University of Ghana http://ugspace.ug.edu.gh CHARACTERIZATION AND HETEROSIS AMONG EXTRA-EARLY MATURING ORANGE MAIZE INBRED LINES UNDER DROUGHT AND STRIGA INFESTATION By Noudifoulè TCHALA (10508999) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE AWARD OF DOCTOR OF PHILOSOPHY IN PLANT BREEDING DEGREE WEST AFRICA CENTRE FOR CROP IMPROVEMENT COLLEGE OF BASIC AND APPLIED SCIENCES UNIVERSITY OF GHANA LEGON DECEMBER, 2018 University of Ghana http://ugspace.ug.edu.gh DECLARATION I hereby declare that except for references to work 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. ……………………………………… TCHALA Noudifoulè (Student) ……………………………………… PROF. Isaac ASANTE (Supervisor) ……………………………………… DR. John Saviour Yaw ELEBLU (Supervisor) ……………………………………… DR. Beatrice Elohor IFIE (Supervisor) ……………………………………… DR. Baffour BADU-APRAKU (Supervisor) i University of Ghana http://ugspace.ug.edu.gh ABSTRACT Striga hermonthica and drought are major threats to maize (Zea mays L.) productivity and production in savannas of sub-Saharan Africa (SSA), a sub-region known to be plagued with vitamin A deficiency (VAD). Incorporating both drought tolerance and Striga resistance into high yielding orange maize cultivars for Striga endemic and drought-prone areas of the savanna agro- ecologies of SSA will increase acceptability of orange cultivars by farmers and aid in alleviating VAD and food shortage in the sub-region. The specific objectives of this research were to determine: (i) the performance of extra-early maturing maize inbred lines for tolerance to Striga and drought, (ii) the genetic diversity in extra-early maturing orange maize inbred lines using SNP- based DArTseq markers, (iii) the combining abilities of extra-early maturing orange maize inbred lines and heterosis for tolerance to Striga and drought, (iv) the performance and stability of extra- early maturing hybrids for tolerance to Striga and drought, (v) the combining abilities of extra- early maturing orange maize inbred lines and heterosis for carotenoids in maize kernels. One hundred and eighty inbred lines comprising 152 orange inbreds and 28 yellow lines were evaluated under Striga infestation, managed drought stress, and optimal environments at Ikenne, Abuja and Mokwa in Nigeria using a 12 x 15 alpha lattice design. Thirty-four (34) out of 180 inbreds evaluated and 32 out of 152 orange inbreds combined Striga resistance and drought tolerance, using base indices for selection. Twenty-four (24) of the 34 selected based on the indices were also selected by the multivariate best linear unbiased predictors (BLUPs) across all environments. The genetic purity and diversity among the 152 orange inbreds were assessed using 4620 polymorphic SNPs. The results revealed that 92% of the inbreds were pure with heterogeneity < 5% while the remaining 8% had heterogeneity ranging from 5.1 to 20.2%. Roger’s genetic distance for about 71% of the pairs of lines fell between 0.2001 and 0.2500. Ninety-two percent of the pairs of inbreds ii University of Ghana http://ugspace.ug.edu.gh also showed relative kinship values ranging from 0.300 to 0.500. The population structure analysis using STRUCTURE and neighbour-joining clustering assigned 71% of the inbreds in 4 distinct groups. Fifteen inbreds selected among the 152 evaluated plus TZdEEI 7 and TZdEEI 12 were used to generate 136 diallel single cross hybrids which were evaluated together with four experimental hybrid checks under Striga-infested, drought stress, and optimal environments at three locations in Nigeria (Ikenne, Abuja, and Mokwa) in 2016 and 2017 (total of 11 environments). The experimental design used was a 10 x 14 alpha lattice. General and specific combining ability components of the genetic variance were significantly different from zero for grain yield and most of the traits. Additive and non-additive genetic effects were both important with a predominance of the latter in controlling most of the measured traits including grain yield under Striga-infested, drought stress, and across test environments. However, additive genetic effects were found to be the primary type of gene action for the staygreen characteristic and Striga resistance indicator traits, suggesting that selection for these traits could easily be done based on predictions of GCA alone. Using base indices, 26% of the hybrids combined Striga resistance with drought tolerance. Stability assessment of the top 26 hybrids across test environments based on their genetic value indicated that TZEEIOR 12 x TZEEIOR 196 was the most stable, combining resistance to Striga and tolerance to drought with grain yield of 3885 kg ha-1 and 5411 kg ha-1 across environments and under optimal conditions, respectively. Hayman diallel analysis revealed predominance of dominant alleles in the parents with the ratio of dominant to recessive alleles being greater than 2 for β-carotene (2.36). Also, at the loci exhibiting dominance, the effects of dominant alleles were predominantly negative. In conclusion, dominance with negative genetic effect was found to be the gene action for carotenoids accumulation in the set of inbreds used. iii University of Ghana http://ugspace.ug.edu.gh DEDICATION To my wife, Pawoubadi LAKOUNYO. To my children, Bolanigni Ulrich and Ningnè Félicité. To my father, Evignénou Seth and my mother, Koulikpama Akoua. iv University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGMENTS I would like to thank the West Africa Agricultural Productivity Program-TOGO (WAAPP-TOGO) for providing the scholarship throughout the period of my study. I am also grateful to DAAD for their financial support, and through them, a special thank you to Professor Nohr Donatus for providing funds for my lab analysis in the University of Hohenheim and to Julian Wald and Alex Koza for their technical support. I would like to sincerely thank Professor Eric Danquah who gave me the opportunity to be partly supported by DAAD. I would like to also thank my supervisors at WACCI, Professor Isaac Asante, Dr John Eleblu, and Dr Beatrice Elohor Ifie, for their contribution to the completion of this study. I also thank them for reading all parts of this thesis and for making numerous suggestions for improvement and clarity. I am very grateful to Dr Baffour Badu-Apraku for accepting me into his project, and for his supervision, and ideas during the past three years. I also thank him for providing all the facilities needed and for the quality of the training under his supervision at IITA. I thank Professor Hans-Peter Piepho and Dr Jens Möhring for reviewing my scripts for data analysis and for providing me with the necessary documentation. My IITA-MIP colleagues over the past three years were a frequent source of support and inspiration, particularly Dr Konate Laban, Ebenezer Obeng-Bio, and many others. I would like to thank my parents for always encouraging me to do what I loved. Finally, my lovely wife Jeannine Lakounyo provided me with support and understanding, and I thank her with all my heart. v University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENTS DECLARATION............................................................................................................................ I ABSTRACT .................................................................................................................................. II DEDICATION............................................................................................................................. IV ACKNOWLEDGMENTS ........................................................................................................... V TABLE OF CONTENTS ........................................................................................................... VI LIST OF TABLES ...................................................................................................................... XI LIST OF FIGURES ................................................................................................................... XV LIST OF ABBREVIATIONS ............................................................................................... XVII CHAPTER ONE ........................................................................................................................... 1 GENERAL INTRODUCTION ................................................................................................ 1 CHAPTER TWO .......................................................................................................................... 5 2.0 LITERATURE REVIEW ................................................................................................... 5 2.1 THE COMPLEX “STRIGA-DROUGHT” AND MAIZE YIELD STABILITY IN THE SAVANNAS OF WEST AFRICA ............................................................................................................................. 5 2.2 OVERVIEW OF AGRICULTURAL DROUGHT AND STRIGA HERMONTHICA .............................. 7 2.2.1 Definition of drought in agriculture ............................................................................ 7 2.2.2 Overview of Striga hermonthica ................................................................................... 7 2.3 BREEDING TOLERANCE IN MAIZE FOR DROUGHT ................................................................ 9 2.3.1 Requirements for breeding for tolerance to drought ................................................... 9 2.3.2 Plant physiological variables addressed in breeding for tolerance to drought......... 10 2.3.3 The use of secondary traits ......................................................................................... 10 2.4 BREEDING FOR RESISTANCE/TOLERANCE IN MAIZE TO STRIGA ........................................ 12 2.4.1 Genetics of resistance to Striga .................................................................................. 12 2.4.1.1 Resistance/tolerance in maize ............................................................................ 12 2.4.1.2 Gene action/ inheritance of resistance to Striga ............................................... 12 2.4.1.3 Genes sources ...................................................................................................... 13 2.5 VITAMIN A DEFICIENCY IN SSA AND BIOFORTIFICATION IN MAIZE ................................ 13 vi University of Ghana http://ugspace.ug.edu.gh 2.5.1 The scope of the problem ............................................................................................ 13 2.5.2 Provitamin A biofortification in maize ....................................................................... 15 2.5.3 Achievement towards conventional breeding for carotenoids in maize .................... 16 2.6 MOLECULAR CHARACTERISATION IN PLANT BREEDING ................................................... 18 CHAPTER THREE .................................................................................................................... 20 3.0 PERFORMANCE OF EXTRA-EARLY MATURING MAIZE INBRED LINES UNDER STRIGA- INFESTED, DROUGHT STRESS AND OPTIMAL ENVIRONMENTS ................................................. 20 3.1 INTRODUCTION ................................................................................................................... 20 3.2 MATERIALS AND METHODS ............................................................................................. 21 3.2.1 Genetic materials...................................................................................................... 21 3.2.2 Experimental design and field trial management ................................................... 21 3.2.2.1 Experimental design ........................................................................................... 21 3.2.2.2 Management of drought trials ........................................................................... 21 3.2.2.3 Management of trials under Striga infestation ................................................ 22 3.2.2.4 Management of trials under rain-fed conditions ............................................. 22 3.2.3 Data collection and measured traits ........................................................................ 23 3.2.3.1 Flowering and plant morphological traits ........................................................ 23 3.2.3.2 Ear related traits ................................................................................................. 23 3.2.3.3 Grain yield estimation ........................................................................................ 24 3.2.3.4 Striga resistance and drought tolerance indicator traits................................. 24 3.2.4 Data analysis ............................................................................................................ 24 3.2.4.1 Model description................................................................................................ 24 3.2.4.2 Random error variance structure ..................................................................... 25 3.2.4.3 Estimation of additive genetic variance ............................................................ 26 3.2.4.4 Significance test of random factors ................................................................... 27 3.2.4.5 Estimation of narrow sense heritability and breeding values ......................... 27 3.2.4.6 Base index for selection and correlation among measured traits ................... 28 3.3 RESULTS .............................................................................................................................. 28 3.3.1 Genetic variability among inbreds for the measured traits ....................................... 28 3.3.3 Genetic correlations among traits under stress environments .................................. 36 3.3.3 Performance of inbred lines under stress and across test environments .................. 39 vii University of Ghana http://ugspace.ug.edu.gh 3.3.3.1 Index-based performance of inbreds ................................................................ 39 3.3.3.2 BLUPs-based performance of inbreds .............................................................. 42 4.5 DISCUSSION AND CONCLUSIONS ...................................................................................... 44 CHAPTER FOUR ....................................................................................................................... 48 4.0 GENETIC DIVERSITY AND HOMOGENEITY OF EXTRA-EARLY MATURING ORANGE MAIZE INBRED LINES ............................................................................................................................ 48 4.1 INTRODUCTION ................................................................................................................... 48 4.2 MATERIALS AND METHODS ............................................................................................. 49 4.2.1 Plant materials ......................................................................................................... 49 4.2.2 Leaf samples collection and DNA extraction.......................................................... 49 4.2.3 Diversity Array Technology sequencing (DArTseq) genotyping and analysis ...... 49 4.2.5 Summary statistics and cluster analyses ................................................................. 