total number of grains produced and their average weight (Peltonen-Sainio et al., 2007). Final grain size and number are interrelated since genetic variation in seed size is often compensated by an adjustment in seed number (Egli, 2006; Sadras, 2007). Grain yield improvement remains one of the major breeding objectives of many cereal improvement programs. Thus the study of association among traits becomes a prerequisite to develop comprehensive selection criteria to achieve this objective (Ezeaku and Mohammed, 2006). Sorghum grain yield has been reported to have low heritability (Bello et al., 2007). The indirect selection for yield related characters with high heritability might be more effective than direct selection for yield. Grain size and number have been targeted for breeding to improve sorghum grain yield (Pushpendra et al., 2006). The genetic control of the number and growth of branches is very important owing to its effect on grain yield (Kellogg, 2007). Sorghum grain yield improvement involves the dissection of yield components such as the number of branches per University of Ghana http://ugspace.ug.edu.gh 2 panicle, the length of the panicle, number and weight of the grain. All of these yield components are under genetic control and are affected either positively or negatively by panicle architecture. Sorghum panicle architecture which includes panicle shape, compactness, branching pattern and seed density is a backbone for targeting grain yield improvements in sorghum. Primarily determined by branching patterns, panicle architecture affect the number of seeds (Zhao et al., 2006), by the developmental fate of shoot apical meristems and variation in branching patterns leading to architectural diversity (Prusinkiewicz et al., 2007). The regulatory mechanism affecting panicle branching may also influence the number of spikelets and consequently has a direct effect on grain production. Therefore sorghum panicle architecture should be of primary interest to breeders for enhancing yield potential. Rice and maize are the leading cereals studied most often to characterize the number of genes involved in the control of inflorescence architecture (Bommert et al., 2005) and inflorescence morphology which is determined by the architecture of the underlying gene regulatory networks (Prusinkiewicz et al., 2007). Sorghum, that shows tremendous variations in panicle architecture, has been little studied (Brown et al., 2006). This may be because of its complex structure than other crops, more diverse and therefore difficult to phenotype. Variation in sorghum panicle architecture is particularly pronounced for the length of the primary branches, which are generally much longer at the base and shorter at the top of the panicle; the position of the node bearing the longest primary branches; the number of primary and secondary branches; rachis diameter, which is mostly bigger at the base of the panicle; number of grain per primary branch that becomes fewer from the base to the top of the panicle and the number and length of the internodes. Unfortunately, the genetic basis for this variation is little understood. Brown et al. (2006) reported that variation in sorghum panicle architecture results from differences in branching, elongation and branch abortion and indicated that the number of branches at each order of branching varies. The University of Ghana http://ugspace.ug.edu.gh 3 need to improve the understanding of the genetic basis of this striking variation in panicle architecture is necessary to better guide breeding strategies directed towards improving sorghum grain yield. Breeding efforts in Mali have been primarily for higher grain yield potential through conventional breeding; hybrid development using male sterility brought some progress in grain yield improvement. Sorghum yield potential, however, compared to other cereals such as maize and rice still remains consistent and low in Mali. Lately the integration of molecular markers into breeding programs has offered promising alternatives to improve the potential yield using approaches such as Marker Assisted Recurrent Selection (MARS) in populations derived from elite by elite crosses. Molecular markers are useful tools to dissect quantitative variation of highly polygenic traits such as panicle architecture traits and grain yield into Quantitative Traits Loci (QTL). The QTLs/markers information can be manipulated in breeding programs as Mendelian factors providing a way to improve breeding efficiency for those traits. QTL mapping provides estimates for number, position, effects, and interactions between QTLs. Therefore, it gives more information about the genetic architecture of quantitative traits. For sorghum panicle architecture traits, such information should help in designing breeding strategies to use for targeting valuable traits in breeding programs. The identification of candidate genes based on protein sequence similarity in related species is the first step in deciphering gene function in crops. Such approach can be used in sorghum by using rice and maize genome databases to propose a list of genes involved in the molecular control of sorghum inflorescence architecture. Very few studies have been done on sorghum panicle architecture. These studies focused on the identification of QTLs for panicle traits, genetic correlation and broad sense heritability estimation (Pereira et al., 1995; Hart et al., 2001; Brown et al., 2006; Srinivas et al., 2009). The research University of Ghana http://ugspace.ug.edu.gh 4 questions, however, are to address the relationship between panicle architecture and grain yield; the determination of the pattern of panicle branching and the number of grains at different positions (base, medium and top) on the primary branches and candidate genes controlling sorghum panicle architecture. The detailed dissection of panicle traits in relation to grain yield was an important starting point. The genetic control of panicle architecture and the relationships between panicle architecture in sorghum and related cereal crops for candidate genes should be an innovative investigation. The objectives of this study were to: 1.) assess the extent of genetic diversity in panicle architecture and yield in an F4 population of sorghum; 2.) identify and map QTLs associated with panicle architecture and grain yield in sorghum; and 3.) identify sorghum candidate genes involved in inflorescence architecture using previously published evidences in rice and maize. University of Ghana http://ugspace.ug.edu.gh 5 CHAPTER TWO 2.0. LITERATURE REVIEW 2.1. Sorghum 2.1.1. Economic importance Sorghum (Sorghum bicolor (L). Moench) is the fifth most important cereal grain crop in the world after wheat, rice, maize and barley. Its origin is diversely discussed; but Haussmann et al. (2002) claimed that sorghum originated from West Africa and is used as staple food in the semi-arid tropics of Africa and Asia. Sorghum is cultivated under rain fed and irrigated conditions in Africa, Asia, South America, and the United States (House, 1985). In Mali, based on tonnage, sorghum is the second most important cereal after millet. Sorghum is used traditionally but lately its industrial utilization has become important. Sorghum is processed into food and also used in the brewing industry for making beers. It is also used as a fodder in the green or dry forms and feed crop. Sorghum constitutes raw material for biofuel, starch, dextrose syrup, fiber, alcohol, soft porridge and malt extract production using wet-milling processes at industrial scale (Liang and Gao, 2001). Therefore, sorghum presents many advantages to economic development and constitutes the cornerstone of food security for millions of poor people in the arid and semi-arid areas of the world. Breeding objectives depend on the socioeconomic and production contexts. In industrialized countries such as the United States, Europe, Argentina, and Brazil, the production of sorghum is mainly obtained from hybrids and destined for animal feed. The principal breeding criteria focus on adaptation to mechanized cultivation and high yield performance. In tropical agriculture zones, countries of West, Central and East Africa and India, sorghum cultivation destined for human consumption is mainly achieved using pure lines as varieties. Selection objectives concern the exploitation of local varieties or introduced materials. According to House (1985), a number of University of Ghana http://ugspace.ug.edu.gh 6 sorghum traits were changed through breeding processes. These changes concerned the decrease in rachis internode length and mainly the increase in number of primary, secondary branches per panicle and seed size. Also breeding goals for sorghum shifted from wide adaptability to specific regional adaptation. In Africa, the focus was on breeding for high yielding material for good grain quality and/ or forage by developing multi-purpose varieties. Photoperiod sensitivity in sorghum plays important role in its adaptation and affect the production. Drought, high temperature and nitrogen deficiency are the major abiotic constraints. Diseases such as grain mold, rust, smut, anthracnose, ergot and bacterial streak, insect pests such as the shootfly and stalk borer; parasitic weeds such as Striga are the main biotic constraints. Therefore, breeding objective for introduced material, resistance to Striga, diseases and insect pests were addressed. Later on, the development of early maturing material, improved grain quality and adaptability to drought-prone environments were of interest. Lately breeding efforts have been driven by the integration of molecular tools that speed up selection potential at early generation and the development of improved varieties. Therefore, the main variety types developed by breeding programs are breeding lines, although photoperiod sensitive hybrids are also being developed. 2.1.2. Different types of sorghum and varietal assortment Sorghum [Sorghum bicolor (L.) Moench] belongs to the tribe Andropogoneae; group Sorghastrae; and genus Sorghum (Doggett, 1988). It has been classified into 31 species by Snowden (1936). De Wet and Harlan (1971), Harlan and De Wet (1972) and De Wet (1978) simplified the classification by defining three species: Sorghum halepense which is a perennial tetraploid; Sorghum propinquum, a perennial diploid; and Sorghum bicolor that is an annual diploid with haploid number of 10 chromosomes. Sorghum bicolor is composed of three subspecies: bicolor, University of Ghana http://ugspace.ug.edu.gh 7 arundinaceum and drummondii. Cultivated sorghum belongs exclusively to the subspecies bicolor and spikelet, inflorescence and plant characteristics were traditionally used to classify this subspecies. Harlan and de Wet (1972) proposed the classification of Sorghum bicolor into five basic races named bicolor, guinea, caudatum, kaffir and durra and 10 intermediate races obtained from the combination of the different basic races. Sorghum panicle types (Figure 2.1) show correlation with spikelet morphology. Therefore, the races bicolor and guinea spikelets were found generally on the more open panicles (types 2, 3, and 4); kafir and durra spikelets were associated with more compact panicle types (types 5, 6, and 7). Caudatum spikelets were observed in wide range of panicle types. Figure 2.1: Head types of cultivated sorghum. Type I consists for wild races and is considerably more diffuse than type 2 [Original picture from Harlan and de wet, (1972)] The bicolor race has long, clasping glumes, elongate seed, and open panicles. Long glumes and open panicles are noted in the guinea race; this race is more evolved than the bicolor race. Turtle- backed grains is the characteristic of the caudatum race. Intermediate races involving caudatum race are in general high yielding. Therefore many derivative hybrids from caudatum race were developed in Africa, particularly in Mali. The spikelet characteristics of kaffir race show less University of Ghana http://ugspace.ug.edu.gh 8 specialization than those of the guinea, caudatum, and durra. However, kaffir panicle is generally semi-compact to compact. The durra race exhibit specialization for panicle morphology than all the cultivated sorghums. Figure 2.2: Sorghum races A = bicolor, B = durra, C = caudatum, D = guinea and E = kafir (Picture adapted from Harlan and de wet, 1972). 2.1.3. Types of panicles in sorghum Rangaswani Ayyangar and Rajabhooshanam (1938) defined four types of sorghum panicles. The type I comprises conical panicles characterized by a steady decrease of the primary branches from the first to the last node at the top. An example is the race guinea. The type II comprises ovoid, ellipsoid and obovate panicles in which primary branches increase in length in the first few nodes and then decrease and an example is the race durra. The type III comprises cylindrical panicles where the length of the primary branch fluctuates within narrow limits from node to node; an example is the race caudatum. The type IV comprises truncate and corymbiform panicles. The length of the primary branches increase steadily from the first to the last node and an example is the broom-corn. Sorghum descriptor that includes twelve types of panicles based on the shape and the compactness of the panicle. University of Ghana http://ugspace.ug.edu.gh 9 The characteristics of the different panicle types are: very lax panicle (typical of wild sorghum); very loose erect primary branches; very loose drooping primary branches; loose erect primary branches; loose drooping primary branches; semis-loose erect primary branches; semis-loose drooping primary branches; semi-compact elliptic; compact elliptic; compact oval; half broom corn and broom corn (IBPGR & ICRISAT, 1993). 2.1.4. Sorghum genome and relationship with members of the Poaceae family Paterson et al. (2009) studied the sorghum genome in relation to other members of the Poaceae family showed that Sorghum bicolor (L.) Moench genome physical size is 730 megabases (Mb). It is essentially composed of heterochromatin that accounts for 62% (460 Mb) of the genome while in rice, heterochromatin occupies 15% (63 Mb) of the genome. This difference between sorghum and rice involved long terminal repeat (LTR) retro transposons. The proportion of DNA transposons in the genome account for 7.5% in sorghum, 2.7% in maize and 13.7% in rice. Sorghum genome exhibits intermediate content of retro transposons compared to those of maize and rice. Sorghum shows about 55%, maize 79% and rice twenty six percent 26% of retro transposons. Sorghum resembles rice in having a higher ratio of gypsy-like to copia-like elements than maize. Most paralogues in sorghum are proximally duplicated, including 5,303 genes in 1,947 families of genes. The number and sizes of sorghum gene families are similar to those of rice (Figure 2.2). About 24% of gene families belong to sorghum and rice. Seven percent of gene families are unique to sorghum. University of Ghana http://ugspace.ug.edu.gh 10 Figure 2.3: Orthologous gene families between sorghum, Arabidopsis, Rice and poplar. [Original picture from Paterson et al. (2009)]. The numbers of gene families (clusters) and the total numbers of clustered genes are indicated for each species and species intersection. 2.2. Flowering in sorghum 2.2.1. Floral biology: inflorescence biology and morphology Singh et al. (1997) described the process of sorghum panicle formation, development and the different floral organs. Sorghum panicle is formed by the gradual differentiation of the vegetative primordium or the growing tip into the reproductive primordium (Figure 2.4). University of Ghana http://ugspace.ug.edu.gh 11 Figure 2.4: Gradual primordial differentiation into sorghum inflorescence. [Original picture from Singh et al. (1997)] Sorghum panicle development starts by the elongation of the shoot apex into the rachis of the panicle. The rachis tapers off towards the top and is grooved longitudinally. The rachis elongates after increasing in dimension and forms branches and branchlets through the formation of primary, secondary and tertiary branch primordia. Two paired spikelet primordia are formed from the development of the tips of the tertiary branches primordium. One of the paired spikelet is hermaphrodite and the other staminate. Often, one hermaphrodite and two staminate spikelets can be formed. Spikelets and florets development in the panicle continue to be covered by the flag leaf. According to House (1985), the primordial differentiation into floral parts may take about 30 days after sowing. Sorghum panicles present striking variation in morphology. They can be compact, semi-compact to open. Singh et al. (1997) describe the following: the peduncle elongation forces the panicle out of the leaf sheath after flag leaf unfolding. Variation in shape, size, and length of sorghum panicles is related to the variation in rachis, branch and internode length associated with the angle of University of Ghana http://ugspace.ug.edu.gh 12 branching. Sorghum spikelet development is basipetal meaning that the upper region of the panicle develops earlier than those in the lower. A raceme consists of one or several spikelets; one spikelet of a raceme is always sessile and the other pedicellate. The length of the raceme varies according to the number of nodes and the length of the internodes. The shape of sessile spikelets ranges from lanceolate to almost round or ovate. There are two glumes which vary from hairy to non-hairy. The glumes are hard and tough with nerves, and are obscure except near the tip. The lower glume is enclosed by the upper glume with its membranous margin. The seed may be enclosed by the glume or may protrude from it either partially or completely. The number of sessile spikelets per panicle in cultivated sorghum varies from 2000 to 4000 (House, 1985). 2.2.2. Flower emergence: from floral induction to flower development The Shoot Apical Meristem (SAM) undergoes several changes (floral induction, initiation and differentiation) to become an inflorescence meristem which is composed of several meristems (BM, branch meristem; FM, floral meristem; SM, spikelet meristem; SPM, spikelet pair meristem) (Barazesh and McSteen, 2008). 2.2.2.1. Floral induction Floral induction is a qualitative change where the transition of meristem development from vegetative to reproductive phase occurs. Floral induction is a separate event from floral development in most of cultivated sorghum, it requires a short-day signal from the leaves (Morgan and Quinby, 1987) for sorghum sensitive to day length. University of Ghana http://ugspace.ug.edu.gh 13 2.2.2.2. Floral initiation Verbeke and Heupel. (1990) studied sorghum floral initiation and floral differentiation and found that sorghum is useful for the study of floral initiation because it exhibits a determinate growth habit in which the single apical meristem first produces leaves and then is completely converted to the production of floral primordia. Morphological changes that occur at the shoot apex in sorghum correspond to an increased mitotic activity in the apical meristem that initiate changes from the vegetative to the floral state (Figure 2.5). The apical meristem appears as a dome covered by leaf initials in the vegetative state (fig. 2.5a, 2.5b). Leaf primordia start as a series of single ridges (Fig. 2.5a) which continue to be deposited by the meristem (Fig. 2.5b) until a floral bud forms (Fig. 2.5c- 2.5g). The stage shown in Figure 2.5c is equivalent to the stage 2 described by Lane. (1963). This scale has been used to indicate the stage of apical development in a wide variety of cereal crops (Large, 1954; Lane, 1963; Williams and Morgan, 1979). Floral initiation is the culmination of meristematic growth in sorghum. The first microscopically visible sign of floral initiation is a smooth apical dome not covered by leaf initials (Fig. 2.5c). In sorghum, initiation and floral development are separate events (Morgan and Quinby, 1987). 2.2.2.3. Floral differentiation Floral differentiation starts with morphological changes of meristems with the formation of the apex dome (Fig 2.5e to 2.5g) and continues until the formation of floral buds. Floral primordial appears in the meristem and then advance progressively. The floral differentiation ends with the development of the primordia of floral organs (Verbeke and Heupel., 1990). Floral differentiation in sorghum terminates leaf differentiation and thereby regulates plant size (Pao and Morgan, 1986). Almost all internodal elongation in sorghum occurs during floral University of Ghana http://ugspace.ug.edu.gh 14 differentiation (Morgan et al., 1977). This process leads to the formation of the different parts of the panicle architecture. However, at this stage one cannot tell which form (open or compact) of the panicle will be formed. Figure 2.5: Scanning electron micrographs of Sorghum bicolor apices showing the transition from vegetative to floral state. 1a and 1b are vegetative; leaf primordial are evident at the arrow. 1c through 1g are floral, with floral primordial evident at the arrow heads. 1a through 1g are all printed at same magnification, with magnification bar shown in 1e. 1a through 1d boxed areas on left enlarged 2x on right (Original picture from Verbeke et al., 1990). 2.3. Panicle architecture The sorghum panicle is composed mainly of the rachis, branches and spikelet pairs (Figure 2.2). The rachis has several nodes spaced all along the rachis. The space between two nodes is called internode. The diameter of the rachis become progressively slender from the base to the top. Primary branches developed directly on the rachis are composed of two parts: the first part is the University of Ghana http://ugspace.ug.edu.gh 15 sterile portion corresponding to the distance to the first secondary branch and the second part is the fertile portion that bear secondary, tertiary branches and spikelet pairs. The position of the longest primary branch is variable and the spikelet pair on which the sorghum grain is formed can be borne on the rachis, primary, secondary and tertiary branches. Figure 2.6: Components of sorghum panicle architecture traits investigated (modified, Witt-Hmon et al., 2013). 2.3.1. Principles of plant architecture: growth, elongation, branching Plants display a variety of architecture defined by the degree of branching, internodal elongation and shoot determinacy (Wang and Li., 2006). University of Ghana http://ugspace.ug.edu.gh 16 2.3.1.1. Growth The shoot apical meristem generates plant components such as leaves, stems and branches in an indeterminate growth and differentiation manner. Therefore, the main activities of the shoot apical meristem during plant development is to maintain the pluripotent stem cells, to form lateral organs and stems that determine different plant species architecture (Foucher et al., 2003). Axillary meristems are major determinants of plant architecture. They are involved in changes in cell proliferation and growth. This process is under gene expression and hormonal control (Tanaka et al., 2013). Branch formation involves two steps: initiation of new axillary meristem followed by their subsequent growth and development. Phytohormone biosynthesis and response play an important role in growth of the inflorescence shoot. Mutations like AUXIN RESISTANT1 (AXR1) affect branching of the inflorescence shoot (Rouse et al., 1998; Stirnberg et al., 1999), while gibberrellic acid biosynthesis and response mutants, like ga5 and GIBBERELLIC ACID INSENSITIVE1 (GAI), affect floral internode growth (Kobayashi et al., 1994; Peng et al., 1997). 2.3.1.2. Elongation Stem elongation determines plant height and affects plant architecture and grain yield. The developmental transition from vegetative to reproductive stage induces internode elongation. The elongated internodes are the response to the growth and development of intercalary meristems in the internodes when the transition begins. Gibberellins belong to the family of tetracyclic diterpenoids; they play important roles in plant growth and developmental processes, especially in stem elongation (Hooley, 1994; Swain and Olszewski., 1996). Brassinosteroids (BRs) are natural plant growth hormones that promote compounds similar to steroids. They affect plant growth and development at very low concentrations (Clouse, 1996; Clouse and Sasse., 1998). The role of University of Ghana http://ugspace.ug.edu.gh 17 brassinosteroids was confirmed in rice stem elongation (Yamamuro et al., 2000). However, the role of such natural hormone is not reported in sorghum. 2.3.1.3. Branching Caraglio and Barthélémy (1997) described plant architectural variation in terms of the orientation of branches (orthotropic or plagiotropic); type of branching (monopodial or sympodial); degree of lateral shoot development as a function of their position on the mother branch (acrotony, mesotony or basitony); type of meristematic activity (rhythmic or continuous); number of internodes per growth unit; leaf arrangement (phyllotaxis) and position of reproductive organs on the branches (terminal or lateral). Plant architecture is a dynamic concept reflecting plant development over time. Halle et al. (1978) claim that branching patterns such as the position, the form resulting from the expression of meristematic activities contribute to plant architecture. Similarly grass inflorescence architecture is determined mainly by the branching pattern, from which the spikelet meristems initiate (Tanaka et al., 2013). 2.3.2. Application of plant architecture principles on sorghum panicles Doust and Kellogg (2002) have demonstrated that meristem and its activities play a central role in the generation of varied inflorescence architecture of plants. Plant architecture modeling proposed by Barthélémy and Caraglio, (2007) can be transposed to inflorescences (Jabbour and Citerne, 2010). In sorghum, the description of branching topology refers to the arrangement of branches on the rachis and their positions. Description of the geometry of sorghum panicles includes the lengths of internodes and branches; rachis diameter and the magnitudes of the branching angles. University of Ghana http://ugspace.ug.edu.gh 18 Singh et al. (1997) found that the grand growth stage in sorghum is due to cell elongation and growth during the vegetative phase. They also indicated that floral initiation leading to inflorescence development involves: - elongation of the apical meristem - differentiation of primary branches primordia on the floral apex - differentiation of secondary branches primordia - development of secondary and tertiary branches primordia - elongation of the panicle - formation of the panicle branches - formation of fertile (sessile) and sterile (pedicellate) spikelet on the branches. 2.4. Genetic control of panicle development 2.4.1. Genes involved in panicle development 2.4.1.1. Inflorescence meristem development The transition from vegetative to reproductive state is the first step in inflorescence development. LEAFY (LFY) act upstream in a transcriptional framework by regulating the transition to flowering in the expression pattern of the floral homeotic ABC genes (Weigel et al., 1992). It also acts on the expression pattern in spikelet pair meristem and floral meristem in the developing floral organ primordia. In rice, LEAFY HULL STERILE1 (LHS1)/OsMADS1 is involved in spikelet meristem identity, because loss-of-function mutants showed defected identity of the lemma/palea (Jeon et al., 2000). A putative orthologous of LEAFY was mapped on chromosome 6 in sorghum by Brown et al. (2006). University of Ghana http://ugspace.ug.edu.gh 19 2.4.1.2. Spikelet pair and branch meristem development The gene BARREN STALK1 (BA1) in maize controls the early developmental switch that is involved in the initiation of the primary branches. The BA1 gene in maize encodes for orthologous basic helix–loop–helix transcription factors. It is expressed between pre-existing and newly initiated meristems; for instance between the shoot apical and axillary meristems, inflorescence meristem and primary branch meristem, primary branch meristem and spikelet pair meristem, upper floral meristem and lower floral meristem (Komatsu et al., 2003a). Ramosa 1 (ra1) and ramosa 2 (ra2) play important roles in branch determinacy; only the expression pattern of ramosa 2 (ra2) is conserved in sorghum, rice, barley and maize (Bortiri et al., 2006). 2.4.1.3. Spikelet meristem development Branch meristems induced by genes such as BIF2, BA1 and LAX acquire new identities and initiate spikelet meristems. FRIZZY PANICLE (FZP) in rice and BRANCHED SILKLESS1 (BD1) in maize and sorghum act to regulate meristem identity during the transition from spikelet meristem to floral meristem (Colombo et al., 1998; Chuck et al., 2002; Komatsu et al., 2003b). FZP and BD1 are orthologs; they encode an ethylene-responsive element-binding factor. They are expressed in analogous patterns at the junction of spikelet meristem and rudimentary glumes in rice, and inner/outer glumes in maize. Other genes regulating spikelet meristem determinacy include maize and sorghum INDETERMINATE SPIKELET1 (IDS1) genes (Chuck et al., 1998; Brown et al., 2006). The spikelet meristem first initiates a lateral floral meristem, and then converts it into the second floral meristem (Irish, 1997a; 1997b; 1998). Comparative analyses of BD1 and FZP indicate that the genetic control of spikelet determinacy is similar in rice, sorghum and maize (Chuck et al., 2002; Komatsu et al., 2003b). Similarly the analyses of BA1 and LAX suggest that their gene University of Ghana http://ugspace.ug.edu.gh 20 functions are conserved in branch meristem initiation. Therefore, spikelets that bear sorghum grain formation are affected by the branching. 2.4.2. Hormonal control of inflorescence development The control of branching by axillary meristems is under hormonal, environmental, developmental, and genetic control. Hormones play a critical role in regulating branching (McSteen and Leyser, 2005; Beveridge, 2006; Ongaro and Leyser, 2008). Auxin is required for branch meristem initiation during both vegetative and inflorescence development (Benjamins and Scheres, 2008). Cytokinin regulates meristem size and hence indirectly affects branching (Shani et al., 2006; Kyozuka, 2007). Plants are able to produce organs throughout their lifetime because the meristem reserves a pool of undifferentiated cells to maintain the meristem as lateral organ primordia are produced and the balance between meristem maintenance and differentiation controls meristem size (Barazesh and McSteen, 2008). In sorghum, there is evidence for developmental regulation of the amounts of sterols and pentacyclic triterpenes in leaves which may be correlated with floral initiation (Heupel et al., 1987). 