50 4.2.6 Population structure and principal component analyses ....................................... 51 4.3 RESULTS ........................................................................................................................... 52 4.3.1 Description of DArTseq SNP markers in the Population ......................................... 52 4.3.2 Genetic purity of the inbred lines ............................................................................... 54 4.3.3 Genetic distance and relatedness among the inbreds ................................................ 55 4.3.4 Population structure analysis ..................................................................................... 57 4.3.5 Principal coordinate analysis and Distance-based grouping .................................... 59 4.4 DISCUSSIONS ..................................................................................................................... 62 CHAPTER FIVE ........................................................................................................................ 67 5.0 ESTIMATES OF COMBINING ABILITIES AND HETEROSIS OF EXTRA-EARLY MATURING ORANGE INBRED LINES AND PERFORMANCE OF THEIR HYBRIDS ............................................ 67 5.1 INTRODUCTION ................................................................................................................... 67 5.2 MATERIALS AND METHODS .............................................................................................. 68 5.2.1 Genetic materials...................................................................................................... 68 5.2.2 Field Evaluation and data collection ...................................................................... 68 5.2.3 Data Analysis and genetic parameter estimate ....................................................... 70 5.2.3.1 Model description................................................................................................ 70 5.2.3.2 Genetic components variance structures .......................................................... 71 viii University of Ghana http://ugspace.ug.edu.gh 5.2.3.3 Significance tests ................................................................................................. 72 5.2.3.4 Genetic value and heritability estimates ........................................................... 72 5.2.3.5 Relative importance of GCA and SCA ............................................................. 73 5.2.3.6 Heterotic Groupings of parental inbreds .......................................................... 74 5.2.3.7 Inbred testers identification ............................................................................... 75 5.2.3.8 Heterosis estimates and analysis ........................................................................ 75 5.2.3.9 Genetic correlation among test environments and stability analysis of hybrid performance..................................................................................................................... 76 5.3 RESULTS .............................................................................................................................. 77 5.3.1 Genetic analysis of performance of extra-early maturing maize inbred lines under contrasting environments .................................................................................................... 77 5.3.2 Performance in crosses of the 17 inbred parents ...................................................... 87 5.3.2.1 GCA effects and GCA-based heterotic grouping of the inbred parents ........ 87 5.3.2.2 Identification of Testers ...................................................................................... 99 5.3.2.3 Heterosis and SCA effects of inbred parents for grain yield .......................... 99 5.3.3 Performance of hybrids and stability analysis ......................................................... 105 5.3.3.1 Index-based performance under Striga environments .................................. 105 5.3.3.2 Index-based performance under drought environments .............................. 107 5.3.3.3 Performance of hybrids across test environments. ........................................ 109 5.3.3.4 Genetic correlation and heritability among test environments for grain yield ......................................................................................................................................... 111 5.3.3.5 Stability of the performance of hybrids across test environments ............... 112 5.4 DISCUSSION AND CONCLUSION ......................................................................................... 117 CHAPTER SIX ......................................................................................................................... 121 6.0 ESTIMATE OF COMBINING ABILITIES AND HETEROSIS FOR CAROTENOIDS IN MAIZE KERNELS OF EXTRA-EARLY ORANGE INBREDS ...................................................................... 121 6.1. INTRODUCTION ........................................................................................................... 121 6.2. MATERIALS AND METHODS ....................................................................................... 122 6.2.1 Plant material ............................................................................................................ 122 6.2.2 Sample preparation ................................................................................................... 122 6.2.3. Carotenoid extraction and quantification ............................................................... 123 ix University of Ghana http://ugspace.ug.edu.gh 6.3 RESULTS ............................................................................................................................ 125 6.3.1 Carotenoids profile of the selected parental lines .................................................... 125 6.3.2 Heterosis for carotenoid accumulation in maize kernels ........................................ 131 6.3.3 Gene action and inheritance of carotenoids in 17 orange maize inbred lines ....... 136 6.4 DISCUSSION AND CONCLUSION ......................................................................................... 139 CHAPTER SEVEN ................................................................................................................... 144 7.0 CONCLUSIONS AND RECOMMENDATIONS ........................................................ 144 REFERENCES .......................................................................................................................... 148 x University of Ghana http://ugspace.ug.edu.gh LIST OF TABLES Table Title Page Table 3.1 Variance components and heritability estimates of measured traits of extra- early maturing maize inbreds under Striga infestation at Abuja and Mokwa, 2016-2017………………………………………………………………….. 30 Table 3.2 Variance components and heritability estimates of measured traits of extra- early maturing maize inbreds under drought stress at Ikenne during the 2016/2017 and 2017/2018 dry seasons…………………………………….. 32 Table 3.3 Variance components and heritability estimates of measured traits of extra- early maturing maize inbreds under five optimal environments at Abuja, Mokwa and Ikenne, 2016-2017……………………………………………. 34 Table 3.4 Variance components and heritability estimates of measured traits of extra- early maturing maize inbreds across test environment…………………….. 35 Table 3.5 Genetic correlations among grain yield and other traits across Striga- infested environments at Abuja and Mokwa, 2016-2017………………….. 37 Table 3.6 Genetic correlations among grain yield and other traits across drought stress environments in Ikenne during 2016/2017 and 2017/2018 dry seasons…… 38 Table 3.7 Performance of index-based top 20% and worst 5 inbred lines evaluated across Striga environments in Abuja, 2016 and Mokwa, 2017……………. 40 Table 3.8 Performance of index-based top 20% and worst 5 inbred lines evaluated across drought stress environments at Ikenne, during 2016/2017 and 2017/2018 dry seasons……………………………………………………... 41 Table 3.9 Pedigree BLUPs-based and multivariate BLUPs-based performances of the top 12% of the inbreds across pedigree under Striga-infested, drought and across test environments, 2016-2017……………………………………… 43 Table 5.1 Characteristic of the selected 17 extra-early maturing maize inbred lines used in the diallel crosses with their reaction to drought stress and Striga infestation in 2015…………………………………………………………. 69 Table 5.2 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 79 xi University of Ghana http://ugspace.ug.edu.gh selected inbreds under four Striga-infested environments in Nigeria, 2016- 2017………………………………………………………………………... Table 5.3 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across two drought environments in Nigeria, 2016-2017… 81 Table 5.4 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across five optimal environments in Nigeria, 2016-2017… 83 Table 5.5 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across 11 test environments in Nigeria, 2016-2017……… 86 Table 5.6 General combining ability effects of inbred parents for grain yield and other traits across drought stress environments in Nigeria, 2016-2017………….. 88 Table 5.7 General combining ability effects of inbred parents for grain yield and other traits across Striga-infested environments in Nigeria, 2016-2017…………. 92 Table 5.8 General combining ability effects of inbred parents for grain yield and other traits across test environments in Nigeria, 2016-2017…………………….. 96 Table 5.9 Means squares of 17 extra-early maturing inbred parents with their 136 diallel hybrids for grain yield across environments under Striga infestation, drought stress, and optimal conditions in Nigeria, 2016-2017…………….. 100 Table 5.10 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across two Striga-infested environments of top 20% and 3 worst extra-early maturing single cross hybrids selected based on the specific heterosis in Abuja in 2016 and Mokwa in 2017……………………………. 101 Table 5.11 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across two drought stress environments of top 20% and 3 worst extra-early maturing single cross hybrids selected based on the specific heterosis at Ikenne during 2016/2017 and 2017/2018 dry seasons………… 103 Table 5.12 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across three optimal environments of top 20% and 3 worst extra- 104 xii University of Ghana http://ugspace.ug.edu.gh early maturing single cross hybrids selected based on the specific heterosis at Ikenne in 2016 and 2017 and at Mokwa in 2017………………………… Table 5.13 Performance of index-based top 20% and 3 worst extra-early maturing single cross hybrids across Striga-infested environments at Abuja and Mokwa in 2016 and 2017………………………………………………….. 106 Table 5.14 Index-based performance of top 20% and 3 worst extra-early maturing single cross hybrids across drought-stress environments at Ikenne during 2016/2017 and 2017/2018 dry seasons…………………………………….. 108 Table 5.15 Performance of the top 20 and 5 worst extra-early maturing single cross hybrids across test environments based on their genetic values across 11 test environments…………………………………………………………... 110 Table 5.16 Spearman's rho correlation coefficients among the four parameters used for selecting the F1 hybrids and their grain yield across 11 test environments… 111 Table 5.17 Genetic correlation among test environments and heritability for grain yield 112 Table 6.1 Derived ratios used to assess inbreds and hybrids in 2018 and their significance………………………………………………………………… 124 Table 6.2 Mean squares of different compounds of carotenoids in 18 tropical inbreds lines………………………………………………………………………... 126 Table 6.3 Profile of the 17 parental inbred lines and one checks in carotenoid compounds and derived parameters and ratios…………………………….. 127 Table 6.4 Analysis of variance for carotenoid compounds, provitamin A (PVA), and total carotenoids (TC) of selected extra-early maturing orange maize…….. 132 Table 6.5 Minimum, maximum, and average heterosis (H) and heterobeltiosis (Hb) for carotenoids compounds, provitamin A (PVA), and total carotenoids (TC) of selected extra-early maturing orange maize………………………. 133 Table 6.6 Estimates of heterobeltiosis (upper triangle) and heterosis (lower triangle) effects (%) for β-cryptoxanthin of 136 F1's derived from half diallel crosses among 17 extra-early maturing orange maize inbreds……………………… 134 Table 6.7 Estimates of heterobeltiosis (upper triangle) and heterosis (lower triangle) effects for β-carotene of 136 F1's derived from half diallel crosses among 17 extra-early maturing orange maize inbreds…………………………….. 135 xiii University of Ghana http://ugspace.ug.edu.gh Table 6.8 Genetic parameter estimates for carotenoid compounds, provitamin A carotenoids (PVA), and total carotenoids (TC) of 17 selected extra-early maturing orange maize inbreds and their 136 hybrids……………………… 137 Table 6.9 General combining ability (GCA) estimates of selected extra-early maturing orange maize for different carotenoid compounds and derived traits………………………………………………………………………... 138 xiv University of Ghana http://ugspace.ug.edu.gh LIST OF FIGURES Figure Title Page Figure 4.1 Summary statistics of the 4620 SNPs used for characterising the 152 orange maize inbred lines from IITA-MIP breeding programs…………… 53 Figure 4.2 Summary of the heterogeneity of 152 inbred lines based on 4,620 polymorphic SNPs………………………………………………………... 54 Figure 4.3 Distribution of pairwise (A) Rogers genetic distance and (B) relative kinship among 152 orange maize inbred lines based on 4,620 SNPs……………………………………………………………………… 56 Figure 4.4 Analysis of the population structure of 152 orange maize inbred lines, (A) Estimated Δk over ten repeats of STRUCTURE analysis; (B) Population structure assessed by STRUCTURE for k =2, k =3, and k = 4…………… 58 Figure 4.5 Contribution of the axis to the total variation explained in the genotypic data set……………………………………………………………………. 60 Figure 4.6 Population structure of 152 orange maize inbred lines using 4,620 SNP markers and the a priori information from STRUCTURE analysis in Principal Coordinate Analysis (PCoA)…………………………………… 60 Figure 4.7 Neighbour-joining grouping of 152 orange maize inbred lines based on Rogers genetic distance calculated from 4,620 SNP markers……………. 61 Figure 5.1 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) under drought environments……………………………………………………………... 89 Figure 5.2 Heterotic grouping of inbred parents under drought environments using HGCAMT-Ward.D2……………………………………………………... 90 Figure 5.3 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) under Striga environments……………………………………………………………... 93 Figure 5.4 Heterotic grouping of inbred parents under Striga-infested environments using HGCAMT-Ward.D2………………………………………………. 94 xv University of Ghana http://ugspace.ug.edu.gh Figure 5.5 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) across test environments……………………………………………………………... 97 Figure 5.6 Heterotic grouping of inbred parents across test environments using HGCAMT Ward.D2……………………………………………………… 98 Figure 5. 7 A ‘which won where’ genotype main effect plus genotype x environment biplot of 30 extra-early maturing maize hybrids evaluated for grain yield across Striga infested and optimal environments in 2016 and 2017……… 113 Figure 5. 