2.5. Genetic maps and QTL mapping in sorghum 2.5.1. Molecular markers In sorghum, the commonly used marker systems include Simple Sequence Repeats (SSRs) or microsatellites, Diversity Array Technology (DArTs) and Single Nucleotide Polymorphism (SNPs). Now SNP are largely preferred to SSR. They can be automated for large datasets, such as Genotyping by Sequencing (GBS), multiplexing. Molecular techniques have been widely used to monitor differences in DNA sequence. SSRs were the most popularly used technique due to the University of Ghana http://ugspace.ug.edu.gh 21 ease in detection and automation. However, the adoption of the SNP marker system is now highly preferred due to the amount of sequence information. A good molecular marker must be: - Co-dominant: it is possible to discriminate between heterozygous and homozygous; - Highly polymorphic: reflect variability among genotypes of a cross; - Specific to a locus - Neutral: allelic substitution don’t have any selective effect on other genetic marker; - Abundant: they are found in great number in the genomes of plants, both in non-coding and coding regions. Genetic markers have great potential in studying complex traits. They reveal allelic polymorphisms and act as signals for target genes between genotypes (Collard et al., 2005). The following factors are important in marker selection: i) reliability and reproducibility, ii) easy technical procedure, iii) cost, iv) level of polymorphism, and v) required DNA quality and quantity. Progress has been achieved by gathering information on the number of loci involved in trait expression, their location along the chromosome, and the relative contribution to trait expression of each locus (Maradiaga, 2003). Sorghum genetic maps Sorghum is a diploid grass with 10 haploid chromosomes (2n=2x=20). Genetic maps provide the understanding of evolution, genetic diversity and phylogeny relevant in revealing the locations of QTLs influencing all measurable traits (Paterson, 1995). In sorghum, several linkage maps have been developed. The first genetic map of sorghum was created by Hulbert et al. (1990) using 55 F2 plants and maize DNA probes. Ragab et al. (1994) constructed a linkage map with 15 linkage groups, with a length of 633 cM and average distance of 8.9 cM. Pereira et al. (1994) developed a University of Ghana http://ugspace.ug.edu.gh 22 sorghum linkage map with 10 complete linkage groups using maize and sorghum probes. Subudhi and Nguyen (2000) aligned the 10 linkage sorghum groups using information generated from a RIL population, sorghum and maize probes, as well as cereal anchors from three different linkage maps (Chittenden et al., 1994; Ragab et al., 1994; Xu et al., 1994). Menz et al. (2002) constructed a 1713 cM high-density map using 2454 AFLPs, 136 SSRs previously mapped in sorghum, and 203 cDNAs and genomic clones from rice, barley, oat, and maize. Wu and Huang (2006) developed a sorghum genetic map using SSR markers. The number of linkage groups in sorghum can be visualized by FISH markers (Kim et al., 2005), where the designation of the chromosomes can be LG for linkage groups (Zhin-Ben et al., 2006) or SB for Sorghum bicolor (Kim et al., 2005) followed by either A, B,C, D or 1, 2, 3, 4. The arrangements of the 10 LGs or SBs, depend on chromosome length, ranging from 5.11 to 2.94 μm with an estimated DNA content ranging from 119.3 to 68.6 Mbp (Kim et al., 2005). Gloria et al. (2008) developed a molecular map for the bloom cuticle locus and fill the gap in sorghum chromosome 10. Sorghum consensus map developed by integrating the colinearity of six independent sorghum component maps into a single reference resource that contains SSRs, AFLPs, and high-throughput Diversity Array Technology (DArT) markers. This consensus map spanning 1603.5 cM, consists of 2029 unique loci (1190 DArT loci and 839 other loci) (Mace et al., 2009). An expanded consensus map was developed through the addition of 1243 markers comprising additional 888 DArTs, 229 SSRs, 81 RFLPs, and the position of 45 genes were established on the map (Mace and Jordan, 2011). The sequencing of the sorghum genome using the accession BTx623 has provided the opportunity for whole-genome sequencing approaches (Paterson et al., 2009). Therefore, an ultra-high-quality SNP map was constructed and used for QTL analysis of eight agronomically important traits under two contrasting photoperiods (Zou et al., 2012). University of Ghana http://ugspace.ug.edu.gh 23 2.5.3. Principle of QTL mapping QTL mapping is a statistical association between phenotypes of interest and molecular markers in a segregating population. It is a linkage-based method for QTL detection in the population derived from cross between two parents (Dudley, 1993). The objective of QTL mapping is to identify neutrally inherited markers that are close to the genetic causatives or genes controlling complex quantitative traits (Ross-Ibarra et al., 2007). QTL mapping requires population development such as F2 or other generation of segregation, back cross (BC), double haploid (DH), recombinant inbred line (RIL) and near isogenic line (NIL) populations, derived from the genetic hybridization of two parental genotypes with an alternative trait of interest. It also requires phenotyping of a large number of progenies or lines that are segregating for a trait of interest under different environmental conditions. Finally genotyping of the population using a set of polymorphic DNA markers that differentiate the parental genotypes and segregate among the progenies in a mapping population must be done. The two parental genotypes are screened with markers for polymorphism. The selected polymorphic markers are then used to screen the progenies of the two parents. The genotypic data generated are used to construct the genetic map of the population representing the order and position of the different markers along the chromosomes. Genetic map construction is based on the assessment of the recombination rates between marker loci so that markers along the same chromosome are statistically correlated with phenotypic characteristics of the progenies in the population. QTL detection is based on the association of genotypic and phenotypic scores of the progenies in the population. Therefore, markers that are genetically linked to a QTL influencing the trait of interest will segregate more frequently with trait values. The precision of QTL mapping depends on the genetic background, the size of the mapping population and a number of marker loci used (Ibrokhim et al., 2008). QTL mapping provides information on the genetic architecture of complex traits such as location, the number of QTLs and University of Ghana http://ugspace.ug.edu.gh 24 magnitude of their estimated additive, dominance and epistasis effects (Holland, 2007). A number of QTL mapping methods have been developed; they range from the simplest method of single- marker analysis (Sax, 1923) to more sophisticated methods such as interval mapping (Haley and Knott, 1992), multiple regression (Whittaker et al., 1996), and composite interval mapping (Zeng, 1994). Software packages for QTL mapping include MAPMAKER/QTL (Lincoln et al., 1993), QTL Cartographer (Basten et al., 1994), PLABQTL (Utz and Melchinger, 1996), QGene (Nelson, 1997), TASSEL (Buckler, 2007) and R/QTL (Broman and Sen, 2009). 2.5.4. Review of QTL in sorghum Many QTLs studies have been undertaken in sorghum on different traits including grain and panicle traits (Pereira et al., 1995; Rami et al., 1998; Hart et al., 2001; Brown et al., 2006; Feltus et al., 2006; Murray et al., 2008; Srinivas et al., 2009; Zou et al., 2012; Morris et al., 2013; Nagaraja et al., 2013). However, the relation between panicle traits and grain yield as well as the branching pattern were not clearly addressed in these studies. Also very few candidate genes involved in sorghum panicle architecture have been identified in these studies. Therefore, the need to extend QTLs studies to these gaps should be of interest. Other QTLs studies were performed in sorghum with focus on stem morphology (Lin et al., 1995; Pereira et al., 1995; Hart et al., 2001; Feltus et al., 2006; Brown et al., 2008; Murray et al., 2008; Shiringani, 2009), leaf morphology (Hart et al., 2001; Feltus et al., 2006), maturity (Crasta et al., 1999; Chantereau et al., 2004; Hart et al., 2001; Brown et al., 2006), stem composition (Murray et al., 2008; Ritter et al., 2008; Shiringani, 2009), stay-green drought tolerance (Tuinstra et al., 1996; Crasta et al., 1999; Subudhi and Nguyen, 2000; Tao et al., 2000; Kebede et al., 2001; Sanchez et al., 2002 ; Hausmann et al., 2002; Harris et al., 2007), fertility restoration (Klein et al., 2001; Jordan et al., 2010), aluminium tolerance (Magalhaes University of Ghana http://ugspace.ug.edu.gh 25 et al., 2004), and biotic stress resistance (Klein et al., 2001; Tao et al., 2003; Mohan et al., 2009; Perumal et al., 2009) and photoperiod (Chantereau et al., 2004; Murphy et al., 2011). 2.5.5. QTL mapping for inflorescence traits in cereals 2.5.5.1. Sorghum Much of the natural variation in inflorescence shape observed are actually due to the cumulative effect of several loci. The study of quantitative trait loci (QTL) is an important field of cereal genetics aimed at yield improvement. Quantitative studies have been energized recently by the advancement of genomic tools. Several QTLs associated with panicle traits such as panicle length, seed branch length, length of sterile portion of seed branch, number of seed branches per panicle, and 100-seed weight were identified by Pereira et al. (1995). Rami et al. (1998) reported QTLs for panicle length. Hart et al. (2001), on the basis of the map positions, mapped the same QTL for panicle length mapped by Pereira et al. (1995). Using two sorghum inbred lines with different inflorescences, Brown et al. (2006) mapped QTLs for number of primary, secondary and tertiary branches, branch length, and rachis diameter. Srinivas et al. (2009) reported QTLs for grain yield, panicle length, panicle weight, seed weight and number of primary branches. The QTLs for number of branches identified by Brown et al. (2006) and those of Srinivas et al. (2009) were not related. Primary, secondary, and tertiary branching in sorghum are largely under separate genetic control. There is no overlap between QTL detected for primary and secondary branch number (Bommert et al., 2005). Morris et al. (2013) found in sorghum co-localization between QTLs for panicle length and candidate genes INDETERMINATE 1 (ID1) and ABERRANT PANICLE ORGANIZATION 1 (APO1). Very little information on sorghum panicle branching pattern was addressed in those previous studies. However, many architectural QTLs were detected and this information will be University of Ghana http://ugspace.ug.edu.gh 26 used in the present study to assess the consistency of QTLs for panicle traits in sorghum. The results of this study will hopefully fill the gap by extending the knowledge on the genetic control of the branching pattern in sorghum. 2.5.5.2. Rice Rice panicle characteristics such as the number of primary and secondary branches influence spikelet number per panicle (Lin et al., 1996; Wu et al., 1996; Xiao et al., 1996). QTL mapping is the first step to dissect Mendelian factors underlying rice panicle architecture (Yano and Sasaki, 1997). Many studies on QTL mapping for spikelet number per panicle (Lin et al., 1996; Lu et al., 1996; Xiao et al., 1996; Zhuang et al., 1997; Redona and Mackill, 1998; Yagi et al., 2001 and Fujita et al., 2012) and for the number of primary branches per panicle (Lin et al., 1996; Wu et al., 1996; Sasahara et al., 1999) have been conducted on rice using mapping populations derived from inter-specific crosses, indica-japonica inter-subspecific crosses and indica-indica crosses. Yamagishi et al. (2002), detected four and two QTLs respectively for the number of primary and secondary branches per panicle. Also, they detected one QTL for average number of spikelets on one secondary branch. Yamamoto et al. (2007), identified QTLs for five morphological components of panicle architecture including number of spikelets per panicle, number of primary branches per panicle, average number of secondary branches per primary branch, panicle length, and primary branch average length. Guo and Hong. (2010) detected QTLs for the number of primary branches per panicle, number of secondary branches per panicle, secondary branch distribution density, number of spikelets per panicle and spikelet density. The QTLs studies on rice panicle may serve as a guideline in defining architectural traits in sorghum. University of Ghana http://ugspace.ug.edu.gh 27 2.5.5.3. Maize Maize tassel architecture is relevant to hybrid seed production and an amenable system for QTL analysis. Maize inbreds vary considerably for tassel branch number, branching pattern and spikelet density. Using maize tassels, Upadyayula et al. (2006) detected two QTLs for tassel branch number, five for spikelet pair density on the central spike and two for spikelet pair density on the branches. Mayor (2008) detected QTLs for tassel branch number, tassel length, central spike length and branching zone length. Kamelmanesh et al. (2012) detected six QTLs for tassel branch number. Therefore, branching pattern in maize tassel characterization may be transposed onto sorghum panicles. 2.5.6. Candidate genes In sorghum, Brown et al. (2006) mapped QTL clusters for plant height, rachis length, and branch length close to the candidate gene Dw3 and QTL for primary branch number in the region of the ramose 2 (ra2) gene. The Dw3 gene, encodes for a P-glyocoprotein responsible for auxin transport (Multani et al., 2003) and ramosa 2 gene regulates the pattern of branching. In maize tassel, ramosa 1 (ra1) was mapped closely to a QTL for branch number, thick tassel dwarf1 (td1), which is in the same region as the QTL that controls ear weight, tassel branch angle, and spikelet pair density on primary branches, ra2 which maps to the region with a QTL for kernel number per row, and fasciated ear2 (fea2), which localizes to a region with a QTL for branch number (Upadyayula et al., 2006). Ramosa 2 (ra2) expression pattern is conserved in sorghum, rice, maize and barley suggesting its important role in grass inflorescence architecture (Bortiri et al., 2006). In rice, FRIZZY PANICLE (FZP) acts as a positive regulator of floral meristem identity to suppress the formation of axillary meristems of rice spikelet. The LAX gene encodes a putative transcriptional regulator that contains a bHLH domain, and has been identified as a regulator controlling the rachis- University of Ghana http://ugspace.ug.edu.gh 28 branch meristem initiation and/or maintenance during rice reproductive development (Komatsu et al., 2003a). The rice cytokinin oxidase, OsCKX2, regulates the development of inflorescence meristems by affecting the cytokinin contents in the rice vascular system of developing culms, suggesting the important role of cytokinin in forming the inflorescence architecture (Ashikari et al., 2005). Together, RA1, RA2, and RA3 coordinate to regulate meristem identity and determinacy in the maize inflorescence. Therefore, inflorescence architecture genes that map in the vicinity of QTL that harbor beneficial traits are of significant importance in crop improvement programs. Most of the studies on genes function involved in inflorescence architecture were on rice and maize. Therefore, genes playing important roles in rice or maize inflorescence architecture may be used in sorghum as candidate genes to access their presence and suggest a possible function in sorghum. University of Ghana http://ugspace.ug.edu.gh 29 CHAPTER THREE 3.0. PHENOTYPING FOR SORGHUM PANICLE ARCHITECTURE AND YIELD IN TIANDOUGOU/LATA-3 F4 POPULATION 3.1. Introduction Sorghum (Sorghum bicolor) is an important food and fodder crop. It is a traditional staple cereal crop in Mali. Variation in sorghum panicle architecture can be used to establish linkages with grain yield potential, and indirect selection may be targeted if panicle traits are found to be associated with grain yield. Therefore, comprehensive knowledge of panicle architecture diversity and genetic relationships between panicle traits and grain yield are invaluable aid in crop improvement strategies for current and future breeding programs. According to Grenieret et al. (2001), measurement of morphological variation is the most easily obtained indicator of genetic diversity. Morphological characters are ecologically adaptive and good indicators of genetic variation, local differentiation, or ecotypes and can be used to classify the phenotypic diversity. Phenotyping of sorghum panicle architecture traits involves a methodological dissection of the panicle into its structural units such as branching pattern, rachis elongation and grain distribution. This should provide better understanding and elucidate relationships between panicle architecture and grain yield. Phenotyping sorghum panicle architecture also examine the heritability, correlation and principal component contribution traits into panicle architecture variability. Heritability is a useful quantitative parameter, which considers the role of heredity and environment in determining the expression of a character (Kukadia et al., 1983). Grain yield is a complex trait, dependent on many characters. Yield potential including a desirable combination of traits has always been the major University of Ghana http://ugspace.ug.edu.gh 30 objective of sorghum breeding programs (El-Din et al., 2012). The ability of the breeder to identify superior genotypes within segregating populations is the key to the success in any plant breeding program. Trait heritability estimation is a first step towards assessing the amount of genetic variation present in a breeding population. Regardless of the type of heritability estimate, heritability broadly defined, is the proportion of observable field variation that is due to genetic factors (Nyquist, 1991). Several studies have been conducted on sorghum populations to estimate heritability, as well as phenotypic and genotypic correlations among traits (Sanchez-Gomez, 2002; Tariq et al., 2007; Warkad et al., 2008; Ayala, 2011, Ali et al., 2012; Nagaraja et al., 2013). Discrepancies in heritability estimates for grain yield may be due to factors such as different estimation methodologies, type of population (wide crosses), diversity of environments, and generation of evaluation. Broad-sense heritability for some panicle architecture traits such as number of primary and secondary branches, rachis diameter, and primary branch length were estimated in different sorghum populations (Pereira et al., 1995; Brown et al., 2006 and Srinivas et al., 2009). Heritability in certain cases is similar to repeatability assessed by multiple measurements of the trait on each individual provided by temporal and spatial repetition (Falconer and Mackay., 1996). Genetic improvement for quantitative traits depends upon the nature and amount of variability present in the genetic stock and the extent to which the desirable traits are heritable (Chavan et al., 2010). In addition to heritability, the knowledge of genetic variability existing among different parameters contributing to yield is also an important criterion for yield enhancement. Although sorghum in Mali possesses a wide range of genetic variability, the improvement of landraces for yielding ability has not been substantial. Hybridization has boosted yield levels of sorghum, besides improving other characters like resistance to biotic and abiotic stresses. University of Ghana http://ugspace.ug.edu.gh 31 Correlation analysis provides information on the interrelationships of important plant characters and leads to a directional model for direct and/or indirect improvement in grain yield (Khan et al., 2004). Principal component analysis (PCA) is used to reveal the pattern of character variation among individual accessions in a population. It involves a procedure that transforms total variation of original characters into a smaller number of uncorrelated new characters, so called principal components (Johnson, 1998). Phenotyping sorghum panicle architecture traits become more useful to address the variability, heritability, correlation among panicle traits to grain yield. Therefore, the objectives of this research were 1.) to assess genetic variability in sorghum panicle architecture traits 2.) to estimate broad sense heritability of panicle architecture traits in F4 families, 3.) to determine correlations among panicle architecture traits in relation to grain yield and 4.) to estimate grain yield using highly heritable and easily measured panicle architecture traits. 3.2. Materials and Methods Phenotyping of sorghum panicle architecture traits was conducted at the research station of Sotuba (12°39’47’’North 7°54’50’’ West), located at Bamako, Mali in 2011 and 2012. Two sowing dates corresponding to different conditions of photoperiod were used in 2011 to phenotype and characterize the effect of delaying sowing date on panicle architecture and grain yield since sorghum reacts to photoperiod. The first sowing date coded (SB1) was performed at the beginning of the rainy season (mid- June) and the second sowing date coded (SB2) was performed 26 days after the first sowing at mid-July. In 2012, one sowing date was performed at the same sowing date as the first sowing date of 2011. Therefore, because of this similarity with SB1, the first sowing date of 2012 was coded (SB1’). University of Ghana http://ugspace.ug.edu.gh 32 3.2.1. Plant Materials A bi-parental population was obtained from the cross between two contrasting parents (Tiandougou and Lata-3) for panicle architecture traits. The female parent, Tiandougou, is a short caudatum- guinea type from the IER breeding programs (Plate 3.1a). It is a dual-purpose variety with compact panicle and high yield. Lata-3 (Plate 3.1b) used as a male is a Guinea type, with open and drooping primary branches panicles also high yielding and intermediate height from ICRISAT, Mali with open and drooping primary branches panicles also high yielding and intermediate height. The population was advanced to the F3 generation by single seed descent. Each single F3 plant was selfed and the corresponding F4 panicle harvested. In order to constitute a stock of seeds representative of each the original F3 plant, 10 F4 plants were grown for each F4 family, self- pollinated and F5 seeds harvested as a bulk (F3:5). This multiplication was conducted in 2010 during the rainy season. A mapping population composed of 401 F4 families was used for phenotyping sorghum panicle architecture traits and grain yield. Plate 3.1b: Open and drooping primary branches panicle of the Male parent panicle (Lata-3) Plate 3.1a: Compact panicle of the Female parent (Tiandougou) University of Ghana http://ugspace.ug.edu.gh 33 3.2.2. Experimental Design The experimental design was an augmented design with 29 blocks and 16 entries composed of 14 test genotypes and 2 checks per block. A total of 464 entries included 404 unique F4 families and 60 checks that were composed of 30 of each parent Tiandougou and Lata-3. The F4 families were randomized and the two parents (checks) were maintained in each block. Six grams of seed for each of the 464 entries were weighed and treated with Calthio C against pest before planting. The seedbed was plowed and compound fertilizer (DAP) applied at 150 kg/ha before ridges formation. Each entry was sown on two rows with 10 hills per row. Spacing between rows was 0.75 m and between hills on the row was 0.40 m. Standard cultural practices, including fertilization, weed and insect control, were followed at each sowing date to minimize exogenous variability that would otherwise mask variability due to genetic differences of the germplasm evaluated. 3.2.3. Data collection One representative panicle for each of the 404 F4 families and for the 60 checks were harvested, sun dried and their traits measured. Panicle architecture complexity was disaggregated into geometrical and topological variables adapted from the model used by Segura et al. (2006). In addition, traits related to grain yield such as number of grains were collected. For the strategy of phenotyping, three organization levels were considered (Figure 3.1): first, different types of axes (peduncle, panicle, rachis, primary and secondary branches and internodes); second, different places in the panicle (secondary, median and penultimate nodes and longest primary branch) and third, different types of measured variables (geometry, topology and grain yield and its components). Geometrical variables were based on the length and the diameter of different axes. Topological variables concerned number and position of internodes and branches as well as branching density. Grain yield components variables mainly concerned weight and number of University of Ghana http://ugspace.ug.edu.gh 34 grains. These measured variables were used to calculate other variables as close as possible to biological factors such as rachis conicity, slenderness and volume; primary and secondary branch density and number of grains per panicle. Figure 3.1: Methodological approach for measuring panicle architecture traits A total of 48 panicle architecture traits were evaluated during the first and second phenotyping seasons in 2011 (SB1 and SB2) and 2012 (SB1’). The data collected were structured according to the model used by Segura et al. (2006). Geometrical variables that account for 24 variables composed of 16 measured and eight calculated variables. Topological variables were 16 with nine calculated traits. Grain yield and its components account for eight including two calculated variables (Table 3.1). University of Ghana http://ugspace.ug.edu.gh 35 Table 3.1: Data on sorghum panicle architecture variables with code collected in 2011 (SB1 and SB2) and 2012 (SB1’) at Sotuba research station in Mali (modified Segura et al., 2006) Variables Formula SB1 SB2 SB1' Geometry Plant height (mm) PH x x x Panicle length (mm) PA_L x x x Peduncle length (mm) PE_L x x x Rachis length (mm) RA_L x x x Mean internode length (mm) IN_L RA_L/PA_IN_N x x x Maximum internode length (mm) IN_L_max x x x Rachis base diameter (at first internode) (mm) RA_B_Dia x x x Rachis top diameter (at last internode) (mm) RA_T_Dia x x x Rachis mean diameter (mm) RA_Dia (RA_B_Dia + RA_T_Dia) / 2 x x x Rachis conicity RA_coni (RA_B_Dia - RA_T_Dia) / RA_L x x x Rachis slenderness RA_slend RA_L / RA_Dia x x x Rachis volume (mm 3 ) RA_Vol RA_L(B_area*+T_area*)/2 x x x Primary branch length at second node (mm) PB_L_SN x x x Primary branch length at median node (mm) PB_L_MN x x x Primary branch length at penultimate node (mm) PB_L_PN x x x Length of the longest Primary Branch (mm) PB_L_max x x x Secondary branch length (mm) SB_L x x Length of the longest secondary branch (mm) SB_L_max x x Distance to the first Secondary Branch at second node (mm) SB_DF_SN x x x Distance to the first Secondary Branch at median node (mm) SB_DF_MN x x x Distance to the first Secondary Branch at penultimate node (mm) SB_DF_PN x x x Primary branch branching zone at second node (mm) PB_BZ_SN PB_L_SN - SB_DF_SN x x x Primary branch branching zone at median node (mm) PB_BZ_MN PB_L_MN - SB_DF_MN x x x Primary branch branching zone at penultimate node (mm) PB_BZ_PN PB_L_PN - SB_DF_PN x x x Topology Number of internodes per panicle PA_IN_N x x x Maximum number of primary branch per node Nb_PB_max x x x Number of primary branches per panicle PA_NPB x x x Average number of primary branches per node av_NPB PA_NPB / PA_IN_N x x x Primary branch density per panicle PB_Dens PA_NPB / RA_L x x x Number of secondary branches on the longest primary branch Nb_SB_LPB x x Number of Secondary Branches at second node Nb_SB_SN x x x Number of Secondary Branches at median node Nb_SB_MN x x x Number of Secondary Branches at penultimate node Nb_SB_PN x x x Secondary branches density at second node SB_Dens_SN Nb_SB_SN / PB_BZ_SN x x x Secondary branches density at median node SB_Dens_MN Nb_SB_MN / PB_BZ_MN x x x Secondary branches density at penultimate node SB_Dens_PN Nb_SB_PN / PB_BZ_PN x x x Relative position of the longest internode on the rachis Rel_Pos_IN_L_max Pos_IN_L_max / PA_IN_N x x x Relative position of the node with the maximum number of primary branches Rel_Pos_NPB_max Pos_NPB_max / PA_IN_N x x x Relative position of the longest primary branch Rel_Pos_PB_L_max Pos_PB_L_max / PA_IN_N x x x Relative position of the longest secondary branch on the longest primary branch Rel_Pos_SB_L_max Pos_SB_L_max / Nb_SB_LPB x x Yield Panicle grain yield (g) PA_GY x x x Panicle weight (g) PA_W x x x Panicle harvest index PA_HI PA_GY / PA_W x x x Thousand Grain Weigth TGW x x Number of grains per panicle NG_PA PA_GY / (TGW/1000) x x Number of grains at second node NG_SN x x x Number of grains at median node NG_MN x x x Number of grains at penultimate node NG_PN x x x Formula are detailed for calculated variables * B_area = π (RA_B_Dia / 2)²; T_area = π (RA_T_Dia / 2)² University of Ghana http://ugspace.ug.edu.gh 36 3.2.4. Data analysis The objectives of these analyze were to determine variance components due to genotype and blocks and to test the significance of the different panicle traits in the three sowing dates in order to generate adjusted means for each trait to be used for QTLs detection. Patterson and Thompson, (1971) proposed the mixed model, REML (Residual Maximum Likelihood) to perform analysis of variance for incomplete block design such as augmented design. The principle relies on random and fixed term. Therefore, variance component analysis was used in Genstat software to establish statistical differences between the F4 families. Two models were automatically fitted. The first model was used to estimate variance components in which the F4 families were fitted as random term (model 1). A second model was fitted in which the F4 families were used as fixed terms (model 2). This was to produce the adjusted means for each phenotypic trait and to test the significance of differences among F4 families underlying on the statistic proposed by Wald (1943) and F statistic for fixed effects. The Student T Test was performed to compare the mean of the parents to the means of the F4 progenies using the software MINITAB version 16. The linear model for augmented design was as follows: yij = μ+ αi + βj + φk+βφjk+ εijk Where Yij is the phenotypic value μ is the population mean; αi is the effect due to the i-th genotype test; βj is the effect due to the j-th genotype check; φk is the effect due to the k-th block; βφjk is the effect due to the interaction of the i-th genotype check with the j-th block; εijk is the error associated with the ij-th individual observations. University of Ghana http://ugspace.ug.edu.gh 37 3.2.5. Correlation, Principal Component Analysis and ANOVA for panicle traits The adjusted means of data for the F4 families was used to perform all other analysis such as Pearson correlation to determine association among traits, and Principal Component Analysis (PCA) to evaluate contribution of traits. Pearson correlation with significant test was computed on the phenotypic and genotypic data for all the panicle architecture traits and individually on each subgroup of traits such as geometry, topology and yield to generate the different correlation matrices. PCA based on the correlation coefficient was computed first for all traits and for geometry, topology and yield. Shapiro test of normality (Shapiro and Wilk, 1965) prior to the linear model of ANOVA (type III) was conducted on the phenotypic data. The linear model is: Yij = µ + Gi + Dj Where: Yij = phenotypic value µ = population mean Gi = genotypic factor Dj = sowing date factor 3.2.6. Calculation of Broad sense heritability According to Knapp et al. (1985), broad-sense heritability of phenotypic means can, in each case, be expressed as follows: H² = 1 – (1/F) Where F is the Fisher statistic for the genotypic effect of the considered model in the ANOVA. University of Ghana http://ugspace.ug.edu.gh 38 3.2.7. Construction of linear model and BLUP extraction A mixed model was fitted with genotype as random effects and sowing dates as fixed effects, the residual was analyzed and the Shapiro normality test applied prior to the Best Linear Unbiased Predictor (BLUP) extraction for each trait. 3.2.8. Analysis on BLUP Pearson correlation with a significant test was computed on the BLUP data to determine the genotypic correlation matrix using the Rcorr function of the Hmisc R package. Principal Component Analysis (PCA) was computed using the PCA function of the FactoMineR R package. 