8 Mean vs. stability view of genotype main effect plus genotype x environment biplot of grain yield of 30 selected extra-early maturing maize hybrids evaluated across Striga-infested and optimal environments in 2016 and 2017…………………………………………………………. 115 Figure 5. 9 Discriminating vs. representativeness view of GGE biplot of 30 selected extra-early maturing hybrid maize for grain yield across Striga-infested and optimum environments in 2016 and 2017…………………………… 116 Figure 6.1 Carotenoids profile of 18 maize inbred lines using Principal Components Analysis (PCA) biplot……………………………………………………. 129 Figure 6.2 Optimal number of clusters suggested by Elbow method (A) and carotenoids profile-based grouping of 18 maize inbred lines (B)………… 130 xvi University of Ghana http://ugspace.ug.edu.gh LIST OF ABBREVIATIONS CIMMYT International Maize and Wheat Improvement Center DNA Deoxyribonucleic acid FAO Food and Agriculture Organization of the United Nations GD Genetic distance GGE Genotype plus genotype by environment interaction IITA International Institute of Tropical Agriculture MAS Marker-assisted selection MIP Maize improvement program NJ Neighbour-joining trees PIC Polymorphic information content SNP Single nucleotide polymorphisms SSA sub-Saharan Africa UPGMA Unweighted pair group method with arithmetic average WA West Africa WAP Week after planting WCA West and Central Africa xvii University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE GENERAL INTRODUCTION Maize (Zea mays L.) is the predominant staple crop in most countries of sub-Saharan Africa (SSA) where vitamin A deficiency is a serious human health problem (WHO, 2009). The prevalence of vitamin A deficiency was estimated to be among the highest (48%) in SSA with an occurrence of 95% of reported death from diarrhea and measles in 2013 as a result of the weakened immune system due to vitamin A deficiency (Stevens et al., 2015). Defined as a liver retinol reserve of < 0.1 µmol g-1liver (Tanumihardjo, 2011), vitamin A deficiency is most severe among pre-school aged children and pregnant women (Rice et al., 2004). In SSA, the annual per capita food maize consumption averages 36 kg (Atlin et al., 2011) making it the most consumed cereal in the sub- region. Besides, maize is a carotenogenic species with natural variation of total carotenoids ranging from 0.00 to 19 μg g−1 DW (HarvestPlus, 2017). Therefore, maize has been targeted as a crop for provitamin A biofortification as a more sustainable and cost-effective way to deal with vitamin A deficiency in SSA. So far, two genes out of 58 known to play key roles in biosynthesis and accumulation of carotenoids in maize have been successfully utilized in marker-assisted breeding. Combinations of favourable alleles at both lycE and crtRB1 loci allowed increases up to 22.6 µg g-1 (Menkir et al., 2017) mostly in maize inbreds. At the same time, the current released cultivars contain an average of 6 to 8 µg g-1 of provitamin A (HarvestPlus, 2017) which is far below the breeding target of 15 µg g-1. Hence, in-depth understanding of the mode of inheritance of such a complex quantitative trait in maize is needed to identify a better choice of breeding scheme and decide whether to go for open pollinated varieties (OPVs) or hybrids using the readily available superior inbreds as parents. 1 University of Ghana http://ugspace.ug.edu.gh The savanna agro-ecology of sub-Saharan Africa is characterized by high solar radiation and low night temperatures, which are known to represent favourable conditions for C4 plants such as maize. Nowadays, savanna agro-ecology is considered as the maize grain basket of the whole sub- region due to the introduction and the wide spread use of extra-early and early maturing maize. However, maize production is constrained by several biotic and abiotic stresses in savanna zones, which under field conditions occur most often simultaneously. Prominent among these threats are drought and the parasitic weed, Striga hermonthica Del. Benth. In the marginal areas where the annual rainfall is mostly below 500 mm, maize yield loss from drought stress may be much higher than the average estimated losses of 15% of total production per annum. Grain yield losses can even be greater if the drought stress occurs at the most drought-sensitive stages of crop growth, such as flowering and grain filling periods. For example, drought stress can reduce yield by 21% when it occurs at the grain filling period, by 50% at flowering (Denmead and Shaw, 1960), and by as much as 90% (NeSmith and Ritchie, 1992) when it occurs a few days before tassel emergence to the beginning of grain filling period. Furthermore, 30 to 80% of maize yield losses have been estimated to occur in 2.5 million hectares infested by Striga in SSA (AATF, 2006). Global warming and the accompanying increase in unpredictable intensity and frequency of rainfall patterns, in addition to the fact that more than 80% of maize in SSA is rainfed (Edmeades et al., 2017), call urgently for more effective improvement of maize yields under drought stress. Breeding maize for resistance to Striga and tolerance to drought is an effective means of combating these threats and tremendous efforts towards this has resulted in varieties resistant or tolerant to each stress (Badu-Apraku and Fakorede, 2017). However, extra-early orange maize cultivars combining resistance to Striga and tolerance to drought with elevated levels of carotenoids are lacking. 2 University of Ghana http://ugspace.ug.edu.gh Success in breeding maize to address the increasing demand for resistant/tolerant cultivars to adverse environmental changes reside in the availability of maize genetic variability and its efficient use through different methods. To this end, inbred lines are routinely developed from numerous source populations and used to develop new cultivars or base populations. Hence, these inbred lines constitute a sample of the existent maize genetic variability. Exploring this variability aids in achieving breeding goals such as developing new varieties that combine desirable multiple traits. Molecular characterisation of genotypes under selection using SNPs has the advantage of assessing the variability at the base level and provides the basis of differentiating and classifying genotypes into heterotic groups (Yang et al., 2011; Dao et al., 2014; Wu et al., 2016). It also aids in identifying genetic purity of genotypes (Semagn et al., 2012; Ertiro et al., 2017), especially advanced maize inbred lines, to assure their identity and guarantee the heterotic effects in crosses. In order to maximize the gains in selection for increased levels of carotenoids, resistance to Striga, and tolerance to drought in hybrid development, in-depth understanding of inheritance and heterosis for provitamin A carotenoids and the aforementioned major constraints are needed. The main objective of this research was to analyse the genetic tolerance of maize inbred lines and hybrids to Striga and drought. The specific objectives of this research were to study: i. the performance of extra-early maturing maize inbred lines for tolerance to Striga and drought; ii. the molecular diversity in extra-early maturing orange maize inbred lines using SNP markers; iii. the combining abilities of extra-early maturing orange maize inbred lines and heterosis for tolerance to Striga and drought; 3 University of Ghana http://ugspace.ug.edu.gh iv. the performance and stability of extra-early maturing hybrids in tolerance to Striga and drought; v. the combining abilities of extra-early maturing orange maize inbred lines and heterosis for carotenoids in maize kernels. 4 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO 2.0 LITERATURE REVIEW 2.1 The complex “Striga-drought” and maize yield stability in the savannas of West Africa Savanna zones are said to comprise about 75% of the total land mass of West Africa (WA) (Badu- Apraku and Fakorede, 2017). They are, in addition, characterized by higher solar radiation with increased daytime temperature and lower night time temperature (Kassam et al., 1975), which are ideal for increased photosynthesis rate of up to 20-40% (Kassam and Kowal, 1973) compared to the forest agro-ecologies which are characterized by cloudy cover especially during the growing season. These are probably reasons why a large proportions of the maize productions in West and Central Africa (WCA) has shifted to the savanna agro-ecology of WCA (Bolaños and Edmeades, 1996; Badu-Apraku and Fakorede, 2017). Adoption of maize in the savannas has been made possible by the development and availability of extra-early and early-maturing varieties in addition to the fact that maize yields are higher than those of the traditional crops namely, sorghum (Sorghum bicolor L.) and millet (Pennisetum typhoides L.). Despite the appropriate environmental conditions that the savannas offer for increased maize production and productivity, the combined effect of drought and Striga, under field conditions, can be disastrous, thereby hindering food security of the sub-region. Drought and Striga are the major abiotic and biotic constraints to maize production in WCA, especially, in savannas zones of West Africa (WA). They play a key role in determining maize yield stability, productivity, and production in this area of maize production considered as the grain basket of WA. On one hand, more than 80% of agriculture in SSA is rainfed (Edmeades et al., 2017). Consequently, rainfall represents one of the most important factors of all climatic variables in 5 University of Ghana http://ugspace.ug.edu.gh agricultural activity. Updated and precise global yield losses due to drought are lacking. However, it is commonly admitted that drought causes about 17% yield losses per year in maize in the tropics (Edmeades et al., 1999) while losses can approach 60% in individual seasons as reported by Rosen and Scott (1992), in regions such as Southern Africa. On the other hand, under maize cultivation, the most deleterious effects of Striga in SSA have been estimated to occur in 2.5 million hectares with grain losses of 30 to 80%, representing a value of approximately US $1 billion per year (AATF, 2006). Fifteen countries among which eight of eastern and southern Africa (Malawi, Kenya, Tanzania, Zimbabwe, Mozambique, Ethiopia, Uganda, Zambia) and seven of western Africa (Nigeria, Ghana, Benin, Togo, Cameroun, Cote d’Ivoire, Burkina Faso) account for 92% of the continent’s Striga infested maize fields. The levels of infestation are often so high that maize can suffer total yield loss, thus compelling farmers to abandon their fields. The Striga problem is intimately associated with low levels of soil fertility which results in the sub-region from many factors including intensification and the reduced fallow periods (Schmidt-Vogt et al., 1991). In summary, it is clear that Striga and drought individually pose an ominous obstacle to maize production in an area known for its potentially high productivity and production. Moreover, it is obvious that the combined effect of the two stress factors will hamper all efforts at alleviating food insecurity, hunger and hidden hunger in SSA. Breeding for combined drought tolerance and Striga resistance in maize more efficiently under the changing environments become imperative because there are no economically viable technologies to facilitate crop production under these stresses (Farooq et al., 2009). 6 University of Ghana http://ugspace.ug.edu.gh 2.2 Overview of agricultural drought and Striga hermonthica 2.2.1 Definition of drought in agriculture Agricultural drought is defined on the basis of soil water deficit and occurs when there is not enough soil moisture to support crop production (Modanesi et al., 2018). Different models predicted a substantial impact on soil moisture and temperature conditions with increased frequency and duration of droughts as a consequence of climate change (Stocker et al., 2013). In such a context, the vulnerability of agriculture in SSA, a sub-region which is already known to be threatened by recurrent drought, is expected to further increase (Kamali et al., 2018). However, in a previous study, Lobell et al. (2011) reported that an increment of 1°C in mean temperatures will result in about 65% and 100% maize yield losses under well-watered conditions and in drought- prone areas, respectively. The most severe effects of drought stress on maize plant occur when drought coincides with 7 to 10 days period before and after flowering (Edmeades et al., 2017). Before flowering, under stress, ear growth slows down more than tassel growth resulting in a delay in silk emergence relative to pollen shed, giving rise to what is known as anthesis-silking interval (ASI) (Edmeades et al., 2017). When drought occurs at flowering period, complete abortion of ears can be observed which consequently may lead to barrenness. Also, when the stress extends throughout grain filling, reduced number of kernels per ear and poor grain filling are observed (Edmeades et al., 2017). 2.2.2 Overview of Striga hermonthica The genus Striga, previously assigned to the family Scrophulariaceae, is now grouped within the family Orobanchaceae Vent. based on several recent phylogenetic evidence (Spallek et al., 2013). It comprises about 30 obligate root-parasitic annual plants, commonly known as witchweeds presumably because most of their life cycles occur underground, the period during which severe 7 University of Ghana http://ugspace.ug.edu.gh symptoms appear before the parasite emerges (Spallek et al., 2013). Approximately, 80% of the characterized Striga species are known to be endemic to Africa and more precisely, originated from the region between the Semien Mountains of Ethiopia and the Nubian Hills of Sudan (Atera et al., 2011) which is also described as the centre of origin of sorghum (Sorghum bicolor L.). S. hermonthica, S. asiatica, S. gesnerioides, and to some extent S. aspera and S. forbesi Benth are, in decreasing order, the five Striga species of economic importance in SSA which cause serious damage to the sub-region cereal production (Spallek et al., 2013). Striga hermonthica (Delile) Benth. is an obligate hemiparasite. That is, S. hermonthica is photosynthetically active but the establishment of parasitism with a host is essential to its survival (Spallek et al., 2013). It is also obligately allogamous and closely related to S. aspera with which it can intermate and produce viable and virulent offspring (Safa et al., 1984). This exchange and recombination of genetic material between S. hermonthica and S. aspera make their respective populations extremely diverse and continually changing. S. hermonthica infests 57% of the total area in SSA under cereal production (Sauerborn, 1991). In Togo, Mali and Nigeria, maize growing areas infested with S. hermonthica represent about 30–40% (De Groote et al., 2008). This large coverage of infestation which results in about 30–90% yield loss (Musyoki et al., 2015) makes S. hermonthica the most economically important parasitic weed. Yield losses largely depend on Striga density (Mbuvi et al., 2017), host species and genotype, land use system, soil nutritional status and rainfall patterns (Atera et al., 2013). The geographic distribution and level of infestation is projected to steadily increase in SSA (Ejeta and Gressel, 2007) because there are no means to properly control the parasite and also because its seeds are easily propagated due to their tiny size (Atera et al., 2012). Furthermore, habitat suitable for its growth are predicted to expand as a consequence of changes in climatic conditions (Mohamed et al., 2006). 