3.2.9. Contribution of panicle traits to panicle grain yield A multivariate regression was computed to establish the contribution of panicle architecture traits to the grain yield according to a linear model. University of Ghana http://ugspace.ug.edu.gh 39 3.3. Results 3.3.2. Experimental design data analysis REML component analysis revealed significant differences between the 401 F4 families for a great number of traits (Table 3.2). Table 3.2: F probability (Fpr) of REML analysis on 401 F4 families across three sowing dates at Sotuba research station, Mali, 2011 and 2012 Panicles traits FprSB1 FprSB2 FprSB1’ Plant height (cm) < 0.001 < 0.001 <0.001 Peduncle length (cm) < 0.001 < 0.001 <0.001 Panicle length (cm) < 0.001 < 0.001 <0.001 Rachis length < 0.001 < 0.001 <0.001 Length of the longest Primary Branch 0.062 0.021* 0.002** Average primary branch length 0.007** < 0.001 0.002** Length of Primary Branch at second node 0.153 0.001** 0.051 Length of Primary Branch at median node < 0.001 0.497 0.056 Length of Primary Branch at penultimate node < 0.001 < 0.001 <0.001 Average internode length 0.004** < 0.001 0.020* Rachis base Diameter 0.999 0.131 <0.001 Rachis top Diameter 0.210 0.563 <0.001 Number of Primary Branches per panicle < 0.001 < 0.001 <0.001 Average primary branches per node < 0.001 < 0.001 <0.001 Number of internodes per panicle 0.005** 0.005** <0.001 Number of Secondary Branches at second node 0.483 0.014* 0.099 Number of Secondary Branches at median node 0.050 0.002** 0.077 Number of Secondary Branches at penultimate node < 0.001 < 0.001 0.009** Average secondary branch per primary branch 0.091 0.026* 0.049* Number of grains at second node < 0.001 0.012* <0.001 Number of grains at median node < 0.001 < 0.001 0.046* Number of grains at penultimate node < 0.001 < 0.001 <0.001 Average number of grains per primary branch < 0.001 < 0.001 <0.001 Grain yield per panicle (g) 0.047* 0.147 0.005** Panicle weight (g) 0.122 0.356 0.017* Panicle Harvest index 0.081 0.053 <0.001 NB: Fpr = F probability of REML analysis; SB1 = first sowing date of 2011; SB2 = second sowing date of 2011; SB1’ = first sowing date of 2012; <0.001 = highly significant at P<0.001; ** = highly significant at P< 0.01; * = significant at P< 0.05. University of Ghana http://ugspace.ug.edu.gh 40 In the second sowing date of 2011, grain yield per panicle did not show significant differences among the progenies. However, number of secondary branches at second and median nodes as well as the length of the primary branches at the second node show significant differences among the progenies compare to the first sowing dates in 2011 and 2012. 3.3.3. Performance of the F4 families and their parents The Student T Test revealed that grain yield per panicle and length of primary branch at median node were reduced compared to the two parents in the second sowing date in 2011. The F4 families performed better than the parent Lata-3 in all the sowing dates for number of grains and the length of the primary branch at the penultimate node. Highly significant differences (P < 0.001) between the mean of the parent Tiandougou and the mean of the F4 families were observed for the length of the longest primary branch, number of secondary and grain on primary branch at second node as well as rachis length across the different sowing dates. Peduncle length, number of internode and rachis top diameter were reduced in the F4 families compared to the parent Tiandougou in the second sowing date of 2011 (Table 3.3). University of Ghana http://ugspace.ug.edu.gh 41 Table 3.3: Mean comparison of the parents and F4 progenies, T test probability of the parents Lata- 3 and Tiandougou vs progenies across the three sowing dates at Sotuba research station, Mali Traits Mean F4 PSB1 Lata PSB2 Lata PSB1’ Lata PSB1 Tiand PSB2 Tiand PSB1’ Tiand Plant height (cm) 172.4 1 1 0.836 0.000 0.000 0.138 Peduncle length (cm) 37.0 1 1 1 0.000 0.086 0.000 Panicle length (cm) 33.2 1 1 1 0.158 0.000 0.000 Rachis length 30.2 1 1 1 0.000 0.000 0.000 Length of the longest Primary Branch 10.0 1 1 1 0.000 0.000 0.000 Length of Primary Branch at second node 8.5 0.000 1 1 0.14 0.000 0.000 Length of Primary Branch at median node 7.8 1 0.000 1 1 0.000 1 Length of Primary Branch at penultimate node 3.8 0.000 0.000 0.000 0.376 1 1 Rachis base Diameter 7.1 1 1 0.000 1 0.018 1 Rachis top Diameter 0.8 0.001 0.675 0.197 1 0.000 0.68 Number of internode per panicle 15.5 1 1 1 0.000 1 0.000 Number of Primary Branches per panicle 100.2 0.943 0.000 0.994 0.537 0.989 0.000 Number of Secondary Branches at second node 7.9 1 1 0.955 0.000 0.833 0.000 Number of Secondary Branches at median node 5.8 0.026 0.000 1 0.026 0.000 0.000 Secondary Branches at penultimate node 2.7 0.000 0.000 0.000 0.000 0.001 1 Number of grain at second node 66.5 0.000 1 0.000 0.000 0.000 0.000 Number of grain at median node 37.7 0.000 0.000 1 0.928 0.529 1 Number of grain at penultimate node 9.9 0.000 0.000 0.000 1 1 1 Grain yield per panicle (g) 69.0 1 0.039 1 1 0.000 1 Panicle weight (g) 92.2 1 0.000 1 1 1 0.743 Panicle Harvest index 0.8 1 1 0.000 0.000 0.000 0.000 PSB1 Lata = T test probability for Lata-3 vs progenies in the first sowing date in 2011; PSB2 Lata = T test probability for Lata-3 vs progenies in the second sowing date in 2011; PSB1’Lata = T test probability for Lata-3 vs progenies in the first sowing date in 2012; PSB1 Tiand = T test probability for Tiandougou vs progenies in the first sowing date in 2011; PSB2 Tiand = T test probability for Tiandougou vs progenies in the second sowing date in 2011; PSB1’Tiand = T test probability for Tiandougou vs progenies in the first sowing date in 2012. 3.3.4. Analysis of Variance (ANOVA) Analysis of variance showed significant differences between the genotypes and between the sowing dates for most of the sorghum panicle architecture and grain yield related traits (Table 3.4). University of Ghana http://ugspace.ug.edu.gh 42 Table 3.4: ANOVA for sorghum panicle traits and grain yield and its components Traits Factors Sum Sq Df F value Pr(>F) Significance PH Genotype 130295324 400 10.696 < 2.20E-16 *** Date 29257835 2 480.37 < 2.20E-16 *** PA_L Genotype 1312511 400 3.4362 < 2.20E-16 *** Date 85253 2 44.638 < 2.20E-16 *** IN_L Genotype 8615 400 2.3151 < 2.20E-16 *** Date 1348.3 2 72.467 < 2.20E-16 *** PB_L_SN Genotype 199210 400 2.1354 < 2.20E-16 *** Date 9822 2 21.057 1.232E-09 *** PB_L_MN Genotype 131922 400 1.7239 6.205E-11 *** Date 5808 2 15.18 3.397E-07 *** PB_L_PN Genotype 122931 400 1.6182 6.994E-09 *** Date 10378 2 27.321 3.374E-12 *** SB_L Genotype 10257.6 400 1.7033 6.481E-08 *** Date 12.5 1 0.8301 0.3628 not SB_DF_SN Genotype 23831 400 1.453 5.68E-06 *** Date 238 2 2.8976 0.05575 not SB_DF_MN Genotype 21548.4 400 1.5094 6.386E-07 *** Date 59.2 2 0.8293 0.4368 not SB_DF_PN Genotype 10598.8 400 1.1519 0.049613 * Date 285.5 2 6.2059 0.002118 ** RA_B_Dia Genotype 844.76 400 2.132 < 2.20E-16 *** Date 60.52 2 30.551 1.664E-13 *** RA_T_Dia Genotype 49.107 400 1.0288 0.3681 not Date 9.298 2 38.96 < 2.20E-16 *** PA_IN_N Genotype 3898.3 400 2.5296 < 2.20E-16 *** Date 24.1 2 3.1329 0.04413 * PA_NPB Genotype 698771 400 5.0133 < 2.20E-16 *** Date 9822 2 14.093 9.696E-07 *** Nb_SB_MN Genotype 759.64 400 1.1226 0.08864 not Date 439.9 2 130.01 < 2.20E-16 *** PA_GY Genotype 290136 400 1.8368 3.071E-13 *** Date 17225 2 21.81 6.037E-10 *** PA_W Genotype 481772 400 1.8352 3.364E-13 *** Date 72990 2 55.608 < 2.20E-16 *** TGW Genotype 4235.3 400 1.6271 8.026E-07 *** Date 611.4 1 93.957 < 2.20E-16 *** NG_PA Genotype 486873026 400 1.6671 2.446E-07 *** Date 1073313 1 1.47 0.2261 not NB: Sum Sq = Sum of square; Df = degree of freedom; Pr(>F) = probability greater than F University of Ghana http://ugspace.ug.edu.gh 43 In general, traits at the top of the panicle were not heritable; therefore, no need to phenotype them. The number of secondary branches at the basal part of the panicle appeared to be just informative. The traits secondary branch length, distance to the first secondary branch at second and median nodes, primary branch density and number of grains per panicle were not affected by the sowing date. Variables such as secondary branch zone at penultimate node and panicle harvest index did not appear to be under genetic control. 3.3.5. Broad sense heritability estimates The Broad-sense heritability estimates ranged from 0.03 to 90.7 % (Figure 3.2). High heritability estimates were observed for the characters plant height (90.7 %), number of primary branches per panicle (80.1 %), primary branch density (80.5 %), peduncle length (79.9 %), panicle length (70.9%) and number of nodes per panicle (60.5%). Figure 3.2: Broad sense heritability estimates for panicle architecture traits and grain yield University of Ghana http://ugspace.ug.edu.gh 44 Moderate heritability estimates were exhibited for the traits: internode length (56.8%), rachis base diameter (53.1%), grain yield per panicle (45.6 %), panicle weight (45.5%), average number of grains per primary branch (42.8%), secondary branch length (41.3%), number of grains per panicle (40.0%) and 1000 grain weight (38.5%), whereas moderate to low heritability estimates were recorded for secondary branching variables such as number and position at median and penultimate nodes (22.7%). Very low heritability estimates were noted for secondary branch density at the penultimate node, distance to the first secondary branch at the penultimate node, panicle harvest index and rachis top diameter. A gradient of genetic variability was observed for the length of the different types of axes such as stem, peduncle, panicle, rachis, primary and secondary branches at different locations in the panicle like the rachis, internode, second, median and penultimate nodes (Fig 3.3). The same tendency was observed for the number of branches where heritability estimates decreased progressively for the following traits: number of primary branches per panicle, maximum number of primary branches per node, number of secondary branches on the longest primary branches and number of primary and secondary branches at second, median and penultimate nodes. Therefore, moderate heritability estimated at median node and low heritability estimated at penultimate node. University of Ghana http://ugspace.ug.edu.gh 45 Figure 3.3: Gradient in broad sense heritability for length of panicle traits 3.3.6. Principal Component Analysis Three dimensional (3D) observations of selected sorghum panicle architecture variables and grain yield per panicle indicated the contribution of each dimension. According to Sneath and Sokal, (1973) eigenvalues > 1 were considered significant, whereas Hair Jr et al. (1998) suggested that component loadings greater than 0.30 were considered to be meaningful. Therefore three axes with eigenvalues greater than 1 were considered in this study. The first dimension explained 29.83% (Figure 3.4) 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 S te m l en g th P ed u n cl e le n g th P an ic le l en g th R ac h is l en g th In te rn o d e le n g th L en g th o f th e lo n g es t P ri m ar y B ra n ch P ri m ar y b ra n ch l en g th a t se co n d n o d e P ri m ar y b ra n ch l en g th a t m ed ia n n o d e P ri m ar y b ra n ch l en g th a t p en u lt im at e n o d e S ec o n d ar y b ra n ch l en g th a t se co n d n o d e S ec o n d ar y b ra n ch l en g th a t m ed ia n n o d e S ec o n d ar y b ra n ch l en g th a t p en u lt im at e n o d e B ro ad -s e n s h e ri ta b ili ty University of Ghana http://ugspace.ug.edu.gh 46 Figure 3.4: Contribution of sorghum panicle architecture traits to the dimension 1 and 2 The second one explained 17.73% while the third explained 10.08% (Figure 3.4). The following variables made a substantial contribution to the first dimension: number of grains on primary branches at median nodes and primary branch length at median node and number of secondary branches at second node. The variables with high contributions to the second dimension were the University of Ghana http://ugspace.ug.edu.gh 47 number of primary branches and internodes per panicle, rachis base diameter and panicle grain yield. Figure 3.5: Contribution of sorghum panicle architecture traits to the dimension 2 and 3 The third dimension was influenced by the high contribution of primary branch length and number of grains at second node. University of Ghana http://ugspace.ug.edu.gh 48 The PCA result indicated that the number of primary branches per panicle is determined by the number of internodes on the panicle, rachis basal diameter affect the grain yield per panicle. The number of grains per panicle is more relevant in grain yield per panicle than the thousand grain weight. The number of grains on a particular primary branch is mainly determined by the number of secondary branches rather than the length of the primary branch. 3.3.7. Correlation among traits Highly significant positive correlations ranging from 0.31 and 0.70 were detected between panicle length and the length of primary branches, number of internodes per panicle, number of secondary branches and internodes length with high coefficient of correlation (Table 3.5). The number of internodes per panicle was highly significant and positively correlated to the number of primary branches (r = 0.55) and panicle grain yield (r = 0.31). The number of primary branches per panicle was highly significant and positively correlated with grain yield per panicle with (r = 0.49) and highly negatively correlated to the length of the internodes (- 0.39). Rachis base diameter was highly significant and positively correlated to the number of primary branches per panicle and the grain yield per panicle with the coefficient of correlation (r = 0.56 and 0.53, respectively). Three sorghum panicle architecture traits were highly significantly positively correlated to grain yield per panicle with the coefficient of correlation ranged from 0.30 to 0.53. Rachis base diameter presented the highest correlation followed by the number of primary branches per panicle and the number of internodes per panicle (Table 3.5). University of Ghana http://ugspace.ug.edu.gh 49 Table 3.5: Correlation among sorghum panicle architecture traits and grain yield per panicle PA_L PB_L_max PB_L_SN IN_L RA_B_Dia PA_IN_N PA_NPB Nb_SB_LPB PB_L_max 0.706 0.000 PB_L_SN 0.594 0.804 0.000 0.000 IN_L 0.489 0.465 0.444 0.000 0.000 0.000 RA_B_Dia 0.079 0.031 -0.080 -0.136 0.116 0.535 0.110 0.006 PA_IN_N 0.350 0.159 0.097 -0.565 0.241 0.000 0.001 0.053 0.000 0.000 PA_NPB 0.056 -0.148 -0.179 -0.391 0.566 0.551 0.260 0.003 0.000 0.000 0.000 0.000 Nb_SB_LPB 0.315 0.372 0.304 0.015 0.066 0.272 0.044 0.000 0.000 0.000 0.764 0.188 0.000 0.377 PA_GY 0.149 0.184 0.056 -0.120 0.533 0.307 0.440 0.224 0.003 0.000 0.265 0.016 0.000 0.000 0.000 0.000 NB: first value in each cell indicates the coefficient of correlation and the second value is the probability. High coefficient of correlation are indicated in bold. 3.3.8. Transgressive segregation Transgressive segregations were observed for most of the panicle architecture traits, grain yield and its related traits. The Figures 3.6 to 3.8 show the distribution for four traits in the three sowing dates. University of Ghana http://ugspace.ug.edu.gh 50 2101801501209060 80 70 60 50 40 30 20 10 0 Number of primary branches F r e q u e n c y Mean 96.88 StDev 26.97 N 406 1 0 11 0 4 22 13 16 21 23 41 54 72 76 53 23 3 Normal Both parents Histogram of Number of primary branches per panicle_ SB1 1351201059075604530 50 40 30 20 10 0 Grain yield (g) F r e q u e n c y Mean 69.54 StDev 19.69 N 406 1 0 1 4 3 5 1111 14 24 29 31 44 27 39 42 39 28 17 22 12 11 Normal Both parents Histogr m of Grain yield er panicle_SB1 907560453015 80 70 6 50 40 3 20 10 0 Number of grains F r e q u e n c y Mean 38.16 StDev 12.63 N 406 1 00 1 5 10 18 27 28 35 65 73 63 47 22 10 1 Normal Both parents Histogram of average number of grain per primary branch_SB1 Figure 3.6: Distribution of four panicle traits showing transgressive segregation in the first sowing date (SB1), 2011, Sotuba, Mali 9.759.008.257.506.756.005.254.50 50 40 30 20 10 0 Rachis base diameter F r e q u e n c y Mean 6.748 StDev 1.062 N 406 111 4 5 8 9 17 16 14 31 32 33 43 46 36 23 24 22 18 6 9 2 5 Normal Histogram of R chis base diameter_SB1 Lata 3 Tiandougou University of Ghana http://ugspace.ug.edu.gh 51 2402101801501209060 70 60 50 40 30 20 10 0 Number of primary branches F r e q u e n c y Mean 103.6 StDev 29.66 N 406 1 0 1 2 0 2 1 7 10 15 22 20 36 52 54 66 59 37 16 5 Normal Lata 3 Tiandougou Histogram of Number of primary branches per panicle_SB2 1201059075604530 50 40 30 20 10 0 Grain yield (g) F r e q u e n c y Mean 64.06 StDev 18.78 N 406 2 1 2 5 4 6 12 16 14 30 26 38 47 49 38 40 31 1717 8 1 2 Normal Both Parents Histogr m of grain yield er panicle_SB2 10.59.07.56.04.53.01.50.0 80 70 60 50 40 3 20 10 0 Rachis base diameter F r e q u e n c y Mean 7.198 StDev 1.138 N 406 1 0 4 6 23 28 62 69 76 66 40 22 4 0 1 2 1 00000 1 Normal Histogram of Rachis base diameter_SB2 Tiandougou Lata 3 907560453015 80 70 60 50 40 3 20 10 0 Number of grains F r e q u e n c y Mean 35.56 StDev 12.43 N 406 1 000 1 5 3 8 26 19 33 64 69 78 39 40 15 4 1 Normal Histogram of Average number of grain per primary branch_SB2 Tiandougou Lata 3 Figure 3.7: Distribution of four panicle traits showing transgressive segregation in the second sowing date (SB2), 2011, Sotuba, Mali University of Ghana http://ugspace.ug.edu.gh 52 24020016012080400 70 60 50 40 30 20 10 0 Number of primary branches F r e q u e n c y Mean 100.0 StDev 31.27 N 394 11 000 1 0 2 33 13 11 18 19 27 41 50 66 61 42 26 3 5 000 1 Normal Histogram of Number of primary branches per panicle_SB1' Tiandougou Lata 3 1801501 0960300 60 50 40 30 20 10 0 Grain yield per panicle (g) F r e q u e n c y Mean 73.39 StDev 28.19 N 394 1 00 4 1 8 14 19 37 46 54 57 52 38 29 14 16 3 0 1 Normal Histogr m of grain yield per panicle_SB1' Tiandougou Lata 3 10.59.07.56.04.53.0 70 60 50 40 30 20 10 0 Rachis base diameter F r e q u e n c y Mean 7.233 StDev 1.343 N 393 1 33 18 22 37 57 51 60 66 30 21 9 5 4 2 3 0 1 Normal Histogram of Rachis base diameter_SB1' Tiandougou Lata 3 12010590756045315 70 60 5 40 30 20 10 0 Number of grains F r e q u e n c y Mean 40.31 StDev 15.70 N 394 11 00 111 2 6 10 5 12 0 25 5353 67 52 30 25 13 3 Normal Histogram of Average number of grain per primary branch_SB1' Tiandougou t Figure 3.8: Distribution of four panicle traits showing transgressive segregation in the first sowing date (SB1’), 2012, Sotuba, Mali University of Ghana http://ugspace.ug.edu.gh 53 3.3.9. Estimation of grain yield per panicle using panicle architecture traits A multivariate regression, correlating panicle architecture traits with the grain yield per panicle, indicated that the number of internodes per panicle and rachis base diameter explained the grain yield per panicle with a determination coefficient (R²) of 31.8%. The regression equation was: PA_GY = - 0.000 + 7.76 RA_B_Dia + 1.23 PA_IN_N, where: PA_GY was grain yield per panicle; RA_B_Dia was rachis base diameter and PA_IN_N was number of internodes per panicle. Therefore, it could be possible to have an estimate of the grain yield per panicle by measuring those easily measurable traits. 3.3.10. Branch length, grain number and rachis diameter at different positions on panicle Primary and secondary branches, number of grains on primary branches and rachis diameter decrease from the base of the panicle toward the top of the panicle (Figure 3.7) University of Ghana http://ugspace.ug.edu.gh 54 PB_L_PNPB_L_MNPB_L_SN 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 D a t a 95% CI for the Mean Primary branches length at different positions in the panicle NG_PNNG_MNNG_SN 1.0 0.5 0.0 -0.5 -1.0 D a t a 95% CI for the Mean Grain number at different posi ion in the panicle Nb_SB_PNNb_SB_MNNb_SB_SN 2.0000E-09 1.0000E-09 0000E+00 -1.000E-09 -2.000E-09 D a t a 95% CI for the Mean Secondary branches on primary branch at different positions RA_T_DiaRA_DiaRA_B_Dia 0.04 0.03 0.02 .01 0.00 -0.01 -0.02 -0.03 -0.04 -0.05 D a t a 95% CI for the Mean Rachis diameter t different positions in the panicle Figure 3.9: Primary and secondary branches length; grain number and rachis diameter at second, median and penultimate nodes University of Ghana http://ugspace.ug.edu.gh 55 3.4. Discussion Most of cultivated sorghum are photoperiod sensitive (Chantereau et al., 2004). The F4 families performed better in the first sowing date than in the second one. Grain yield was reduced in the second sowing date in 2011 because of photoperiod that shortened the vegetative cycle and affected the potential yield. Consequently, highly significant differences were observed among the genotypes owing to the delay of the sowing date that resulted in high variability in the population. The reduction in grain yield resulted from the decrease in some panicle architecture traits such as the length of the primary branches at the median node, number of primary branches and nodes per panicle. The weight of the panicle which is one of the principal grain yield components was also affected negatively by the delay in the sowing date. A shorter vegetative cycle affects panicle elongation and branching. The ANOVA results confirmed the effect of the sowing date on most of sorghum panicle architecture traits. Broad sense heritability for grain yield was moderate (46%) in this study. This is similar to the observation made by Addissu (2011) who reported 47% broad sense heritability for grain yield per plant. Brown et al. (2006), Murray et al. (2008) and Srinivas et al. (2009) reported moderate broad sense heritability estimates of 56%, 57% and 65%, respectively, for grain yield per panicle. However, Ali et al. (2012) reported high broad sense heritability estimates of (98%) for grain yield per plant. Warkad et al. (2008) also reported high broad sense heritability for grain yield per plant (86.0%), Godbharle et al. (2010) using elite ‘B’ and ‘R’ lines and Chavan et al. (2010) reported broad sense heritability estimates of up to (95 %) for grain yield per plant in F5 families of sorghum. Therefore, this discrepancy in heritability estimates for grain yield per panicle may be explained by the interaction between the genotypes and their environments (GxE). High heritability was observed for panicle length. This is in agreement with that obtained by Ali et al. (2012) who found, after the first cycle of selection, high broad sense heritability of 91% for University of Ghana http://ugspace.ug.edu.gh 56 panicle length. Chavan et al. (2010) reported 87% broad sense heritability in F5 families in sorghum for panicle length. Rami et al. (1998) also reported a high broad sense heritability (83%) for panicle length. Moderate heritability was noted for panicle weight. Srinivas et al. (2009) also reported moderate (55%) broad sense heritability for panicle weight. Warkad et al. (2008) reported moderate broad sense heritability for number of nodes per panicle (34%). That is in agreement with the result obtained in this study. High heritability observed for the number of primary branches per panicle in this study is in agreement with the finding of Godbharle et al. (2010), Chavan et al. (2010) and Brown et al. (2006) with broad sense heritability estimates of 96%, 93% and 94%, respectively for number of primary branches per panicle. However, Srinivas et al. (2009) reported relatively high (74%) broad sense heritability for the number of primary branches per panicle whereas Warkad et al. (2008) reported a moderate broad sense heritability estimate of 55% for the number of primary branches per panicle. These differences may be explained by the genetic background of the genotypes and the sowing dates. Number of branches was reduced in the late sowing date. Moderate heritability estimates were recorded for the number of secondary branches in this study while Brown et al. (2006) found relatively high broad sense heritability (72%) for number of secondary branches. The number of grains per primary branch was influenced by other factors such as soil fertility. Godbharle et al. (2010) reported high broad sense heritability (95%) for number of grains per primary branch while moderate heritability was detected in this study. Brown et al. (2006) reported high broad sense heritability for rachis diameter (90%). However, in this study, rachis diameter showed moderate broad sense heritability. In this study, principal component analysis revealed some important sorghum panicle architecture traits with positive contribution. The traits included panicle length, number of grains per panicle, the lengths of primary and secondary branches, number of secondary branches and grains per University of Ghana http://ugspace.ug.edu.gh 57 primary branch at median and penultimate node. Bucheyeki et al. (2009) also reported high loading of grain number per panicle that is consistent with the findings of this study. Principal component analysis displayed relationship between grain yield and some of the panicle architecture traits such as rachis base diameter, number of secondary branches on the basal primary branches and the number of primary branches per panicle. Emphasis should be placed on these traits in relation to their heritability and correlation for better understanding of the basic architectural principles such as branching pattern, elongation and relation to grain yield. This will provide the basis for developing breeding strategies to improve grain yield potential in sorghum. Correlations among traits determine the genetic relatedness that provide the basis of breeding strategies for direct or indirect selection. The phenotype of a plant is the result of interaction of a large number of factors. Therefore, final yield is the total sum of several component characters. The phenotypic and genotypic correlation coefficients among different characters revealed that genotypic correlation coefficients were higher than phenotypic correlations. In general, the signs and magnitudes of correlations at both phenotypic and genotypic level are similar. Positive correlations were observed between grain yield and its components. Brown et al. (2006) found similar significant positive correlations between panicle grain yield and number of primary branches per panicle, plant height, 100 seed weight and seeds per panicle. The coefficient of correlation ranged from 0.3 to 0.82. Similar high positive correlation (0.82) was also found between panicle grain yield and grain number per panicle by Brown et al. (2006). Positive association of grain yield with panicle length, plant height, number of primary branches per panicle, and number of grains per panicle have been reported by Mahajan et al. (2011). Panicle length and number of grains per panicle had greater importance in increasing grain yield per panicle (El-din et al., 2012). For efficient selection in sorghum including panicle architecture traits, the following consideration should be kept in mind for instance, panicle length is proportional to the length of the rachis and University of Ghana http://ugspace.ug.edu.gh 58 affects rachis slenderness and volume. Longer panicles have long internodes, long primary and secondary branches and poorly correlated to grain yield per panicle. However, increasing the number of primary branches per panicle increases the number of grains and consequently grain yield per panicle. Increasing the number of internodes per panicle may increase the number of primary and secondary branches at the base and middle of the panicle and, subsequently, the panicle grain yield may be improved. The bigger the rachis base diameter, the more the primary and secondary branches the more the grains and subsequently high grain yield per panicle would result. Phenotyping specific axes and in specific parts on the panicle is important to fully account for genetic variability in sorghum panicle architecture; for instance, length data collected on primary branches at second node were much more informative than on secondary branches at the penultimate node. Number of primary branches per panicle was highly heritable compared to the dissected pattern of branching. However, the number of branches at the second node affects the number of grains more than the branch at the top of the panicle. This suggests that the length and the number of primary branches may play an important role in capturing panicle architecture and act as a good predictor to select for grain yield improvement. Therefore, a simple strategy of sorghum panicle architecture phenotyping should rely on the characterization of panicle elongation by measuring the length of the panicle, longest primary branch and basal primary branches. Branching intensity is also an important trait to assess architectural variability in sorghum panicle by measuring the number of primary branches per panicle and the average number of primary branches per node. Rachis basal diameter influences the branching that plays a major role in grain setting and consequently the grain yield. University of Ghana http://ugspace.ug.edu.gh 59 3.5. Conclusion Delay in sowing date negatively affected panicle architecture trait and grain yield. Most of the sorghum panicle architecture traits were heritable in the F4 families derived from Tiandougou/Lata- 3. Panicle length, primary branches length at second and median nodes, secondary branch length, number of primary branches per panicle, number of secondary branches on primary branch at second nodes and position of the longest primary branch had positive contribution to sorghum panicle architecture variability. Other architectural traits affecting grain yield such as, number of primary branches per panicle, number of secondary branches on primary branches at second node and rachis base diameter have also a role to play in the variability of panicle architecture variability. Transgressive segregation was identified for most of the panicle architecture traits. Across the three sowing dates, the parent Lata-3 panicles consistently displayed more nodes per panicle and more secondary branches per primary branch than Tiandougou. Panicle, peduncle and primary branches at the base and middle of the panicle were longer in Lata-3 while Tiandougou displayed more primary branches and more grains per primary branch at the top of the panicle. Nevertheless, panicle grain yield, number of grains per panicle, grain weight, rachis base diameter and number of primary branches per panicle were similar in the two parents. Additive, non-epistasis effects and less environmental effect may explain the stability of the genotypes across the different sowing dates. Branching and grain number decrease progressively from the base to the top of the panicle. Causal relationships were detected between architecture traits and grain yield per panicle. Therefore, indirect selection responses between number of primary branches per panicle or number of internodes per panicle and rachis base diameter on grain yield per panicle were promising and need to be further investigated. University of Ghana http://ugspace.ug.edu.gh 60 CHAPTER FOUR 4.0. QUANTITATIVE TRAIT LOCUS (QTL) ANALYSIS OF PANICLE ARCHITECTURE IN THE TIANDOUGOU/ LATA-3 BREEDING POPULATION 4.1. Introduction Molecular markers are useful tools for dissecting the variation of complex traits, such as grain yield and its components into simply inherited factors. A quantitative trait locus (QTL) represents the statistical association observed in a segregating population between the quantitative variation of a given trait and the genotype (homozygous for one of the two parental alleles or heterozygous) of progenies at a given locus as revealed by molecular markers. In addition to the identification of the most probable location in the genome of such QTL, molecular markers allow estimation of their genetic effects and interactions among them. The incorporation through various ways of such information in a breeding program provides a strong basis for improving and optimizing plant breeding process. The identification of QTLs affecting sorghum panicle architecture will permit an understanding of the genetic factors, their interactions, and elucidate their relationship to grain yield. It will provide useful information for the definition of target genotypes to breed for in the breeding population. Indeed, the Marker Assisted Recurrent Selection (MARS) approach constitutes a framework for the valorization of QTL results together with other traits such as grain quality and plant adaptation. QTL studies on the panicle architecture in sorghum were mainly done using Recombinant Inbreds Lines (RILs) population. The first reports of genomic regions associated with grain yield and grain yield components were reported by Tuinstra et al. (1996), while genomic regions controlling University of Ghana http://ugspace.ug.edu.gh 61 sorghum panicle architecture traits were reported by Pereira et al. (1995); Rami et al. (1998); Hart et al. (2001); Brown et al. (2006) and Srinivas et al. (2009). The objectives of this research were (1) to identify QTL responsible for the phenotypic variation of sorghum panicle architecture traits in F3 segregating population, (2) to determine the consistency of the detected QTL across different sowing dates, and (3) to determine genomic regions with pleiotropic effects on two (or more) different traits. 