8 University of Ghana http://ugspace.ug.edu.gh Germination of Striga seeds requires a preconditioning phase which is characterized by high temperature (20-40°C), adequate moisture for 7-14 days (Parker and Riches, 1993) and germination stimulants known as Strigalactones which are exuded from a susceptible host. Strigalactones are plant hormones which are synthesized from carotenoids and serve as regulators in several developmental processes to adapt root architecture to nutrient availability. They also play key role in plant-arbuscular mycorrhizal (AM) fungus symbiosis (Charnikhova et al., 2017). In maize, in addition to the classical strigolactones such as 5deoxystrigol (5-DS) and sorgomol (Yoneyama et al., 2015), two chemical structures for the strigolactones were recently described and termed zealactones (Charnikhova et al., 2017). Once a germinated Stiga plant is able to establish a xylem-to-xylem connections with its host, it extracts water and nutrients throughout its life cycle. This results in a range of drought-like and nutrients deficiency symptoms on the host. Commonly, the infested host plants become stunted, show chlorosis, and can even die in severe cases (Dörr, 1997). It has been estimated that each single plant of S. hermonthica can cause 5% yield loss on a host plant (Parker and Riches, 1993). 2.3 Breeding tolerance in maize for drought 2.3.1 Requirements for breeding for tolerance to drought Breeding for drought tolerance, like any other trait, requires two basic things (Edmeades, 2013): genetic variation and selection environment. The genetic variation should be identifiable and heritable within a breeding population. It is usually revealed in the significant levels of interactions between genotype and the stress level in response to drought stress (Edmeades, 2013). The choice of selection environment is made among the target environments and must ensure that stress intensity, timing, and frequency can be reliably managed to expose genetic variation for traits season after season (Barker et al., 2005). However, in most maize target environments, drought 9 University of Ghana http://ugspace.ug.edu.gh occurs randomly and so breeders nowadays prefer the “hotspot” approach in which the stress is managed in the environments through irrigation. Managed drought stress environments are rain- free testing sites where water stress is induced and its timing and intensity are managed through irrigation. 2.3.2 Plant physiological variables addressed in breeding for tolerance to drought Genetic improvement of grain yield in a water-limited environment targets at least one of the three variables reviewed by Passioura (1977) which are: (i) the amount of water captured by the plant (W), (ii) the efficiency with which that water is converted to biomass (water use efficiency, WUE), and (iii) the harvest index (HI), in other words, the proportion of biomass forming grain. For instance, genetic control of root morphology (root depth and root biomass) is known to be the most effective means of increasing W. However, reports have shown that there is genetic variation for root depth and that to increase the amount of water captured by the roots, deeper roots are preferred to roots biomass increase (Edmeades et al., 2006). In addition to increasing the amount of water captured, the more the roots go deeper into the soil, the better the plant stand (resistance to root lodging). Maize breeding also targets delayed leaf senescence also known as “staygreen” characteristic or leaf death as the best way of increasing WUE but, at the same time this trait seems to also assure root health and increases duration of kernel filling (Edmeades, 2013). However, a weak association between grain yield and staygreen characteristic has been reported under drought (Chapman and Edmeades, 1999). 2.3.3 The use of secondary traits Emphasis on secondary traits in breeding for tolerance to drought stems from the fact that heritability and genetic variance for grain yield decline under stress as interplant and interplot variability that occurs under stress increase (Edmeades, 2013). The importance of the secondary 10 University of Ghana http://ugspace.ug.edu.gh traits can be revealed by examining their genetic correlations with GY under drought stress, or by estimating their correlated response after selecting for grain yield under stress (Edmeades et al., 1997), and also through path analysis (Badu-Apraku et al., 2011; Talabi et al., 2017). Edmeades et al. (2017) defined an appropriate secondary trait as a trait that is (1) genetically associated with grain yield under drought; (2) highly heritable; (3) stable and feasible to measure; and, (4) not associated with yield loss under ideal growing conditions. Commonly used secondary traits under drought are reduced barrenness, ASI, staygreen characteristic, and leaf rolling (Edmeades et al., 2017). The barrenness is expressed in terms of number of ears per plant (EPP) while the staygreen characteristic is assessed through senescence. Some other traits such as osmotic adjustment and ability to remobilize stem reserves are also often used in studies of drought tolerance mechanism (Ludlow and Muchow, 1990). However, in a breeding programme, the criteria of cost and easy to be measured attached to a trait is of great importance and most of these traits do not meet (Chapman and Edmeades, 1999). It has also been pointed out that, traits that are known to increase the current photosynthesis rate are more valuable than that related to remobilization of the reserves (Schussler and Westgate, 1995; Westgate et al., 1997). Furthermore, development of tassel, roots, and stems during flowering and grain filling periods can negatively affect ear development and therefore, decrease in their demand of resources during the periods can favour the development of ears and result in increased grain yield under stress. In addition to the commonly used secondary traits, potential traits to be considered in breeding for tolerance to drought in maize are disease resistance and ear aspect. Because the capability of a plant to withstand any stress decreases when diseased while ear aspect directly affects grain yield. 11 University of Ghana http://ugspace.ug.edu.gh 2.4 Breeding for resistance/tolerance in maize to Striga 2.4.1 Genetics of resistance to Striga 2.4.1.1 Resistance/tolerance in maize With regard to the type of mechanism involved in the interaction between maize plant and Striga, three types of genotypes can be distinguished (Berner et al., 1995). Low stimulators are genotypes that are less efficient in stimulating Striga to germinate. They are also known as pre-attachment resistant genotypes. Their resistance may involve the lgs1 gene found in sorghum (Gobena et al., 2017). This gene is located in a conserved region; therefore, studies are underway to use reverse genetics to find lgs1 in maize. Genotypes that stimulate Striga to germinate and allow it to attach, but slow its growth, delay its emergence, and reduce its vigour are known as post-attachment resistant genotypes. Finally, genotypes that stimulate Striga to germinate and allow it to attach, grow, and reproduce normally, but do not suffer much from the “intoxication effect” are known as tolerant genotypes. In field conditions, resistance is known to be associated with number of emerged Striga plants and yield under infestation while tolerance is associated with Striga damage, number of ears per plant, and yield under infestation. These traits have been combined together in a base index for selecting performant genotypes under Striga infestation, in the IITA Maize Improvement Programme (MIP). 2.4.1.2 Gene action/ inheritance of resistance to Striga Studies on the genetics of maize to Striga hermonthica have yielded contradictory reports on the gene action involved in maize resistance to Striga even though all agreed that resistance is quantitatively inherited. For instance, Kim (1994); Berner et al. (1995); and Badu-Apraku (2007) found that additive gene effects were more important than nonadditive gene effects for host plant damage and grain yield under Striga infestation. On the contrary, Kim (1991) and Gethi and Smith 12 University of Ghana http://ugspace.ug.edu.gh (2004), for example, concluded that nonadditive gene effects were more important than additive gene effects for host plant damage while additive gene effects were predominant for number of emerged Striga plants. Moreover, moderate heritability estimates were reported for host plant damage and grain yield while low heritability estimates were reported for Striga emergence under S. hermonthica infestation (Badu-Apraku et al., 2007). For instance, Akanvou et al. (1997) estimated narrow-sense heritability of 0.33, 0.14, and 0.32, respectively for Striga damage, number of emerged Striga plants, and grain yield under Striga infestation. 2.4.1.3 Genes sources At the beginning of the research in maize resistance to Striga hermonthica in IITA, the initial screening for resistance allowed identification of resistant inbred lines among the existing germplasm which were used as resistant gene sources to improve two composites : TZL Comp.1 and TZE Comp.5 (Badu-Apraku and Fakorede, 2017). These two materials have been the basis of the great achievement in developing and delivering resistant open-pollinated varieties, inbred lines and hybrids in all maturity groups. Furthermore, resistance identified in the perennial teosinte (Zea diploperennis) and in improved temperate maize have been used to increase the genes for resistance to Striga, which has been found to be quantitatively inherited. 2.5 Vitamin A deficiency in SSA and Biofortification in maize 2.5.1 The scope of the problem The incidence of vitamin A (VAD) deficiency was estimated to be among the highest (48%) in SSA with an occurrence of 95% of the reported death from diarrhea and measles in 2013 (Stevens et al., 2015). For instance, more than 7.2 million pregnant women and approximately 127 million pre-school aged children have been reported vitamin A deficient in developing countries (West , 2002). About 5-10 million children develop xerophthalmia from VAD every year and up to 13 University of Ghana http://ugspace.ug.edu.gh 500,000 among them become blind (Sommer, 1995). Also, VAD account for more than 600,000 of childhood deaths (West and Darnton-Hill, 2008). Children affected by VAD often develop corneal blindness and their growth is retarded (West and Darnton-Hill, 2008). VAD also affects the immune system and can therefore result in several diseases such as measles and diarrhea which can lead to an increased risk of mortality (Rice et al., 2004). Two main reasons explain the high prevalence indices of VAD in developing countries: the almost exclusive consumption of white maize compared to the yellow maize and poverty. For human consumption in most parts of Africa there is a high preference of white maize compared to yellow maize (De Groote et al., 2008, Pillay et al., 2011). Also, during hunger periods the imported relief food were mainly from yellow maize, therefore, people perceive yellow maize to be inferior to white maize in addition to its unfavourable taste and texture (Muzhingi et al., 2008). However, several studies have proved the contrary. For instance, in a study conducted in Mozambique, consumers preferred orange maize to white maize because of its aroma (Stevens and Winter- Nelson, 2008). More recently, another study conducted in rural Zambia demonstrated that when nutrition information is properly provided, consumers show higher preference to orange than white maize (Meenakshi et al., 2012). These studies in summary highlight the importance of appropriate consumer education in the efforts towards alleviation of VAD through adoption and consumption of orange maize in Africa. Many other strategies such as supplementation, food fortification, and diet diversification have been effective in addressing VAD (Mora, 2003) their impacts remain negligible due to the very high rate of poverty and other complicating factors in developing countries (Graham et al., 2001). 14 University of Ghana http://ugspace.ug.edu.gh 2.5.2 Provitamin A biofortification in maize Biofortification is a cost-effective, sustainable, and long-term means of breeding nutrients into food crops. It can help to alleviate VAD more efficiently because a provitamin A biofortified staple crop like maize will increase the daily micronutrient intakes throughout the lifecycle of individuals (Bouis et al., 2011). Currently, agronomic, conventional breeding, and transgenic biofortification are three common approaches used. However, the conventional breeding biofortification is the most used in maize as regard the provitamin A because of its naturally great diversity in provitamin A carotenoids content. Maize is a carotenogenic species with a known genetic diversity of carotenoids content and profiles (Burt et al., 2011). Carotenoids represent more than 600 characterized structures of pigments synthesized in plant (Messias et al., 2015). They are essential for growth and development, play key roles in the process of photosynthesis, protect plants against photooxidative damage, and are also precursors of abscisic acid synthesis (Gallagher et al., 2004). Carotenoids are responsible for the observed range of colours (light yellow to dark orange) in maize kernels. They are primarily concentrated in the vitreous portion of the endosperm (Weber, 1987). Several studies found association between darker orange colour and higher total carotenoids, particularly lutein and zeaxanthin in maize (Burt et al., 2011; Almeida Rios et al., 2014) but not necessarily with higher provitamin A carotenoids (Harjes et al., 2008). However, the estimated β-carotene level in majority of yellow maize is on average 1.7 µg g-1 (Harjes et al., 2008). The phytoene formation from geranylgeranyl pyrophosphate is the first step in biosynthesis of carotenoids. This first step is mediated by phytoene synthase (PSY). Three genes are known to encode PSY in maize. They are psy1, psy2, and psy3. The gene psy1 also known as Y1 has been reported to highly determine carotene synthesis and levels of carotenoids in the maize endosperm 15 University of Ghana http://ugspace.ug.edu.gh (Li et al., 2008). For instance, white endosperm phenotype was observed to be due to a loss-of- function allele of psy1 which prevents accumulation of carotenoids in the endosperm (Gallagher et al., 2004). Also, high polymorphism has been reported for the psy1 in different varieties of maize (Palaisa et al., 2003). The triploid maize endosperm (3n = 30) results in four possible phenotypic classes at the Y1 locus depending on the number of dominant (Y1) and recessive (y1) alleles which are: y1y1y1, y1y1Y1, y1Y1Y1, and Y1Y1Y1. 