4.2. Material and methods 4.2.1. Plant material A mapping population of 401 F4 families obtained from a cross between Tiandougou and Lata-3 was used for QTL detection. The population is described in the chapter III. 4.2.2. Leaf sampling and DNA extraction A sample of leaf tissue (400g) was collected for each of the 401 families during the F4 to F5 multiplication in 2010. For each family, a bulk of 10 F4 plants were sampled to represent the parental F3 line. Leaf samples of the parental lines, Tiandougou and Lata-3, were collected during the initial crossing made to produce the population. Leaf samples were collected in individual zipper storage plastic bags together with silica gel and allowed to dry during shipment to Montpellier, France. DNA extraction was performed at CIRAD, Montpellier following the MATAB isolation protocol as described by Risterucci et al. (2000). University of Ghana http://ugspace.ug.edu.gh 62 4.2.3. Genotyping DNA of the F3 population and the 2 parents were sent to LGC genomics (formerly KBioscience) for SNP genotyping using KASPTM technology. Additional genotyping was performed at CIRAD Montpellier on another five (5) SNP markers using the same KASPTM technology on the same DNA samples (Appendix 4.1). 4.2.4. Phenotyping A great number of panicle architecture traits related to geometry, topology and grain yield and its components were collected. However, focus was on selected traits based on the result of Pearson correlation and principal component analysis. These traits were grain yield per panicle, number of primary branches per panicle, primary branch length, number of grains per panicle, number of internodes per panicle, number of grains per primary branch, panicle length, panicle weigh and rachis base diameter (refer to Chapter III). 4.2.5. Construction of Genetic maps Genetic markers were chosen for their position on the genome to get a good coverage of the genome and a good distance between markers for QTL detection. A total of 228 polymorphic markers (225 markers from LGC Genomics and 3 markers from the CIRAD lab) were used for the construction of genetic map. The genotypes of the individuals were coded as A for the allele of the parent Lata- 3, B for the allele of the parent Tiandougou, and H for the heterozygous genotype. Linkage groups were determined by results of pair-wise comparisons at a minimum likelihood of odds (LOD) value of 3. The best order was determined by comparing the goodness of fit of the resulting map for each tested order using a threshold of 0.5 and 1.0 for the linkage groups and the loci, respectively. The functions order sequences, check inversion, ripple sequences and draw maps were used to elaborate University of Ghana http://ugspace.ug.edu.gh 63 the preliminary genetic map. Kosambi mapping function was used to translate recombination frequencies into map distances. The genetic map obtained with Mapdisto software was drawn using spiderMap software. The 10 linkage groups were assigned to sorghum chromosomes and named SBI-01 to SBI-10, according to the a priori known physical position of the SNP markers on the genome sequence. The constructed genetic map was compared to the sorghum consensus map published by Mace et al. (2009). 4.2.6. QTL Detection QTL detection was performed on trait mean values in each individual sowing date and BLUP values that integrate all sowing dates. QTL analyses were performed using Simple interval mapping (SIM) with the regression method described (Halley and Knott, 1992) using the software R/qtl (Broman and Sen, 2009). A thousand-permutation test was applied to each trait variable to determine the LOD threshold (P = 0.05) for considering significant levels of identified QTLs (Churchill and Doerge, 1994). SIM detection was followed by the “refine qtl” procedures that refine the position of the detected QTLs using an iterative scan of QTL positions using a multiple QTL model. Support interval of LOD was calculated for each QTL to obtain a 95% confidence interval. Adjacent QTLs on the same chromosome were considered different when the support intervals were not overlapping. Attention was also accorded to declare two peaks in the same chromosome for the same trait as two QTLs only if the distance between the QTLs was greater than 20 cM otherwise the higher peak was considered for better estimation of the QTL position (Parth et al., 2008; Ungerer et al., 2002). The percentage of phenotypic variance, the additive and dominance effects attributable to an individual QTL were estimated using an additive multi-QTL model involving all QTLs detected for a given trait, using the function “fitqtl” of the R/qtl package. An average level of dominance for QTL (d/a) was calculated as the ratio of dominance effect (d) over University of Ghana http://ugspace.ug.edu.gh 64 additive effect (a). The QTL action was determined according to guidelines presented by (Stuber et al. 1987): additive QTL action (A) = 0 to 0.20; partial dominance (PD) = 0.21 to 0.80; dominance (D) = 0.81 to 1.20; and over dominance (OD) = >1.20. According to R/qtl documentation the “additive effect is derived from the coding scheme -1/0/+1 for genotypes AA/AB/BB, and so is half the difference between the phenotype averages for the two homozygotes”. For a given QTL, if the additive effect is positive, the increasing allele came from the Tiandougou parent; if negative the increasing allele came from Lata-3 parent. QTLs were designated with italicized symbols composed of a Q followed by trait name, a hyphen and the chromosome on which the QTL is detected. Serial numbers were used to design more than one QTL controlling the same trait detected on the same chromosome. For instance, QPA_L-SBI- 03-2 refers to the second panicle length QTL detected on SBI-03. 4.2.7. Digenic epistasis detection The digenic epistasis interaction was analyzed within F4 families using bi-parental populations (BIP) mode in QTLIciMapping software version 3.2. The ICIM-EPI (Inclusive Composite Interval Mapping for epistasis mapping method was used with a step size of 5 cM and a probability in stepwise regression of 0.0001. Deletion was selected as the means for dealing with missing phenotypic data and the LOD threshold was fixed for all traits at 5. 4.2.8. QTLs overlapping confidence interval across studies A sorghum consensus map (Mace et al., 2009), with multiple anchor points to the whole genome sequence was used to identify QTLs for the same trait overlapping confidence intervals across four (4) studies on sorghum panicle architecture traits previously reported. University of Ghana http://ugspace.ug.edu.gh 65 4.3. Results 4.3.1. Summary of QTLs detected in Tiandougou/Lata-3 F4 families population Simple Interval Mapping followed by refine QTLs revealed significant QTLs for sorghum panicle architecture and grain yield (Table 4.1). Table 4.1: Summary of significant QTLs detected for sorghum panicle architecture traits and grain yield in Tiandougou / Lata-3 F4 families using Simple Interval Mapping followed by refine QTLs Chrs Nb_M T_L (cM) Av_DM (cM) N_Q_ PVE>7 T_Q Add PD Dom Over Dom Tiand Lata 1 29 178.3 6.1 5 18 5 11 2 0 10 8 2 25 169.1 6.8 15 33 12 14 6 1 15 18 3 26 163.7 6.3 16 47 20 21 5 1 23 24 4 20 135.6 6.7 1 13 2 11 0 0 2 11 5 22 119.7 5.4 0 2 0 0 1 1 1 1 6 20 110.1 5.5 15 27 13 11 2 1 10 17 7 21 123.1 5.9 2 17 6 9 1 1 11 6 8 16 115.4 7.2 0 1 0 0 1 0 0 1 9 23 118.5 5.1 1 4 1 2 0 1 1 3 10 26 133.8 5.1 0 8 3 5 0 0 5 3 Total 228 1362.3 6.01 53 164 62 80 17 5 75 89 NB: Chrs = chromosomes; Nb_M = Number of Markers; T_L (cM) = Total length (centiMorgan); Av_DM (cM) = Average Distance between Markers (centiMorgan); N_Q_PVE>7 = Number of QTLs with Percentage of variation explained superior to 7%; T_Q = Total QTLs; Add = Additive; PD = Partial Dominance; Dom = Dominance; Over Dom = Over Dominance; Tiand = Tiandougou; Lata = Lata-3. The ten chromosomes were covered with 228 SNP markers (Appendix 4.2). The average distance between markers was 6.01 cM. The total length of the constructed genetic map was 1362.3 cM that was shorter compare to the sorghum consensus genetic maps (1603.5 cM). A total of 164 QTLs were detected with phenotypic variation explained in general low. Only 53 QTLs presented the phenotypic variation explained more than 7% (Appendix 4.3). Non-additive gene actions was predominant in the control of sorghum panicle architecture. The parent Lata-3 contributed University of Ghana http://ugspace.ug.edu.gh 66 favorable alleles that increased the trait value than the parent Tiandougou. The SBI-02, SBI-03 and SBI-06 bear a great number of QTLs for panicle architecture traits and grain yield. 4.3.2. QTLs detected for the most important sorghum panicle architecture traits and grain yield in Tiandougou/Lata-3 F4 families using Simple Interval Mapping followed by refine QTLs Panicle grain yield, panicle length, rachis base diameter, number of primary branches per panicle, primary branch length, number of grains per primary branch, panicle weight and number of internodes per panicle were the most important traits related to sorghum panicle architecture and grain yield (refers ACP result in chapter III). 4.3.2.1. Panicle grain yield Two QTLs were detected for panicle grain yield on the SBI-03 and SBI-06 (Figure 4.1). The parent Lata-3 contributed the allele that increased the trait value in the QTL detected on SBI-06 with 11.86% of phenotypic variation explained. The parent Tiandougou contributed to the second QTL with 8.55% of phenotypic variation explained. Panicle grain yield was governed by non-additive gene action. University of Ghana http://ugspace.ug.edu.gh 67 Figure 4.1: QTLs detected for panicle grain yield in Tiandougou/Lata-3 F4 families population 4.3.2.2. Number of primary branches per panicle Six QTLs were identified to control number of primary branches per panicle in the Tiandougou/Lata-3 population (Figure 4.2). The parent Tiandougou contributed the alleles University of Ghana http://ugspace.ug.edu.gh 68 increasing the trait value for four QTLs, while the parent Lata-3 provided increasing allele in two QTLs. Figure 4.2: QTLs detected for number of primary branches per panicle in Tiandougou/Lata-3 F4 families population University of Ghana http://ugspace.ug.edu.gh 69 4.3.2.3. Number of internodes per panicle Three (3) QTLs for the number of internodes per panicle were identified (Figure 4.3). The percentage variation explained by the QTLs ranged from 2.95 (QPA_IN_N-SBI-02-2) to 7.92% (QPA_IN_N-SBI-02-1). Figure 4.3: QTLs detected for number of internodes per panicle in Tiandougou/Lata-3 F4 families population University of Ghana http://ugspace.ug.edu.gh 70 The parent Lata-3 contributed to beneficial alleles increasing trait values for the QTL on the SBI- 02 with 7.92 % of phenotypic variation explained. Tiandougou contributed to the two QTLs with partial dominance and additive gene action. 4.3.2.4. Rachis base diameter A total of six QTLs were detected for rachis base diameter (Figure 4.4). Figure 4.4: QTLs detected for rachis base diameter in Tiandougou/Lata-3 F4 families population University of Ghana http://ugspace.ug.edu.gh 71 All the QTLs have partial dominance gene action and the parent Tiandougou contributed for beneficial allele except the major QTL in which Lata-3 contributed for favorable allele. The percentage of variation explained by those QTLs ranged from 2.67 (QRA_B_Dia-SBI-10) to 16.60% (QRA_B_Dia-SBI-06). 4.3.2.5. Panicle length Four QTLs on SBI01, SBI02, SBI03, and SBI04 were detected for panicle length (Figure 4.5). The percentage of variation explained by those QTLs ranged from 6.74 (QPA_L-SBI-01) to 11.19% (QPA_L-SBI-03). The parent Lata-3 contributed mostly alleles that increase trait values of three QTLs. Among those three QTLs, two QTLs displayed partial dominance gene action while one QTL showed additive effect. The parent Tiandougou contributed a beneficial allele increasing trait values for only one partial dominance QTL of panicle length. One major QTL for panicle length was identified (QPA_L_SBI-03) on SBI-03 at the position 143 cM. The flanking markers were SB03137 on left and SB03164 at right. The percentage of variation explained by the QTL was 11.19%. The parent Lata-3 allele increased the trait value with partial dominance gene action. University of Ghana http://ugspace.ug.edu.gh 72 Figure 4.5: QTLs detected for panicle length in Tiandougou/Lata-3 F4 families population 4.3.2.6. Average number of grains per primary branch A total of four QTLs were detected for the average number of grains (Figure 4.6). The percentage of phenotypic variation explained by these QTLs ranged from 3.40 (QAv_NG-SBI-01) to 6.66% University of Ghana http://ugspace.ug.edu.gh 73 (QAv_NG-SBI-02). Two QTLs have partial dominance gene effect and another two QTLs have additive gene action. Each parent contributed in alleles increasing trait value for two QTLs. Figure 4.6: QTLs detected for average number of grains per primary branch in Tiandougou/Lata-3 F4 families population 4.3.3. Consistent QTLs detected in Tiandougou/Lata-3 F4 population across the three sowing dates in the combined data (BLUP) Consistent QTLs across sowing dates and in BLUP data were indicated in Table 4.1 and Table 4.2. University of Ghana http://ugspace.ug.edu.gh 74 Table 4.2: Topological consistent QTLs detected in Tiandougou/Lata-3 population Traits Sowing dates Chr Position cM LOD PVE Add Dom Dom/Add Gene action Dir Parent BLUP 6 76 31.914 20.186 -10.97 2.451 -0.13 PD Lata SB1 6 79 14.9 13.11 -11.58 1.97 -0.17 A Lata PA_NPB SB2 6 72 18.69 17.9 -14.73 4.1 -0.28 PD Lata SB1’ 6 79 16.49 15.02 -13.25 3.76 -0.28 PD Lata BLUP 6 62 12.833 12.46 -141.5 6.899 -0.05 A Lata NG_PA SB1 6 79 8.52 6.54 -256.3 187.22 -0.73 PD Lata SB1’ 6 69 8.24 8.61 -385.9 -10.15 0.03 A Lata SB1 1 15 10.22 9.93 0.87 -0.47 -0.54 PD Tiand PA_IN_N SB2 1 8 11.4 11.42 0.93 -0.22 -0.24 PD Tiand SB1’ 1 9 9.57 10.23 0.93 -0.04 -0.04 A Tiand SB2 2 116 8.03 7.23 -5.13 2.75 -0.54 PD Lata PB_BZ_SN SB1’ 2 141 9.53 10.24 -7.05 -1.38 0.20 A Lata BLUP 6 72 28.82 18.35 -0.036 -0.009 0.03 A Lata SB1 6 79 14.05 11.74 -0.04 0 0.00 A Lata PB_Dens SB2 6 76 14.17 13.6 -0.04 0.008 -0.20 A Lata SB1’ 6 79 16.33 13.72 -0.04 0.009 -0.23 PD Lata BLUP 6 67 24.551 19.352 -1.25 0.07 -0.06 A Lata SB1 6 69 17.19 14.87 -1.67 -0.23 0.14 A Lata Nb_PB_max SB2 6 70 16.31 14.39 -1.92 0.86 -0.45 PD Lata SB1’ 6 87 9.4 9.92 -1.89 0.77 -0.41 PD Lata NB: PA_NPB = Number of Primary Branches per panicle; NG_PA = Number of Grains per panicle; PA_IN_N = Number of Internodes per panicle; PB_BZ_SN = Primary branch Zone at second node; PB_Dens = Primary Branch Density; Nb_PB_max = Maximum number of primary branches per node; SBI = Chromosome; LOD = Logarithm of Odd; PVE = Percentage of Variance Explained; Add = Additive; Dom = Dominance; Additive gene action (A) if Dom/Add = 0 to 0.20; partial dominance (PD) if Dom/Add = 0.21 to 0.80; dominance (D) if Dom/Add = 0.81 to 1.20; and overdominance (OD) if Dom/Add > 1.20. Additive effects are associated with the allele from Tiandougou. Thus, a negative value means that the Tiandougou allele decreases the value of the trait. Dir_ Parent = Direction of response of the parent whose additive value of a marker allele increased the value of the trait. (Tiand = Tiandougou; Lata = Lata-3). One consistent QTL for Number of Primary Branches per panicle (QPA_NPB_SBI-06) was detected on SBI-06 in both combined analyses (BLUP) and individual analysis of the three sowing dates. The percentage of variation explained ranged from 13.11 to 20.18%. The parent Lata-3 allele increased the trait value with partial dominance gene action. University of Ghana http://ugspace.ug.edu.gh 75 One consistent QTL for primary branch density on SBI-6. The percentage of variation explained ranged from 11.74 to 18.35%. The major QTL had additive gene action with the contribution of Lata-3 increasing trait value. A consistent QTL for the maximum number of primary branches per node was identified on SBI- 06. The phenotypic variance explained by the QTL ranged from 9.92 to 19.35%. The parent Lata- 3 contributed favorable alleles that increased the trait value with additive gene action. However, one consistent QTL for number of internodes per panicle was detected in each the sowing dates on SBI-01. This QTL was not detected in the combined analysis. Also one consistent QTL was detected on SBI-06 in the combined and individual sowing date analyses excluding the second sowing date (SB2) in 2011. One consistent QTLs for panicle grain yield was detected in each of the sowing dates and in the combined data on SBI-06 with the percentage of variation explained ranging from 4.7 (SB1) to 11.86% (BLUP). The parent Lata-3 contributed to beneficial allele that increased the trait value. The date from sowing to flag leaf appearance that gave indication on flowering time displayed one major consistent QTL in the first and second sowing dates in 2011 on SBI-03. The QTL explained 45.64% and 53.88% of phenotypic variation respectively for SB1 and SB2. The QTL had partial dominance gene action and the parent Tiandougou allele increased the flowering time in the two sowing dates. One consistent QTL was identified for rachis base diameter on SBI-06. The percentage of variation explained by the QTL ranged from 6.17% (SB1’) to 16.60% (BLUP). The parent Lata-3 contributed alleles increasing the trait value with both additive and non-additive gene actions. University of Ghana http://ugspace.ug.edu.gh 76 Table 4.3: Geometrical and grain yield consistent QTLs detected in Tiandougou/Lata-3 population Traits Sowing dates Chr Position LOD PVE Add Dom Dom/Add Gene action Dir Parent BLUP 6 66 12.08 11.863 -3.1024 0.9234 -0.30 PD Lata SB1 6 82.8 4.77 4.7 -4.53 4.96 -1.09 OD Lata PA_GY SB2 6 71 6.7 6.86 -5.84 0.46 -0.08 A Lata SB1’ 6 54 6.26 5.8 -7.96 4.04 -0.51 PD Lata SB1 3 74 53.08 45.64 4.67 -2.36 -0.51 PD Tiand SFD SB2 3 71 67.39 53.88 5.02 -2.46 -0.49 PD Tiand BLUP 3 71,2 28.305 20.436 -25.389 -2.722 0.11 A Lata SB1 3 73 14.6 13.97 -27.23 -5.82 0.21 PD Lata PE_L SB2 3 73 19.67 17.59 -35.93 -3.23 0.09 A Lata SB1’ 3 73 9.8 9.04 -25.87 -9.01 0.35 PD Lata BLUP 7 94 60.265 41.128 -238.01 67.857 -0.29 PD Lata SB1 7 95 51.79 42.57 -316.76 106.23 -0.34 PD Lata PH SB2 7 96 30.97 32.17 -219.56 33.92 -0.15 A Lata SB1’ 7 96 29.69 29.57 -219.02 48.64 -0.22 PD Lata BLUP 6 67 20.363 16.602 -0.2342 0.01118 -0.05 A Lata SB1 6 71 8.02 8.5 -0.34 0.21 -0.62 PD Lata RA_B_Dia SB2 6 70 14.7 13.01 -0.45 0.12 -0.27 PD Lata SB1’ 6 69 7.12 6.17 -0.42 0.07 -0.17 A Lata BLUP 2 125 12.861 8.735 9.34 -1.90 -0.20 A Tiand RA_coni SB1 2 128 6.51 23.34 0.001 0 0.00 A Tiand SB1’ 2 140 8.1 7.53 0.001 0 0.00 A Tiand BLUP 2 138 16.918 11.701 -3.39 -6.13 0.18 A Lata RA_slend SB1 2 130 8.03 7.59 -5.3 -0.9 0.17 A Lata SB1’ 2 142 10.45 9.65 -6.56 -0.89 0.14 A Lata BLUP 3 134 21.016 14.895 -3.67 -7.43 0.00 A Lata RA_slend SB1 3 138 9.48 9.13 -5.68 1 -0.18 A Lata SB2 3 135 6.01 5.87 -4.45 -1.06 0.24 PD Lata NB: PA_GY = Grain Yield per panicle; SFD = Flowering time; PE_L = Peduncle Length; PH = Plant Height; RA_B_Dia = Rachis base Diameter; RA_coni = Rachis conicity; RA_Slend = Rachis Slenderness; SBI = Chromosome; LOD = Logarithm of Odd; PVE = Percentage of Variance Explained; Add = Additive; Dom = Dominance; Additive gene action (A) if Dom/Add = 0 to 0.20; partial dominance (PD) if Dom/Add = 0.21 to 0.80; dominance (D) if Dom/Add = 0.81 to 1.20; and overdominance (OD) if Dom/Add > 1.20. Additive effects are associated with the allele from Tiandougou. Thus, a negative value means that the Tiandougou allele decreases the value of the trait. Dir_ Parent = Direction of response of the parent whose additive value of a marker allele increased the value of the trait. (Tiand = Tiandougou; Lata = Lata-3). The traits such as peduncle length, plant height, rachis conicity and rachis slenderness also displayed consistent QTLs across the sowing dates. University of Ghana http://ugspace.ug.edu.gh 77 4.3.4. Consistent major QTLs cluster One cluster of QTLs for peduncle length and QTLs for flowering time was detected on SBI-03 (Figure 4.7) Figure 4.7: Major consistent QTL cluster on SBI-03 NB: Blue color= SB1; Green color = SB2; Red color= SB1’ and white color= BLUP One major cluster composed by the QTLs for panicle grain yield, number of primary branches per panicle, maximum primary branches per node, primary branches density, number of grains per panicle and rachis base diameter (Figure 4.8). 3 0.0 SB03002 3.4 SB03006 17.2 SB03019 27.3 SB03030 32.2 SB03032 36.5 SB03037 41.6 SB03042 46.6 SB03045 49.6 SB03048 60.4 SB03060 69.6 SB03068 70.0 SB03069 70.2 SB03070 71.6 SB03075 79.9 SB03085 86.5 SB03091 90.7 SB03094 105.2 SB03102 111.1 SB03107 120.0 SB03115 132.5 SB03131 135.6 SB03137 142.7 SB03149 149.8 SB03158 156.4 SB03164 162.7 SB03169 P E _ L _ B L U P L o d = 2 8 . 3 / R 2 = 2 0 . 4 % P E _ L _ S B 1 L o d = 1 4 . 6 / R 2 = 1 4 % P E _ L _ S B 1 ' L o d = 9 . 8 / R 2 = 9 % P E _ L _ S B 2 L o d = 1 9 . 7 / R 2 = 1 7 . 6 % R A _ s l e n d _ B L U P L o d = 2 1 / R 2 = 1 4 . 9 % R A _ s l e n d _ S B 1 L o d = 9 . 5 / R 2 = 9 . 1 % R A _ s l e n d _ S B 2 L o d = 6 / R 2 = 5 . 9 % S F D _ S B 1 L o d = 5 3 . 1 / R 2 = 4 5 . 6 % S F D _ S B 2 L o d = 6 7 . 4 / R 2 = 5 3 . 9 % University of Ghana http://ugspace.ug.edu.gh 78 Figure 4.8: Major consistent QTLs cluster on SBI-06 NB: Blue color= SB1; Green color = SB2; Red color= SB1’ and white color= BLUP 6 0.0 SB06008 13.4 SB06014 18.8 SB06015 24.3 SB06027 28.7 SB06034 54.0 SB06057 59.8 SB06065 64.2 SB06071 64.9 SB06074 74.2 SB06085 74.6 SB06087 82.8 SB06098 85.2 SB06103 86.7 SB06106 94.0 SB06113 99.1 SB06122 101.7 SB06127 108.7 SB06140 110.1 SB06145 N b _ P B _ m a x _ B L U P L o d = 2 4 . 6 / R 2 = 1 9 . 4 % N b _ P B _ m a x _ S B 1 L o d = 1 7 . 2 / R 2 = 1 4 . 9 % N b _ P B _ m a x _ S B 1 ' L o d = 9 . 4 / R 2 = 9 . 9 % N b _ P B _ m a x _ S B 2 L o d = 1 6 . 3 / R 2 = 1 4 . 4 % N G _ P A _ B L U P L o d = 1 2 . 8 / R 2 = 1 2 . 5 % N G _ P A _ S B 1 L o d = 8 . 5 / R 2 = 6 . 5 % N G _ P A _ S B 1 ' L o d = 8 . 2 / R 2 = 8 . 6 % P A _ G Y _ B L U P L o d = 1 2 . 1 / R 2 = 1 1 . 9 % P A _ G Y _ S B 1 L o d = 4 . 8 / R 2 = 6 . 3 % P A _ G Y _ S B 1 ' L o d = 6 . 3 / R 2 = 5 . 8 % P A _ G Y _ S B 2 L o d = 6 . 7 / R 2 = 6 . 9 % P A _ N P B _ B L U P L o d = 3 1 . 9 / R 2 = 2 0 . 2 % P A _ N P B _ S B 1 L o d = 1 4 . 9 / R 2 = 1 3 . 1 % P A _ N P B _ S B 1 ' L o d = 1 6 . 5 / R 2 = 1 5 % P A _ N P B _ S B 2 L o d = 1 8 . 7 / R 2 = 1 7 . 9 % P B _ D e n s _ B L U P L o d = 2 8 . 8 / R 2 = 1 8 . 4 % P B _ D e n s _ S B 1 L o d = 1 4 / R 2 = 1 1 . 7 % P B _ D e n s _ S B 1 ' L o d = 1 6 . 3 / R 2 = 1 3 . 7 % P B _ D e n s _ S B 2 L o d = 1 4 . 2 / R 2 = 1 3 . 6 % R A _ B _ D i a _ B L U P L o d = 2 0 . 4 / R 2 = 1 6 . 6 % R A _ B _ D i a _ S B 1 L o d = 8 / R 2 = 8 . 5 % R A _ B _ D i a _ S B 1 ' L o d = 7 . 1 / R 2 = 6 . 2 % R A _ B _ D i a _ S B 2 L o d = 1 4 . 7 / R 2 = 1 3 % University of Ghana http://ugspace.ug.edu.gh 79 4.3.5. Pleiotropic effects Many QTLs influencing multiple traits were observed at different positions on SBI-01, SBI-02, SBI-03, SBI-04 and SBI-06 (Table 4.4-6). Some of the most important traits affected by the pleiotropic QTLs were grain yield per panicle, rachis mean diameter, average primary branch length and panicle weight that were all affected by the same pleiotropic QTL on SBI-06 at the position 70 cM. Pleiotropic QTLs were detected which affect rachis base diameter, number of primary branches per panicle, number of grains per panicle and maximum number of primary branches per node. Rachis length, the longest secondary branch, average primary branch length and number of internodes per panicle were affected by pleiotropic QTLs. Table 4.4: Pleiotropic QTLs detected on SBI-01 and SBI-02 QTLs Position cM Traits influenced Q P H -S B I- 0 2 Q R A _ sl en d -S B I- 0 2 Q P B _ L _ P N -S B I- 0 2 Q A v_ L P B -S B I- 0 2 Q S B _ L _ m a x- S B I- 0 2 Q P B _ B Z _ S N -S B I- 0 2 Q P A _ IN _ N -S B I- 0 1 Q P B _ D en s- S B I- 0 1 Q P B _ D en s- S B I- 0 2 Q P o s_ N P B _ m a x- S B I- 0 1 Q A v_ N G -S B I- 0 2 Q B en d in g -S B I- 0 2 QIN_L-SBI-1 7 x QRA_L-SBI-1 9 x QAv_NG-SBI-1 12.7 x QPA_IN_N-SBI-02 118 x QSB_L-SBI-02 127 x QRA_L-SBI-02 128 x x QPB_L_SN-SBI-02 140 x QPA_L-SBI-02 141 x x Qav_NPB-SBI-02 146 X QSC-SBI-02 147 x On SBI-01, pleiotropic effect was noted between the QTL for internode length and the position of the maximum number of primary branches per node. The traits rachis length and number of University of Ghana http://ugspace.ug.edu.gh 80 internodes per panicle were affected by the same QTL. Average number of grains per primary branch and primary branch density show pleiotropic QTL at the position 12.7 cM. On SBI-02, seven pleiotropic QTLs were detected. Number of internodes and primary branch length at penultimate node; secondary branch length and plant height; rachis length, average primary branch length and the longest secondary branch were affected by the same QTL. Primary branch length at second node and primary branch zone at second node; average number of primary branch and panicle bending; panicle shape and compactness and primary branch density, displayed pleiotropic effects. A pleiotropic QTL affected panicle length, rachis slenderness and average number of grains per primary branch. Table 4.5: Pleiotropic QTLs detected on SBI-03 and SBI-04 QTLs Position cM Traits influenced Q R A _ L -S B I- 0 3 Q R A _ L -S B I- 0 4 Q R A _ co n i- S B I- 3 Q R A _ co n i- S B I- 4 Q P B _ L _ m a x- S B I- 0 3 S B _ L _ m a x- S B I- 0 4 P B _ B Z _ S N -S B I- 0 3 P B _ B Z _ M N -S B I- 0 3 N b _ P B _ m a x- S B I- 0 3 P A _ N P B -S B I- 0 3 N b _ S B _ M N -S B I- 0 3 N b _ S B _ P N -S B I- 0 3 S B _ D en s_ S N -S B I- 0 3 N G _ P A -S B I- 0 3 N G _ S N -S B I- 0 3 QNG_MN-SBI-03 14 x x QAv_LPB-SBI-03 135 x QRA_slend-SBI-03 137 x QRA_B_Dia-SBI-03 138 x QPA_L-SBI-03 143 x x QIN_L-SBI-03 144 x x x QPA_W-SBI-03 146.3 x QPB_L_SN-SBI-03 148 x QPA_L-SBI-04 55 x x QIN_L_max-SBI-04 56.5 x On SBI-03, the QTL for the number of grains at median node influenced the QTLs for the number of secondary branches at median and penultimate nodes. The QTL for internode length showed University of Ghana http://ugspace.ug.edu.gh 81 pleiotropic effect with the QTLs for the longest primary branch, primary branch zone at second node and number of grains at second node. A pleiotropic QTL for panicle length affects rachis length and rachis conicity. The traits average primary branch length and primary branch zone at median node; Rachis slenderness and maximum number of primary branch per node; rachis base diameter and number of primary branch per panicle; panicle weight and number of grains per panicle were affected by the same QTLs. On SBI-04, pleiotropic QTL for panicle length influenced rachis length and rachis conicity. The longest internode and the longest secondary branch were influenced by the same pleiotropic QTL. Table 4.6: Pleiotropic QTLs detected on SBI-06 and SBI-07 QTLs Position cM Traits influenced Q P E _ L -S B I- 0 7 Q R A _ _ D ia -S B I- 0 6 Q A v_ L P B -S B I- 0 6 Q N b _ P B _ m a x- S B I- 0 6 Q S B _ D en s_ S N -S B I- 0 6 Q P A _ W -S B I- 0 6 Q N G _ P A -S B I- 0 6 Q B en d in g -S B I- 0 7 QRA_B_Dia-SBI-06 69 x x x QPA_GY-SBI-06 70 x x x QSC-SBI-07 91 x QPH-SBI-07 95 x Two pleiotropic QTLs were detected on SBI-06 and another two on SBI-07. On SBI-06, rachis base diameter, maximum number of primary branches per node, secondary branch density at second node and number of grains per panicle were affected by the same pleiotropic QTL at the position 69 cM. The QTL for grain yield per panicle displayed pleiotropic effect on rachis mean diameter, average primary branch length and panicle weight. University of Ghana http://ugspace.ug.edu.gh 82 On SBI-07, panicle compactness and panicle bending were under the same pleiotropic QTL. Also plant height and peduncle length were influenced by a pleiotropic QTL at the position 95 cM. 4.3.6. Clustered heatmap analysis The coded color heat map shows five main genomic regions corresponding to the SBI-01, SBI-02, SBI-03, SBI-04 and SBI-06 in which QTLs for several variable clusters. However, individual QTLs were spread along different chromosomes. Seven clusters or group of variables can be distinguished as follows: The first cluster (Group-1) shows predominant contribution of Lata-3 alleles (Blue color) in the major QTLs on the SBI-02, SBI-03 and SBI-04. This group was composed of traits corresponding mainly to panicle, primary branches and internode elongation including the longest primary branch and the longest internode (Figure 4.9). These two variables represent a potential of elongation that will be expressed in function of the environment. If the conditions are favorable, the elongation will be at maximum. This cluster indicated relationship between panicle length, basal primary branch length and the longest primary branch and internode. The second cluster (Group-2) indicated the contribution of Lata-3 in majors QTLs while Tiandougou contributed exclusively in minors QTLs. This group of traits was constituted by the QTLs related mainly to secondary branching and elongation, number of internodes per panicle and grain number on basal branches of the panicle. In the third cluster, QTLs for peduncle length was observed with Lata-3 alleles in relation to primary branch zone at the middle of the panicle on SBI-03 and SBI-07. University of Ghana http://ugspace.ug.edu.gh 83 Figure 4.9: QTLs Clustered heat map for sorghum panicle architecture traits in F4 families derived from bi-parental population Group-3 Group-7 Group-6 Group-1 Group-5 Group-4 Group-2 SBI-07 SBI-08 SBI-06 SBI-05 SBI-04 SBI-03 SBI-02 SBI-01 SBI-10 SBI-09 Lata-3 alleles= Blue color Tiand alleles= Red color University of Ghana http://ugspace.ug.edu.gh 84 The fourth cluster presents major QTLs for both parents with more contribution of the parent Lata- 3. This cluster was composed of QTLs related to grain weight, secondary branching and primary branch elongation. In the fifth cluster (Group-5), few QTLs composed of variables at penultimate node and position of the longest primary branch were observed. However, some individual majors QTLs with contribution of Lata-3 and minors QTLs with contribution of Tiandougou were dispersed along all chromosomes except SBI-05 and SBI-08. The sixth cluster (Group-6) shows grouping of many majors QTLs such as grain yield with the contribution of the two parents at three chromosomes (SBI-02, SBI-03 and SBI-06). This cluster was composed of rachis diameter, branching pattern and grain yield and its components. The seventh cluster (Group-7) indicated the contribution of alleles from Tiandougou on SBI-03 and Lata-3 on SBI-06 to QTLs for number and density of primary branches. 4.3.6. Epistasis effect of QTLs detected in F4 families derived from the cross between Tiandougou and Lata-3 Epistasis plays an important role in genetic determination of sorghum panicle architecture. Several epistasis interactions were observed for sorghum panicle traits, grain yield and its components on all the chromosomes at different positions in the genome (Figure 4.10) University of Ghana http://ugspace.ug.edu.gh 85 Figure 4.10: Digenic epistasis detected for number of grains on primary branch at median node (NG_MN) using QTL IciMapping. The most important traits displaying major digenic epistasis effects with percentage of variance explained more than 7% are presented in Table 4.7. University of Ghana http://ugspace.ug.edu.gh 86 Table 4.7: Significant digenic epistasis QTL detected in F4 families derived from Tiandougou x Lata-3 using BLUP data TraitName Chrs Left_Marker1 Chrs Left_Marker2 PVE Add by Add IN_L 4 SBI-04122 10 SBI-10043 7.6523 0.082 RA_Dia 2 SBI-02013 8 SBI-08028 7.6068 -0.0586 RA_Vol 2 SBI-02086 3 SBI-03164 7.228 122.8724 PB_L_max 1 SBI-01103 9 SBI-09091 7.3482 2.3068 Av_LPB 7 SBI-07067 10 SBI-10012 9.3613 -0.7289 SB_L 2 SBI-02013 3 SBI-03102 13.0152 0.041 SB_L_max 2 SBI-02013 3 SBI-03102 16.8535 0.166 SB_DF_SN 2 SBI-02040 9 SBI-09003 7.6676 -0.3041 SB_DF_MN 6 SBI-06098 8 SBI-08042 7.1256 -0.4199 Av_SB_DF 2 SBI-02013 3 SBI-03002 22.1997 0.1669 Av_SB_DF 4 SBI-04090 5 SBI-05105 11.3372 -0.0288 Pos_NPB_max 1 SBI-01062 6 SBI-06106 7.2599 -0.1476 Pos_NPB_max 8 SBI-08042 10 SBI-10064 7.0829 -0.3316 NG_MN 3 SBI-03006 9 SBI-09014 18.03 -0.2119 NG_MN 4 SBI-04136 7 SBI-07067 15.3263 -0.097 NG_MN 4 SBI-04136 9 SBI-09014 24.6544 -1.0632 Av_NG 1 SBI-01038 2 SBI-02013 18.5469 0.2117 Av_NG 2 SBI-02013 3 SBI-03137 18.3073 -0.3306 Av_NG 2 SBI-02013 6 SBI-06127 17.0226 0.0658 NB: PVE = Percentage of Variance Explained; addxadd = additive x additive; chrs = chromosomes A total of nineteen significant digenic epistatic interactions (additive by additive) were detected for 13 traits using the BLUP extracted from all three sowing dates indicating that two loci interactions were widespread in the entire genome. A total of three digenic epistasis were found in number of grains per primary branch at median node and average number of grains per primary branch. Two were found in average distances to the first secondary branch on the primary branch and position of the maximum number of primary branches per node. One digenic epistasis QTLs was found in each of the following traits: internode length, rachis mean diameter, rachis volume, the longest primary branch, average primary branch length, secondary branch length, the longest secondary branch on the longest primary branch, distance to the first secondary branch at second and median nodes. University of Ghana http://ugspace.ug.edu.gh 87 Internode length, the one significant additive x additive interactions, was detected between SBI-04 and SBI-10 with the effect of 0.082 explaining 7.65% of phenotypic variance. For rachis mean diameter, additive x additive interaction was observed between SBI-02 and SBI-08 with negative effects of -0.05 and 7.60% of phenotypic variance explained by the epistasis QTL. Positive Epistasis effect of 122. 