2.5.3 Achievement towards conventional breeding for carotenoids in maize The advent of genomics and bioinformatics have enabled the identification of additional genes in the maize carotenoid biosynthetic pathway (Wurtzel et al., 2012). For example, Owens et al. (2014) performed a genome-wide association study in a panel of maize inbreds ranging from light yellow to dark orange in grain colour and found the existence of 58 candidate genes involved in biosynthesis of carotenoids in maize. Among them, eight candidate genes, y1, zds1, lcyE, crtRB3, lut1, crtRB1, zep1, and ccd1, have been reported to play key roles in carotenoids synthesis and accumulation and are all in chromosome regions associated with QTL for carotenoids (Chander et al., 2008; Kandianis, 2010; Chandler et al,. 2013). Except for crtRB3 and lut1, these genes were also associated with QTL for the intensity of the orange colour (Chandler et al., 2013). Several studies revealed significant allelic variation for lycopene epsilon cyclase (lcyE) (Harjes et al., 2008) and β-carotene hydroxylase 1 (crtRB1) (Yan et al., 2010) which are known to significantly influence synthesis and accumulation of carotenoid in maize grains. There are, respectively, four and three polymorphism sites at lcyE and crtRB1 known to be correlated with high levels of β- carotene. Variation at crtRB1 include 5’TE, InDel4, and 3’TE with the largest effect attributable to the rare allele (insertion of 206 bp at 5’TE), found only in temperate germplasm at 2.9% (Yan et al., 2010). 16 University of Ghana http://ugspace.ug.edu.gh Babu et al. (2013) validated the effects of 2 polymorphisms (lcyE 5′TE, lcyE 3′Indel) of lcyE and the crtRB1-3′TE of crtRB1 in 26 diverse tropical genetic backgrounds. When the transcript level of lcyE is reduced (favourable alleles), there is an increase of β-branch carotenoids. Also, favourable alleles of crtRB1 with reduced transcript levels decrease hydroxylation of β-carotene. The increase of β-branch carotenoids and the decrease in hydroxylation of β-carotene result in higher provitamin A carotenoids levels in maize kernels (Harjes et al., 2008; Yan et al., 2010). For example, the levels of β-carotene and total provitamin A content have been shown to increase to up to two-tenfold as a result of CrtRB1-3′TE alone while favourable alleles at lcyE can reduce up to 30% the ratio of levels of α- to β-branch. Azmach et al. (2013) also reported higher levels of provitamin A in tropical inbreds that possessed favourable alleles at the crtRB1-5′TE and 3′TE. Fraser and Bramley (2004) showed that lcyE controls the zeaxanthin/lutein ratio and that it is a key gene determining the provitamin A content in maize. The use of markers in selection for carotenoids has been shown to be effective. Indeed, the use of markers with appropriate breeding scheme have moved the existing variability of β-carotene, 0.24 to 8.80 μg g−1 , (Ortiz-Monasterio et al., 2007, p.) to levels matching or exceeding the breeding target in tropical maize inbreds. Up to 13.6 μg g−1 of β-carotene was reported with favourable alleles at lcyE (Harjes et al., 2008). Combinations of favourable alleles at both lcyE and crtRB1 allowed increases up to 17.25 μg g−1 (Azmach et al., 2013), from 15 to 20 μg g−1 (Babu et al., 2013) and more recently, up to 22.6 μg g−1 (Menkir et al., 2017) mostly in inbred lines. Despite efforts that have been made to address the provitamin A deficiency in developing countries through the development of maize cultivars with a significant increase of provitamin A content, the currently released cultivars contain only 6 to 8 μg g−1 of provitamin A (HarvestPlus, 2017). 17 University of Ghana http://ugspace.ug.edu.gh Furthermore, the predominance of additive genetic effects and high heritability were reported (Egesel et al., 2003) and yet, OPVs with high-level of provitamin A are lacking. On the contrary, Halilu et al. (2016) found non-additive genetic effects to be predominant for all carotenoids and their finding supports that of Burt et al. (2011) who reported heterosis in carotenoids to be a rare phenomenon. The controversial report on the genetic basis underlying carotenoids accumulation in maize kernels suggest that different gene actions exist depending on the genetic background of the provitamin A materials. Moreover, the cost attached to the use of makers and the High- Performance Liquid Chromatography (HPLC) based carotenoids quantification have led to the extensive use of colour rating in some breeding programmes as an alternative. In fact, maize grain colour, or more precisely, endosperm colour have been reported to be highly correlated with the accumulation of β-carotene (Johnson and Miller 1938; Chander et al., 2008) and more recently, with β-cryptoxanthin and zeaxanthin (Venado et al., 2017). Conversely, Egesel et al. (2003); Harjes et al. (2008) reported no correlation between endosperm colour and β-carotene accumulation in maize. Even though Chander et al. (2008) recommended the use of colour rating along with specific markers for successful breeding, it is clear that its exclusive use could create controversy. 2.6 Molecular characterisation in plant breeding Genetic variability is a prerequisite for a successful maize breeding to address the increasing demand for resistant genotypes to adverse environmental changes and consumer preferences. To this end, inbred lines are routinely developed from numerous source populations improved through different methods (Hallauer, 1990). These developed inbred lines are used subsequently to form new cultivars or base populations, constitute a sample of the genetic diversity of maize. Thus, 18 University of Ghana http://ugspace.ug.edu.gh genetic variability assessment among inbred lines has become an unavoidable step in breeding, prior to performing hybrids and maize cultivars development. Many tools have been used to characterize and assess diversity in maize inbred lines. But nowadays, the availability and accessibility at relatively low cost of Single Nucleotide Polymorphism (SNP) markers has popularized their use in different molecular studies. SNPs are discovered by different platforms among which the most commonly used are genotyping by sequencing (GBS) and diversity array technology sequencing (DArTSeq), which are all sequence- based technologies. Pro and cons of each of the two technologies have been highlighted by Chen et al. (2016). GBS generates a very high density of markers (>800,000 SNPs) compared to DArTSeq which yields markers with a number ranging from 50,000 to 350,000 SNPs (Sansaloni et al., 2011). On the contrary, with regard to the coverage of the genotyping, DArTSeq have much higher coverage than GBS (<0.5X) and lower levels of missing data (<20%) in comparison with GBS (> 50%), in the case of maize. Beside these differences, the aforementioned methods have significantly contributed to today’s large-scale use of SNP makers in a range of studies for investigating the molecular basis of phenotypic variations in plants such as diversity and association mapping studies, and genomic selection. Molecular-based diversity study using SNPs has the advantage of exploring the variation between genotypes at the base level and provides a means of differentiating cultivars and classifying inbred lines into heterotic groups (Yang et al., 2011; Dao et al., 2014; Wu et al., 2016); identifying gaps and redundancy in germplasm collections (Semagn et al., 2012; Ertiro et al., 2017). It also allows the understanding of the genetic changes that occur in the process of germplasm conservation, or regeneration, or during breeding. Furthermore, molecular-based diversity study is a means of identifying novel and superior alleles for improvement of agronomic traits. 19 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE 3.0 Performance of extra-early maturing maize inbred lines under Striga-infested, drought stress and optimal environments 3.1 Introduction Stress from Striga infestation and drought constitutes the most important biotic and abiotic factors limiting maize production and productivity in SSA. Developing cultivars for resistance to these stresses in order to stabilize maize yield production in the sub-region is the major strategy of the Maize Improvement Program (MIP) at the International Institute of Tropical Agriculture (IITA). Per se performance evaluation of inbred lines is an important step in plant breeding. It allows the identification of elite lines to be used as parents in planned crosses to develop promising hybrids. But, direct selection for improved performance under stresses, such as Striga and drought, based on grain yield alone has been reported inefficient (Edmeades, 2013). Therefore, efficient improvement of maize resistance/tolerance to these stresses has often been based on the use of highly heritable secondary traits for which genetic variability increases under stress conditions (Edmeades et al., 2017). Alternatively, the use of breeding values through the numerator relationship matrix derived either from pedigree (A) or molecular markers (G) has been proposed for predicting the performance of genotypes that would be used as parents for developing new genotypes (Piepho et al., 2008). Several studies have proven the importance of the genetic variance-covariance structure in predicting genotype performance. Bromley et al. (2000) showed, for instance, that for inbred lines of maize ignoring pedigree relationships results in a reduction in estimates of genetic variance. The objectives of this study were to (i) determine the performance of extra-early maturing maize inbred lines for resistance to Striga and tolerance to drought, and (ii) predict their breeding values using the correlation among their genetic variances. 20 University of Ghana http://ugspace.ug.edu.gh 3.2 Materials and Methods 3.2.1 Genetic materials The genetic material used in the present study consisted of 152 orange and 28 yellow extra-early (80-85 days to maturity) maturing maize inbred lines. The orange inbreds were selected out of 253 inbreds developed from 2009 TZEE-OR1 STR in IITA-Maize Improvement Programme (MIP) based on their performance under Striga infestation and drought stress during 2014 and 2015 field evaluations. The 28 yellow inbreds including five checks (TZdEEI 1, TZdEEI 7, TZdEEI 9, TZdEEI 12, and TZdEEI 13) were also developed in IITA-MIP from different populations. 3.2.2 Experimental design and field trial management 3.2.2.1 Experimental design Three different field trials with regard to the management were conducted using a 12 x 15 alpha lattice design with two replications. Single-row plots each 3 m long with a spacing of 0.75 m between two adjacent rows and 0.40 m between plants within the row were used. Three seeds were planted per hill, and the seedlings were later thinned to two per hill about 2 weeks after emergence to give a final population density of about 66,667 plants ha-1. 3.2.2.2 Management of drought trials The drought trials were evaluated at Ikenne (6º53’N, 3º42’E, 60 m altitude, 1200 mm annual rainfall), under managed drought during the January-May dry seasons of 2016/2017 and 2017/2018. The managed drought was achieved using irrigation system. The field was irrigated using sprinkler irrigation system for the first three weeks. Plants were irrigated using a sprinkler irrigation system, which provided 17 mm of water each week. Then, from 22 days after planting (DAP) until maturity, drought was induced by withdrawing irrigation water. 21 University of Ghana http://ugspace.ug.edu.gh 3.2.2.3 Management of trials under Striga infestation The Striga-infested trials were evaluated at Mokwa (9°18’ N, 5°4 E, 457 m altitude, 1100 mm rainfall) and Abuja (9°16’ N, 7°20’ E, 300 m altitude, 1500 mm rainfall) in the southern Guinea Savanna of Nigeria during the rainy season in 2016 and 2017. The artificial infestation of the fields was done following the method developed by IITA’s maize programme (Kim, 1991; Kim and Winslow, 1991). Striga seeds were collected from sorghum fields around the experimental sites at the end of the growing season. The cleaned, sieved and preconditioned Striga seeds were mixed with finely sieved sand in the ratio of 1:99 (seed: sand) by weight, at planting. The scoops, which ensure about 5000 germinable Striga seeds when filled, were used to infest each planting hole. To avoid the negative effect of fertilizers on the germination ability of the Striga seed, fertilizer application was delayed until about 30 days after planting. The fertilizer NPK 15-15- 15 was used at 30 kg N ha-1, 26 kg P ha-1, 50 kg K ha-1. Hand weeding was used to control weeds other than Striga. 3.2.2.4 Management of trials under rain-fed conditions Trials planted under rain-fed conditions were evaluated at Abuja (2016), Mokwa (2016 and 2017), and Ikenne (2016, 2017) during the June-October rainy season of each year. In these trials, nitrogen was applied at 90 kg N ha-1, which is considered as the optimal level of nitrogen under maize cultivation. Therefore, these trials were referred to as “optimal” in the rest of this thesis. The optimal trials received at each location, 60 kg N ha-1, 60 kg P ha-1 and 60 kg K ha-1 at 2 weeks after planting (WAP) with an additional 30 kg N ha-1 top-dressed at 4 WAP. In all trials except those under Striga infestation, Primextra (active ingredient: Atrazine) as pre- emergence and Paraquat (active ingredient: Gramoxone) as post-emergence herbicides were used to control the weeds. The application of these herbicides was done at 5 L/ha each. 22 University of Ghana http://ugspace.ug.edu.gh 3.2.3 Data collection and measured traits 3.2.3.1 Flowering and plant morphological traits The number of days from planting to when 50% of the plants had emerged silks and had shed pollen was recorded as days to silking (DYS), and days to anthesis (DYA), respectively. The difference between DYS and DYA was calculated as the anthesis-silking interval (ASI). The plant and the ear heights were measured as the distance from the base of the plant to the height of the first tassel branch (PLHT) and the node of insertion of the upper ear (EHT). The overall architecture and the appeal to sight of plants in a plot were assess using a scale of 1 to 9 and recorded as plant aspect (PASP) where 1 represents excellent overall phenotypic appeal and 9 represents extremely poor overall phenotypic appeal. The percentage of plants leaning more than 30° from the vertical, in a plot was recorded as root lodging (RL) while the percentage of plants broken at or below EHT was recorded as stalk lodging (SL). 3.2.3.2 Ear related traits The extent to which ears are covered by the husk was rated using a scale of 1 to 9 as husk cover (HC) where 1 represents ears with very tight husk extending beyond the tips and 9 represents ears with exposed tips. The overall appearance of the ears without the husks was assessed as ear aspect (EASP) using a scale of 1 to 9 where 1 represents uniform and fully filled ears with no diseases and 9 represents the case where no ear is produced. The number of ears per plant (EPP) was calculated by dividing the total number of ears harvested per plot by the number of plants in a plot at harvesting. 