87 was noted for rachis volume between SBI-02 and SBI-03 with the percentage of variation explained by the epistasis QTL of 7.22%. The one significant additive x additive interaction detected for the longest primary branch explained 7.34% of phenotypic variance with 2.30 interaction effects. Average primary branch length displayed negative additive x additive interaction between SBI-07 and SBI-10 with the percentage of phenotypic variance of 9.36% explained. Between SBI-02 and SBI-03 epistasis effect was noted for secondary branch length with 13.01% of phenotypic variance explained. For the longest secondary branch on the longest primary branch the one significant epistasis QTL was detected on the SBI-02 and SBI-03 with the percentage of variance explained of 13.85%. The additive x additive interaction observed for distance to the first secondary branch at second and median nodes were between SBI-02 and SBI-09 at second node and between SBI-06 and SBI-08 at median node with respectively 7.66 and 7.12% of phenotypic variance explained. The total phenotypic variation explained by interactions for average distance to the first secondary branch on the primary branch and position of the maximum number of primary branch per node were 33.54% and 14.34% respectively from two digenic pairs; number of grains on primary branch at median node and average number of grains per primary branch were respectively 58.01% and 53.88% from three digenic pairs. In general, for the traits with major digenic epistasis interaction, the total phenotypic variations explained by the interactions were greater than that of corresponding main effects for the different University of Ghana http://ugspace.ug.edu.gh 88 traits. The importance of additive x additive interaction effects in total genetic effects may be trait- dependent. 4.3.7. Identification of QTLs with overlapping confidence interval for sorghum panicle architecture traits across reported studies QTLs with overlapping confidence interval for panicle architecture traits from four previous studies and the present study were identified for the following traits: primary branch length, secondary branch number at second node, distance to the first secondary branch on the primary branch, panicle length, panicle weight and thousand grain weight (Table 4.8). Most of the QTLs identified were composed of two individual QTLs for the concerned traits. Primary branch length presented the maximum QTL with overlapping confidence interval with four individual QTLs on SBI-03. The present study on panicle architecture using F4 families confirmed results from Pereira et al. (1995) and Brown et al. (2006) on SBI-03 two QTLs with overlapping confidence interval for primary branch length. University of Ghana http://ugspace.ug.edu.gh 89 Table 4.8: QTLs with overlapping confidence interval identified from five (5) studies on sorghum panicle architecture traits. Panicle traits Number of QTLs Chrs Mean CI position CI Start CI End LOD PVE Sources Primary branch length 2 SBI- 01 69.5 60.3 78.7 6.6 19.0 Pereira et al. (1995) 65.8 61.5 70.1 7.0 13.8 Brown et al. (2006) Primary branch length 4 SBI- 03 22.5 0.7 44.3 2.6 8.0 Pereira et al. (1995) 13.3 8.7 17.8 6.5 13.0 Brown et al. (2006) 13.0 6.0 16.0 4.4 4.1 This study 14.0 8.0 26.0 4.5 4.5 This study Panicle length 2 SBI- 02 141.0 137.7 144.3 4.9 14.7 Srinivas et al. (2009) 141.0 125.0 147.0 8.6 6.9 This study Panicle length 2 SBI- 03 155.0 135.6 174.4 3.0 9.0 Pereira et al. (1995) 143.0 135.0 147.0 13.4 11.2 This study Panicle weight 2 SBI- 06 84.0 80.7 87.3 4.9 14.7 Srinivas et al. (2009) 70 64.9 78.0 12.8 12.3 This study University of Ghana http://ugspace.ug.edu.gh 90 For panicle length, this study showed two QTLs with overlapping confidence interval with the study of Srinivas et al. (2009) on SBI-02 and Pereira et al. (1995) on SBI-03. QTLs with overlapping confidence interval were recorded on SBI-06 for panicle weight in the Srinivas et al. (2009) study and in this study. However this study showed no QTL with overlapping confidence interval with those already identified (result not shown) for primary branch length on SBI-02 (Pereira et al., 1995 and Brown et al., 2006) and thousand seed weight on SBI-07 (Rami et al., 1998). Distance to the first secondary branch or the sterile portion of the seed branch seems to be affected by the genetic background of the different parents used in the population. Indeed QTLs with overlapping confidence interval were identified for each specific study. 4.4. Discussion A great number of QTLs were identified for sorghum panicle architecture traits in different studies. Some interesting QTLs detected across studies were used in the current discussion. Pereira et al. (1995) identified three QTLs for primary branch number per panicle on SBI-02, SBI-07 and SBI- 08. Brown et al. (2006) reported two QTLs for primary branch number on SBI-03 and SBI-06 and Srinivas et al. (2009) identified four QTLs for primary branch number per panicle on SBI-01, SBI- 05, SBI-07 and SBI-08 that did not relate to the QTLs detected by Brown et al. (2006). In this study, QTLs for primary branch number were detected on SBI-02, SBI-03, SBI-06 and SBI-07. One common QTL for primary branch number detected on SBI-03 seems to be similar to the one detected by Brown et al. (2006) based on the position of the QTLs in the genome. Three other common QTLs on the SBI-02, SBI-06 and SBI-07 across studies were different and situated very far away from each other in the genome. This result may suggest that depending on the genetic background. QTLs for the number of primary branches per panicle were less closely related. Therefore, there were different loci regulating the primary branching pattern in sorghum panicle University of Ghana http://ugspace.ug.edu.gh 91 resulting in their involvement in panicle architecture. Rachis diameter QTLs were detected on SBI- 02, SBI-03, SBI-06, SBI-07 and SBI-10 in this study. Brown et al. (2006) reported one QTL for rachis diameter on SBI-06. However the one common QTLs on SBI-06 from this study and the study of Brown et al. (2006) were respectively detected at the position 69 and 55. These positions were close and it is possible to assume that they were the same QTLs. Brown et al. (2006) reported four QTLs for primary branch length, two on SBI-01 and one on each of SBI-03 and SBI-07 while this study detected five QTLs two on SBI-03 and one on each SBI-02, SBI-06 and SBI-07. The QTLs detected on SBI-03 and SBI-07 in the two studies were different. Thousand seed weight QTLs were identified by Rami et al. (1998) on SBI-01, SBI-03 and SBI-07, Murray et al. (2008) also reported three QTLs for thousand seed weight on SBI-01, SBI-06 and SBI-08. In this study two different QTLs on SBI-03 and SBI-07 other than those detected by Rami et al. (1998) for thousand seed weight were found based on the position in the genome. Feltus et al. (2006) found one QTLs on SBI-10 for panicle weight while Srinivas et al. (2009) reported one QTL for panicle weight on SBI-06. This study identified two QTLs for panicle weight on SBI-01 and SBI-06. The QTLs on SBI-06 from Srinivas and this study were close in the genome respectively at position 84 and 70 cM. The proximity of location on the chromosome may suggest that they are the same QTL. Rami et al. (1998) reported four QTLs for Number of kernels per panicle. Two QTLs were detected on SBI-01, one on SBI-02 and another one on SBI-07 while this study identified two QTLs on SBI-02 and SBI-06. The QTLs detected on SBI-02 in these studies were at different positions on the chromosome. Rami et al. (1998) and Pereira et al. (1995) found two QTLs for panicle length, this study identified one QTL on SBI-01. These QTLs were spread along SBI-01. On SBI-02, this study detected one QTL for panicle length that was not similar to any of the QTLs detected and reported in the literature. Rami et al. (1998) found two QTLs for panicle length while Parh (2005) and Srinivas et al. (2009) reported one QTL for panicle length on University of Ghana http://ugspace.ug.edu.gh 92 SBI-02. Pereira et al. (1995) and Rami et al. (1998) also reported one QTL for panicle length. This study identified one QTL on SBI-03 that is close to the QTL detected by Pereira et al. (1995). There were no QTLs detected in this study on SBI-06, SBI-07, SBI-08, SBI-09 and SBI-10. Parh (2005) and Srinivas et al. (2009), however, found two QTLs for panicle length on SBI-06. Rami et al. (1998) and Pereira et al. (1995) reported one QTL for panicle length on SBI-06. Hart et al. (2001), Srinivas et al. (2009), Rami et al. (1998) and Pereira et al. (1995) reported one QTL for panicle length on SBI-07. Hart et al. (2001) reported one QTL for panicle length on SBI-08 and SBI-09. Rami et al. (1998) and Pereira et al. (1995) reported one QTL for panicle length on SBI-10. QTLs for grain yield per panicle were reported by Ritter et al. (2008) in R9188/R9403463-2-1 sweet sorghum population on SBI-03 at the position 133 cM in the genome and Srinivas et al. (2009) in 296B/IS18551 sorghum on SBI-06 at 83.5 cM. In this study two QTLs were detected for grain yield per panicle on SBI-03 at 146.3 cM and SBI-06 at 70 cM in Tiandougou/Lata-3 population. The QTL for grain yield per panicle detected on SBI-06 in this study is likely the same detected by Srinivas et al. (2009) based on their positions in sorghum genome. Nagaraja Reddy et al. (2013) detected one major QTL for grain yield per panicle on SBI-06 and one minor QTL on SBI-03. The position of these QTLs are not related to the QTLs for grain yield per panicle detected in this study. Traits correlated are known generally to be mapped together (Hittalmani et al., 2002), Lin et al. (1995) and Klein et al. (2008) detected pleiotropic QTLs for maturity and height. Zou et al. (2012) detected pleiotropic QTLs for heading date, number of nodes, plant height and stem diameter in long and short day photoperiod. Pleiotropic QTLs in this study suggest that increasing grain yield may be obtained by positively influencing primary branch length, rachis diameter and the weight of the panicle. Nagaraja Reddy et al. (2013) also found pleiotropic QTL affecting grain yield per panicle and panicle weight. Rachis growth and primary branches pattern are under the same genetic control. The number of University of Ghana http://ugspace.ug.edu.gh 93 internodes per panicle determine rachis, primary and secondary branches elongation and flowering time might be affected by the mechanism regulating peduncle elongation. Srinivas et al. (2009) reported the QTL cluster for plant height, days to anthesis, green leaf area at maturity, panicle length, grain yield, panicle weight, and seed weight. In this study, QTLs clustered heatmap confirmed the link between panicle grain yield and some panicle architecture traits such as the number of primary branches per panicle and rachis base diameter. Indeed this cluster points out the traditional relationship between grain yield and grain yield components like panicle weight and number of grains per panicle. There was somehow a link between plant growths in term of plant height that may also affect the grain yield. Therefore, the improvement of sorghum grain yield should combine certain grain yield components as well as associating some of panicle architecture traits such as secondary branch density at the base of the panicle, rachis base diameter and the number of primary branches per panicle. Epistasis refers to the phenotypic effects of interactions among alleles at multiple loci (Xing et al., 2002). The advent of molecular makers and their utilization has revealed that epistastatic interactions play an important role on the genetic basis of quantitative traits (Paterson et al., 1991; Li et al., 1997). Liao et al. (2001) suggests that there are three types of epistasis affecting complex traits: (1) interactions between QTL, (2) interactions between QTL and modifying loci, and (3) interaction between complementary loci. Shiringani (2009) reported epistasis in seven out of ten traits studied, Digenic epistasis interactions were observed to be very important in the pattern of the number of grains per primary branches and the length of the secondary branches in this study. Routman and Cheverud. (1997) reported that epistasis in both negative and positive directions is important for the breeding populations even if the net effects are not observed. Overlapping confidence interval of QTLs across studies on sorghum panicle architecture denote the consistency University of Ghana http://ugspace.ug.edu.gh 94 of these QTLs and may be used in QTL Meta- analysis to declare them as Meta-QTL by refining their position and by projecting onto the unique consensus genetic map. 4.5. Conclusion Both additive and non-additive gene effects control most of sorghum panicle architecture traits. Additive gene action was determinant in the control of the number of internodes per panicle, primary branches zone and maximum primary branches per node, while the length and number of secondary branches were governed by non-additive gene effects. Grain yield and some of its components such as panicle weight and thousand grain weight were under non additive gene action. Number of grains per panicle was predominantly governed by additive gene effects; however, the number of grains per primary branches at different positions in the panicle was controlled by both additive and non-additive gene action. Most of the major QTLs were consistently detected across the three sowing dates. QTLs clustered for grain yield per panicle, rachis diameter, number of primary branches per panicle and secondary branch density at the base of the panicle were observed in this study. Therefore it is possible to improve grain yield through indirect selection for these traits. Fourteen out of forty eight traits studied showed digenic epistasis interaction indicating the important role of epistasis in the regulation of the pattern of internode length, primary and secondary branches length and the number of grains per primary branch that showed the maximum percentage of variation explained (36%). QTLs showing multiple effects or pleiotrophism were detected on six chromosomes out of ten. The major pleiotropic QTL affecting grain yield per panicle, rachis mean diameter, average primary branch length and panicle weight was found at 70 cM on SBI-06. Enhancement of grain yield- related traits could be faster if pleiotropic effects favor the positive alleles. University of Ghana http://ugspace.ug.edu.gh 95 QTLs with overlapping confidence intervals were consistently identified across studies for primary branch length, panicle length and panicle weight. The flanking markers both left and right SB06074 and SB06098 of grain yield per panicle QTL on SBI-06 were consistently associated with QTLs for panicle weight, secondary branch length and primary branch density, whereas the flanking markers both left and right SB03149 and SB03164 of grain yield per panicle QTL on SBI-03 were consistently associated with QTLs for thousand grain weight, internode length and the longest secondary branch. University of Ghana http://ugspace.ug.edu.gh 96 CHAPTER FIVE 5.0. IDENTIFICATION OF CANDIDATE GENES INVOLVED IN SORGHUM PANICLE ARCHITECTURE IN MAIZE AND RICE 5.1. Introduction Sorghum panicle architecture is determined by meristem activities during the reproductive phase. Axillary meristems specify the formation of branches and spikelets which are defined as short branch floret-bearing branches in the panicle. Axillary meristem initiation and development are regulated by developmental, environmental, and hormonal controls (Barazesh and McSteen, 2008). Understanding the regulation of axillary meristem formation is essential to improving grain crops as panicle structure directly affects grain yield. The genes required to maintain spikelet meristem identity belong to the SEPALLATA group. Sorghum MADS box genes SbMADS1 and SbMADS2 are the two MADS box genes whose cDNAs have been isolated from sorghum (Greco et al., 1997). PANICLE PHYTOMER2 (PAP2 or rice MADS34) controls loss of function in a panicle composed of many primary, but few secondary, branches because many of the early spikelets are transformed into rachis branches (Kobayashi et al., 2010; Gao et al., 2010). Also the APETALA2 (AP2) family genes, such as AP2-like genes, have been implicated in spikelet determinacy and the transition to the floral meristem. These include BRANCHED SILKLESS1/FRIZZY PANICLE1/BRANCHED FLORETLESS 1 (BD1/FZP1/BFL1) (Komatsu et al., 2003b; Chuck et al., 2002), INDETERMINATE SPIKELET 1 (ZmIDS1/OsIDS1) (Chuck et al., 2008; Chuck et al., 1998) and SISTER OF INDETERMINATE SPIKELET1/SUPER NUMERARY BRACT1 (ZmSID1/OsSNB1) (Chuck et al., 2008; Lee et al., 2007). It has been reported that OsSNB1 and University of Ghana http://ugspace.ug.edu.gh 97 OsIDS1 act synergistically to regulate the transition of spikelet meristem into floral meristem and to control inflorescence architecture in rice (Lee and An, 2012). Several downstream genes have been identified as being involved in spikelet meristem identity in rice; these include ABERRANT PANICLE ORGANIZATION1 (APO1) (which encodes an F-box protein related to Arabidopsis UFO) (Sreenivasulu and Schnurbusch, 2012). The panicles of APO1 loss-of-function mutants form fewer spikelets than the wild-type panicles. Elevated transcript levels of APO1 in the rice stems or panicles are accompanied by a higher growth rate of the inflorescence meristem and by enlarged vascular bundles, thus producing a higher number of spikelets per panicle (Terao et al., 2010; Ikeda-Kawakatsu et al., 2009). Spikelet, the primary panicle architecture unit in sorghum, development is dictated by several specific genes. Spikelets bear one to several florets and each floret develops a set of grain contributing to yield potential and final grain number. The ability to manipulate the spikelet formation process may allow for re-engineering of panicle architecture in a way that increases yield potential (Sreenivasulu and Schnurbusch, 2012). Sorghum, together with maize and sugarcane, belongs to the group of grasses that develop spikelet pairs and the ramosa pathway is particularly important (Kellogg, 2007; McSteen et al., 2007; Vollbrecht et al., 2005). In maize, analyses of the highly branched mutants Zmramosa1 (ra1), Zmra2 and Zmra3 have revealed a pathway that negatively regulates inflorescence branching (Bortiri et al., 2006; Vollbrecht et al., 2005). Only ramosa2, which is a putative LOB domain- containing a transcription factor, is well conserved across the Poaceae family (Kellogg, 2007). In rice, MONOCULM 1, MOC1 (Li et al., 2003), LAX PANICLE 1, LAX1 (Komatsu et al., 2003a), LAX PANICLE 2, LAX2 (Tabuchi et al., 2011) and SHORT PANICLE 1, SP1 (Li et al., 2009), all regulate the degree of panicle branching, as well as panicle size. It was established that LAX1 and LAX2 synergistically function in regulating the axillary meristem and thereby panicle branching University of Ghana http://ugspace.ug.edu.gh 98 (Tabuchi et al., 2011). Higher numbers of branches and spikelets have been associated with increased grain number associated with the recessive Gn1a (=OsCKX2) allele in rice (Ashikari et al., 2005). Huang et al. (2009) reported that Dense and Erect panicle1 (DEP1) acts upstream of OsCKX2 to control cytokinin homeostasis in the panicle meristem (Huang et al., 2009). To date, only a few number of genes have been implicated in controlling panicle architecture in sorghum (Brown et al., 2006). The need to identify and characterize the function of novel panicle architecture genes cannot be ever emphasized. Therefore, the objectives of this study were 1). to identify candidate genes regulating sorghum panicle architecture. 2). to establish protein sequence similarity with sorghum, maize and rice genes and 3). to determine co-localization between sorghum candidate gene predictions with QTLs detected for panicle architecture traits in Tiandougou/ Lata-3 segregating population. 5.2. Methods An inventory of genes known to be involved in panicle architecture in rice and maize was made based on the literature and their homologs were retrieved from the sorghum genome sequence based on protein sequence similarity. The hypothesis underlying this work was that genes involved in panicle architecture in species closely related to sorghum, such as rice (Oryza sativa) and maize (Zea mays), have conserved similar functions in Sorghum and could be good candidate genes for the control of panicle architecture in sorghum. Rice and maize protein sequences corresponding to genes involved in meristem regulation during inflorescence development, specification of the spikelet meristem identity and plant growth and development were retrieved from the NCBI (National Center for Biotechnology Information) database (http://www.ncbi.nlm.nih.gov/). A total of 43 protein sequences were mined from the University of Ghana http://ugspace.ug.edu.gh 99 NCBI database, including 12 accessions common to rice and maize; 21 accessions specific to rice and 10 accessions specific to maize. Sorghum homologs for these genes were searched in silico within the sorghum whole-genome sequence v.1.4. (Paterson et al., 2009) using BLASTP (Basic Local Alignment Search Tool for Protein) versus sorghum gene predictions (amino acid) (http://www.phytozome.net/search.php?show=blast&method=Org_Sbicolor_v1.4). Twenty five (25) gene predictions having the best BLAST (Basic Local Alignment Search Tool) expected values were selected for each rice and maize genes searched. The chromosome and their positions were noted in order to identify the best rice and maize proteins related to each sorghum gene prediction (E-value <1E-50). Alignments and phylogenetic tree analyses were carried out to determine the relationship between sorghum gene predictions and maize and rice genes. Predicted sorghum genes were named based on their similarity with rice and maize proteins. Deduced amino acid sequences were analyzed using the Phylogeny.fr platform (http://www.phylogeny.fr) using the one click mode that includes the pipeline chaining programs: MUSCLE for multiple alignment (Edgar, 2004), PhyML for tree building (Guindon and Gascuel, 2003; Anisimova and Gascuel, 2006), and TreeDyn for tree drawing (Chevenet et al., 2006). The tree building was based on an approximation of the maximum likelihood ratio test. A physical map was generated that included the positions in Megabases (Mb) of the predicted genes and of the SNP markers present on the genetic map. The positions of the QTLs detected on the Tiandougou/Lata-3 genetic map without candidate genes were then compared with the physical map. University of Ghana http://ugspace.ug.edu.gh 100 5.3. Results The availability of sorghum genome sequences allowed the search for candidate genes involved in meristem regulation during inflorescence development and specification of spikelet meristem identity in relation to plant growth and development. Therefore, co-localization of candidate genes and QTLs detected for panicle architecture traits that provide new information on putative genetic control of sorghum panicle architecture could be established. However, it is necessary to note that co-localization between candidate genes and QTLs clusters was not always evident for a large number of candidate genes involved in inflorescence architecture. A total of 42 candidate genes for inflorescence architecture, growth and development were selected on the basis of their known function in rice and maize. The list of selected candidate genes included 17 genes related to meristem regulation during inflorescence development (Table 5.1); 15 related to specification of spikelet meristem identity (Table 5.2) and 10 related to plant growth and development (Table 5.3). The candidate gene identification gave an indication of the position in the genome of the 43 genes involved in inflorescence, growth and development and enabled determination of predicted genes in sorghum named on the basis of similarity with rice and maize genes. Emphasis has been maintained on 25 candidate genes showing high sequence similarity and co-localization with the QTLs detected for sorghum panicle architecture traits in relation to the three traits concerned, which are meristem regulation during inflorescence development, specification of spikelet meristem identity and plant growth and development. 5.3.1. Candidate genes related to meristem regulation during inflorescence development The focus on candidate genes for meristem regulation during inflorescence development (Table 5.1), was with regard to Aberrant spikelet and panicle1 (ASP1) in rice which encodes a transcriptional co-repressor important in branching pattern, transition from the branch meristem to University of Ghana http://ugspace.ug.edu.gh 101 the spikelet meristem and stem cell maintenance in both the branch meristem and the spikelet meristem. The function of this gene is also closely associated with auxin action. Therefore, in rice, ASP1 regulates various aspects of developmental processes and physiological responses as a transcriptional co-repressor (Yoshida et al., 2012). The corresponding genes to ASP1 in maize is RAMOSA ENHANCER LOCUS2 (REL2) which is a recessive mutation affecting both tassel and ear branching (Soderlund et al., 2009). These candidate genes co-localized with the QTLs for average primary branch length and density on SBI-07 at the position 5.31 cM. This suggest that the sorghum homologue of ASP1 (SbASP1) may be responsible for branch pattering during panicle development and elongation. Lax panicle1 (LAX1) encodes a basic helix-loop-helix (bHLH) domain transcription factor. In rice, LAX1 regulates the number of branches as a mutation affecting LAX1 presents fewer branches in the inflorescence (Komatsu et al., 2003a). LAX1 protein sequence showed high sequence similarity to the sorghum Sb03g038820 predicted gene. Sorghum LAX1 (SbLAX1) on SBI-03 co-localize with the QTLs for number of primary branches per panicle; maximum number of primary branches per node and rachis base diameter. This result suggests that LAX1 in sorghum may regulate primary branch number and also play an important role in rachis size. The rice Gn1a/OsCKX2 (Grain number 1a/Cytokinin oxidase 2) gene, which encodes a cytokinin oxidase, has been identified as a major quantitative trait locus contributing to grain number improvement in rice. The candidate gene GRAIN NUMBER 1a (GN1a)/osCKX2 in rice accession Q4ADV8 (Ashikari et al., 2005) had sequence similarity with sorghum gene prediction (Sb03g002810). GN1 encodes cytokinin oxidase/dehydrogenase, an enzyme that catalyzes the degradation of cytokinin, implying the involvement of that phytohormone in panicle branching (Ashikari et al., 2005). Sorghum GRAIN NUMBER 1a (SbGN1a) on SBI-03 at the position 2.72 cM co-localized with the QTLs for average grain number per primary branch and the QTLs for University of Ghana http://ugspace.ug.edu.gh 102 grain number on the primary branch at median and penultimate nodes. This suggests that this gene in sorghum regulates grain number at different stages of primary branches on the main rachis and may act upstream of grain production at a different stratum of branching. University of Ghana http://ugspace.ug.edu.gh 103 Table 5.1: List of candidate genes related to meristem regulation during inflorescence development Sorghum name Sorghum gene prediction # Chr Position Reference gene in Oryza sativa Reference gene in Zea mays QTL co- location References Oryza name Oryza accession # E-value Zea name Zea accession # E-value Sbra2 Sb03g004590 3 4,862 ramosa2 (Zmra2) ABC54560 4,20E- 102 Av_LPB, PB_L_MN Bortiri et al. 2006 Sbra3 Sb02g039820 2 73,973 ramosa3 (Zmra3) ABD92779 3,50E- 149 Satoh-Nagasawa et al. 2006 Sbsra3 Sb02g039810 2 73,961 Sister of ramosa3 DAA63682 2,00E- 160 Sbfl Sb06g027340 6 56,245 floricaula (zfl1, zfl2) AAO43174 8,70E- 132 Nb_SB_SN Bomblies et al. 2003 Sbfea2 Sb04g025240 4 55,066 fasciated ear2 (fea2) AAL17871 0,00E+00 Taguchi-Shiobara et al. 2001 SbTu1 Sb04g033930 4 63,833 Tunicate1 (Tu1) NP_001105148 5,10E- 106 PH Mendez-Vigo et al. 2013 SbASP1 Sb07g004180 7 5,318 ABERRANT SPIKELET AND PANICLE1 (ASP1) BAL44266 0,00E+00 ramosa enhancer locus2 (rel2) NP_001167872 0,00E+00 Av_LPB, PB_Dens Yoshida et al. 2012; Soderlund et al. 2009 Sbbif2 Sb08g021520 8 53,232 PINOID (OsPID) Q2QM77 2,50E- 135 barren inflorescence2 (bif2) ABR13340 2,50E- 130 McSteen et al. 2007 SbLAX1 Sb03g038820 3 66,623 LAX PANICLE 1 (LAX1) Q7XAQ6 6,00E-47 PA_NPB, Nb_PB_max, RA_B_Dia Komatsu et al. 2003a SbLAX2 Sb06g014520 6 40,193 LAX PANICLE 2 (LAX2) BAL03066 2,70E-19 Tabuchi et al. 2011 Sbba1 Sb02g026970 2 62,203 BARREN STALK (ba1) AEL78914 7,30E-72 PA_IN_N Gallavotti et al. 2011 SbAPO1 Sb10g026580 10 56,003 ABERRANT PANICLE ORGANIZATION1 (APO1) BAF75467 3,60E- 129 PE_L, PH Ikeda et al. 2007 SbMOC1 Sb10g023950 10 52,722 MONOCULM 1 (MOC1) BAD35485 6,80E- 136 SbCKX2 Sb03g002810 3 2,722 GRAIN NUMBER 1a (GN1a)/OsCKX2 Q4ADV8 4,40E- 150 Av_NG, NG_MN, NG_PN Ashikari et al. 2005 SbDEP1 Sb02g025860 2 60,87 DENSE AND ERECT PANICLE 1 (DEP1) ACI25447 2,60E-33 Nb_SB_LPB Huang et al. 2009 SbIPA1 Sb07g027740 7 62,69 IDEAL PLANT ARCHITECTURE (IPA1) ADJ19220 2,50E- 105 RA_Bdia, RA_Dia, RA_Slend Jiao et al. 2010 SbGS3 Sb01g032830 1 55,779 GRAIN LENGTH AND WEIGHT PROTEIN (GS3) ABC84855 1,20E-32 Fan et al. 2006 University of Ghana http://ugspace.ug.edu.gh 104 5.3.2. Candidate genes related to specification of the spikelet meristem identity For candidate genes related to specification of the spikelet meristem identity (Table 5.2), focus has been on rice FRIZZY PANICLE (FZP), an ortholog of the maize BRANCHED SILKLESS 1 (BD1), that regulates spikelet meristem identity (Chuck et al., 2002; Komatsu et al., 2003b) and has sequence similarity with sorghum prediction Sb02g042400 on 2 and co-localize with the QTLs for the number of primary branches per panicle in this study. This result may indicate that spikelet meristem development may act on the whole process of branch formation. Maize INDETERMINATE SPIKELET1 (IDS1) regulates spikelet-meristem determinacy (Chuck et al., 1998). Spikelet of the IDS1 mutant are indeterminate, producing extra florets. SISTER OF IDS1 (SID1), a paralog in maize, is important for floral meristem initiation and spikelet regulation (Chuck et al., 2008). OsIDS1 positively regulates panicle branching by repressing spikelet meristem identity (Yeon Lee and An, 2012). In this study, INDETERMINATE SPIKELET1 (OsIDS1; ZmIDS1) candidate genes in rice and maize presented sequence similarity with sorghum gene prediction (Sb01g003400) on SBI-01 at the position 2.73 which co-localized with the QTLs for average grain number per primary branch and thousand grain weight. The OsFOR1 (Oryza sativa floral