23 University of Ghana http://ugspace.ug.edu.gh 3.2.3.3 Grain yield estimation For the drought trials, the grain yield was adjusted to 15% moisture and computed as follows: (100−m) 10000 GY = Gwt × × 85 (rwl×Φ) Where, GY = Grain yield in kg ha-1, Gwt= shelled grain weight per plot (kg), m = grain moisture content at harvest, rwL = row length in m, Φ = inter-plant spacing in a row. The grain yield in optimal and Striga-infested environments was also adjusted to 15% moisture and computed assuming 80% of shelling percentage for the genotypes as follows: (100 − m) 10000 GY = fwt × × × 0.8 85 (rwL × Φ) Where, GY = Grain yield in kg ha-1, fwt= ears weight at harvest per plot, in kg, m = grain moisture content at harvest, rwL = row length in m, Φ = inter-plant spacing in a row. 3.2.3.4 Striga resistance and drought tolerance indicator traits The staygreen characteristic (STGC) was rated under drought stress at 70 days after planting (DAP) using a scale of 1 to 9 where 1 is when dead leaf area represents up to 10% of the total leaf area and 9 is when dead leaf area falls between 90 to 100%. Under Striga-infestation, the damage syndrome on the plants in a plot and the number of emerged Striga plants were recorded at 8 and 10 WAP. The Striga damage syndrome was rated on plot basis using the scale of 1 to 9 described by Kim (1994). 3.2.4 Data analysis 3.2.4.1 Model description The data was analysed using the mixed model equation (Equation 3.1) in ASReml-R (Butler et al., 2009; Gilmour et al., 2015) statistical package. All the factors in the model were considered as random. 24 University of Ghana http://ugspace.ug.edu.gh The mixed model equation is given as follow: Y = Xβ + ZEnvuEnv + ZEnv:RepuEnv:Rep + ZEnv:Rep:BlkuEnv:Rep:Blk + ZPeduPed + ZEnv:Ped uEnvxPed + e (Equation 3.1) Where: Y is the vector of unadjusted observations. The X-matrix and Z-matrices are incidence matrices belonging to their respective components, β is a vector of intercepts. uEnv, uEnv:Rep, uEnv:Rep:Blk and uPed are vectors of random effects of environments, replications within environments, blocks within replications and genetic effects, respectively. uEnvxPed and e are respectively the interaction terms of the genetic effect with environments and the random error. The random effects in the model were assumed to follow a multivariate distribution with means and variances defined respectively by the equations 3.2 and 3.3. uEnv 0 u Env:Rep 0 u 𝐸 Env:Rep:Blk = 0 (Equation 3.2) uPed 0 uEnvxPed 0 [ 𝑒 ] [0] u 2Env 𝜎 0 0𝐸𝑛𝑣 0 0 0 u 2 Env:Rep 𝜎 0 𝑅𝑒𝑝 0 0 0 0 2 u 𝜎 0 0 0 𝑉𝑎𝑟 Env:Rep:Blk = 0 0 𝐵𝑙𝑘 (Equation 3.3) uPed 0 0 0 𝜎 2 𝑃𝑒𝑑 0 0 u 𝑛 2EnvxPed 0 0 0 0 𝐼𝐸𝑛𝑣 ⊗𝑗 𝐼𝑃𝑒𝑑𝜎𝑃𝑒𝑑 0 [ 𝑒 ] [ 0 0 0 0 0 𝑅′] Where ⨂ represent the Kronecker direct product (Cullis and Gleeson, 1991). 3.2.4.2 Random error variance structure A heterogeneous variance structure of the random error was fitted as a direct sum of the identity variance matrices with dimension the number of environments using the “at” variance model function in ASReml-R. 25 University of Ghana http://ugspace.ug.edu.gh The heterogeneous variance structure is given by the equation 3.4. 𝜎2𝑒𝑗 ⋯ 0 𝑅′ =⊕𝑛𝑗 𝐼𝐸𝑛𝑣𝜎 2 𝑒 = [ ⋮ ⋱ ⋮ ] (Equation 3.4) 0 ⋯ 𝜎2𝑒𝑛 Where ⨁ represents the Kronecker direct sum (Cullis and Gleeson, 1991). 3.2.4.3 Estimation of additive genetic variance The genetic relatedness between inbreds was used in the genetic variance structure through pedigree information. In this case, the numerator relationship matrix (A) was created by ASReml- R based on the pedigree file. The A matrix indicates the additive relationship between the inbreds and allows accurate prediction of breeding values. The diagonals of the matrix are 1+F, where F is the inbreeding coefficient, and off-diagonals are twice the kinship coefficients. Assuming absence of dominance and epistasis, the variance-covariance matrix for genetic effects (breeding values) used is given by the equation 3.5. 𝜑𝑖𝑖 ⋯ 𝜑𝑖𝑗 𝜎2𝑃𝑒𝑑 = 𝐴. 𝜎 2 ⋮ ⋱ ⋮ 2 𝑎 = [ ] 𝜎𝑎 (Equation 3.5) 𝜑𝑗𝑖 ⋯ 𝜑𝑖𝑖 Where 𝜎2𝑎 is the additive genetic variance; 𝜑𝑖𝑖 = 1 + 𝐹 th 𝑖 with 𝐹𝑖 = inbreeding coefficient of the i genotype; 𝑎𝑖𝑗 = 𝑎𝑗𝑖. In the analysis of data from each trial type (Striga-infested, drought stress and optimal), the variance-covariance structure in equation 3.5 was used to fit interactions of genetic effects with environment. But, in the combined analysis for all test environments, the loglikelihood of the model fails to converge when the equation 3.5 was used to fit interactions. So, the identity matrix (id (ped)) was rather used assuming independence among individual genotypes. 26 University of Ghana http://ugspace.ug.edu.gh 3.2.4.4 Significance test of random factors To determine the significance of a random factor in the model, the log-likelihoods of models with and without the appropriate random factor were computed. The p-value of each random factor was then computed using a chi-square test of a statistic equals to twice the difference in likelihoods with degrees of freedom equal to the number of additional parameters in the more complex model. 3.2.4.5 Estimation of narrow sense heritability and breeding values The narrow-sense heritability was calculated on entry mean basis unbiased by environment, replicates, and blocks variances using the additive genetic variance estimated through the A matrix (section 3.2.4.3). The formula used is given by Equation 3.6 and was implemented in ASReml-R using the “pin” function. σ2 σ2 h2 = a = a2 2 2 𝑛 2 (Equation 3.6) σ σ2 σ ∑ σP σ2+ Env Rep σBlk 𝑗 e a + + +n r b n Where n, r, and b are respectively the number of environments, replicates and blocks; σ2a is the additive genetic variance estimated through the numerator relationship matrix (A). The Best linear unbiased predictions (BLUPs) based on pedigree information (breeding values) and BLUPs based on multivariate analysis were extracted using the “Predict” function in ASReml- R. 27 University of Ghana http://ugspace.ug.edu.gh 3.2.4.6 Base index for selection and correlation among measured traits The base indices developed by Menkir and Kling (2007) were used for selection of inbred lines under Striga infestation (Equation 3.7) and drought stress (Equation 3.8). IS = 2 × GY + EPP − (SDR8 + SDR10) − 0.5 × (ESP8 + ESP10) (Equation 3.7) Where, IS is the base index under Striga infestation, GY is grain yield under Striga infestation, EPP is number of ears per plant, SDR8 and SDR10 are Striga damage at 8 and 10 WAP, ESP8 and ESP10 are number of emerged Striga plants at 8 and 10 WAP, and ID = 2 × GY + EPP − ASI − PASP − EASP − STGC (Equation 3.8) Where, ID is the base index under drought stress, GY is grain yield under drought; EPP, is ears per plant; STGC is staygreen characteristic; ASI, is anthesis-silking interval; PASP, is plant aspect; and EASP, is ear aspect. The genetic correlation coefficients among traits were calculated using META macro in SAS v9.4 (Vargas et al., 2013) in order to validate the use of the traits involved in the base indices. 3.3 Results 3.3.1 Genetic variability among inbreds for the measured traits Table 3.1 presents the variance components and narrow-sense heritability estimates for measured traits of the inbreds across Striga-infested environments. The analysis indicated that the genotype- environment interaction (GEI) variance was not significantly different from zero for most of the traits including grain yield except for the traits with high sensitivity to environmental changes such as anthesis-silking interval, plant, and ear heights. The GEI effects for Striga damage at 8 and 10 weeks after planting (WAP), and ear aspect were found to be significantly different from zero at α (p<0.05 and p<0.001), despite the variance among the two environments was not significant. 28 University of Ghana http://ugspace.ug.edu.gh On the contrary, the variability among inbreds was significantly different from zero (p<0.001) for all traits, indicating a high discrepancy in the response to Striga infestation probably due to the observed differences in the correlated additive genetic variances. The additive genetic variance contribution to the total variance ranged from zero for root lodging to 72% for days to silking. It accounted for 70% for grain yield, and 69% and 67% for Striga damage at 8 weeks after planting and 10 weeks after planting, respectively. The proportion of genotype x environment interaction varied from zero to 49% for ear height. In general, environmental contribution to the total variation was low with the highest observed for anthesis-silking interval (20%) and root lodging (33%). Narrow-sense heritability estimates varied from zero for ear height and root lodging to 48% for grain yield and days to 50% anthesis. The proportion of variation due to additive genetic variance among inbreds was moderate for Striga damage at 8 WAP (42%) and 10 WAP (45%) but low for number of emerged Striga plants at 8 and 10 WAP with values of 11% and 14%, respectively. 29 University of Ghana http://ugspace.ug.edu.gh Table 3.1 Variance components and heritability estimates of measured traits of extra-early maturing maize inbreds under Striga infestation at Abuja and Mokwa, 2016-2017 Anthesis silking Days to 50% Days to 50% Plant height, Source of variation Grain Yield, kgha-1 Interval silking anthesis cm Environment (Env) 10550.69*** (0.01 ± 0.02) 2.81*** (0.20 ± 0.26) 8.99*** (0.13 ± 0.18) 3.01*** (0.06 ± 0.10) 0.00*** (0.0 ± 0.0) Replication (Env) 242.04ns 0.21** 2.10*** 0.92** 137.95*** Block (Env x Rep) 44337.56*** 0.00ns 0.34ns 0.42* 0.00ns Entry (ped) 1067534.9*** (0.70 ± 0.07) 4.28*** (0.31 ± 0.14) 47.08*** (0.67 ± 0.14) 34.58*** (0.72 ± 0.09) 94.85*** (0.09 ± 0.13) Env x Entry 117192.6ns (0.08 ± 0.05) 2.62*** (0.19 ± 0.11) 0.00ns (0.0 ± 0.0) 0.00ns (0.0 ± 0.0) 354.18** (0.35 ± 0.14) Env_1 Residual 307022.29 4.81 15.77 13.32 581.87 Env_2 Residual 248202.67 2.87 8.55 4.69 264.40 ℎ2𝐴 0.48±0.07 0.19±0.07 0.46±0.07 0.48±0.06 0.05±0.06 Ear height, Striga damage Striga damage Number of emerged Number of emerged Source of variation Cm at 8 WAP At 10WAP Striga plants 8 WAP Striga plants 10 WAP Environment (Env) 0.00*** (0.0 ± 0.0) 0.00***(0.0 ± 0.0) 0.03*** (0.01 ± 0.03) 0.00*** 0.00*** (0.0 ± 0.0) Replication (Env) 21.55*** 0.00ns 0.00ns 14.52*** 21.80*** Block (Env x Rep) 0.00ns 0.08** 0.08*** 4.92*** 7.61*** Entry (ped) 0.00*** (0.0 ± 0.0) 2.47*** (0.69 ± 0.07) 2.91*** (0.67 ± 0.07) 37.26*** (0.18 ± 0.05) 58.79*** (0.33 ± 0.09) Env x Entry 124.23*** (0.49 ± 0.09) 0.22***(0.06 ± 0.05) 0.45* (0.10 ± 0.06) 0.00ns 0.00ns (0.0 ± 0.0) Env_1 Residual 133.36 0.83 0.94 124.91 128.62 Env_2 Residual 81.25 0.83 0.73 21.13 48.81 ℎ2𝐴 0.0±0.0 0.42±0.06 0.45±0.06 0.11±0.03 0.14±0.04 Source of variation Root logging Stalk logging Husk cover Ear per plant Ear aspect Environment (Env) 75.53*** (0.33 ± 0.4) 0.00ns (0.0 ± 0.01) 0.00ns (0.0 ± 0.0) 0.03***(0.11 ± 0.14) 0.00***(0.0 ± 0.02) Replication (Env) 32.78*** 0.00ns 0.00ns 0.00ns 0.00ns Block (Env x Rep) 1.94ns 5.36ns 0.02ns 0.01** 0.18*** Entry (ped) 0.00ns (0.0 ± 0.0) 35.02ns (0.15 ± 0.10) 1.32*** (0.58 ± 0.08) 0.17*** (0.59 ± 0.11) 2.66*** (0.55 ± 0.08) Env x Entry 0.75ns (0.0 ± 0.01) 0.00ns (0.0 ± 0.0) 0.17ns (0.08 ± 0.07) 0.00ns (0.0 ± 0.0) 0.18*(0.04 ± 0.04) Env_1 Residual 195.54 197.42 0.39 0.06 1.09 Env_2 Residual 43.05 193.61 1.14 0.09 2.59 ℎ2𝐴 0.0±0.0 0.04±0.03 0.29±0.06 0.34±0.06 0.26±0.06 “*”, “**”, “***”, = significance at p(𝛼 =0.05), p(𝛼 =0.01), and p(𝛼 =0.001), respectively; “ns” = non-significant. Values in parenthesis are contributions of different sources of variance to the total variance with their standard errors; ℎ2𝐴 = narrow-sense heritability. 30 University of Ghana http://ugspace.ug.edu.gh Table 3.2 presents the variance components and narrow-sense heritability estimates for measured traits of the inbreds across drought environments. Results indicated a similarity between the two years, with respect to the management of the drought stress, which resulted in GEI variance being non-significantly different from zero for grain yield, plant aspect, number of ears per plant, and staygreen characteristic. Furthermore, the variance due to environmental effect was low with the highest contribution being 10% for ears per plant and staygreen characteristic and the genotype- environment interaction with a range of zero to 29% for anthesis-silking interval. Similar to Striga- infested environments, the inbreds showed different responses to drought stress with the variance of the additive genetic component being significantly (p<0.001) different from zero. The contribution of the additive genetic variance to the total variation ranged from 25% for root lodging to 77% for ear height. The contribution of the additive genetic variance to the total variation was high for grain yield (73%), and moderate for ear aspect (52%), staygreen characteristic (41%), ears per plant (41%), plant aspect (33%), and for anthesis-silking interval (26%). Narrow sense heritability estimates ranged from 8% for root lodging to 47% for ear height. Moderate heritability estimates were observed for grain yield (41±7%), staygreen characteristic (25±7%) and ear aspect (25±6%) while low additive genetic effects were found for anthesis-silking interval (12±6%), plant aspect (13±4%), and number of ears per plant (16±5%). 31 University of Ghana http://ugspace.ug.edu.gh Table 3.2 Variance components and heritability estimates of measured traits of extra-early maturing maize inbreds under drought stress at Ikenne during the 2016/2017 and 2017/2018 dry seasons Source of Grain Yield, Days to 50% Days to 50% Anthesis silking Plant height, Ear height, variation kgha-1 silking anthesis Interval Cm cm Environment (Env) 0.005**(0.00 ± 0.00) 1.61***(0.05 ± 0.15) 0.29*** (0.01 ± 0.08) 0.00*** (0.0± 0.0) 0.00*** (0.0 ± 0.0) 0.00***(0.0 ± 0.0) Replication (Env) 2821.9** 2.44*** 1.46*** 0.16* 28.20*** 6.89*** Block (Env x Rep) 0.001ns 0.02ns 0.30ns 0.00ns 7.71ns 0.93ns Entry (ped) 204064.3***(0.73 ± 0.06) 10.10***(0.35 ± 0.10) 11.35*** (0.46 ± 0.10) 3.15*** (0.26 ± 0.13) 612.96***(0.66 ± 0.08) 267.53***(0.77 ± 0.05) Env x Entry 0.000ns (0.0 ± 0.0) 1.56*** (0.06 ± 0.05) 1.88*** (0.07 ± 0.06) 3.27* (0.29 ± 0.12) 78.21*(0.08 ± 0.) 0.00ns (0.0 ± 0.0) Env_1 Residual 80633.5 14.19 9.31 5.23 238.02 92.08 Env_2 Residual 68954.1 12.63 10.50 5.06 198.46 59.65 ℎ2𝐴 0.41±0.07 0.15±0.05 0.22±0.06 0.12±0.06 0.41±0.07 0.47±0.06 Source of Staygreen Root lodging Husk cover Plant aspect Ear per plant Ear aspect variation characteristic Environment (Env) 5.53*** (0.07 ± 0.09) 0.38*** (0.08 ± 0.14) 0.00***(0.0 ± 0.0) 0.00***(0.10 ± 0.19) 0.00***(0.0 ± 0.0) 0.40*** (0.10 ± 0.19) Replication (Env) 0.00ns 0.02* 0.43*** 0.01*** 0.44*** 0.31*** Block (Env x Rep) 0.30ns 0.00ns 0.05ns 0.00ns 0.06ns 0.05* Entry (ped) 22.74***(0.25 ± 0.08) 1.75*** (0.34 ± 0.13) 1.01***(0.33 ± 0.09) 0.05***(0.41 ± 0.13) 3.07***(0.52 ± 0.09) 1.79*** (0.41 ± 0.13) Env x Entry 0.00ns (0.0 ± 0.0) 1.91*** (0.38 ± 0.13) 0.00ns (0.0 ± 0.0) 0.00ns (0.14 ± 0.09) 0 (0.0 ± 0.01) 0.44ns (0.14 ± 0.09) Env_1 Residual 21.88 1.55 0.79 0.07 2.33 1.14 Env_2 Residual 80.76 0.52 2.18 0.06 2.01 1.05 ℎ2𝐴 0.08±0.03 0.25±0.09 0.13±0.04 0.16±0.05 0.25±0.06 0.25±0.07 “*”, “**”, “***”, = significance at p(α=0.05), p(α=0.01), and p(α=0.001), respectively; “ns” = non-significant. Values in parenthesis are contributions of different sources of variance to the total variance with their standard errors; ℎ2𝐴 = narrow-sense heritability. 32 University of Ghana http://ugspace.ug.edu.gh Table 3.3 presents the variance components and narrow-sense heritability estimates for measured traits of the inbreds across optimal environments. There were significant differences among random error variances for grain yield. There were also high and significant variability of the GEI. Despite this variability among environments for grain yield, the narrow sense heritability was moderate with a value of 49±5%. All the measured traits showed high narrow sense heritability estimates except for ASI (18±4%) and husk cover (7±5%). Only the additive genetic component of variance for root lodging was not significantly different from zero. Table 3.4 presents the variance components and narrow-sense heritability estimates for measured traits of the inbreds across test environments. There was high variability among test environments (Striga-infested, drought, and optimal) with the variance due to environment significantly (p<0.001) different from zero for all the traits. Consequently, the genotype-environment interaction variance was also found to be significantly (p<0.001) different from zero for all the measured traits except for stalk lodging. However, the contribution of the genotype-environment variance to the total variance of measured traits was very low with 17% as the highest for husk cover. This indicated consistency of the genotypic response to test environments despite the high heterogeneity of random error. Thirty (30%) percent of the total variation was due to additive genetic effects among the inbreds for grain yield and 47%, 43% and 29%, respectively, for ear aspect, ears per plant, and anthesis-silking interval. The narrow sense heritability estimates were very low for measured traits used in the combined analysis of variance with the highest value of 26% for plant and ear heights. The observed heritability estimate for grain yield was 8%. 33 University of Ghana http://ugspace.ug.edu.gh Table 3.3 Variance components and heritability estimates of measured traits of extra-early maturing maize inbreds under five optimal environments at Abuja, Mokwa and Ikenne, 2016-2017 Grain Yield, Days to50% Days to 50% Anthesis-silking Plant height, Ear height, Source of variation kgha-1 silking Anthesis Interval Cm cm Environment 715242*** (0.19±0.16) 0.81***(0.02±0.03) 0.00***(0.0±0.0) 0.50***(0.11±0.12) 107.01***(0.05±0.05) 107.01ns (0.14±0.12) Replication (Env) 2735ns 0.07ns 0.06ns 0.04ns 7.77ns 7.77ns Block (Env x Rep) 44942*** 0.75*** 0.42*** 0.08** 41.64*** 41.64ns Entry (ped) 2342034*** (0.62±0.13) 29.94***(0.79±0.04) 25.50***(0.82±0.03) 1.92*** (0.43±0.10) 1748.97***(0.81±0.05) 1748.97***(0.72±0.11) Env x Entry 347857***(0.09±0.03) 2.70*(0.07±0.03) 2.84**(0.09±0.03) 0.55ns (0.12±0.06) 22.41ns (0.01±0.01) 22.41ns (0.02±0.02) Env_1 Residual 305857 4.62 2.58 1.32 188.63 188.63 Env_2 Residual 556727 3.07 2.19 1.08 312.48 312.48 Env_3 Residual 153501 2.91 2.35 1.77 153.69 153.69 ℎ2𝐴 0.49±0.06 0.57±0.05 0.63±0.04 0.18±0.04 0.56±0.05 0.51±0.05 Source of variation Root lodging Stalk lodging Husk cover Ears per plant Plant aspect Ear aspect Environment 1.52*** (0.05±0.05) 0.00***(0.00±0.01) 0.97***(0.18±0.18) 0.03***(0.10±0.09) 0.26***(0.05±0.06) 1.13***(0.14±0.13) Replication (Env) 0.00ns 0.48* 0.02ns 0.00ns 0.00ns 0.01ns Block (Env x Rep) 0.00ns 0.00ns 0.05* 0.00ns 0.08*** 0.07*** Entry (ped) 0.00ns (0.0±0.0) 191.66***(0.78±0.04) 0.56***(0.10±0.08) 0.24***(0.69±0.08) 3.40***(0.69±0.07) 5.18*** (0.65±0.11) Env x Entry 0.00ns (0.0±0.0) 16.22*(0.07±0.03) 2.99***(0.54±0.14) 0.00ns (0.00±0.00) 0.46**(0.09±0.04) 0.64**(0.08±0.03) Env_1 Residual 4.52 27.99 0.58 0.03 0.41 0.78 Env_2 Residual 65.02 28.38 0.93 0.09 0.83 1.11 Env_3 Residual 15.91 54.75 1.43 0.10 0.98 0.85 ℎ2𝐴 0.00±0.00 0.39±0.05 0.07±0.05 0.34±.005 .046±0.05 0.42±0.05 “*”, “**”, “***”, = significance at p(𝛼 =0.05), p(𝛼 =0.01), and p(𝛼 =0.001), respectively; “ns” = non-significant. Values in parenthesis are contributions of different sources of variance to the total variance with their standard errors; ℎ2𝐴 = narrow-sense heritability. 34 University of Ghana http://ugspace.ug.edu.gh Table 3.4 Variance components and heritability estimates of measured traits of extra-early maturing maize inbreds across test environment Source of variation Grain Yield, kgha-1 Days to 50% silking Anthesis-silking Interval Plant height, cm Ear height, cm Environment 355983.70*** (0.35 ± 0.14) 12.47*** (0.23 ± 0.11) 1.83*** (0.23 ± 0.11) 337.43***(0.18 ± 0.09) 59.22***(0.10 ± 0.06) Replication (Env) 3919.63*** 1.25*** 0.10** 58.78*** 10.64*** Block (Env x Rep) 9045.00*** 0.60*** 0.07* 19.41*** 4.93*** Entry (ped) 308488.48*** (0.30 ± 0.08) 30.03*** (0.56 ± 0.09) 2.31*** (0.29 ± 0.07) 1211.26***(0.63 ± 0.08) 435.66***(0.71 ± 0.06) Env x Entry 74581.25*** (0.07 ± 0.02) 1.23*** (0.02 ± 0.01) 0.33*** (0.04 ± 0.01) 55.90***(0.03 ± 0.01) 18.09***(0.03 ± 0.01) Env_1 Residual 294543.73 15.13 5.21 600.91 160.40 Env_2 Residual 230165.90 7.55 2.85 187.30 51.17 Env_3 Residual 75684.00 14.36 5.39 231.24 83.74 Env_4 Residual 24047.32 9.80 5.56 125.43 38.35 Env_5 Residual 367562.46 3.80 1.11 154.54 81.86 Env_6 Residual 674433.12 2.50 0.93 277.63 165.14 Env_7 Residual 161999.93 2.45 1.54 126.61 43.89 ℎ2𝐴 0.08 ±0.02 0.21 ±0.03 0.05 ±0.01 0.26 ±0.03 0.26 ±0.03 Source of variation Root lodging Stalk lodging Husk cover Ears per plant Ear aspect Environment 18.05***(0.21 ± 0.13) 14.49***(0.06 ± 0.04) 0.71***(0.29 ± 0.12) 0.05*** (0.19 ± 0.10) 0.87*** (0.17 ± 0.09) Replication (Env) 8.05*** 0.92*** 0.02* 0.00*** 0.19*** Block (Env x Rep) 0.00ns 0.02*** 0.02** 0.00*** 0.08*** Entry (ped) 1.11ns (0.02 ± 0.01) 137.17***(0.49 ± 0.06) 0.45***(0.19 ± 0.06) 0.11***(0.47 ± 0.08) 2.18*** (0.43 ± 0.07) Env x Entry 0.01***(0.0 ± 0.0) 10.96ns (0.05 ± 0.01) 0.41***(0.17 ± 0.03) 0.02***(0.07 ± 0.01) 0.43*** (0.08 ± 0.02) Env_1 Residual 195.89 191.84 0.31 0.05 0.92 Env_2 Residual 45.31 181.10 0.95 0.08 2.29 Env_3 Residual 25.67 57.61 1.53 0.06 2.26 Env_4 Residual 87.82 113.73 0.39 0.04 1.63 Env_5 Residual 4.45 23.89 0.48 0.02 0.63 Env_6 Residual 63.99 24.26 0.89 0.07 0.82 Env_7 Residual 15.94 39.37 1.32 0.09 0.71 ℎ2𝐴 0.00 ±0.0 0.08 ±0.02 0.04 ±0.01 0.12 ±0.02 0.10 ±0.02 “*”, “**”, “***”, = significance at p(α =0.05), p(α =0.01), and p(α =0.001) respectively; “ns” = non-significant. Values in parenthesis are contributions of different sources of variance to the total variance with their standard errors; ℎ2𝐴 = narrow-sense heritability. 35 University of Ghana http://ugspace.ug.edu.gh 3.3.3 Genetic correlations among traits under stress environments Genetic correlation coefficients among traits under Striga-infested environments are presented in Table 3.5. Grain yield had highly significant positive genetic association with ears per plant (0.98) and plant height (0.91). In addition, negative correlations were detected between grain yield and ear aspect (-0.98), husk cover (-0.68), Striga damage at 8 WAP (-0.63) and Striga damage at 10 WAP (-0.59). Moderate genetic correlations were observed between grain yield and other traits. A strong positive correlation was observed between Striga damage at 8 WAP and 10 WAP (0.98) and between number of emerged Striga plants at 8 WAP and 10 WAP (0.96). Table 3.6 presents the genetic correlation coefficients among traits under drought-infested environments. Low to high genetic associations wer observed between grain yield and other measured traits. The genetic correlation between grain yield and ears per plant was strong and positive (0.84) but grain yield had significant negative genetic correlation with ear aspect (-0.75) and plant aspect (-0.64). The staygreen characteristic and anthesis-silking interval showed moderate correlations with grain yield. Moreover, a highly significant positive genetic correlation was observed between plant aspect and ear aspect (0.86) and between plant aspect and the staygreen characteristic (0.67), all of which are included in the selection index. 36 University of Ghana http://ugspace.ug.edu.gh Table 3.5 Genetic correlations among grain yield and other traits across Striga-infested environments at Abuja and Mokwa, 2016-2017 Trait variable 1 2 3 4 5 6 7 8 9 10 11 12 13 1 Grain yield 1 2 50% days to -0.4*** 1 silking 3 Anthesis- -0.3*** 0.22** 1 silking interval 4 Plant height 0.91*** 0.39*** 0.98*** 1 5 Striga damage -0.63*** 0.23** 0.28*** -0.82*** 1 8 WAP 6 Striga damage -0.59*** 0.1ns 0.28*** -0.78*** 0.96*** 1 10 WAP 7 Number of -0.39*** -0.03ns 0.24** 0.28*** 0.5*** 0.67*** 1 Emerged Striga plants 8WAP 8 Number of -0.58*** 0.08ns 0.41*** 0.59*** 0.38*** 0.61*** 0.96*** 1 Emerged Striga plants 10WAP 9 Root lodging -0.55*** -0.78*** 0.98*** 0.58*** 0.98*** 0.98*** 0.98*** 0.42*** 1 10 Stalk lodging -0.5*** 0.22** 0.08ns -0.45*** 0.73*** 0.76*** 0.86*** 0.94*** 0.98*** 1 11 Husk cover -0.68*** 0.1ns 0.12ns -0.68*** 0.98*** 1 0.7*** 0.49*** 0.98*** 0.7*** 1 12 Ears per plant 0.98*** -0.34*** -0.36*** 0.84*** -0.76*** -0.84*** -0.58*** -0.64*** -0.58*** -0.98*** -0.98*** 1 13 Ear aspect -0.98*** 0.4*** 0.34*** -0.41*** 0.77*** 0.81*** 0.44*** 0.53*** 0.98*** 0.46*** 0.85*** -0.98*** 1 37 University of Ghana http://ugspace.ug.edu.gh Table 3.6 Genetic correlations among grain yield and other traits across drought stress environments in Ikenne during 2016/2017 and 2017/2018 dry seasons Traits variable 1 2 3 4 5 6 7 8 9 10 11 12 13 1 Grain yield 1 50% days to 2 anthesis -0.4*** 1 50% days to 3 silking -0.55*** 0.78** 1 Anthesis- silking 4 interval -0.26** -0.31** 0.35*** 1 5 Plant height 0.16* -0.04ns 0.01ns 0.09ns 1 6 Ear height 0.18* 0.26** 0.18* -0.06ns 0.8*** 1 7 Root lodging -0.09ns -0.17* -0.36*** -0.31*** 0.12ns -0.04ns 1 8 Stalk lodging -0.16* -0.26** -0.25** 0.02ns -0.06ns -0.12ns 0.1ns 1 9 Husk cover -0.18* 0.41*** 0.3*** -0.17* -0.4*** -0.04ns -0.7*** -0.04ns 1 10 Plant aspect -0.64*** 0.35*** 0.61*** 0.41*** -0.39*** -0.25** -0.05ns 0.21* 0.5*** 1 Ears per 11 plant 0.84*** -0.3*** -0.6*** -0.51*** 0.17* 0.2* 0.07ns -0.16* -0.27*** -0.87*** 1 12 Ear aspect -0.75*** 0.38*** 0.63*** 0.39*** -0.21** -0.19* -0.17* 0.13ns 0.31*** 0.86*** -0.85*** 1 Staygreen 13 characteristic -0.27** -0.23** -0.1ns 0.21** -0.39*** -0.31*** -0.05ns 0.43*** 0.28*** 0.67*** -0.32*** 0.42*** 1 38 University of Ghana http://ugspace.ug.edu.gh 3.3.3 Performance of inbred lines under stress and across test environments 3.3.3.1 Index-based performance of inbreds Table 3.7 presents the performance of the top 20% and five worst inbreds under Striga infestation. The grain yield across Striga-infested environments ranged from zero for TZEEIOR 37 and TZEEIOR 87 to 2946 kg ha-1 for TZEEIOR 113 with an average of 931 kg ha-1. Striga damage at 8 and 10 WAP ranged from 2 to 7 and 3 to 7 with a mean of 4 and 5, respectively. Similarly, number of emerged Striga plants at 8 and 10 WAP ranged from zero to 41 and from zero to 32, respectively. The anthesis-silking interval ranged from zero to 10 days. Number of ears per plant obtained ranged from zero to 2 and ear aspect varied from 3 to 9. Among the selected inbreds under Striga infestation, that is high yielding inbreds with few number of emerged Striga plants and reduced Striga damage, TZEEIOR 113, TZEEIOR 214, TZEEIOR 189, TZEEIOR 221, TZdEEI 7, TZEEIOR 38, TZEEIOR 12, TZEEIOR 76 exceeded the average yield by 216%, 171%, 167%, 138%, 129%, 122%, 109%, and 104%, respectively. Table 3.8 presents the performance of the top 20% and five worst inbreds under drought stress. The grain yield ranged from zero for TZEEIOR 174 to 1287 kg ha-1 for TZEEIOR 109 with a mean of 317 kg ha-1. The plant aspect ranged from 4 to 8 with a mean of 6. The anthesis-silking interval varied from zero to 10 days. Also, on the average, inbreds showed staygreen characteristic of 4 while the range was 2 to 7. The outstanding inbreds exceeding the mean grain yield under drought stress by at least over 140% were TZEEIOR 109 (306%), TZEEIOR 38 (201%), TZEEIOR 11 (180%), TZEEIOR 91 (155%), TZEEIOR 145 (144%), TZEEIOR 195 (143%), and TZEEIOR 42 (140%). The inbreds TZEEIOR 113 and TZEEIOR 76 produced 130% and 138% over the average grain yield under drought stress. These inbreds along with TZEEIOR 221 and TZEEIOR 38 showed outstanding performance under both stress conditions. 39 University of Ghana http://ugspace.ug.edu.gh Table 3.7 Performance of index-based top 20% and worst 5 inbred lines evaluated across Striga environments in Abuja, 2016 and Mokwa, 2017 Grain Number of emerged Anthesis Striga Damage Ears per Ear Inbreds yield Striga plants Silking Index -1 plant aspect Kg ha 8 WAP 10 WAP 8 WAP 10 WAP Interval TZEEIOR 189 2490 3 4 10 12 2 3 0 12.92 TZEEIOR 221 2220 3 3 0 6 1 6 2 12.42 TZdEEI 7 2135 2 3 8 9 1 4 1 11.73 TZEEIOR 41 1689 2 3 3 5 1 3 3 9.50 TZEEIOR 223 1474 3 3 1 3 1 4 1 9.47 TZEEIOR 113 2946 5 6 6 9 1 4 2 8.89 TZdEEI 12 1763 3 4 7 11 1 4 1 8.54 TZEEIOR 38 2067 3 4 7 14 1 4 1 8.37 TZEEIOR 197 1802 3 4 2 6 1 4 3 8.36 TZEEIOR 243 1570 2 4 3 7 1 4 0 7.96 TZEEIOR 213 1809 3 4 2 5 1 4 1 7.86 TZEEIOR 219 1608 3 4 2 3 1 4 1 7.67 TZEEIOR 195 1760 3 4 4 5 1 5 2 7.59 TZEEIOR 222 1161 2 3 0 0 1 5 2 7.58 TZEEIOR 76 1900 4 5 1 4 1 5 1 7.55 TZEEIOR 214 2519 4 5 10 12 1 4 3 7.35 TZEEIOR 42 1541 3 4 1 6 1 5 3 7.32 TZEEIOR 140 1745 4 4 4 5 1 5 3 6.69 TZEEIOR 244 1229 2 4 7 12 1 4 3 6.31 TZEEIOR 196 1651 4 4 3 4 1 5 4 6.19 TZEEIOR 109 1720 3 5 3 7 1 5 3 6.14 TZEEIOR 252 1646 3 4 4 9 1 4 2 6.00 TZEEIOR 251 1192 2 3 7 12 1 4 1 5.76 TZEEI 79 1225 3 4 4 10 1 5 5 5.28 TZEEIOR 12 1945 4 5 8 9 1 5 4 5.02 TZEEIOR 253 1452 3 5 6 7 1 4 1 4.95 TZEEIOR 145 1535 4 5 2 8 1 5 3 4.94 TZEEIOR 248 1362 4 4 4 5 1 5 2 4.76 TZEEIOR 43 1157 3 4 0 0 1 6 4 4.58 TZEEIOR 217 1301 3 4 3 7 1 5 3 4.56 TZEEIOR 245 1188 2 4 10 13 1 5 3 4.50 TZEEIOR 146 1176 4 4 9 10 1 4 2 4.14 TZEEIOR 154 1816 5 6 9 10 1 5 2 4.06 TZEEIOR 212 1099 3 4 2 5 1 5 0 3.96 TZEEIOR 240 1227 4 5 2 6 1 5 1 3.89 TZEEI 74 935 4 4 0 1 1 6 2 3.72 TZEEIOR 36 0.00 5 6 4 5 0 8 4 -8.58 TZEEIOR 238 224 5 7 17 20 0 8 5 -8.91 TZEEIOR 118 428 7 7 8 8 0 7 7 -9.02 TZEEIOR 86 542 5 6 26 32 0 7 0 -9.70 TZEEIOR 87 0.00 6 6 8 19 0 9 3 -10.41 Mean 931 4 5 7 10 1 6 3 Minimum 0.00 2 3 0 0 0 3 0 Maximum 2946 7 7 41 32 2 9 7 S.E 405 1 1 6 7 0 1 2 S.E = standard error 40 University of Ghana http://ugspace.ug.edu.gh Table 3.8 Performance of index-based top 20% and worst 5 inbred lines evaluated across drought stress environments at Ikenne, during 2016/2017 and 2017/2018 dry seasons Grain Plant Ears per Ear Stay Anthesis- Inbreds Yield, Aspect plant Aspect green Silking Index kgha-1 (1-9) (1-9) (1-9) interval TZEEIOR 109 1288 4 1 4 3 3 16.85 TZEEIOR 38 954 4 1 4 3 3 13.16 TZEEIOR 91 808 5 1 5 3 1 11.84 TZEEIOR 161 705 4 1 5 4 1 11.01 TZEEIOR 29 682 4 1 5 3 2 10.55 TZEEIOR 11 888 5 1 5 4 4 9.54 TZEEIOR 126 421 4 1 4 3 2 9.20 TZEEIOR 35 542 5 1 5 4 1 9.17 TZdEEI 13 490 4 1 4 4 2 8.99 TZEEIOR 145 773 6 1 4 5 2 8.46 TZEEIOR 32 495 4 1 5 3 3 8.33 TZEEIOR 42 762 5 1 6 4 3 8.30 TZEEIOR 53 513 5 1 5 5 1 8.25 TZEEI 64 511 5 0 5 3 2 8.14 TZEEIOR 113 687 5 0 5 4 3 7.92 TZdEEI 9 641 6 1 6 3 2 7.72 TZEEIOR 76 755 5 1 6 4 3 7.42 TZEEIOR 5 675 5 1 6 4 3 7.20 TZEEIOR 100 674 6 1 6 3 3 7.15 TZEEIOR 195 770 5 1 4 5 6 6.84 TZEEIOR 123 280 4 1 6 3 2 6.77 TZEEIOR 24 418 5 1 5 4 3 6.69 TZEEIOR 203 574 4 0 5 3 5 6.37 TZEEIOR 102 477 5 1 5 3 5 6.27 TZEEIOR 31 541 5 1 6 4 4 5.83 TZEEIOR 10 474 6 0 5 4 1 5.53 TZEEIOR 85 375 5 0 5 3 2 5.41 TZEEIOR 197 513 5 1 6 4 4 5.36 TZEEIOR 45 475 6 0 5 3 2 5.30 TZEEIOR 232 181 5 1 7 3 0 5.05 TZEEIOR 167 402 5 1 5 4 3 4.87 TZEEIOR 108 528 5 0 6 3 5 4.80 TZEEIOR 104 625 6 1 6 4 4 4.77 TZEEIOR 221 730 5 0 6 5 3 4.76 TZEEIOR 125 371 5 0 5 3 4 4.63 TZEEIOR 97 462 5 0 4 4 4 4.63 TZEEIOR 191 49 6 0 8 5 7 -9.30 TZEEIOR 87 46 7 0 8 6 6 -9.91 TZEEIOR 135 58 7 0 8 7 6 -10.82 TZEEI 96 110 8 0 8 5 10 -11.66 TZEEIOR 174 0.00 7 0 9 5 9 -13.48 Mean 317 6 0 6 4 4 Minimum 0.00 4 0 4 2 0 Maximum 1288 8 1 9 7 10 S.E 196 1 0 1 1 2 S.E = standard error 41 University of Ghana http://ugspace.ug.edu.gh 3.3.3.2 BLUPs-based performance of inbreds Table 3.9 presents the pedigree-based BLUPs (breeding values) and multivariate BLUPs of grain yield under Striga-infested, drought stress and across test environments. The top 10 outstanding inbreds based on the estimated breeding values for grain yield under Striga infestation were: TZdEEI 7 (1910±322 kg ha-1), TZdEEI 12 (1712±322 kg ha-1), TZEEIOR 214 (1574±168 kg ha- 1), TZEEIOR 221 (1522±188 kg ha-1), TZEEIOR 109 (1453±211 kg ha-1), TZEEIOR 213 (1432±163 kg ha-1), TZEEIOR 189 (1421±180 kg ha-1), TZEEIOR 113 (1418±162 kg ha-1), TZEEIOR 38 (1372±209 kg ha-1), and TZEEIOR 215 (1371±168 kg ha-1). Under drought stress, the best 10 inbreds in the pedigree based on the predicted additive genetic effects of grain yield were as follows: TZEEIOR 109 (744±96 kg ha-1), TZdEEI 12 (727±133 kg ha-1), TZEEIOR 38 (674±95 kg ha-1), TZdEEI 7 (612±133 kg ha-1), TZEEIOR 130 (529±99 kg ha-1), TZdEEI 9 (517±133 kg ha-1), TZdEEI 13 (502±134 kg ha-1), TZEEIOR 5 (488±103 kg ha-1), TZEEIOR 76 (488±83 kg ha-1), and TZEEIOR 37 (487±83 kg ha-1). Selection based on the BLUPs from test environments, including the optimal, revealed inbreds TZdEEI 7, TZEEIOR 109, TZdEEI 12, TZEEIOR 221, TZEEIOR 195, TZEEIOR 189, TZEEIOR 12, TZEEIOR 236, TZEEI 73, and TZEEIOR 196 to be the best 10 performers. Results also showed a general tendency of low breeding values of the inbreds under drought stress compared to those predicted under Striga infestation. Moreover, ranking of inbreds based on BLUPs of grain yield from multivariate analysis identified 7 out of 10 best-selected inbreds based on breeding values across test environments. 