organ regulator 1) gene encodes a protein that contains a leucine-rich repeat (LRR) domain and suggests that OsFOR1 plays a role in the formation and/or maintenance of floral organ primordia. In this study, the candidate gene FLORAL ORGAN REGULATOR1 (OsFOR1) in rice accession AA017320 showed sequence similarity with sorghum gene prediction (Sb02g036750) and co-localize with QTLs for panicle shape and compactness and average primary branches per node. This result suggests that SbFOR1 may play a role in the organization of the panicle that determines sorghum panicle shape and compactness. University of Ghana http://ugspace.ug.edu.gh 105 Table 5.2: List of candidate genes related to specification of spikelet meristem identity Sorghum name Sorghum gene prediction # Chr Position Reference gene in Oryza sativa Reference gene in Zea mays QTL co- location References Oryza name Oryza accession # E-value Zea name Zea accession # E-value Specification of the spikelet meristem identity SbBD1 Sb02g042400 2 76,137 FRIZZY PANICLE (OsFZP) BAC79264 4,20E-53 BRANCHED SILKLESS1 (ZmBD1) NP_001105200 1,00E-60 PA_NPB Chuck et al. 2002; Komatsu et al. 2003b SbIDS1 Sb01g003400 1 2,731 INDETERMINATE SPIKELET 1 (OsIDS1) BAC21448 3,20E- 123 INDETERMINATE SPIKELET 1 (ZmIDS1) AAC05206 0,00E+00 Av_NG, NG_MN, NG_PN, PA_HI, TGW Chuck et al. 1998 SbSID1 Sb02g007000 2 8,952 SUPERNUMERARY BRACT 1 (OsSNB1) ABD24033 2,70E- 144 SISTER OF INDETERMINATE SPIKELET 1 (ZmSID1) NP_001139539 9,70E- 180 Lee et al. 2007 SbMFO1 Sb04g031750 4 61,676 MOSAIC FLORAL ORGANS1 (MFO1) ACV89861 4,00E- 118 bearded-ear1 (bde) NP_001105332 4,20E- 139 Thompson et al. 2009 SbLHS1 Sb01g042840 1 66,077 LEAFY HULL STERILE1 (LHS1) Q10PZ9 2,20E- 101 Jeon et al. 2000 SbMADS3 Sb03g002525 3 2,406 OsMADS3 Q40704 8,60E- 111 zmm2, zmm23 CAA57074 2,90E- 109 Kang et al. 1995 SbMADS58 Sb09g006360 9 9,847 OsMADS58 BAE54300 4,40E-96 zag1 NP_001105321 1,30E- 139 Schmidt et al. 1993 SbMADS13 Sb08g006460 8 10,063 OsMADS13 Q2QW53 1,50E-91 zag2 CAA56504 9,20E- 140 Theissen et al. 1995 SbDL Sb01g042850 1 66,095 DROOPING LEAF (OsDL) AAR84663 4,60E-95 DROOPING LEAF (ZmDL) ACG32674 2,40E- 106 SbFON1 Sb03g037580 3 65,504 FLORAL ORGAN NUMBER1 (OsFON1) Q5Z9N5 1,50E- 101 Suzaki et al. 2004 SbFOR1 Sb02g036750 2 71,135 FLORAL ORGAN REGULATOR 1 (OsFOR1) AAO17320 1,00E- 142 PB_Dens, Av_NPB, Shape and compactness Jang et al. 2003 SbLP Sb04g009700 4 12,281 LARGER PANICLE (LP) ADQ13183 0,00E+00 SbAP1 Sb02g001090 2 0,956 APETALA (AP1) NP_001105333 2,10E- 102 Mena et al. 1995 SbCEN- like2 Sb05g003200 5 3,587 TFL1/CEN-like AF159883 2,10E-92 Zhang et al. 2005 SbCEN- like1 Sb08g003210 8 3,572 TFL1/CEN-like AF159883 6,10E-94 Zhang et al. 2005 University of Ghana http://ugspace.ug.edu.gh 106 5.3.3. Candidate genes related to plant growth and development Candidate genes related to plant growth and development, HIGH-TILLERING DWARF1 (OsCCD7) in rice (Umehara et al., 2008), displayed sequence similarity with sorghum gene prediction (Sb06g024560) on chromosome 6. This sorghum gene prediction co-localized with the QTLs for number of primary branches per panicle and average primary branch per node. SbCCD7 may play a role in the primary branch production. According to Zou et al. (2006), RNA gel blot analysis showed that the HIGH-TILLERING DWARF1 (HTD1) gene was expressed in all tissues examined. The levels of expression were generally higher in aerial tissues (leaf, stem and panicle) than in root tissue. Gibberellins are growth factors with a tetracyclic diterpenoid structure that are essential regulators of diverse growth and developmental processes of plants (Davies, 1995). The candidate gene Gibberellin 20-oxidase (OsGA20ox) in rice presented sequence similarity with sorghum gene prediction (Sb03g041900) on SBI-03 and co-localized with QTLs for internode length, primary branch length at second node, and primary and secondary branch density. This suggests the role played by Gibberellin 20 oxidase in sorghum panicle elongation and branch development. Toyomasu et al. (1997) demonstrated in rice that the OsGA20ox transcript level was controlled in a negative feedback manner by the level of bioactive Gibberellin. University of Ghana http://ugspace.ug.edu.gh 107 Table 5.3: List of candidate genes related to plant growth and development Sorghum name Sorghum gene prediction # Chr Position Reference gene in Oryza sativa Reference gene in Zea mays QTL co- location References Oryza name Oryza accession # E-value Zea name Zea accession # E-value Plant growth and developement SbCCD7 Sb06g024560 6 53,677 HIGH- TILLERING DWARF1 (OsCCD7) Q7XU29 0,00E+00 Av_NPB, PA_NPB Umehara et al. 2008 SbCCD8 Sb03g034400 3 62,608 HIGH- TILLERING DWARF1 (OsCCD8b) Q8LIY8 0,00E+00 PB_BZ_MN Umehara et al. 2008 SbTB1 Sb01g010690 1 9,506 TEOSINTE BRANCHED 1 (Oryza latifolia) BAM13331 4,60E-54 TEOSINTE BRANCHED 1 (ZmTB1) AFI70995 5,20E-79 SbPLA1 Sb01g022690 1 29,065 PLASTOCHRON1 (PLA1) Q7Y1V5 0,00E+00 Miyoshi et al. 2004 SbPLA2 Sb03g043230 3 70,604 PLASTOCHRON2 (PLA2) BAE79763 8,10E- 171 TERMINAL EAR1 AAK29419 0,00E+00 Kawakatsu et al. 2006 SbPLA3 Sb01g005710 1 4,17 PLASTOCHRON3 (PLA3) BAH22722 0,00E+00 Kawakatsu et al. 2009 SbDWARF8 Sb01g010660 1 9,419 DWARF8 AAL10393 0,00E+00 SbGA20ox Sb03g041900 3 69,376 Gibberellin 20- oxidase (OsGA20ox) AAM56041 1,00E- 166 IN_L, PB_L_SN, PB_Dens, SB_DENS_SN Spielmeyer et al. 2002 SbGA20ox Sb09g020760 9 50,182 Gibberellin 20- oxidase (OsGA20ox) AAM56041 9,70E- 127 Spielmeyer et al. 2002 SbGA20ox Sb01g000580 1 0,462 Gibberellin 20- oxidase (OsGA20ox) AAM56041 2,10E-95 Spielmeyer et al. 2002 NB: sorghum candidate genes that show meaningful co-localization with panicle architecture QTL are shown in bold University of Ghana http://ugspace.ug.edu.gh 108 5.3.4. Phylogenetic analysis A phylogenetic analysis of the candidate genes indicated the presence of 13 distinct clades within the sorghum genome for meristem regulation during inflorescence development (Fig 5.1). Twelve clades for specification of spikelet meristem identity and five separate clades for plant growth and development with one sorghum predicted gene per clade except the clade TFL1/CEN- like where two paralogous predicted gene copies were observed and the clade Gibberellin 20 oxidase presented three paralogous sorghum genes prediction. In many clades, sorghum predicted genes share more homology with maize than rice. University of Ghana http://ugspace.ug.edu.gh 109 Figure 5.1: Phyllogenetic tree for candidate genes involved in inflorescence architecture University of Ghana http://ugspace.ug.edu.gh 110 5.4. Conclusion Protein sequence similarities with sorghum for a number of candidate genes in rice and maize involved in meristem regulation during inflorescence development, specification of spikelet meristem identity and plant growth were established. Several of these candidate genes co-localize closely with QTLs for relevant inflorescence architecture traits in a sorghum mapping population derived from the cross between Tiandougou (caudatum-guinea type) and Lata3 (guinea type). The results of this study suggest promising confirmation of the function of the candidate genes in rice and maize in relation to the QTLs detected for panicle architecture traits. BLAST may give indication of possible colinearity and homology of genes that can serve as a starting point for further investigation of the effectiveness of studying candidate genes in related species such as rice and maize to elucidate orthologous roles and function of these genes in sorghum and provide the basic understanding of sorghum grain yield improvement through panicle architecture. Candidate gene cloning studies in large genetic background of sorghum germplasm may contribute to a better understanding of the relationship between genes involved in sorghum panicle architecture and grain yield. The candidate genes identified in sorghum and localized with QTLs detected in this study for panicle architecture may help design breeding strategies based on genetic and physical architecture to improve sorghum productivity. University of Ghana http://ugspace.ug.edu.gh 111 CHAPTER SIX 6.0. GENERAL DISCUSSION 6.1. Phenotyping for sorghum panicle architecture and yield in Tiandougou/Lata-3 F4 population The relationship between panicle architecture and grain yield, the determination of the pattern of panicle branching and the number of grains at different positions such as the basal, medium and the top of the panicle were investigated in this study. The relationship between sorghum panicle architecture and grain yield was established through the identification of some highly heritable architectural and easy measurable traits. The number of internodes per panicle and the basal diameter of the panicle contribute to thirty two percent (32%) of grain yield per panicle. Branching and grain pattern at different positions on the panicle were characterized to capture panicle architecture variability. Gradual decrease was seen in the length of the primary branch, the number of secondary branches and number of grains from the base of the panicle toward the top of the panicle. Significant variation existed between the F4 families for panicle architecture traits in the second sowing date compared to the first sowing date in 2011. This is common in sorghum because of photoperiod differences and late sowing reduction of sorghum grain yield per panicle. Broad sense heritability estimated for panicle length, panicle weight, the number of primary branches per panicle, and the number of nodes per panicle corroborate most previous reports. Different degrees of heritability for number of secondary branches, number of grains per primary branches and rachis diameter were noted across studies. This difference may be explained by the dependence of some of these traits on others. For instance, number of secondary branches depends on the length of the primary branch. Broad sense heritability estimated in this study for grain yield University of Ghana http://ugspace.ug.edu.gh 112 per panicle corroborates some previous findings. Any differences may be explained by difference in the genetic background, the statistical method of estimation and interaction of genotype with environment. The existence of decreasing gradient for the length and number of primary branches from the base of the panicle toward the top of the panicle may refine previous findings by giving meaningful information for phenotyping strategies. Correlation between grain yield per panicle and a number of sorghum panicle architecture traits agrees with most previous findings. However, this study extends the correlation of grain yield per panicle to the pattern of branching and number of grains at different positions in the panicle base, median and top. Genetic variability in sorghum panicle architecture traits is important for the diversity in sorghum. Variability is expressed in terms of elongation, branching pattern and growth. The length of the panicle, rachis, primary and secondary branches at different positions in the panicle in relation to the number of primary and secondary branches per panicle at different positions plays an important role in capturing panicle architecture. Rachis diameter at different positions along the rachis influence the branching pattern, number of grains and consequently the grain yield per panicle. The contribution of panicle length, primary and secondary branches length and the number of primary branches per panicle to sorghum panicle variability agree with the previous findings. This study established other contributing traits such as the variation in length and number of primary branches at different positions in the panicle as well as the position of the longest primary branch. These traits characterized sorghum panicle architectural variability by identifying an evolutionary trend from the base toward the top of the panicle. Grain yield per panicle was associated with the number of primary branches per panicle and rachis base diameter explaining important sorghum panicle architecture variation. Transgressive segregation in sorghum University of Ghana http://ugspace.ug.edu.gh 113 panicle architecture traits and grain yield is common for quantitative traits in a bi parental population. In sorghum, sowing date affects panicle architecture traits and grain yield. Indeed, sorghum is photoperiod sensitive for floral initiation. This physiological mechanism makes sorghum one of the most adapted crops to climate change. Variability in sorghum panicle architecture is expressed according to the sowing date that determines the vegetative cycle of the genotype. The longer the vegetative cycle, the more the dry matter which is accumulated and which is further mobilized for grain formation, thus affecting grain yield. Delay in sowing date influences both vegetative and reproductive growth. Sorghum panicle architecture traits were influenced by the late sowing date in term of elongation, branching pattern and growth. Consequently the potential yield is negatively affected. This implies an optimum sowing date is necessary for a genotype to be able to express its full yielding potential. In Mali, at Bamako area the optimum sowing date should be in June. Branching pattern indicated decreasing heritability of branch length and number from the base to the top of the panicle. This implies that variability within the panicle and the pattern of this variation may be under different genetic control, therefore, the phenotyping of sorghum panicle architecture should focus on the basal zone of the panicle followed by the middle part of the panicle. Sorghum panicle architecture traits such as rachis base diameter and number of internodes per panicle contributed more (31.8%) to the grain yield per panicle than other traits. They were followed in importance by the number of primary branches per panicle, number of secondary branches on primary branches at second node. This demonstrated the contribution of sorghum panicle architecture to grain yield, and suggests that grain yield per panicle can be estimated based on architecture traits. Sorghum panicle architecture may be used in breeding programs to improve the grain yield. The ideotype genotype to be targeted in breeding should have big basal rachis diameter with many primary branches. University of Ghana http://ugspace.ug.edu.gh 114 The results of this study open up ground for further research on sorghum panicle architecture such as phenotyping to capture variability mainly at the basal part of the panicle. Designing breeding strategies using contrasting parents for rachis basal diameter and number of primary branches per panicle to elucidate the contribution of these architectural traits in grain yield through indirect selection should be successful. 6.2. QTLs analysis of panicle architecture in the Tiandougou/Lata-3 breeding population The genetic mechanisms underlying branching patterns were also studied. Sorghum panicle architecture was dissected into individual QTL which clearly establish causal relationships between grain yield and panicle architecture traits by identifying the genomic region and deciphering the control of sorghum panicle branching patterns. Both additive and non-additive gene effects control most sorghum panicle architecture traits. Additive gene action was determinant in the control of the number of internodes per panicle, primary branches zone and maximum primary branches per node. These traits will be easy to fix in a breeding program at early generations. Panicle traits such as the length and number of secondary branches were governed by non-additive gene effect. These findings extend the genetic mechanism controlling sorghum panicle traits. Grain yield per panicle, panicle weight, and thousand grain weight were under non additive gene action. This results agree with previous findings. Number of grains per panicle was predominantly governed by additive gene effect, however, the number of grains per primary branch at different positions in the panicle was controlled by both additive and non-additive gene action. It is possible to breed for traits controlled by both additive and non-additive gene effects, however, the selection will be delayed to later generations. University of Ghana http://ugspace.ug.edu.gh 115 The position of the QTL for primary branch number detected on SBI-03; rachis base diameter on SBI-06; panicle weight on SBI-06; panicle length on SBI-03; and grain yield per panicle on SBI- 06 corroborate previous findings (Brown et al., 2006; Srinivas et al., 2009). However, the position in sorghum genome of the QTLs for primary branch length, thousand seed weight, Number of grains per panicle, panicle length conflict with previous findings. This discrepancy may be explained by the different genetic backgrounds and the architectural variability in cultivated sorghum. This study extends the previous findings for the novel QTL for panicle length on SBI- 04. The flanking markers both left and right SB06074 and SB06098 reveal that secondary branch length and primary branch density from the parent Lata-3 were linked to grain yield per panicle. While internode length, the longest secondary branch and number of secondary branches at penultimate node were associated to grain yield per panicle through the parent Tiandougou with the flanking markers SB03149 and SB03164. This marker information can be used to design breeding strategies targeting grain yield improvement through the recombination of these traits based on the relatedness of F4 families to each parent to be selected for further crossing. The QTL for grain yield per panicle on SBI-06 reveals pleiotropic effect on panicle weight, rachis diameter and average primary branch length. This result corroborates and extends the findings of Nagaraja Reddy et al. (2013). Therefore, sorghum improvement for grain yield should focus on these architectural traits such as secondary branch density at the base of the panicle, rachis base diameter and the number and length of primary branches per panicle. These architectural traits should be considered in the breeding program for further investigations to elucidate their impact on grain yield using molecular breeding approach. One parent to be used should have a big rachis base diameter with long primary branches and the other parent should have many primary branches. QTL region for grain yield on SBI-06 can be targeted with the QTL region for other favorable traits such as the number of primary branches per panicle to build an ideal genotype. In this study, digenic University of Ghana http://ugspace.ug.edu.gh 116 epistasis controlled the pattern of the grain number on primary branches at different positions in the panicle and the length of the primary and secondary branches as well as internode length. This finding extends the role of epistasis in sorghum particularly for panicle architecture traits. QTLs with overlapping confidence intervals for primary branch length, panicle length and panicle weight corroborate previous findings and may constitute potential Meta-QTLs. Therefore, Meta-analysis studies to determine the position of the Meta- QTLs and project them onto the consensus genetic map should be done. 6.3. Identification of candidate genes involved in sorghum panicle architecture in maize and rice Candidate genes controlling sorghum panicle architecture traits were addressed in this study. Sequence similarity of candidate genes involved in inflorescence architecture in rice and maize with predicted genes in sorghum were evaluated. Co-localization between sorghum predicted genes and QTLs detected for panicle architecture traits was investigated to characterize the candidate genes and infer their role in sorghum based on known functions in rice or maize. Promising co- localization based on sequence similarity between QTLs for sorghum panicle architecture and candidate genes involved in inflorescence architecture in rice or/and maize suggested that: Sorghum homologue of ABERRANT SPIKELET AND PANICULE1 genes (SbASP1) on SBI-07 at the position 5.31 cM may regulate branching pattern during panicle development and elongation. LAX PANICLE1 (LAX1) homologue in sorghum (SbLAX1) on SBI-03 at the position 66.62 cM is suggested to regulate primary branch number and rachis size. Sorghum GRAIN NUMBER 1a (SbGN1a) on SBI-03 at the position 2.72 cM may regulates grain number on primary branches at different positions in the panicle (second, median and penultimate nodes). INDETERMINATE SPIKELET1 (OsIDS1; ZmIDS1) candidate genes in rice and maize homologues in sorghum University of Ghana http://ugspace.ug.edu.gh 117 (SbIDS1) on SBI-01 at the position 2.73 cM may regulate grain size in relation to grain number. FLORAL ORGAN REGULATOR1 homologue in sorghum (SbFOR1) on SBI-02 at the position 71.13 cM may influence the general structure of the panicle such as the shape and compactness. GIBBERELLIN 20-OXIDASE (SbGA20ox) in sorghum on SBI-03 at the position 69.37 cM may play an important role in panicle elongation and branch development. Few studies were done on candidate genes controlling sorghum panicle architecture therefore, these findings extend the information that had been obtained earlier by other workers. There is a need to undertake further research on these candidate genes to elucidate and establish their functions in sorghum. University of Ghana http://ugspace.ug.edu.gh 118 GENERAL CONCLUSIONS AND RECOMMENDATIONS Sorghum panicle architecture traits show tremendous variation in branching, elongation and growth. Most of sorghum panicle architecture traits were heritable in the F4 families derived from Tiandougou/Lata-3. The basal parts of the panicle show high heritability compared to the top parts. Rachis base diameter and number of internodes per panicle were highly significant and positively correlated to grain yield per panicle. Indirect selection for these traits is feasible to improve grain yield. Panicle length, primary and secondary branches length at second, number of primary branches per panicle and position of the longest primary branch showed positive contributions to sorghum panicle architecture variability. Both additive and non-additive gene effects control most of sorghum panicle architecture traits. Grain yield per panicle was governed by non-additive gene action. Number of grains per panicle was under additive gene effect. However, at different positions, second, median and penultimate nodes, it was controlled by both additive and non- additive gene effects. The SBI-02, SBI-03 and SBI-06 were hot spots of major QTLs for panicle architecture traits and grain yield. The flanking markers (SB03149 - SB03164; SB06074 - SB06098) of grain yield per panicle QTL respectively on SBI-03 and SBI-06 were associated with architectural QTLs for secondary branch length, primary branch density, internode length and the longest secondary branch. Pleiotropic QTL affecting grain yield per panicle influences primary branches length and rachis diameter. This arrangement may be correlated to heterosis. Therefore, rachis base diameter and number and length of primary branches can be used in both conventional and molecular breeding programs to improve sorghum grain yield. Digenic epistasis determined the pattern of the grain number at different positions in the panicle and the length of the secondary branches. QTLs for panicle length, primary branch length and panicle weight were stable across different studies. University of Ghana http://ugspace.ug.edu.gh 119 Sorghum homologues of ABERRANT SPIKELET AND PANICULE1 genes (SbASP1) on SBI-07; LAX PANICLE1 (LAX1) gene (SbLAX1) on SBI-03; GRAIN NUMBER 1a (SbGN1a) gene on SBI- 03; INDETERMINATE SPIKELET1 gene (SbIDS1) on SBI-01; FLORAL ORGAN REGULATOR1 gene (SbFOR1) on SBI-02; GIBBERELLIN 20-OXIDASE (SbGA20ox) on SBI-03 could play important role in the control of sorghum panicle architecture traits. However, further study should confirm and validate the function of these candidate genes in sorghum. RECOMMENDATIONS Development of varieties in sorghum, based on panicle architecture traits should be addressed by other workers interested in sorghum cultivar development. Indirect selection for panicle architecture traits correlated to grain yield in conventional breeding should be initiated. At the molecular level, genomic regions related to grain yield and those contributing architectural traits should be combined in the ideal genotype by using the MBDT software to build ideal genotypes and OPTIMAS software should be used to determine different parents among the F4 families to be cross to reach this ideal genotype. Information on QTLs for panicle architecture traits can be used in the first and second cycles of recombination in MARS breeding approach for sorghum grain yield improvement. It is valuable to address further studies to determine the relevance of panicle architecture traits in biotic and abiotic stresses such as diseases affecting sorghum panicle (smut, ergot) and drought for better food security in Mali. Also, it is recommended to undertake further studies such as loss of function, in situ hybridization and Northern blot analysis for expression patterns to decipher the role of candidate genes involved in panicle architecture identified in sorghum and use such information to improve sorghum breeding. University of Ghana http://ugspace.ug.edu.gh 120 REFERENCES Addissu, G. A. (2011). Heritability and genetic advance in recombinant inbred lines for drought tolerance and other related traits in sorghum (Sorghum bicolor). Continental Journal of Agricultural Science, 5 (1), 1 - 9 Ali, H. I., Mahmoud, K. M., & Amir, A. A. (2012). Estimation of Genetic Variability, Heritability and Genetic Advance in Grain Sorghum Population. American-Eurasian Journal of Agriculture & Environmental Science, 12 (4), 414 - 422 Anisimova, M., & Gascuel, O. (2006). Approximate likelihood ratio test for branchs: A fast, accurate and powerful alternative. Systematic Biology, 55(4), 539-52. Ashikari, M., Sakakibara, H., Lin, S.Y., Yamamoto, T., Takashi, T., Nishimura, A., Angeles, E. R., Qian, Q., Kitano, H., & Matsuoka, M. (2005). Cytokinin Oxidase Regulates Rice Grain Production. Science, 309, 741-745. Ayala, A. G. (2011). Heritability and genetic advance in recombinant inbred lines for drought tolerance and other related traits in sorghum (Sorghum bicolor). Continental Journal of Agricultural Science, 5 (1), 1-9. Barazesh, S., & McSteen, P. (2008). Hormonal control of grass inflorescence development. Trends in Plant Science, 13, 656-662. Barthélémy, D., & Caraglio, Y. (2007). Plant Architecture: A Dynamic, Multilevel and Comprehensive Approach to Plant Form, Structure and Ontogeny. Annals of Botany, 99, 375- 407. Basten, C. J., Weir, B. S., & Zeng, Z. B. (1994). QTL Cartographer. Dep. of Statistics, North Carolina State Univ., Raleigh, NC. Bello, D., Kadams, A. M., Simon, S. Y., & Mashi, D. S. (2007). Studies on genetic variability in cultivated sorghum (Sorghum bicolor (L.) Moench) cultivars of Adamawa State Nigeria. American-Eurasian Journal of Agricultural & Environmental Sciences, 2 (3), 297-302. Benjamins, R., & Scheres, B. (2008). Auxin: the looping star in plant development. Annual Review of Plant Biology, 59, 443-465. Beveridge, C. A. (2006). Axillary bud outgrowth: sending a message. Current Opinion in Plant Biology, 9, 35-40 Bomblies, K., Wang, R. L., Ambrose, B. A., Schmidt, R. J., Meeley, R. B., & Doebley, J. (2003). Duplicate FLORICAULA/LEAFY homologs zfl1 and zfl2 control inflorescence architecture and flower patterning in maize. Development, 130, 2385-2395. University of Ghana http://ugspace.ug.edu.gh 121 Bommert, P., Satoh-Nagasawa, N., Jackson, D., & Hirano, H. Y. (2005). Genetics and evolution of inflorescence and flower development in grasses. Plant Cell Physiology, 46, 69-78. Bortiri, E., Chuck, G., Vollbrecht, E., Rocheford, T., Martienssen, R., & Hake, S. (2006). Ramosa2 encodes a LATERAL ORGAN BOUNDARY domain protein that determines the fate of stem cells in branch meristems of maize. Plant Cell, 18, 574–585. Broman, K. W., & Sen, S. (2009). A Guide to QTL Mapping with R/qtl, Statistics for Biology and Health, DOI 10.1007/978-0-387-92125-9 1. Brown, P. J., Klein, P. E., Bortiri, E., Acharya, C. B., Rooney, W. L., & Kresovich, S. (2006). Inheritance of inflorescence architecture in sorghum. Theoretical and Applied Genetics, 113, 931-942. Bucheyeki, T. L., Gwanama, C., Mgonja, M., Chisi, M., Folkertsma, R., & Mutegi, R. (2009). Genetic variability characterisation of Tanzania sorghum landraces based on simple sequence repeats (SSRs) molecular and morphological markers. African Crop Science Journal, 17, 71- 86. Buckler, E. (2007). TASSEL: Trait Analysis by association, Evolution, and Linkage. User manual. http://www.maizegenetics. net/tassel; verified 10 July 2008. Caraglio, Y., & Barthélémy, D. (1997). Revue critique des termes relatifs à la croissance et à la ramification des tiges des végétaux vasculaires. In : Modélisation et simulation de l’architecture des végétaux (D. R. P. Bouchon Jean, Barthélémy Daniel, ed.), pp. 11-87, INRA, Paris Chantereau, J., Trouche, G., Rami, J. F., Deu, M., Barro, C., & Grivet, L. (2004). RFLP mapping of QTLs for photoperiod response in tropical sorghum. Euphytica, 120, 183-194. Chavan, S. K., Mahajan, R. C., & Fatak, S. U. (2010). Genetic variability studies in sorghum. Karnataka Journal of Agricultural Sciences, 23 (2), 322-323. Chevenet, F., Brun, C., Banuls, A. L., Jacq, B., & Chisten, R. (2006). TreeDyn: towards dynamic graphics and annotations for analyses of trees. BMC Bioinformatics, Oct 10, 7, 439. Chittenden, L. M., Schertz, K. F., Lin, Y. R., Wing, R. A., & Paterson, A. H. (1994). A detailed RFLP map of Sorghum bicolor x S. Propinquum, suitable for high-density mapping, suggests ancestral duplication of Sorghum chromosomes or chromosomal segments. Theoretical and Applied Genetics, 87, 925-933. Chuck, G., Meeley, R. B., & Hake, S. (1998). The control of maize spikelet meristem fate by the APETALA2-like gene indeterminate spikelet1. Genes & Development, 12, 1145-1154. Chuck, G., Muszynski, M., Kellogg, E., Hake, S., & Schmidt, R. J. (2002). The control of spikelet meristem identity by the branched silkless1 gene in maize. Science, 298, 1238-1241. University of Ghana http://ugspace.ug.edu.gh 122 Chuck, G., Meeley, R., & Hake, S. (2008). Floral meristem initiation and meristem cell fate are regulated by the maize AP2 genes ids1 and sid1. Development, 135, 3013-3019. Churchill, G. A., & Doerge, R. W. (1994). Empirical threshold values for quantitative trait mapping. Genetics, 138, 963-971. Clouse, S. D. (1996). Molecular genetic studies confirm the role of brassinosteroids in plant growth and development. Plant Journal, 10, 1-8. Clouse, S. D., & Sasse, J. M. (1998). BRASSINOSTEROIDS: essential regulators of plant growth and development. Annual Review of Plant Physiology and Plant Molecular Biology, 49, 427- 451. Collard, B. C. Y., Jahufer, M. Z. Z., Brouwer, J. B., & Pang, E. C. K. (2005). An introduction to markers, quantitative trait loci (QTL) mapping and markers-assisted selection for crop improvement: The basic concepts. Euphytica, 142, 169-196. Colombo, L., Marziani, G., Masiero, S., Wittich, P. E., Schmidt, R. J., Gorla, M. S. & Pe, M. E. (1998). BRANCHED SILKLESS mediates the transition from spikelet to floral meristem during Zea mays ear development. Plant Journal, 16, 355-363. Crasta, O. R., Xu, W.W., Rosenow, D. T. Mullet, J., & Nguyen, H.T. (1999). Mapping of post- flowering drought resistance traits in grain sorghum: Association between QTLs influencing premature senescence and maturity. Molecular and General Genetics, 262, 579-588. Davies, P. J. (1995). Plant Hormones: Physiology, Biochemistry and Molecular Biology. Dordrecht, Netherlands: Kluwer Academic Publishers. De Wet, J. M. J., & Harlan, J. R. (1971). The origin and domestication of Sorghum bicolor (L) Moench. Economic Botany, 25, 128-135. De Wet, J. M. J. (1978). Systematics and evolution of Sorghum sect. Sorghum (Gramineae). American Journal of Botany, 65, 477-484. Doggett, H. (1988). Sorghum, 2nd edn. New York: John Wiley and Sons. Doust, A. N., & Kellogg, E. A. (2002). Inflorescence diversification in the panicoid “bristle grass” clade (Paniceae, Poaceae): evidence from molecular phylogenies and developmental morphology. American Journal of Botany, 89, 1203-22. Dudley, J.W. (1993). Molecular markers in plant improvement: Manipulation of genes affecting quantitative traits. Crop Science, 33, 660-668. Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research, Mar 19, 32(5), 1792-7. (PubMed). University of Ghana http://ugspace.ug.edu.gh 123 Egli, D. B. (2006). The role of seed in the determination of yield of grain crops. Australian Journal of Agricultural Research, 57, 1237-1247. El-Din, A. A. T., Hessein Eatemad, M., & Ali, E. A. (2012). Path Coefficient and Correlation Assessment of Yield and Yield Associated Traits in Sorghum (Sorghum bicolor L.) Genotypes. American Eurasian Journal of Agricultural & Environmental, 12 (6), 815-819. Ezeaku, I. E., & Mohammed, S. G. (2006). Character association and path analysis in grain Sorghum. African Journal of Biotechnology, 5 (14), 1337-1340. Falconer, D. S., & Mackay, T. F. C. (Eds) (1996). Introduction to quantitative genetics (4th edition). Harlow: Longman. Fan, C. C., Xing, Y. Z., Mao, H. L., Lu, T. T., Han, B., Xu, C. G., Li, X. H., & Zhang, Q. F. (2006). GS3, a major QTL for grain length and weight and minor QTL for grain width and thickness in rice, encodes a putative transmembrane protein. Theoretical and Applied Genetics, 112(6), 1164-1171. Feltus, F. A., Hart, G. E., Schertz, K. F., Casa, A. M., Kresovich, S., Abraham, S., Klein, P. E., Brown, P. J., & Paterson, A. H. (2006). Alignment of genetic maps and QTLs between inter- and intraspecific sorghum populations. Theoretical and Applied Genetics, 112, 1295-1305. Foucher, F., Morin, J., Courtiade, J., Cadioux, S., Ellis, N., Banfield, M. J., & Rameau, C. (2003). DETERMINATE and LATE FLOWERING are two TERMINAL FLOWER1/CENTRORADIALIS homologs that control two distinct phases of flowering initiation and development in pea, Plant Cell, 15, 2742-2754. Fujita, D., Tagle, A. G., Ebron, L. A., Fukuta, Y., & Kobayashi, N. (2012). Characterization of near-isogenic lines carrying QTL for high spikelet number with the genetic background of an indica rice variety IR64 (Oryza sativa L.) Breeding Science, 62, 18-26. Gallavotti, A., Malcomber, S., Gaines, C., Stanfied, S., Whipple, C., Kellogg, E., & Schmidt, R. J. (2011). BARREN STALK FASTIGIATE1 is an AT-hook protein required for the formation of maize ears. Plant Cell, 23 (5), 1756-71. Gao, X. C., Liang, W., Yin, C., Ji, S., Wang, H., Su, X., Guo, C., Kong, H., Xue, H., & Zhang, D. (2010). The SEPALLATA-like gene OsMADS34 is required for rice inflorescence and spikelet development. Plant Physiology, 153, 728-740. Gloria, B. B., Cleve, D. F., Veronica, A. M., & Zhanguo, X. (2008). Molecular mapping and characterization of BLMC, a locus for profuse wax (bloom) and enhanced cuticular features of Sorghum (Sorghum bicolor (L.) Moench.). Theoretical and Applied Genetics, 118(3), 423-431 University of Ghana http://ugspace.ug.edu.gh 124 Godbharle, A. R., More, A. W., & Ambekar, S. S. (2010). Genetic Variability and Correlation Studies in elite ‘B’ and ‘R’ lines in Kharif Sorghum. Electronic Journal of Plant Breeding, 1(4), 989-993. Greco, R, Stagi, L., Colombo, L., Angenent, G. C., Sari-Gorla, M., & Pè, M. E. (1997). MADS box genes expressed in developing inflorescences of rice and sorghum. Molecular and General Genetics, 253, 615-623. Grenieret, C., Hamon, P., & Bramel Cox, P. J. (2001). Core Collection of Sorghum: II. Comparison of Three Random Sampling Strategies. Crop Science, 41(1), 241-246. Guindon, S., & Gascuel, O. A. (2003). Simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology, 52(5), 696-704. Guo, Y., & Hong, D. (2010). Novel pleiotropic loci controlling panicle architecture across environments in japonica rice (Oryza sativa L.). Journal of Genetics and Genomics, 37, 533- 544. Hair Jr, J. F., Anderson, R. E., Tatham, R. L., & Black, W.C. (Eds). (1998). Multivariate Data Analysis. London: Prentice-Hall international Haley, C. S., & Knott, S. A. (1992). A simple regression method for mapping quantitative traits loci in line crosses using flanking markers. Heredity, 69, 315-324 Halle, F., Oldeman, R. A. A., & Tomlinson, P. B. (Eds) (1978). Tropical trees and forests: an architectural analysis. Berlin: Springer-Verlag. Harlan, J. R., & de Wet, J. M. J. (1972). A simplified classification of cultivated sorghum. Crop Science, 12, 172-176. Harris, K., Subudhi, P. K., Borrell, A., Jordan, D., Rosenow, D., Nguyen, H., Klein, P., Klein, R., & Mullet, J. (2007). Sorghum stay-green QTL individually reduce post-flowering drought- induced leaf senescence. Journal of Experimental Botany, 58, 327-338. Hart, G. E., Schertz, K. F., Peng, Y., & Syed, N. Y. (2001). Genetic mapping of Sorghum bicolor (L.) Moench QTLs that control variation in tillering and other morphological characters. Theoretical and Applied Genetics, 103, 1232-1242. Hausmann, B. I. G., Mahalakshini, V., Reddy, B. V. S., Seetharama, N., Hash, C. T., & Geiger, H. H. (2002). QTL mapping of stay-green in two sorghum recombinant inbred populations. Theoretical and Applied Genetics, 106, 133-142. Heupel, R. C., Nes, W. D., & Verbeke, J. A. (1987). Developmental regulation of sterol and pentacyclic triterpene biosynthesis and composition: a correlation with sorghum floral University of Ghana http://ugspace.ug.edu.gh 125 initiation. In P. K. STUMPF., J. B. MUDD., & W. D. NES (Eds). The metabolism5 structure, and function of plant lipids. (pp. 53-56). New York: Plenum. Hittalmani, S., Shashidhar, H. E., Bagali, P. G., Ning Huang, Sidhu, J. S., Singh, V. P., & Khush, G. S. (2002). Molecular mapping of quantitative traits loci for plant growth, yield and yield related traits across three diverse locations in a doubled haploid rice population. Euphytica, 125, 207-214. Holland, J. B. (2007). Genetic architecture of complex traits in plants. Current Opinion in Plant Biology, 10, 156-161. Hooley, R. (1994). Gibberellins: perception, transduction and responses. Plant Molecular Biology, 26, 1529–55. House, L. R. (1985). A guide to sorghum breeding. 2nd ed. ICRISAT. Patancheru, India. Huang, X. Z., Qian, Q., Liu, Z. B., Sun, H. Y., He, S. Y., Luo, D., Xia, G. M., Chu, C. C., Li, J. Y., & Fu, X. D. (2009). Natural variation at the DEP1 locus enhances grain yield in rice. Nature genetics, 41(4), 494-497. Hulbert, S. H., Richter, T. E., Axtell, J. D., & Bennetzen, J. L. (1990). Genetic mapping and characterization of sorghum and related crops by means of maize DNA probes. Proceedings of the National Academy of Sciences USA, 87, 4251-4255. Ibrokhim, Y., Abdurakhmonov & Abdusattor, A. (2008). Application of Association Mapping to Understanding the Genetic Diversity of Plant Germplasm Resources. International Journal of Plant Genomics, Article ID 574927, 18 pages. IBPGR & ICRISAT. (1993). Descriptors for sorghum, [Sorghum bicolor (L.)Moench]. Rome, Italie; Patancheru, India: IBPGR; ICRISAT. Ikeda, K., Ito, M., Nagasawa, N., Kyozuka, J. & Nagato, Y. (2007). Rice ABERRANT PANICLE ORGANIZATION 1, encoding an F-box protein, regulates meristem fate. Plant Journal, 51, 1030-1040. Ikeda-Kawakatsu, K., Yasuno, N., Oikawa, T., Iida, S., Nagato, Y., & Maekawa, M. (2009). Expression level of ABERRANT PANICLE ORGANIZATION1 determines rice inflorescence form through control of cell proliferation in the meristem. Plant Physiology, 150, 736-747. Irish, E. E. (1997a). Class II tassel seed mutations provide evidence for multiple types of inflorescence meristems in maize (Poaceae). American Journal of Botany, 84, 1502-1515. Irish, E. E. (1997b). Experimental analysis of tassel development in the maize mutant Tassel seed 6. Plant Physiology, 114, 817-825. University of Ghana http://ugspace.ug.edu.gh 126 Irish, E. E. (1998). Grass spikelets: A thorny problem. BioEssays, 20, 789-793. Jabbour, F., & Citerne, H. L. (2010). Modeling plant and inflorescence. International Journal of Plant Developmental Biology, 4, 38-46 Jang, S., Lee, B., Kim, C., Kim, S. J., Yim, J., Han, J. J…..An, G. (2003). The OsFOR1 gene encodes a polygalacturonase-inhibiting protein (PGIP) that regulates floral organ number in rice. Plant Molecular Biology, 53, 357-369. Jeon, J.S., Jang, S., Lee, S., Nam, J., Kim, C., Lee, S.H……An, G. (2000). Leafy hull sterile1 is a homeotic mutation in a rice MADS box gene affecting rice flower development. Plant Cell, 12, 871-884. Jiao, Y., Wang, Y., Xue, D., Wang, J., Yan, M., Liu, G., Dong, G., Zeng, D., Lu, Z., Zhu, X., Qian, Q., & Li, J. (2010). Regulation of OsSPL14 by OsmiR156 defines ideal plant architecture in rice. Nature Genetics, 42(6), 541-544. Johnson, D. E. (1998). Applied Multivariate Methods for Data Analysts. Duxbury Press, New York. Jordan, D. R., Mace, E. S., Henzell, R. G., Klein, P. E., Klein, R. R. (2010). Molecular mapping and candidate gene identification of the Rf2 gene for pollen fertility restoration in sorghum (Sorghum bicolor (L.) Moench). Theoretical and Applied Genetics, 120, 1279-1287. Kamelmanesh, M. M., Namayandeh, A., Zadehbagheri, M., & Choukan, R. (2012). Analysis of Quantitative Trait Loci of some agronomic traits in maize. International Journal of Agriculture and Crop Sciences, 4 (9), 550-555. Kang, H. G., Noh, Y. S., Chung, Y. Y., Costa, M. A., An, K., & An, G. (1995). Phenotypic alterations of petal and sepal by ectopic expression of a rice MADS box gene in tobacco. Plant Molecular Biology, 29, 1-10. Kawakatsu, T., Itoh, J., Miyoshi, K., Kurata, N., Alvarez, N., Veit, B., & Nagato, Y. (2006). PLASTOCHRON2 regulates leaf initiation and maturation in rice. Plant Cell, 18, 612-625. Kawakatsu, T., Taramino, G., Itoh, J., Allen, J., Sato, Y., Hong, S. K……Nagato, Y. (2009). PLASTOCHRON3/GOLIATH encodes a glutamate carboxypeptidase required for proper development in rice. Plant Journal, 58, 1028-1040 Kebede, H., Subadhi, P. K., Rosenow, D. T., & Nguyen, H. T. (2001). Quantitative trait loci influencing drought tolerance in grain sorghum (Sorghum bicolor L. Moench). Theoretical and Applied Genetics, 103, 266-276. Kellogg, E. A. (2007). Floral displays: genetic control of grass inflorescences. Current Opinion in Plant Biology, 10, 26-31. University of Ghana http://ugspace.ug.edu.gh 127 Khan, M. D., Khalil, I. H., Khan, M. A., & Ikramullah. (2004). Genetic divergence and association for yield and related traits in mash bean. Sarhad Journal of Agriculture, 20, 555-561. Kim, J. S., Klein, P. E., Klein, R. R., Price, H. J., Mullet, J. E., & Stelly, D. M. (2005). Chromosome identification and nomenclature of Sorghum bicolor. Genetics, 169, 1169-1173. Klein, R. R., Rodriquez-Herrera, R., Schlueter, J. A., Klein, P. A., Yu, Z. H., & Rooney, W. L. (2001). Identification of genomic regions that affect grain-mould incidence and other traits of agronomic importance in sorghum. Theoretical and Applied Genetics, 102, 307-319. Knapp, S. J., Stroup, W. W., & Roos, M. W. (1985). Exact confidence intervals for heritability on a progeny mean basis. Crop Science, 25, 192-194. Kobayashi, M., Gaskin, P., Spray, C. R., Phinney, B. O., & Macmillan, J. (1994). The metabolism of gibberellin A20 to gibberellin A1 by tall and dwarf mutants of Oryza sativa and Arabidopsis thaliana, Plant Physiology, 106(4), 1367-1372. Kobayashi, K., Maekawa, M., Miyao, A., Hirochika, H., & Kyozuka, J. (2010). PANICLE PHYTOMER2 (PAP2), encoding a SEPALLATA subfamily MADS-box protein, positively controls spikelet meristem identity in rice. Plant Cell Physiology, 51, 47-57. Komatsu, K., Maekawa, M., Ujiie, S., Satake, Y., Furutani, I., Okamoto, H., Shimamoto, K. & Kyozuka, J. (2003a). LAX and SPA: major regulators of shoot branching in rice. Proceedings of the National Academy of Sciences of the United States of America, 100, 11765-11770. Komatsu, M., Chujo, A., Nagato, Y., Shimamoto, K., & Kyozuka, J. (2003b). FRIZZY PANICLE is required to prevent the formation of axillary meristems and to establish floral meristem identity in rice spikelets. Development, 130, 3841-3850. Kukadia, M. V., Desai, K. V., Desai, M. S., Patel, R. H., & Rajan, K. R. V. (1983). Estimates of heritability and other genetic parameters in sorghum. Sorghum Newsletter, 26, 31. Kyozuka, J. (2007). Control of shoot and root meristem function by cytokinin. Current Opinion in Plant Biology, 10, 442-446 Lane, H. C. (1963). Effect of light quality on maturity in the milo group of sorghum. Crop Science, 3, 496-499. Large, E. C. (1954). Growth stages in cereals: illustration of the Feekes scale. Plant Pathology, 3, 128-129. Lee, D. Y., Lee, J., Moon, S., Park, S. Y., & An, G. (2007). The rice heterochronic gene SUPERNUMERARY BRACT regulates the transition from spikelet meristem to floral meristem. Plant Journal, 49, 64-78. University of Ghana http://ugspace.ug.edu.gh 128 Lee, D. Y., & An, G. (2012). Two AP2 family genes, SUPERNUMERARY BRACT (SNB) and OsINDETERMINATE SPIKELET 1 (OsIDS1), synergistically control inflorescence architecture and floral meristem establishment in rice. Plant Journal, 69, 445-461. Li, S., Qian, Q., Fu, Z., Zeng, D., Meng, X., Kyozuka, J., Masahiko Maekawa, M., Zhu, X., Zhang, J., Li, J., & Wang, Y. (2009). Short panicle1 encodes a putative PTR family transporter and determines rice panicle size. Plant Journal, 58, 592-605. Li, X., Qian, Q., Fu, Z., Wang, Y., Xiong, G., & Zeng, D. (2003). Control of tillering in rice. Nature, 422, 618-621. Li, Z. K., Pinson, S. R. M., Park, W. D., Paterson, A. H., & Stansel, J. W. (1997). Epistasis for three grain yield components in rice Oryza sativa L. Genetics, 145, 453-465. Liang, G. H., & Gao, Z. S. (2001). Phylogenetic analysis and transformation of sorghum. Recent Research, Development of Plant Biology, 1, 17-33. Liao, C. Y., Wu, P., Hu, B., & Yi, K. K. (2001). Effects of genetic background and environment on QTLs and epistasis for rice (Oryza sativa L.) panicle number. Theoretical and Applied Genetics, 103, 104-111. Lin, H. X., Qian, H. R., Zhuang, J. Y., Lu, J., Min, S. K., Xiong, Z. M., Huang, N., & Zheng, K. L. (1996). RFLP mapping of QTLs for yield and related characters in rice (Oryza sativa L.). Theoretical and Applied Genetics, 92, 920-927. Lin, Y. R., Schertz, K. F., & Paterson, A. H. (1995). Comparative analysis of QTLs affecting plant height and maturity across Poaceae, in reference to an interspecific sorghum population. Genetics, 141, 391-411. Lincoln, S. E., Daly, M. J., & Lander, E. S. (1993). Mapping genes controlling quantitative traits using MAPMAKET/QTL version 1.1: A tutorial and reference manual. Whitehead Inst., Cambridge, MA. Lu, C., Shen, L., Tan, Z., Xu, Y., He, P., & Chen, Y. (1996). Comparative mapping of QTLs for agronomic traits of rice across environments using a doubled haploid population. Theoretical and Applied Genetics, 93, 1211-1217. Mace, E. S., Rami, J. F., Bouchet, S., Klein, P. E., Klein, R. R., Kilian, A., Wenz, P., Xia, L., Halloran, K., & Jordan, D. R. (2009). A consensus genetic map of sorghum that integrates multiple component maps and high-throughput Diversity Array Technology (DArT) markers. BMC Plant Biology, 9, 13. Mace, E. S., & Jordan, D. R. (2011). Integrating sorghum whole genome sequence information with a compendium of sorghum QTL studies reveals uneven distribution of QTL and of gene-rich University of Ghana http://ugspace.ug.edu.gh 129 regions with significant implications for crop improvement. Theoretical and Applied Genetics, 123, 169-191. Magalhaes, J. V., Garvin, D. F., Wang, Y. H., Sorrells, M. E., Klein, P. E., Schaffert, R. E., Li, L., Kochian, L.V. (2004). Comparative mapping of a major aluminum tolerance gene in sorghum and other species in the Poaceae. Genetics, 167, 1905-1914. Mahajan, R. C., Wadikar, P. B., Pole, S. P., & Dhuppe, M. V. (2011).Variability, Correlation and Path Analysis Studies in Sorghum. Research Journal of Agricultural Sciences, 2(1), 101-103. Maradiaga, J. L. M. (2003). Quantitative trait loci affecting the agronomic performance of a Sorghum bicolor (L.) Moench recombinant inbred restorer line population. (Published doctoral thesis). A&M University of Texas. Mohan, S. M., Madhusudhana, R., Mathur, K., Howarth, C. J., Srinivas, G., Satish, K., Reddy, R. N., & Seetharama, N. (2009). Co-localization of quantitative trait loci for foliar disease resistance in sorghum. Plant Breeding 128, 532–535. Mayor, M. L. (2008). Genetic analysis of ear development and tassel architecture in maize (Zea mays L. ssp. mays). (Published doctoral thesis). State University of Iowa, Ames McSteen, P., & Leyser, O. (2005). Shoot branching. Annual Review of Plant Biology 56: 353–374 McSteen, P., Malcomber, S., Skirpan, A., Lunde, C., Wu, X. T., Kellogg, E., & Hake, S. (2007). barren inflorescence2 encodes a co-ortholog of the PINOID serine/threonine kinase and is required for organogenesis during inflorescence and vegetative development in maize. Plant Physiology, 144, 1000-1011. Mena, M., Mandel, M. A., Lerner, D. R., Yanofsky, M. F., & Schmidt, R. J. (1995). A characterization of the MADS-box gene family in maize. Plant Journal, 8, 845-854. Mendez-Vigo, B., Martınez-Zapater, J. M., & Alonso-Blanco, C. (2013). The Flowering Repressor SVP Underlies a Novel Arabidopsis thaliana QTL Interacting with the Genetic Background. PLOS Genetics, 9(1), e1003289. Menz, M. A., Klein, R. R., Mullet, J. E., Obert, J. A., Unruh, N. C., & Klein, P. E. (2002). A high density map of Sorghum bicolor (L.) Moench based on 2926 AFLP, RFLP and SSR markers. Plant Molecular Biology, 48, 483-499. Miyoshi, K., Ahn, B. O., Kawakatsu, T., Ito, Y., Itoh, J., Nagato, Y., & Kurata, N. (2004). PLASTOCHRON1, a timekeeper of leaf initiation in rice, encodes cytochrome P450. Proceedings of the National Academy of Sciences of the United States of America, 101, 875880. Morgan, P. W., Miller, F. R., & Quinby, J. R. (1977). Manipulation of sorghum growth and development with gibberellic acid. Agronomy Journal, 69, 789-793. University of Ghana http://ugspace.ug.edu.gh 130 Morgan P. W. & Quinby J. R. (1987) Genetic regulation of development in Sorghum bicolor. IV. GA3 hastens floral differentiation but not floral development under nonfavorable photoperiods. Plant Physiology, 85, 615-620. Morris, G. P., Ramub, P., Deshpande, S. P., Thomas Hashc, C., Shahb, T., Hari, D….. & Kresovicha, S. (2013). Population genomic and genome-wide association studies of agroclimatic traits in sorghum. Proceedings of the National Academy of Sciences, 110(2), 453- 458. Multani, D. S., Briggs, S.P., Chamberlin, M. A., Blakeslee, J. J., Murphy, A. S., Johal, G.S. (2003). Loss of an MDR transporter in compact stalks of maize br2 and sorghum dw3 mutants. Science 302, 81-84. Murphy, R. L., Klein, R. R., Morshige, D. T., Brady, J. A., Rooney, W. L., Miller, F. R., Dugas, D. V., Klein, P. E., Mullet, J. E. (2011). Coincident light and clock regulation of pseudoresponse regulator protein 37 (PRR37) controls photoperiodic flowering in sorghum. Proceedings of the National Academy of Sciences, USA 108, 16469-16474. Murray, S. C., Sharma, A., Rooney, W. L., Klein, P. E., Mullet, J. E., Mitchell, S. E., & Kresovich, S. (2008). Genetic improvement of sorghum as a biofuel feedstock: I. QTL for stem sugar and grain nonstructural carbohydrates. Crop Science, 48, 2165-2179. Nagaraja Reddy, R., Madhusudhana, R., Murali Mohan, S., Chakravarthi, D. V. N., Mehtre, S. P., Seetharama, N., & Patil, J. V. (2013). Mapping QTL for grain yield and other agronomic traits in post-rainy sorghum [Sorghum bicolor (L.) Moench]. Theoretical and Applied Genetics, 126, 1921-1939. Nelson, J. C. (1997). QGENE: Software for marker-based genomic analysis and breeding. Molecular Breeding, 3, 239-245. Nyquist, W. E. (1991). Estimation of heritability and prediction of selection response in plant populations. Critical Reviews in Plant Sciences, 10, 235-322. Ongaro, V., & Leyser, O. (2008). Hormonal control of shoot branching. Journal of Experimental Botany, 59, 67-74 Pao, C. I., & Morgan, P. W. (1986). Genetic regulation of development in Sorghum bicolor. Role of the maturity genes. Plant Physiology, 82, 575-580. Parh, D. (2005). DNA-based markers for ergot resistance in sorghum. (Published doctoral thesis). University of Queensland, Brisbane. Parth, D. K., Jordan, D. R., Aitken, E. A. B., Mace, E. S., Jun-ai, P., McIntyre, C. L., & Godwin, I. D. (2008). QTL analysis of ergot resistance in sorghum. Theoretical and Applied Genetics, 117, 369-382. University of Ghana http://ugspace.ug.edu.gh 131 Paterson, A. H., Damon, S., Hewitt, J. D., Zamir, D., Rabinowitch, H. D., Lincoln, S. E., Lander, E. S., & Tanksley, S. D. (1991). Mendelian factors underlying quantitative traits in tomato: Comparison across species, generations, and environments. Genetics, 127, 181-197. Paterson, A. H. (1995). Molecular dissection of quantitative traits: progress and prospects. Genome research, 5, 321-333. Paterson, A. H., Bowers, J. E., Bruggmann, R., Dubchak, I., Grimwood, J., Gundlach, H...... & Rokhsar, D. S. (2009). The Sorghum bicolor genome and the diversification of grasses. Nature, 457. Patterson, H. D., & Thompson, R. (1971). Recovery of inter-block information when block sizes are unequal. Biometrika, 58, 545-554. Peltonen- Saimio, P., Kangas, A., Salo, Y., & Jauhiamen, L. (2007). Grain number dominates grain weight in temperate cereal yield determination: evidence based on 30 years multilocation trials. Field crops research, 100, 179-188. Peng, J. R., Carol, P., Richards, D. E., King, K. E., Cowling, R. J., Murphy, G. P., Harberd, N. P. (1997). The Arabidopsis GAI gene defines a signaling pathway that negatively regulates gibberellin responses. Genes & Development, 11, 3194-3205. Pereira, M. G., Lee, M., Bramel-Cox, P., Woodman, W., Doebley, J., & Whitkus, R. (1994). Construction of an RFLP map in sorghum and comparative mapping in maize. Genome, 37, 236-243. Pereira, M. G., Ahnert, D., Lee, M., & Klier, K. (1995). Genetic-mapping of Quantitative Trait Loci for panicle characteristics and seed weight in sorghum. Brazilian Journal of Genetics, 18, 249- 257 Perumal, R., Menz, M.A., Mehta, P.J., Katilé, S., Gutierrez-Rojas, L. A., Klein, R. R., Klein, P. E., Prom, L. K., Schlueter, J. A., Rooney, W. L., Magill, C. W. (2009). Molecular mapping of Cg1, a gene for resistance to anthracnose (Colletotrichum sublineolum) in sorghum. Euphytica, 165, 597-606 Prusinkiewicz, P., Erasmus, Y., Lane, B., Harder, L. D., & Coen, E. (2007). Evolution and development of inflorescence architectures. Science, 316, 1452–1456. Pushpendra, K., Gupta, S. R., & Neeraj, K. (2006). Genetic and molecular basis of grain size and grain number and its relevance to grain productivity in higher plant. Genome, 49, 565-571. Ragab, R. A., Dronovalli, S., Maroof Saghai, M. A., & Yu, Y. G. (1994). Construction of a sorghum RFLP linkage map using sorghum and maize DNA probes. Genome, 37, 590-594. University of Ghana http://ugspace.ug.edu.gh 132 Rami, J. F., Dufour, P., Trouche, G., Fliedel, G., Mestres, C., Davrieux, F., Blanchard, P., & Hamon, P. (1998). Quantitative trait loci for grain quality, productivity, morphological and agronomical traits in sorghum (Sorghum bicolor L. Moench). Theoretical and Applied Genetics, 97, 605- 616. Rangaswani Ayyangar, G. N & Rajabhooshanam, D. S. (1939). A preliminary analysis of panicle structure in sorghum-the great millet. Proceedings of the India Academy of Sciences – Section A. Part 3, Mathematical Science, 9, 29-38. Redona, E. D., & Mackill, D. J. (1998). Quantitative trait locus analysis for rice panicle and grain characteristics. Theoretical and Applied Genetics, 96, 957-963. Risterucci, A. M., Grivet, L., N’Goran, J. A. K., Pieretti, I., Flament, M. H., & Lanaud, C. (2000). A high density linkage map of Theobroma cacao L. Theoretical and Applied Genetics, 101, 948-955. Ritter, K. B., Jordan, D. R., Chapman, S. C., Godwin, I. D., Mace, E. S., & McIntyre, C. L. (2008). Identification of QTL for sugar-related traits in a sweet×grain sorghum (Sorghum bicolor L. Moench) recombinant inbred population. Molecular Breeding, 22, 367-384. Ross-Ibarra, J., Morrell, P. L., & Gaut, B. S. (2007). Plant domestication, a unique opportunity to identify the genetic basis of adaptation, Proceedings of the National Academy of Sciences of the United States of America, vol. 104, supplement 1, pp. 8641–8648. Rouse, D., Mackay, P., Stirnberg, P., Estelle, M., & Leyser, O. (1998). Changes in auxin response from mutations in an AUX/IAA gene. Science, 279, 1371-1373. Routman, E. J., & Cheverud, J. M. (1997). Gene effects on a quantitative trait: two locus epistatic effects measured at microsatellite markers and at estimated QTL. Evolution, 51, 1654-1662. Sadras, V. O. (2007). Evolutionary aspects of the trade-off between seed size and number in crops. Field Crops Research, 100, 125-138. Sanchez, A. C., Subudhi, P. K., Rosenow, D. T., & Nguyen, H. T. (2002). Mapping QTLs associated with drought resistance in sorghum (Sorghum bicolor L. Moench). Plant Molecular Biology, 48, 713-726. Sanchez-Gomez, A. (2002). Identification of quantitative trait loci for grain yield in a recombinant inbred B-line population in sorghum. (Published doctoral thesis). A&M University of Texas, College Station. Sasahara, H., Fukuta, Y., & Fukuyama, T. (1999). Mapping of QTLs for vascular bundle system and spike morphology in rice, Oryza sativa L. Breeding Science, 49, 75-81. University of Ghana http://ugspace.ug.edu.gh 133 Satoh-Nagasawa, N., Nagasawa, N., Malcomber, S., Sakai, H., & Jackson, D. (2006). A trehalose metabolic enzyme controls inflorescence architecture in maize. Nature, 441, 227-230. Sax, K. (1923). The association of size differences with seed-coat pattern and pigmentation in Phaseolus vulgaris. Genetics, 8, 552-560. Schmidt, R. J, Velt, B., Mandel, M. A., Mena, M., Hake, S., & Yanofsky, M. F. (1993). Identification and molecular characterization of ZAG1, the maize homolog of the Arabidopsis floral homeotic gene AGAMOUS. Plant Cell, 5, 729-737. Segura, V., Cilas, C., Laurens, F., & Costes, E. (2006). Phenotyping progenies for complex architectural traits: a strategy for 1-year-old apple trees (Malus x domestica Borkh.). Tree Genetics & Genomes, 2, 140-151. Shani, E., Yanai, O., & Ori, N. (2006). The role of hormones in shoot apical meristem function. Current Opinion in Plant Biology, 9, 484-489 Shapiro, S. S., & Wilk, M. B. (1965). An analysis of variance test for normality (complete samples). Biometrika, 52 (3-4), 591-611. Shiringani, A. L. (2009). Identification of genomic regions of Sorghum bicolor (L.) Moench linked to biofuel-related traits in grain x sweet sorghum recombinant inbred lines. (Published doctoral thesis). University of Giessen, Justus-Liebig Singh, F., Rai, K. N., Reddy, Belum V. S., & Diwakar, B. (Eds.). (1997). Development of cultivars and seed product ion techniques in sorghum and pearl millet. Training manual. Training and Fellowships Program and Genetic Enhancement Division, ICRISAT Asia Center, India. Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi -Arid Tropics. 118 pp. (Semi - formal publication). Sneath, P. H. A., & Sokal, R. R. (1973). Numerical taxonomy. New York: Freeman. Snowden, J. D. (1936). Cultivated race of sorghum. London: Adlard and Sons. Soderlund, C., Descour, A., Kudrna, D., Bomhoff, M., Boyd, L., Currie, J…..Yu, Y. (2009) Sequencing, Mapping and Analysis of 27,455 Maize Full-length cDNAs. PLOS Genetics, 5:e10000740. Spielmeyer, W., Ellis, M. H., & Chandler, P. M. (2002). Semi dwarf sd-1 green revolution rice contains a defective gibberellin 20-oxidase gene. Proceedings of the National Academy of Sciences of the United States of America, 99 (13), 9043-9048. Sreenivasulu. N., & Schnurbusch. T. (2012). A genetic playground for enhancing grain number in cereals. Trends in Plant Science, 17, 2. University of Ghana http://ugspace.ug.edu.gh 134 Srinivas, G., Satish, K., Madhusudhana, R., Nagaraja, R., Murali, M., & Seetharama, N. (2009). Identification of quantitative traits loci for agronomically important traits and their association with genic-microsatellite markers in sorghum. Theoretical and Applied Genetics, 118, 1439- 1454. Stirnberg, P., Chatfield, S. P., & Leyser, H. M. O. (1999). AXR1 acts after lateral bud formation to inhibit lateral bud growth in Arabidopsis. Plant Physiology, 121, 839-847. Stuber, C. W., Edwards, M. D., & Wendel, J. F. (1987). Molecular marker facilitated investigations of quantitative trait loci in maize. II. Factors influencing yield and its component traits. Crop Science, 27, 639-648. Subudhi, P. K., & Nguyen, H. T. (2000). New horizons in biotechnology. p. 349-397. In Smith et al. (ed.) Sorghum: Origin, history, technology, and production. John Wiley & Sons, Inc., New York. Suzaki, T., Sato, M., Ashikari, M., Miyoshi, M., Nagato, Y., & Hirano, H. Y. (2004). The gene FLORAL ORGAN NUMBER1 regulates floral meristem size in rice and encodes a leucine-rich repeat receptor kinase orthologous to Arabidopsis CLAVATA1. Development, 131, 5649-5657. Swain, S. M., & Olszewski, N. E. (1996). Genetic analysis of gibberellin signal transduction. Plant Physiology, 112, 11-17. Tabuchi, H., Zhang, Y., Hattori, S., Omae, M., Shimizu-Sato, S., & Oikawa, T. (2011). LAX PANICLE2 of rice encodes a novel nuclear protein and regulates the formation of axillary meristems. Plant Cell, 23, 3276-3287. Taguchi-Shiobara, F., Yuan, Z., Hake, S., & Jackson, D. (2001). The fasciated ear2 gene encodes a leucine-rich repeat receptor-like protein that regulates shoot meristem proliferation in maize. Genes & Development, 15, 2755-2766. Tanaka, W., Pautler, M., Jackson, D., & Hirano, H-Y. (2013). Grass Meristems II: Inflorescence Architecture, Flower Development and Meristem Fate. Plant Cell Physiology 54(3), 313-324 Tao, Y. Z., Henzell, R. G., Jordan, D.R., Butler, D.G., Kelly, A.M., McIntyre, C. L. (2000). Identification of genomic regions associated with stay green in sorghum by testing RILs in multiple environments. Theoretical and Applied Genetics, 100, 1225-1232. Tao, Y. Z., Hardy, A., Drenth, J., Henzell, R. G., Franzmann, B. A., Jordan, D. R., Butler, D. G., & McIntyre, C. L. (2003). Identifications of two different mechanisms for sorghum midge resistance through QTL mapping. Theoretical and Applied Genetics, 107, 116-122. Tariq, M., Awan, S. I., & HAQ, M. I. U. (2007). Genetic Variability and Character Association for Harvest Index in Sorghum under Rainfed Conditions. International journal of agriculture & biology, 3, 470-472. University of Ghana http://ugspace.ug.edu.gh 135 Terao, T., Nagata, K., Morino, K., & Hirose, T. (2010). A gene controlling the number of primary rachis branches also controls the vascular bundle formation and hence is responsible to increase the harvest index and grain yield in rice. Theoretical and Applied Genetics, 120, 875-893. Theissen, G., Strater, T., Fisher, A., & Saedler, H. (1995). Structural characterization, chromosomal localization and phylogenetic evaluation of two pairs of AGAMOUS-like MADS-box genes from maize. Gene, 156, 155-166. Thompson, B. E., Bartling, L., Whipple, C., Hall, D. H., Sakai, H., Schmidt, R., & Hake, S. (2009). bearded-ear encodes a MADS box transcription factor critical for maize floral development. Plant Cell, 21, 2578-2590. Toyomasu, T., Kawaide, H., Sekimoto, H., Von Numers, C., Phillips, A. L., Hedden, P., & Kamiya, Y. (1997). Cloning and characterization of a cDNA encoding gibberellin 20-oxidase from rice (Oryza sativa) seedlings. Plant, 99, 111-118. Tuinstra, M. R., Grote, E. M., Goldsbrough, P. B., & Ejeta, G. (1996). Identification of quantitative trait loci associated with preflowering drought tolerance in sorghum. Crop Science, 36, 1337- 1344. Umehara, M., Hanada, A., Yoshida, S., Akiyama, K., Arite, T., Takeda-Kamiya, N…..Yamaguchi, S. (2008). Inhibition of shoot branching by new terpenoid plant hormones. Nature, 455, 195- 200. Ungerer, M. C., Halldorsdottir, S. S., Modliszekwski, J. L., Mackay, T. F. C., & Purugganan, M. D. (2002). Quantitative trait loci for inflorescence development in Arabidopsis thaliana. Genetics, 160, 1133-1152. Upadyayula, N., da Silva, H. S., Bohn, M. O., Rocheford, T. (2006). Genetic and QTL analysis of maize tassel and ear inflorescence architecture. Theory of Applied Genetics, 112, 592-606. Utz, H. F., &. Melchinger, A. E. (1996). PLABQTL: A program for composite interval mapping of QTL. Institut of Plant Breeding, Seed Science, Population Genetics, University of Hohenheim, Stuttgart, Germany. Verbeke, J. A., & Heupel, R. C. (1990). Transition from vegetative to floral apex in Sorghum bicolor (L.) moench: an allometric index. Botanical Gazette, 151(1), 1-4. Vollbrecht, E., Springer, P. S., Goh, L., Buckler, E. S., & Martienssen, R. (2005). Architecture of floral branch systems in maize and related grasses. Nature, 436, 1119-1126. Wald, A. (1943). Test of statistical hypotheses concerning several parameters when the number of observation is large. Transactions of the American Mathematical Society, 54, 426-482. University of Ghana http://ugspace.ug.edu.gh 136 Wang, Y., & Li, J. (2006). Genes controlling plant architecture. Current Opinion in Biotechnology, 17, 123-29 Warkad, Y. N., Potdukhe, N. R., Dethe, A. M., Kahate, P. A., & Kotgire, R. R. (2008). Genetic variability, heritability and genetic advance for quantitative traits in sorghum germplasm. Agricultural Science Digest, 28 (3), 165-169. Weigel, D., Alvarez, J., Smyth, D. R., Yanofsky, M. F., & Meyerowitz, E. M. (1992). Leafy controls floral meristem identity in Arabidopsis. Cell, 69, 843-859. Whittaker, J. C., Thompson, R., & Visscher, P. M. (1996). On the mapping of QTL by regression of phenotypes on marker type, Heredity, 77, 23-32. Williams, E.A., & Morgan, P. W. (1979). Floral initiation in sorghum hastened by gibberellic acid and far-red light, Planta, 145, 269-272 Witt-Hmon, K. P., Shehzad, T., & Okuno, K. (2013). Variation in inflorescence architecture associated with yield components in a sorghum germplasm. Plant Genetic Resources: Characterization and Utilization, 1-8. Wu, P., Zhang, G., & Huang, N. (1996). Identification of QTLs controlling quantitative characters in rice using RFLP markers. Euphytica, 89, 349-354. Wu, Y. Q., & Huang, Y. (2006). An SSR genetic map of Sorghum bicolor (L.) Moench and its comparison to a published genetic map. Genome, 50, 84-89. Xiao, J., Li, J., Yuan, L., & Tanksley, S. D. (1996). Identification of QTLs affecting traits of agronomic importance in a recombinant inbred population derived from a subspecific rice cross. Theoretical and Applied Genetics, 92, 230-244. Xing, Y. Z., Tan, Y. F., Hua, J. P., Sun, X. L., Xu, C. G., & Zhang, Q. (2002). Characterization of the main effects, epistasis effects and their environmental interactions of QTLs on the genetic basis of yield traits in rice. Theoretical and Applied Genetics, 105, 248-257. Xu, G.W., Magill, C.W. Schertz, K.F., & Hart, G.E. (1994). A RFLP linkage map of Sorghum bicolor (L.) Moench. Theoretical and Applied Genetics, 89, 139-145. Yagi, T., Nagata, K., Fukuta, Y., Tamura, K., Ashikawa, I., & Terao, T. (2001). QTL mapping of spikelet number in rice (Oryza sativa L.). Breeding Science, 51, 53-56. Yamagishi, M., Takeuchi, Y., Kono, I., & Yano, M. (2002). QTL analysis for panicle characteristics in temperate japonica rice. Euphytica, 128: 219-224. University of Ghana http://ugspace.ug.edu.gh 137 Yamamoto, T., Ando, T., & Yano, M. (2007). Genetic dissection and pyramiding of quantitative traits for panicle architecture by using chromosomal segment substitution lines in rice. Research Highlights FY, 2006-2010. Yamamuro, C., Ihara, Y., Wu, X., Noguchi, T., Fujioka, S., Takatsuto, S….. Matsuoka, M. (2000). Loss of function of a rice brassinosteroid insensitive1 homolog prevents internode elongation and bending of the lamina joint. Plant Cell, 12, 1591-1606. Yano, M., & Sasaki, T. (1997). Genetic and molecular dissection of quantitative traits in rice. Plant Molecular Biology, 35, 145-153. Yeon Lee, D., & An, G. (2012). Two AP2 family genes, SUPERNUMERARY BRACT (SNB) and OsINDETERMINATE SPIKELET 1 (OsIDS1), synergistically control inflorescence architecture and floral meristem establishment in rice. The Plant Journal, 69, 445-461. Yoshida, A., Ohmori, Y., Kitano, H., Taguchi-Shiobara, F., & Hirano, H. Y. (2012). ABERRANT SPIKELET AND PANICLE1, encoding a TOPLESS-related transcriptional co-repressor, is involved in the regulation of meristem fate in rice. Plant Journal, 70, 327-339. Zeng, Z. B. (1994). Precision mapping of quantitative trait loci. Genetics, 136:1457-1468. Zhang, S., Hu, W., Wang, L., Lin, C., Cong, B., Sun, C., & Luo, D. (2005). TFL1/CEN-like genes control intercalary meristem activity and phase transition in rice. Plant Science, 168, 1393-1408 Zhao, D. L., Atlin, G. N., Bastiaans, L., & Spiertz, J. H. J. (2006). Developing selection protocols for weed competitiveness in aerobic rice. Field Crops Research, 97, 272-285. Zhin-Ben, Y., Yi, S., Xiao-Hong, L., Wei-Jun, Z., Min, Y., & Li Xia, C. (2006). Advances in genetic mapping of the sorghum genome. Chinese Journal of Agricultural Biotechnology, 3, 155-161. Zhuang, J. Y., Lin, H. X., Lu, J., Qian, H. R., Hittalmani, S., Huang, N., & Zheng, K. L. (1997). Analysis of QTL×environment interaction for yield components and plant height in rice. Theoretical and Applied Genetics, 95, 799-808. Zou, J., Zhang, S., Zhang, W., Li1, G., Chen, Z., Zhai1, W., Zhao, X., Pan, X., Xie, Q., & Zhu, L. (2006). The rice HIGH-TILLERING DWARF1 encoding an ortholog of Arabidopsis MAX3 is required for negative regulation of the outgrowth of axillary buds. The Plant Journal, 48, 687- 696. Zou, G., Zhai, G., Feng, Q., Yan, S., Wang, A., Zhao, Q., Shao, J., Zhang, Z., Zou, J., Han, B., & Tao, Y. (2012). Identification of QTLs for eight agronomically important traits using an ultra- high-density map based on SNPs generated from high-throughput sequencing in sorghum under contrasting photoperiods. Journal of Experimental Botany, 63, 5451-5462. University of Ghana http://ugspace.ug.edu.gh 138 APPENDIXES Appendix 4.1: Genotyping Material and method at CIRAD lab Material  LightCycler® 480 Multiwell Plate 384, white (04 729 749 001)  Plate PCR96 Biorad (réf : HSS9601)  LightCycler® 480 Sealing Foil, (04 729 757 001)  Robot Hamilton Microlab Star  Thermocycler Veriti 384 Well (Applied Biosystems)  LightCycler® 480 II (LC480), Roche Chemical and biological reagents  KASPar 2X Reagent Mix (Kbioscience, PN KBS-1004-001), conserved at -20°C  MgCl2 50mM, Kbiosciences, conserved at -20°C  H2O milliQ Method  Dispense DNA in plate 384 2μL of DNA at 10ng/μl was dispensed into PCR white plate 384 in the following order: A1, B1, A2, and B2.  Preparing and dispensing PCR mix Preparing « assays mix » (mix of primers); for 100μl assay mix:  12 μL primer specific for Allele 1 (100μM)_ 12 μM final  12 μL primer specific for Allele 2 (100μM)_ 12 μM final  30 μL common primer specific (100μM)_ 30 μM final University of Ghana http://ugspace.ug.edu.gh 139  46 μl H2O mQ The mix KASPAR has 1.8mM in MgCl2 of concentration, when the percentage GC of the assay is very low. We need to add 2.2 mM MgCl2 if the percentage of GC is between 33% and 55% and 2.5 mM MgCl2 for GC less than 33%. The plates were sealed with LightCycler® 480 Sealing Foil  PCR Because of the great size of the sample, PCR were directly done in LightCycler® 480 II (LC480), Roche and in Thermocycler Veriti 384 Well (Applied Biosystems) using this program « outside KASPar Project»  94 °C for 15 minutes Hot‐start enzyme activation  94 °C for 20 seconds Touchdown over 65‐57°C for 60 seconds 10 cycles (dropping 0.8°C per cycle)  94°C for 20 seconds; 57°C for 60 seconds; 26 cycles  Reading on LightCycler® 480 II (LC480), Roche  Open the software « LightCycler® 480 SW 1.5 »  Click on « New experiment from template »  Highlight Run template : « read Kaspar Plateform »  Click start Run  Give a name to the plate and folder in which it will be saved.  Coding of the alleles University of Ghana http://ugspace.ug.edu.gh 140 The allale “B” is coded for the parent Tiandougou, “A” allele is noted for the parent Lata-3 and the heterozygous alleles were coded “H”. Then the genotype of the different F4 families for each markers was established. University of Ghana http://ugspace.ug.edu.gh 141 Appendix 4.2: QTLs map for sorghum panicle architecture traits in F4 families from Tiandougou/Lata-3 SB01 0.0 SB01002 1.7 SB01004 12.7 SB01015 17.3 SB01023 18.6 SB01029 20.2 SB01034 25.0 SB01038 29.8 SB01041 45.3 SB01050 53.9 SB01056 60.7 SB01062 66.5 SB01070 68.9 SB01073 77.7 SB01087 81.7 SB01092 89.3 SB01096 90.0 SB01103 102.6 SB01112 108.7 SB01115 122.8 SB01121 126.4 SB01124 132.7 SB01128 135.8 SB01134 141.6 SB01143 142.8 SB01144 154.4 SB01157 160.2 SB01160 163.3 SB01162 171.4 SB01168 A v _ N G L o d = 8 . 8 / R 2 = 8 . 1 % B e n d i n g L o d = 3 . 7 / R 2 = 3 . 4 % I N _ L L o d = 4 . 5 / R 2 = 4 % N b _ S B _ M N L o d = 4 . 6 / R 2 = 4 . 7 % N G _ M N L o d = 6 . 2 / R 2 = 6 . 2 % N G _ P N L o d = 7 / R 2 = 7 . 3 % P A _ H I L o d = 4 / R 2 = 4 . 3 % P A _ I N _ N L o d = 2 1 . 4 / R 2 = 1 9 . 2 % P A _ L L o d = 8 . 3 / R 2 = 6 . 7 % P A _ N P B L o d = 1 8 . 2 / R 2 = 1 0 . 6 % P B _ D e n s L o d = 7 . 4 / R 2 = 4 . 2 % P E _ L L o d = 6 . 1 / R 2 = 3 . 9 % P o s _ N P B _ m a x L o d = 6 . 1 / R 2 = 5 . 9 % R A _ L L o d = 9 . 5 / R 2 = 7 . 9 % R A _ s l e n d L o d = 5 . 1 / R 2 = 3 . 3 % R A _ V o l L o d = 4 . 6 / R 2 = 4 % S B _ D F _ M N L o d = 3 . 6 / R 2 = 4 . 1 % T G W L o d = 3 . 6 / R 2 = 3 . 8 % SB02 0.0 SB02002 3.7 SB02006 6.1 SB02009 9.7 SB02013 43.9 SB02033 48.5 SB02036 56.6 SB02040 70.8 SB02057 74.9 SB02074 76.3 SB02078 81.4 SB02086 98.3 SB02097 98.9 SB02098 103.0 SB02099 112.7 SB02104 117.7 SB02107 122.3 SB02112 124.1 SB02118 132.2 SB02132 136.6 SB02139 144.7 SB02145 150.3 SB02154 155.6 SB02161 8.4 SB02171 169.1 SB02172 A v _ L P B L o d = 6 . 9 / R 2 = 4 . 2 % A v _ N G L o d = 7 . 3 / R 2 = 6 . 7 % a v _ N P B L o d = 5 . 4 / R 2 = 5 . 2 % A v _ S B _ D F L o d = 4 . 7 / R 2 = 5 . 3 % B e n d i n g L o d = 5 . 6 / R 2 = 5 . 2 % B e n d i n g L o d = 5 . 8 / R 2 = 5 . 4 % I N _ L _ m a x L o d = 6 . 1 / R 2 = 5 . 8 % N b _ P B _ m a x L o d = 4 . 2 / R 2 = 3 % N b _ S B _ L P B L o d = 8 . 1 / R 2 = 8 . 6 % N b _ S B _ S N L o d = 9 . 1 / R 2 = 9 . 1 % N G _ S N L o d = 8 . 5 / R 2 = 8 . 4 % P A _ I N _ N L o d = 9 . 5 / R 2 = 7 . 9 % P A _ L L o d = 8 . 6 / R 2 = 6 . 9 % P A _ N P B L o d = 4 . 1 / R 2 = 2 . 2 % P B _ B Z _ P N L o d = 3 . 7 / R 2 = 3 . 4 % P B _ B Z _ S N L o d = 1 4 / R 2 = 1 3 . 7 % P B _ D e n s L o d = 8 . 4 / R 2 = 4 . 8 % P B _ L _ m a x L o d = 9 . 7 / R 2 = 9 . 9 % P B _ L _ P N L o d = 3 . 8 / R 2 = 3 . 8 % P B _ L _ S N L o d = 1 2 . 9 / R 2 = 1 1 . 7 % P E _ L L o d = 4 . 2 / R 2 = 2 . 6 % P H L o d = 6 . 9 / R 2 = 3 . 4 % P o s _ P B _ L _ m a x L o d = 7 . 3 / R 2 = 7 . 7 % R A _ B _ D i a L o d = 4 . 3 / R 2 = 3 . 2 % R A _ c o n i L o d = 1 2 . 9 / R 2 = 8 . 7 % R A _ D i a L o d = 4 . 3 / R 2 = 3 . 4 % R A _ L L o d = 9 . 4 / R 2 = 7 . 8 % R A _ s l e n d L o d = 1 6 . 9 / R 2 = 1 1 . 7 % S B _ D F _ P N L o d = 3 . 5 / R 2 = 3 . 8 % S B _ L L o d = 9 . 7 / R 2 = 9 . 5 % S B _ L _ m a x L o d = 1 0 . 6 / R 2 = 9 . 4 % S h a p e L o d = 7 . 3 / R 2 = 7 . 9 % University of Ghana http://ugspace.ug.edu.gh 142 SB03 0.0 SB03002 3.4 SB03006 17.2 SB03019 27.3 SB03030 32.2 SB03032 36.5 SB03037 41.6 SB03042 46.6 SB03045 49.6 SB03048 60.4 SB03060 69.3 SB03069 69.7 SB03068 69.9 SB03070 71.2 SB03075 79.4 SB03085 86.0 SB03091 90.2 SB03094 104.6 SB03102 110.4 SB03107 120.0 SB03115 129.0 SB03131 132.1 SB03137 139.2 SB03149 146.3 SB03158 152.9 SB03164 159.2 SB03169 A v _ L P B L o d = 1 1 . 7 / R 2 = 7 . 3 % A v _ L P B L o d = 5 . 4 / R 2 = 3 . 3 % A v _ N G L o d = 4 . 3 / R 2 = 3 . 8 % A v _ N G L o d = 4 . 8 / R 2 = 4 . 3 % a v _ N P B L o d = 6 . 2 / R 2 = 6 % I N _ L L o d = 1 4 . 8 / R 2 = 1 4 . 1 % I N _ L _ m a x L o d = 6 . 8 / R 2 = 6 . 5 % N b _ P B _ m a x L o d = 8 / R 2 = 5 . 7 % N b _ P B _ m a x L o d = 5 . 9 / R 2 = 4 . 1 % N b _ S B _ M N L o d = 7 . 8 / R 2 = 8 . 1 % N b _ S B _ P N L o d = 6 . 5 / R 2 = 6 . 6 % N G _ M N L o d = 8 . 2 / R 2 = 8 . 3 % N G _ P A L o d = 8 / R 2 = 7 . 5 % N G _ P N L o d = 3 . 7 / R 2 = 3 . 8 % N G _ S N L o d = 3 . 6 / R 2 = 3 . 5 % N G _ S N L o d = 5 . 1 / R 2 = 5 % P A _ G Y L o d = 7 . 2 / R 2 = 6 . 9 % P A _ I N _ N L o d = 5 . 1 / R 2 = 4 . 1 % P A _ L L o d = 1 3 . 4 / R 2 = 1 1 . 2 % P A _ N P B L o d = 1 5 . 2 / R 2 = 8 . 7 % P A _ N P B L o d = 5 / R 2 = 2 . 7 % P A _ W L o d = 9 . 1 / R 2 = 8 . 6 % P B _ B Z _ M N L o d = 5 . 9 / R 2 = 5 . 9 % P B _ B Z _ M N L o d = 5 / R 2 = 5 % P B _ B Z _ P N L o d = 6 . 2 / R 2 = 5 . 8 % P B _ B Z _ S N L o d = 8 . 1 / R 2 = 7 . 7 % P B _ D e n s L o d = 1 2 . 3 / R 2 = 7 . 1 % P B _ D e n s L o d = 1 4 / R 2 = 8 . 2 % P B _ L _ m a x L o d = 5 . 7 / R 2 = 5 . 7 % P B _ L _ M N L o d = 4 . 4 / R 2 = 4 . 1 % P B _ L _ M N L o d = 6 . 1 / R 2 = 5 . 8 % P B _ L _ M N L o d = 4 . 2 / R 2 = 4 % P B _ L _ P N L o d = 4 . 5 / R 2 = 4 . 5 % P B _ L _ S N L o d = 7 . 3 / R 2 = 6 . 4 % P E _ L L o d = 2 8 . 3 / R 2 = 2 0 . 4 % P H L o d = 1 1 . 7 / R 2 = 5 . 9 % P o s _ N P B _ m a x L o d = 3 . 8 / R 2 = 3 . 7 % R A _ B _ D i a L o d = 6 . 9 / R 2 = 5 . 2 % R A _ c o n i L o d = 5 . 1 / R 2 = 3 . 3 % R A _ c o n i L o d = 8 . 5 / R 2 = 5 . 6 % R A _ D i a L o d = 5 . 4 / R 2 = 4 . 3 % R A _ L L o d = 1 2 . 1 / R 2 = 1 0 . 2 % R A _ s l e n d L o d = 2 1 / R 2 = 1 4 . 9 % S B _ D e n s _ S N L o d = 1 4 . 4 / R 2 = 1 3 . 6 % S B _ L _ m a x L o d = 5 / R 2 = 4 . 3 % S F D L o d = 5 3 . 1 / R 2 = 5 . 2 % SB04 0.0 SB04010 10.6 SB04021 15.4 SB04026 20.2 SB04029 28.9 SB04035 50.9 SB04046 53.1 SB04074 53.7 SB04077 56.4 SB04082 70.6 SB04090 84.7 SB04108 87.1 SB04112 90.4 SB04115 91.3 SB04117 93.6 SB04119 96.6 SB04122 101.4 SB04129 104.8 SB04133 106.4 SB04136 132.2 SB04154 I N _ L L o d = 7 . 5 / R 2 = 6 . 8 % I N _ L _ m a x L o d = 6 . 7 / R 2 = 6 . 4 % P A _ H I L o d = 3 . 5 / R 2 = 3 . 8 % P A _ L L o d = 9 . 7 / R 2 = 7 . 9 % P B _ L _ S N L o d = 5 . 5 / R 2 = 4 . 8 % P E _ L L o d = 5 . 8 / R 2 = 3 . 6 % P H L o d = 5 . 7 / R 2 = 2 . 8 % R A _ c o n i L o d = 5 . 2 / R 2 = 3 . 4 % R A _ c o n i L o d = 5 . 4 / R 2 = 3 . 5 % R A _ L L o d = 8 . 1 / R 2 = 6 . 7 % R A _ s l e n d L o d = 6 . 6 / R 2 = 4 . 3 % R A _ V o l L o d = 4 . 1 / R 2 = 3 . 5 % S B _ L _ m a x L o d = 4 . 8 / R 2 = 4 . 1 % University of Ghana http://ugspace.ug.edu.gh 143 SB05 0.0 SB05006 7.8 SB05011 11.8 SB05014 24.2 SB05022 37.6 SB05032 42.1 SB05037 51.7 SB05048 57.8 SB05068 62.0 SB05087 63.7 SB05090 65.6 SB05093 69.9 SB05097 75.4 SB05105 88.9 SB05115 94.5 SB05121 101.9 SB05131 107.8 SB05138 111.9 SB05150 115.0 SB05159 117.5 SB05163 118.6 SB05166 118.7 SB05166s P B _ B Z _ P N L o d = 4 . 6 / R 2 = 4 . 2 % S B _ L _ m a x L o d = 3 . 7 / R 2 = 3 . 1 % SB06 0.0 SB06008 13.4 SB06014 18.8 SB06015 24.3 SB06027 28.7 SB06034 54.0 SB06057 59.8 SB06065 64.2 SB06071 64.9 SB06074 74.2 SB06085 74.6 SB06087 82.8 SB06098 85.2 SB06103 86.7 SB06106 94.0 SB06113 99.1 SB06122 101.7 SB06127 108.7 SB06140 110.1 SB06145 A v _ L P B L o d = 3 6 . 5 / R 2 = 2 6 . 6 % a v _ N P B L o d = 6 . 6 / R 2 = 6 . 5 % N b _ P B _ m a x L o d = 2 4 . 6 / R 2 = 1 9 . 4 % N b _ S B _ P N L o d = 4 . 7 / R 2 = 4 . 7 % N b _ S B _ S N L o d = 4 . 8 / R 2 = 4 . 7 % N G _ P A L o d = 1 2 . 8 / R 2 = 1 2 . 5 % P A _ G Y L o d = 1 2 . 1 / R 2 = 1 1 . 9 % P A _ N P B L o d = 3 1 . 9 / R 2 = 2 0 . 2 % P A _ W L o d = 1 2 . 8 / R 2 = 1 2 . 3 % P B _ B Z _ M N L o d = 4 . 8 / R 2 = 4 . 7 % P B _ B Z _ P N L o d = 7 / R 2 = 6 . 6 % P B _ D e n s L o d = 2 8 . 8 / R 2 = 1 8 . 4 % P B _ L _ M N L o d = 4 . 8 / R 2 = 4 . 6 % P B _ L _ P N L o d = 5 . 5 / R 2 = 5 . 6 % P H L o d = 7 . 3 / R 2 = 3 . 6 % P o s _ N P B _ m a x L o d = 7 . 6 / R 2 = 7 . 4 % P o s _ N P B _ m a x L o d = 4 . 8 / R 2 = 4 . 6 % R A _ B _ D i a L o d = 2 0 . 4 / R 2 = 1 6 . 6 % R A _ c o n i L o d = 1 1 . 7 / R 2 = 7 . 9 % R A _ D i a L o d = 1 6 . 4 / R 2 = 1 4 % R A _ s l e n d L o d = 7 . 8 / R 2 = 5 . 1 % R A _ V o l L o d = 1 7 . 3 / R 2 = 1 6 . 1 % S B _ D e n s _ M N L o d = 7 . 2 / R 2 = 7 . 9 % S B _ D e n s _ S N L o d = 1 0 . 5 / R 2 = 9 . 7 % S B _ L L o d = 1 0 / R 2 = 9 . 9 % S B _ L _ m a x L o d = 5 . 3 / R 2 = 4 . 5 % University of Ghana http://ugspace.ug.edu.gh 144 SB07 0.0 SB07002 0.7 SB07003 9.9 SB07012 15.3 SB07019 20.9 SB07023 34.6 SB07031 36.2 SB07034 41.0 SB07036 47.5 SB07041 50.0 SB07044 51.7 SB07046b 55.5 SB06045 59.9 SB07059 65.1 SB07067 79.1 SB07077 85.3 SB07085 93.4 SB07090 99.7 SB07096 110.4 SB07108 114.7 SB07115 115.5 SB07120 119.6 SB07125 A v _ L P B L o d = 4 . 8 / R 2 = 2 . 9 % B e n d i n g L o d = 5 / R 2 = 4 . 7 % N b _ P B _ m a x L o d = 5 . 5 / R 2 = 3 . 9 % P A _ N P B L o d = 5 . 6 / R 2 = 3 % P B _ D e n s L o d = 4 . 5 / R 2 = 2 . 5 % P E _ L L o d = 7 . 1 / R 2 = 4 . 6 % P H L o d = 6 0 . 3 / R 2 = 4 1 . 1 % R A _ B _ D i a L o d = 4 . 9 / R 2 = 3 . 6 % R A _ B _ D i a L o d = 5 . 7 / R 2 = 4 . 3 % R A _ c o n i L o d = 4 . 4 / R 2 = 2 . 8 % R A _ D i a L o d = 5 / R 2 = 4 % R A _ D i a L o d = 5 . 2 / R 2 = 4 . 1 % R A _ s l e n d L o d = 3 . 6 / R 2 = 2 . 3 % R A _ s l e n d L o d = 5 . 7 / R 2 = 3 . 7 % R A _ V o l L o d = 4 . 5 / R 2 = 3 . 9 % S h a p e L o d = 4 . 2 / R 2 = 4 . 5 % T G W L o d = 7 . 1 / R 2 = 7 . 5 % SB08 0.0 SB08006 3.9 SB08009 21.3 SB08019 25.9 SB08023 31.6 SB08028 35.4 SB08034 40.8 SB08042 57.6 SB08075 62.2 SB08086 81.9 SB08097 87.3 SB08100 91.1 SB08103 102.4 SB08110 111.1 SB08121 111.5 SB08122 115.4 SB08127 N b _ S B _ S N L o d = 4 . 1 / R 2 = 3 . 9 % University of Ghana http://ugspace.ug.edu.gh 145 SB09 0.0 SB09003 5.5 SB09009 12.9 SB09014 30.6 SB09021 35.2 SB09024 42.6 SB09030 44.6 SB09033 47.1 SB09037 49.5 SB09044 53.3 SB09055 62.7 SB09080 66.1 SB09086 69.6 SB09091 80.0 SB09098 84.8 SB09101 90.7 SB09107 97.1 SB09114 103.0 SB09122 106.7 SB09127 109.8 SB09134 111.8 SB09138 115.1 SB09141 117.5 SB09145 B e n d i n g L o d = 5 . 5 / R 2 = 5 . 1 % N b _ S B _ P N L o d = 4 . 6 / R 2 = 4 . 6 % P B _ B Z _ P N L o d = 5 . 3 / R 2 = 4 . 9 % P E _ L L o d = 1 1 . 7 / R 2 = 7 . 7 % SB10 0.0 SB10001 3.1 SB10004 11.7 SB10012 23.9 SB10021 31.3 SB10030 38.8 SB10033 41.6 SB10038 44.4 SB10041 46.4 SB10043 60.6 SB10055 64.5 SB10058 67.5 SB10062 68.0 SB10064 70.4 SB10072 73.5 SB10079 76.4 SB10083 83.1 SB10091 85.3 SB10095 88.1 SB10098 94.1 SB10103 96.1 SB10104 101.7 SB10108 110.5 SB10116 117.9 SB10121 126.5 SB10131 132.8 SB10138 N b _ S B _ L P B L o d = 4 / R 2 = 4 . 2 % P B _ D e n s L o d = 4 . 3 / R 2 = 2 . 4 % P B _ L _ S N L o d = 3 . 9 / R 2 = 3 . 3 % P E _ L L o d = 5 . 1 / R 2 = 3 . 2 % P H L o d = 4 / R 2 = 1 . 9 % P o s _ P B _ L _ m a x L o d = 3 . 6 / R 2 = 3 . 8 % R A _ B _ D i a L o d = 3 . 6 / R 2 = 2 . 7 % S B _ D F _ P N L o d = 4 . 1 / R 2 = 4 . 4 % University of Ghana http://ugspace.ug.edu.gh 146 Appendix 4.3: Quantitative Trait Loci (QTLs) detected in Tiandougou/Lata-3 F4 families population Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QShcomp- SBI-02 SBI-02 147 141 154 SB02139 SB02161 7.3 7.9 0.5 0.4 0.80 D Tiand QShcomp- SBI-07 SBI-07 91 81 117 SB07077 SB07125 4.2 4.5 0.4 0.0 -0.08 A Tiand QPH- SBI-02 SBI-02 127 123 137 SB02112 SB02145 6.9 3.4 -63.0 -51 0.82 D Lata QPH- SBI-03 SBI-03 74 69.9 85 SB03070 SB03091 11.7 5.9 91.8 -49 -0.54 PD Tiand QPH- SBI-04 SBI-04 125 115 134 SB04136 SB04154 5.7 2.8 -85.8 -6.5 0.08 A Lata QPH- SBI-06 SBI-06 67 63 72 SB06065 SB06085 7.3 3.6 -71.9 11.1 -0.15 A Lata QPH- SBI-07 SBI-07 95 94 96 SB07090 SB07096 60.3 41.1 -238 67.9 -0.29 PD Lata QPH- SBI-10 SBI-10 99 90 113 SB10098 SB10121 4.0 1.9 51.8 19.9 0.38 PD Tiand QSFD- SBI-03 SBI-03 71.2 74 79.4 SB03075 SB03085 53.1 45.6 5.2 -2.6 -0.49 PD Tiand QPA_L- SBI-01 SBI-01 8 3 12.7 SB01004 SB01015 8.3 6.7 7.8 -2.0 -0.26 PD Tiand QPA_L- SBI-02 SBI-02 141 125 147 SB02118 SB02154 8.6 6.9 -8.1 1.0 -0.13 A Lata QPA_L- SBI-03 SBI-03 143 135 147 SB03137 SB03164 13.4 11.2 -9.7 -2.7 0.27 PD Lata QPA_L- SBI-04 SBI-04 55 43 61 SB04035 SB04090 9.7 7.9 -8.3 -1.8 0.22 PD Lata QPE_L- SBI-01 SBI-01 28 21 37 SB01034 SB01050 6.1 3.9 -10.6 4.4 -0.41 PD Lata QPE_L- SBI-02 SBI-02 132.2 112.7 157 SB02104 SB02171 4.2 2.6 -8.9 4.8 -0.53 PD Lata QPE_L- SBI-03 SBI-03 73 71.2 75 SB03075 SB03085 28.3 20.4 -25.4 -2.7 0.11 A Lata QPE_L- SBI-04 SBI-04 15.4 0 20 SB04010 SB04029 5.8 3.6 -10.1 1.6 -0.16 A Lata QPE_L- SBI-07 SBI-07 95 88 108 SB07085 SB07108 7.1 4.6 -11.5 6.0 -0.52 PD Lata QPE_L- SBI-09 SBI-09 51 48 61 SB09037 SB09080 11.7 7.7 14.9 -4.6 -0.31 PD Tiand QPE_L- SBI-10 SBI-10 103 99 130 SB10104 SB10138 5.1 3.2 10.2 -0.7 -0.07 A Tiand QRA_L- SBI-01 SBI-01 9 6 14 SB01004 SB01023 9.5 7.9 8.6 -2.5 -0.29 PD Tiand QRA_L- SBI-02 SBI-02 128 124.1 140 SB02118 SB02145 9.4 7.8 -8.6 2.0 -0.23 PD Lata QRA_L- SBI-03 SBI-03 143 137 151 SB03137 SB03164 12.1 10.2 -9.5 -2.4 0.25 PD Lata QRA_L- SBI-04 SBI-04 55 46 64 SB04035 SB04090 8.1 6.7 -7.8 -1.8 0.23 PD Lata QIN_L- SBI-01 SBI-01 7 1 17.3 SB01002 SB01023 4.5 4.0 -0.4 -0.1 0.30 PD Lata QIN_L- SBI-03 SBI-03 144 141 150 SB03149 SB03164 14.8 14.1 -0.7 -0.1 0.17 A Lata QIN_L-SBI-04 SBI-04 60 55 66 SB04077 SB04090 7.5 6.8 -0.5 0.2 -0.33 PD Lata University of Ghana http://ugspace.ug.edu.gh 147 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QIN_L_max- SBI-02 SBI-02 153 144.7 162 SB02145 SB02171 6.1 5.8 -0.9 0.3 -0.39 PD Lata QIN_L_max- SBI-03 SBI-03 151 143 158 SB03149 SB03169 6.8 6.5 -0.9 -0.2 0.20 A Lata QIN_L_max- SBI-04 SBI-04 56.5 54 64 SB04077 SB04090 6.7 6.4 -0.9 -0.4 0.49 PD Lata QRA_B_Dia- SBI-02 SBI-02 151 9 164 SB02009 SB02171 4.3 3.2 0.1 0.1 0.65 PD Tiand QRA_B_Dia- SBI-03 SBI-03 138 123 144 SB03115 SB03158 6.9 5.2 0.1 0.0 -0.28 PD Tiand QRA_B_Dia- SBI-06 SBI-06 69 67 72 SB06074 SB06085 20.4 16.6 -0.2 0.0 -0.05 A Lata QRA_B_Dia- SBI-07-1 SBI-07 35 27 38 SB07023 SB07036 4.9 3.6 0.1 0.0 0.31 PD Tiand QRA_B_Dia- SBI-07-2 SBI-07 119.6 113 119.6 SB07108 SB07125 5.7 4.3 0.1 0.0 0.41 PD Tiand QRA_B_Dia- SBI-10 SBI-10 62 53 118 SB10043 SB10131 3.6 2.7 0.0 0.2 5.73 PD Tiand QRA_Dia- SBI-02 SBI-02 151 9 163 SB02009 SB02171 4.3 3.4 0.0 0.0 0.42 PD Tiand QRA_Dia- SBI-03 SBI-03 136 131 143 SB03131 SB03158 5.4 4.3 0.1 0.0 -0.21 A Tiand QRA_Dia- SBI-06 SBI-06 70 67 73 SB06074 SB06085 16.4 14.0 -0.1 0.0 -0.08 A Lata QRA_Dia- SBI-07-1 SBI-07 34.6 27 38 SB07023 SB07036 5.0 4.0 0.1 0.0 0.10 A Tiand QRA_Dia- SBI-07-2 SBI-07 119 108 119.6 SB07096 SB07125 5.2 4.1 0.1 0.0 0.22 PD Tiand QRA_coni- SBI-02 SBI-02 129 125 144 SB02118 SB02145 12.9 8.7 0.0 0.0 -0.20 A Tiand QRA_coni-SBI-03-1 SBI-03 117 105 122 SB03102 SB03131 5.1 3.3 0.0 0.0 0.08 A Tiand QRA_coni-SBI-03-2 SBI-03 143 140 151 SB03149 SB03164 8.5 5.6 0.0 0.0 -0.09 A Tiand QRA_coni- SBI-04-1 SBI-04 55 44 62 SB04035 SB04090 5.2 3.4 0.0 0.0 0.59 PD Tiand QRA_coni- SBI-04-2 SBI-04 134.6 125 134.6 SB04136 SB04154 5.4 3.5 0.0 0.0 -0.74 PD Lata QRA_coni- SBI-06 SBI-06 68 58 72 SB06057 SB06085 11.7 7.9 0.0 0.0 -0.03 A Lata QRA_coni- SBI-07 SBI-07 119.6 42 119.6 SB07036 SB07125 4.4 2.8 0.0 0.0 0.76 PD Tiand QRA_Vol- SBI-01 SBI-01 15 9 21 SB01004 SB01038 4.6 4.0 199.7 -69 -0.35 PD Tiand QRA_Vol- SBI-04 SBI-04 64 53 77 SB04046 SB04108 4.1 3.5 -200 78.3 -0.39 PD Lata QRA_Vol- SBI-06 SBI-06 71 68 74 SB06074 SB06085 17.3 16.1 -417 55.0 -0.13 A Lata QRA_Vol- SBI-07 SBI-07 33 25 38 SB07023 SB07036 4.5 3.9 206.8 3.2 0.02 A Tiand University of Ghana http://ugspace.ug.edu.gh 148 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QRA_slend- SBI-01 SBI-01 81.7 0 86 SB01002 SB01096 5.1 3.3 -1.4 1.9 -1.41 PD Lata QRA_slend- SBI-02 SBI-02 141 138 145 SB02139 SB02154 16.9 11.7 -3.4 -0.6 0.18 A Lata QRA_slend- SBI-03 SBI-03 137 134 141 SB03137 SB03158 21.0 14.9 -3.7 0.0 0.00 A Lata QRA_slend- SBI-04 SBI-04 53.2 47 60 SB04035 SB04090 6.6 4.3 -1.9 -0.9 0.48 PD Lata QRA_slend- SBI-06 SBI-06 58 52 71 SB06034 SB06085 7.8 5.1 2.1 0.6 0.30 PD Tiand QRA_slend- SBI-07-1 SBI-07 56 46 64 SB07036 SB07067 3.6 2.3 -1.2 -1.5 1.29 OD Lata QRA_slend- SBI-07-2 SBI-07 119.6 114 119.6 SB07108 SB07125 5.7 3.7 -1.7 -0.9 0.52 PD Lata QPB_L_SN- SBI-02 SBI-02 140 134 143 SB02132 SB02145 12.9 11.7 -3.1 0.4 -0.14 A Lata QPB_L_SN- SBI-03 SBI-03 148 133 152 SB03137 SB03164 7.3 6.4 -2.1 -1.0 0.49 PD Lata QPB_L_SN- SBI-04 SBI-04 57 49 66 SB04035 SB04090 5.5 4.8 -1.8 0.5 -0.26 PD Lata QPB_L_SN- SBI-10 SBI-10 53 43 66 SB10038 SB10062 3.9 3.3 1.6 0.6 0.40 PD Tiand QPB_L_MN- SBI-03-1 SBI-03 13 6 16 SB03006 SB03019 4.4 4.1 3.5 4.2 1.21 D Tiand QPB_L_MN- SBI-03-2 SBI-03 18 14 19 SB03019 SB03030 6.1 5.8 -4.4 -3.7 0.83 D Lata QPB_L_MN- SBI-03-3 SBI-03 132 124 140 SB03115 SB03158 4.2 4.0 -1.0 0.1 -0.11 A Lata QPB_L_MN- SBI-06 SBI-06 85.2 76 89 SB06087 SB06113 4.8 4.6 1.2 -0.1 -0.12 A Tiand QPB_L_MN- SBI-02 SBI-02 118 109 129 SB02099 SB02132 3.8 3.8 0.8 0.8 1.03 D Tiand QPB_L_PN- SBI-03 SBI-03 14 8 26 SB03006 SB03030 4.5 4.5 -1.0 0.6 -0.57 PD Lata QPB_L_PN- SBI-06 SBI-06 85.2 84 91 SB06098 SB06113 5.5 5.6 1.1 0.2 0.22 PD Tiand QPB_L_PN- SBI-02 SBI-02 142 138 148 SB02139 SB02154 9.7 9.9 -3.0 0.5 -0.15 A Lata QPB_L_max- SBI-03 SBI-03 144 128 152 SB03115 SB03164 5.7 5.7 -2.1 -0.1 0.06 A Lata QPB_L_max- SBI-02 SBI-02 128 116 133 SB02104 SB02139 6.9 4.2 0.3 -0.1 -0.34 PD Tiand QAv_LPB- SBI-03-1 SBI-03 23 19 33 SB03019 SB03037 11.7 7.3 0.3 0.1 0.23 PD Tiand QAv_LPB- SBI-03-2 SBI-03 135 131 141 SB03131 SB03158 5.4 3.3 0.2 0.2 0.89 D Tiand QAv_LPB- SBI-06 70 69 72 SB06074 SB06085 36.5 26.6 -0.6 0.0 -0.07 A Lata QAv_LPB- SBI-07 SBI-07 39 29 50 SB07023 SB07044 4.8 2.9 0.2 0.0 0.05 A Tiand QAv_LPB -SBI-02 SBI-02 127 115 146 SB02104 SB02154 9.7 9.5 -0.6 0.1 -0.18 A Lata University of Ghana http://ugspace.ug.edu.gh 149 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QSB_L- SBI-06 SBI-06 75 68 78 SB06074 SB06098 10.0 9.9 0.5 0.1 0.24 PD Tiand QSB_L- SBI-02 SBI-02 128 118 143 SB02107 SB02145 10.6 9.4 -1.1 0.3 -0.24 PD Lata QSB_L_max- SBI-03 SBI-03 126 113 133 SB03107 SB03149 5.0 4.3 -0.7 -0.5 0.72 PD Lata QSB_L_max- SBI-04 SBI-04 56.5 53 65 SB04046 SB04090 4.8 4.1 -0.6 0.4 -0.71 PD Lata QSB_L_max- SBI-05 SBI-05 101.9 90 118.7 SB05115 SB05166s 3.7 3.1 0.5 -0.5 -1.04 D Tiand QSB_L_max- SBI-06 SBI-06 72 47 79 SB06034 SB06098 5.3 4.5 0.7 0.3 0.46 PD Tiand QSB_L_max- SBI-01 SBI-01 157 138 171.4 SB01134 SB01168 3.6 4.1 0.3 0.0 0.09 A Tiand QSB_DF_MN- SBI-02 SBI-02 117.7 41 132.2 SB02013 SB02132 3.5 3.8 0.0 0.0 0.55 PD Tiand QSB_DF_PN- SBI-10 SBI-10 108 92 117 SB10098 SB10121 4.1 4.4 0.0 0.0 -0.79 A Tiand QSB_DF_PN- SBI-02-1 SBI-02 117.7 114 126 SB02104 SB02132 4.7 5.3 0.2 0.2 0.99 D Tiand QAv_SB_DF- SBI-02-2 SBI-02 140 126 143 SB02118 SB02145 14.0 13.7 -3.4 0.7 -0.19 A Lata QPB_BZ_SN- SBI-03-1 SBI-03 144 133 152 SB03137 SB03164 8.1 7.7 -2.4 -0.8 0.32 PD Lata QPB_BZ_SN- SBI-03-2 SBI-03 45 17.2 57 SB03019 SB03060 5.9 5.9 -1.1 0.0 0.02 A Lata QPB_BZ_MN- SBI-03 SBI-03 135 128 143 SB03115 SB03158 5.0 5.0 -1.0 0.0 0.00 A Lata QPB_BZ_MN- SBI-06 SBI-06 88 80 96 SB06087 SB06122 4.8 4.7 0.9 -0.8 -0.84 D Tiand QPB_BZ_MN- SBI-02 SBI-02 121 115 127 SB02104 SB02132 3.7 3.4 0.6 1.2 2.08 OD Tiand QPB_BZ_PN- SBI-03 SBI-03 22 9 28 SB03006 SB03032 6.2 5.8 -1.0 1.0 -1.01 D Lata QPB_BZ_PN- SBI-05 SBI-05 28 18 32 SB05014 SB05032 4.6 4.2 -0.2 -2.1 10.98 OD Lata QPB_BZ_PN- SBI-06 SBI-06 85.2 84 92 SB06098 SB06113 7.0 6.6 1.2 0.1 0.07 A Tiand QPB_BZ_PN- SBI-09 SBI-09 115.1 113 117.5 SB09138 SB09145 5.3 4.9 -0.9 0.7 -0.82 A Lata QPB_BZ_PN- SBI-01 SBI-01 9 7 14 SB01004 SB01023 21.4 19.2 0.6 -0.1 -0.16 A Tiand QPA_IN_N- SBI-02-1 SBI-02 118 115 120 SB02104 SB02112 9.5 7.9 -0.4 -0.1 0.15 A Lata QPA_IN_N- SBI-03 SBI-03 8 0 14 SB03002 SB03019 5.1 4.1 0.3 0.0 -0.05 A Tiand QPA_IN_N- SBI-02-2 SBI-02 130 115 147 SB02104 SB02154 4.2 3.0 0.5 -0.2 -0.51 PD Tiand University of Ghana http://ugspace.ug.edu.gh 150 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QNb_PB_max- SBI-03-1 SBI-03 30 19 33 SB03019 SB03037 8.0 5.7 0.6 -0.2 -0.36 PD Tiand QNb_PB_max- SBI-03-2 SBI-03 137 132 151 SB03131 SB03164 5.9 4.1 0.5 0.6 1.08 D Tiand QNb_PB_max- SBI-06 SBI-06 69 67 71 SB06074 SB06085 24.6 19.4 -1.3 0.1 -0.06 A Lata QNb_PB_max- SBI-07 SBI-07 44 36 50 SB07031 SB07044 5.5 3.9 0.6 -0.1 -0.20 A Tiand QNb_PB_max- SBI-01 SBI-01 11 8 13 SB01004 SB01023 18.2 10.6 7.9 -1.5 -0.19 A Tiand QPA_NPB- SBI-02 SBI-02 166 150 169.1 SB02145 SB02172 4.1 2.2 3.7 0.6 0.16 A Tiand QPA_NPB- SBI-03-1 SBI-03 19 12 22 SB03006 SB03030 15.2 8.7 7.1 -0.4 -0.06 A Tiand QPA_NPB- SBI-03-2 SBI-03 138 132 154 SB03131 SB03169 5.0 2.7 3.9 1.4 0.36 PD Tiand QPA_NPB- SBI-06 SBI-06 78 76 80 SB06087 SB06098 31.9 20.2 -11.0 2.5 -0.22 PD Lata QPA_NPB- SBI-07 SBI-07 49 36 51 SB07031 SB07046b 5.6 3.0 4.1 -0.8 -0.21 A Tiand QPA_NPB- SBI-02 SBI-02 146 140 151 SB02139 SB02161 5.4 5.2 -1.1 0.4 -0.38 PD Lata Qav_NPB- SBI-03 SBI-03 19 15 26 SB03006 SB03030 6.2 6.0 -1.2 -0.1 0.10 A Lata Qav_NPB- SBI-06 SBI-06 86 79 88 SB06087 SB06113 6.6 6.5 1.2 0.4 0.36 PD Tiand Qav_NPB- SBI-01 SBI-01 12.7 8 15 SB01004 SB01023 7.4 4.2 0.0 0.0 -0.10 A Tiand QPB_Dens- SBI-02 SBI-02 147 139 153 SB02139 SB02161 8.4 4.8 0.0 0.0 -0.03 A Tiand QPB_Dens- SBI-03-1 SBI-03 20 15 24 SB03006 SB03030 12.3 7.1 0.0 0.0 0.25 PD Tiand QPB_Dens- SBI-03-2 SBI-03 142 135 150 SB03137 SB03164 14.0 8.2 0.0 0.0 0.15 A Tiand QPB_Dens- SBI-06 SBI-06 77 72 80 SB06074 SB06098 28.8 18.4 0.0 0.0 0.03 A Lata QPB_Dens- SBI-07 SBI-07 38 31 50 SB07023 SB07044 4.5 2.5 0.0 0.0 -0.50 PD Tiand QPB_Dens- SBI-10 SBI-10 126.5 114 132.8 SB10116 SB10138 4.3 2.4 0.0 0.0 -0.60 PD Lata QPB_Dens- SBI-02 SBI-02 110 106 117 SB02099 SB02107 8.1 8.6 -0.2 0.2 -0.80 PD Lata QNb_SB_LPB- SBI-10 SBI-10 32 2 41 SB10001 SB10038 4.0 4.2 -0.1 0.0 -0.09 A Lata QNb_SB_LPB- SBI-02 SBI-02 139 125 143 SB02118 SB02145 9.1 9.1 0.0 0.0 -0.87 D Lata QNb_SB_SN- SBI-06 SBI-06 93 61 97 SB06065 SB06122 4.8 4.7 0.0 0.0 0.09 A Lata QNb_SB_SN- SBI-08 SBI-08 62.2 30 71 SB08023 SB08097 4.1 3.9 0.0 0.0 1.07 D Lata QNb_SB_SN- SBI-01 SBI-01 152 145 167 SB01144 SB01168 4.6 4.7 0.0 0.0 -0.98 D Lata University of Ghana http://ugspace.ug.edu.gh 151 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QNb_SB_MN- SBI-03-1 SBI-03 14 9 19 SB03006 SB03030 7.8 8.1 0.0 0.0 -0.76 PD Lata QNb_SB_MN- SBI-03-2 SBI-03 14 10 21 SB03006 SB03030 6.5 6.6 0.0 0.0 -1.36 OD Lata QNb_SB_PN- SBI-06 SBI-06 90 85 107 SB06098 SB06140 4.7 4.7 0.0 0.0 -0.07 A Tiand QNb_SB_PN- SBI-09 SBI-09 116 113 117.5 SB09138 SB09145 4.6 4.6 0.0 0.0 -0.58 PD Lata QNb_SB_PN- SBI-03 SBI-03 148 143 150 SB03149 SB03164 14.4 13.6 0.0 0.0 0.12 A Tiand QSB_Dens_SN- SBI-06-1 SBI-06 69 59.8 73 SB06065 SB06085 10.5 9.7 0.0 0.0 0.48 PD Lata QSB_Dens_SN- SBI-06-2 SBI-06 80 66 86 SB06074 SB06106 7.2 7.9 0.0 0.0 0.26 PD Lata QSB_Dens_MN- SBI-01 SBI-01 7 1 13 SB01002 SB01023 6.1 5.9 0.4 0.2 0.56 PD Tiand QRel_Pos_IN_L_max- SBI-06 SBI-06 73 67 78 SB06074 SB06098 4.9 5.5 0.0 0.0 -0.70 D Lata QRel_Pos_NPB_max- SBI- 02 SBI-02 116 113 124.1 SB02104 SB02118 11.3 12.2 0.0 0.0 -0.54 PD Tiand QRel_Pos_PB_L_max- SBI-03 SBI-03 147 142 151 SB03149 SB03164 7.2 6.9 2.2 1.0 0.45 PD Tiand QBending- SBI-02-1 SBI-02 102 94 106 SB02086 SB02104 5.6 5.2 0.2 -0.2 -0.73 PD Tiand QBending- SBI-02-2 SBI-02 146 140 166 SB02139 SB02171 5.8 5.4 -0.3 0.0 -0.06 A Lata QBending- SBI-07 SBI-07 91 87 102 SB07085 SB07108 5.0 4.7 -0.2 0.2 -1.17 D Lata QBending- SBI-09 SBI-09 98 86 102 SB09101 SB09122 5.5 5.1 -0.2 -0.3 1.60 OD Lata University of Ghana http://ugspace.ug.edu.gh 152 Appendix 4.3 continued Trait Chrs Position (cM) CI_low (0.95) CI_high (0.95) Left marker Right marker LOD PVE % Add Dom Dom/Add Gene action Dir_ Parents QPA_GY- SBI-06 SBI-06 70 66 79 SB06074 SB06098 12.1 11.9 -3.1 0.9 -0.30 PD Lata QPA_GY- SBI-03 SBI-03 146.3 143 150 SB03149 SB03164 9.1 8.6 3.1 1.3 0.42 PD Tiand QPA_W- SBI-06 SBI-06 70 64.9 78 SB06074 SB06098 12.8 12.3 -4.0 2.4 -0.59 PD Lata QPA_W- SBI-01 SBI-01 5 0 28 SB01002 SB01041 4.0 4.3 0.0 0.0 -0.22 PD Lata QPA_HI- SBI-04 SBI-04 72 62 83 SB04082 SB04108 3.5 3.8 0.0 0.0 0.64 PD Tiand QPA_HI- SBI-01 SBI-01 18.6 11 28 SB01004 SB01041 3.6 3.8 0.2 0.1 0.35 PD Tiand QTGW- SBI-07 SBI-07 90 73 99.7 SB07067 SB07096 7.1 7.5 -0.3 0.1 -0.26 PD Lata QTGW- SBI-03 SBI-03 146.3 143 150 SB03149 SB03164 8.0 7.5 102.3 17.5 0.17 A Tiand QNG_PA- SBI-06 SBI-06 69 62 77 SB06065 SB06098 12.8 12.5 -141 6.9 -0.05 A Lata QNG_PA- SBI-02 SBI-02 126 112 138 SB02099 SB02145 8.5 8.4 -3.3 -0.2 0.06 A Lata QNG_SN- SBI-03-1 SBI-03 49 43 69.7 SB03042 SB03068 3.6 3.5 -1.6 2.6 -1.66 A Lata QNG_SN- SBI-03-2 SBI-03 144 139 159.2 SB03137 SB03169 5.1 5.0 2.5 -0.7 -0.28 PD Tiand QNG_SN- SBI-01 SBI-01 14 8 18.6 SB01004 SB01029 6.2 6.2 -1.3 0.7 -0.49 PD Lata QNG_MN- SBI-03 SBI-03 14 9 19 SB03006 SB03030 8.2 8.3 -1.6 0.5 -0.31 PD Lata QNG_MN- SBI-01 SBI-01 13 9 16 SB01004 SB01023 7.0 7.3 -0.6 0.5 -0.82 D Lata QNG_PN- SBI-03 SBI-03 11 1 21 SB03002 SB03030 3.7 3.8 -0.5 0.3 -0.67 PD Lata QNG_PN- SBI-01 SBI-01 12.7 9 15 SB01004 SB01023 8.8 8.1 -1.3 0.1 -0.04 A Lata QAv_NG- SBI-02 SBI-02 141 135 146 SB02132 SB02154 7.3 6.7 -1.3 0.3 -0.23 PD Lata QAv_NG- SBI-03-1 SBI-03 12 2 54 SB03002 SB03060 4.3 3.8 -1.0 0.1 -0.07 A Lata QAv_NG- SBI-03-2 SBI-03 146 139.2 159.2 SB03149 SB03169 4.8 4.3 1.0 -0.1 -0.09 A Tiand QAv_NG- SBI-01 SBI-01 4 0 12.7 SB01002 SB01015 3.7 3.4 0.2 -0.1 -0.63 PD Tiand University of Ghana http://ugspace.ug.edu.gh