42 University of Ghana http://ugspace.ug.edu.gh Table 3.9 Pedigree BLUPs-based and multivariate BLUPs-based performances of the top 12% of the inbreds across pedigree under Striga-infested, drought and across test environments, 2016-2017 Breeding values for Grain yield, Kgha-1) Multivariate blups ranking Inbreds Striga S.E Inbreds Drought S.E Inbreds Across S.E Inbreds MVBlups S.E TZdEEI7 1910.4 322 TZEEIOR109 744 96 TZdEEI7 1630 271 TZdEEI7 0.81 0.27 TZdEEI12 1711.8 322 TZdEEI12 727 133 TZEEIOR109 1262 250 TZEEI73 0.39 0.27 TZEEIOR214 1574.1 168 TZEEIOR38 674 95 TZdEEI12 1261 271 TZdEEI12 0.32 0.27 TZEEIOR221 1522.1 188 TZdEEI7 612 133 TZEEIOR221 1197 245 TZEEI81 0.31 0.27 TZEEIOR109 1452.7 211 TZEEIOR130 529 99 TZEEIOR195 1195 244 TZEEIOR189 0.24 0.20 TZEEIOR213 1431.7 163 TZdEEI9 517 133 TZEEIOR189 1187 243 TZEEIOR195 0.24 0.20 TZEEIOR189 1420.5 180 TZdEEI13 502 134 TZEEIOR12 1173 237 TZEEIOR221 0.24 0.20 TZEEIOR113 1418.0 162 TZEEIOR5 488 103 TZEEIOR236 1143 254 TZEEIOR197 0.23 0.19 TZEEIOR38 1371.5 209 TZEEIOR76 488 83 TZEEI73 1139 271 TZEEIOR196 0.22 0.19 TZEEIOR215 1370.7 168 TZEEIOR37 487 83 TZEEIOR196 1139 239 TZEEIOR194 0.21 0.19 TZEEIOR154 1351.0 225 TZEEIOR100 478 86 TZEEIOR10 1138 237 TZEEIOR243 0.20 0.20 TZEEIOR223 1333.3 178 TZEEI64 477 133 TZEEIOR197 1137 239 TZEEIOR12 0.20 0.18 TZEEIOR252 1324.6 180 TZEEIOR75 476 83 TZEEIOR14 1132 236 TZEEIOR244 0.19 0.20 TZEEIOR243 1321.5 174 TZEEIOR195 470 83 TZEEIOR11 1127 237 TZEEIOR10 0.19 0.18 TZEEIOR222 1315.3 178 TZEEIOR4 469 113 TZEEIOR13 1122 236 TZEEIOR192 0.19 0.20 TZEEIOR217 1314.3 157 TZEEIOR36 468 83 TZEEIOR8 1121 236 TZEEIOR191 0.19 0.19 TZEEIOR244 1306.8 174 TZEEIOR101 464 90 TZEEIOR192 1117 242 TZEEI63 0.19 0.27 TZEEIOR112 1304.4 162 TZEEIOR11 462 66 TZEEIOR193 1111 242 TZEEIOR193 0.18 0.20 TZEEIOR195 1301.2 184 TZEEIOR35 459 86 TZEEIOR240 1108 252 TZEEIOR14 0.18 0.18 TZEEIOR140 1289.4 182 TZEEIOR113 457 74 TZEEIOR9 1105 236 TZEEIOR222 0.18 0.20 TZEEIOR111 1286.5 162 TZEEIOR12 452 66 TZEEIOR7 1103 238 TZEEIOR13 0.18 0.18 TZEEIOR216 1286.2 157 TZEEIOR112 450 74 TZEEIOR194 1103 239 TZEEI64 0.17 0.27 TZEEIOR253 1271.9 180 TZEEIOR114 448 87 TZdEEI13 1103 271 TZEEIOR8 0.17 0.18 TZEEIOR219 1270.9 184 TZEEIOR32 443 75 TZEEIOR6 1092 237 TZEEIOR223 0.17 0.20 TZEEIOR114 1269.3 194 TZEEIOR31 441 75 TZEEIOR191 1090 239 TZEEIOR252 0.17 0.20 TZEEIOR197 1268.1 159 TZEEIOR91 441 74 TZEEIOR38 1088 250 TZEEIOR253 0.17 0.20 TZEEIOR145 1251.3 186 TZEEIOR111 439 74 TZEEIOR244 1087 242 TZEEIOR240 0.17 0.22 TZEEIOR249 1238.1 200 TZEEIOR34 438 86 TZEEIOR243 1087 242 TZEEIOR11 0.17 0.18 TZEEIOR250 1238.1 200 TZEEIOR33 438 78 TZEEIOR187 1085 256 TZEEIOR245 0.16 0.20 TZEEIOR212 1237.5 163 TZEEIOR29 437 75 TZEEI64 1072 271 TZEEIOR109 0.16 0.21 TZEEIOR251 1236.9 190 TZEEIOR108 426 103 TZEEIOR241 1071 252 TZEEIOR7 0.16 0.18 TZEEIOR245 1232.8 191 TZEEIOR82 425 83 TZEEI81 1069 271 TZEEIOR9 0.16 0.18 TZEEIOR196 1232.6 158 TZEEIOR42 423 68 TZEEIOR219 1066 244 TZEEIOR236 0.16 0.22 S.E, standard error of estimation; MVBlups, multivariate Best Linear Unbiased predictions 43 University of Ghana http://ugspace.ug.edu.gh 4.5 Discussion and Conclusions Differences in the response to Striga infestation and drought stress observed among the inbreds for most of the traits including grain yield is an indication of the existence of genetic variability needed for a successful selection for improvements of the traits. This finding is supported by previous reports of Badu-Apraku et al. (2011) and Badu-Apraku and Oyekunle (2012). In each of the trials, the non-significant genotype-environment interaction (GEI) observed for grain yield and most of the traits indicated similarity in the management and showed that individual inbreds responded similarly in each of the environments. However, the highly significant GEI observed, in the combined analysis, for all the traits except stalk lodging depicted the heterogeneity among environments and indicated a differential response of the inbreds to the different test environments. This result supports the findings of Badu-Apraku et al. (2012a) and highlights the importance of considering stress and non-stress environments in determining tolerant genotypes. Moreover, in the ranking of inbred lines and prediction of response to selection, genetic effects are needed and their estimation in plant breeding is often done ignoring relationships among lines. To ensure accurate and precise estimates of genetic effects and reduce biases in selecting superior lines, in the present study, genetic variance-covariance relationships among the 180 inbred lines were included in the analysis. Comparable narrow sense heritabilities were observed for grain yield under Striga infestation (48%) and drought stress (41%) suggesting possible improvement for grain yield under these stresses. However, across test environments, low narrow sense heritabilities were observed for all traits including grain yield (8%), probably due to the heterogeneity in error variances associated with the differences in the management of the environments. Narrow-sense heritabilities obtained for Striga damage rating at 8 and 10 weeks after planting were respectively, 0.42 and 0.45. In addition, high and negative genetic correlation between each of 44 University of Ghana http://ugspace.ug.edu.gh these traits and grain yield were obtained. A similar correlation between grain yield and number of emerged Striga plants at 8 and 10 weeks after planting were also observed. This finding supports that of Badu-Apraku et al. (2013) and justifies the use of these traits in indirect selection for increased grain yield under Striga environments. Nevertheless, the high and positive genetic correlation between Striga damage at 8 and 10 weeks after planting (0.96**), on one hand, and between number of emerged Striga plants at 8 and 10 weeks after planting (1), on the other hand, indicated that the use of both (8 and 10 WAP) might over-parametrise the base index actually in use with same effects. The use of only one rating (8 or 10 WAP) could allow other important secondary traits to be considered in the index. A resistant genotype to Striga is a genotype which when grown under conditions of Striga infestation, supports significantly fewer Striga plants and has a higher yield than a susceptible cultivar (Menkir, 2006). The same author defined tolerant genotype as the one that shows smaller yield reductions than susceptible cultivars under the same level of infestation. This definition might not suite that of pure pathologists but the base index fulfills both definitions and, thus, its use can allow identification of resistant/tolerant genotypes. Eighty-two (82) out of 180 inbreds and 78 out of 152 orange lines were selected as combining high grain yield, low Striga damage, low number of emerged Striga plants and high number of ears per plant. Under drought stress, 79 out of 180 inbreds (44%) and 72 out of 152 (47%) orange lines showed levels of tolerance to drought by combining high yield, reduced interval of days to flowering and reduced and prolonged stay green characteristic. The orange inbred lines were extracted from a Striga resistant genetic broad base variety and the identification of resistant inbreds among the evaluated lines is not surprising. Also, 47% of the orange lines with some level of drought tolerance were identified. This finding is a proof that the population from which these inbreds were extracted contained some alleles for 45 University of Ghana http://ugspace.ug.edu.gh tolerance to drought. Indeed, Badu-Apraku et al. (2007) included some sources of drought tolerance germplasm during the formation of the population that served as base for extraction of the orange inbreds involved in the present study. The selection using pedigree BLUPs in this study confirmed that resistance/tolerance to Striga and tolerance to drought stress in the inbreds identified by the base indices is genetically controlled and highly heritable. TZEEIOR 189, TZEEIOR 221, TZdEEI 7, TZEEIOR 113, TZdEEI 12, and TZEEIOR 38 were selected both by the index and the pedigree BLUPs among the top ten performers under Striga infestation. TZEEIOR 109, TZEEIOR 38 and TZEEIOR 29 were the three out of the top ten performers which were consistently selected by these methods under drought stress. In addition, four inbreds namely, TZEEIOR 7, TZEEIOR 187, TZEEIOR 249, and TZEEIOR 250 with predicted breeding values for grain yield of 1103 kg ha-1, 1085 kg ha-1, 1028 kg ha-1, and 1028 kg ha-1 respectively, were not among the selected inbreds evaluated in this study. Inbred TZEEIOR 114 was selected as combining Striga resistance and drought stress using the multivariate BLUPs across the pedigree with predicted breeding values of 1269 kg ha-1 and 447 kg ha-1 under Striga infestation and drought stress, respectively. At the same time, inbred TZEEIOR 33 was selected among the best under drought stress with a breeding value of 438 kg ha-1. The six inbreds: TZEEIOR 7, TZEEIOR 187, TZEEIOR 249, TZEEIOR 250, TZEEIOR 114, and TZEEIOR 33 described above, initially not selected from the pedigree, should be tested to confirm their performance and eventually used in population improvement. In conclusion, there was significant genetic variation among the inbred lines evaluated for different traits under different environments. Selection of superior inbred parents for use in hybrid and/or population development is possible. Eighty-two inbreds including 78 orange lines were selected 46 University of Ghana http://ugspace.ug.edu.gh as drought tolerant. Seventy-nine inbreds including seventy 72 orange lines were selected as Striga resistant. About 19% (34 out of 180) of the inbreds combined Striga resistance/tolerance with drought tolerance. Genetic values of the inbreds across pedigrees for tolerance to drought were lower than the genetic values for resistance/tolerance to Striga. Seventy percent (70%) of the selected inbreds combining resistance to Striga and tolerance to drought were also selected by the multivariate BLUPS, thus confirming that their performance was due to genetic effects. 47 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR 4.0 Genetic diversity and homogeneity of extra-early maturing orange maize inbred lines 4.1 Introduction The success of maize breeding in addressing the increasing demand for resistant genotypes to adverse environmental changes, a myriad of consumer preferences, and some hidden nutrition- related issues is based on availability of genetic variability. Inbred lines derived from numerous source populations serve as source of new cultivars and parental inbred lines that constitute a sample of the genetic diversity available in maize. Hence, genetic variability assessment of inbred lines has become a crucial step in breeding, prior to the development of outstanding hybrids and open-pollinated maize cultivars. Many tools have been used to characterise and assess diversity in maize inbred lines. Presently, the availability and accessibility at the relatively low cost of Single Nucleotide Polymorphism (SNP) markers has made their use popular. Molecular-based diversity study using SNPs has the advantage of exploring the variation between genotypes at the base level and provides a means of differentiating cultivars and classifying inbred lines into heterotic groups (Fan et al., 2008; Lu et al., 2009; Dao et al., 2014); identifying gaps and redundancy in germplasm collections (Ertiro et al., 2017; Semagn et al., 2012) ; monitoring genetic shifts that occur during germplasm conservation, regeneration, domestication, and breeding; identifying novel and superior alleles for improvement of agronomic traits; and constructing a representative subset or core collection. The objectives of the present study were to: (i) determine the genetic purity of selected extra-early orange inbred lines and, (ii) classify the selected inbred lines based on the molecular diversity within the population. 48 University of Ghana http://ugspace.ug.edu.gh 4.2 Materials and Methods 4.2.1 Plant materials One hundred and fifty-two extra-early maturing orange inbred lines developed by IITA-MIP constituted the inbred lines used in the study. The set of inbred lines were selected from the inbreds described in chapter three, section 3.2.1 based on their performance under both Striga infestation and drought stress during the 2014 and 2015 field evaluations. 4.2.2 Leaf samples collection and DNA extraction Leaf samples were collected from seedlings at 2-3 leaf-stage grown in the breeding nursery, in Ibadan. For each inbred line, one maize leaf per plant was collected from 10 representative plants, placed in paper envelopes and immediately freeze-dried. Genomic DNA samples were isolated from freeze-dried leaf tissues of each inbred line following the DArT DNA extraction protocol (https://www.diversityarrays.com/ files/DArT_DNA_isolation.pdf). The DNA concentration was obtained by spectrometry measurement using Nanodrop 8000 machine (Thermo Scientific, USA), and DNA quality was confirmed by running DNA samples on 0.8 % agarose gel. Short or degraded DNA was eliminated and DNA concentration of 30 ng μl-1 were used. 4.2.3 Diversity Array Technology sequencing (DArTseq) genotyping and analysis Genotyping by sequencing analysis of the inbred lines was performed using a high-density whole genome profiling of DArT services. DNA samples (100 μl of 50 ng μl-1) were sent to the Integrated Genomic Service and Support (IGSS) platform of BecA-ILRI in Kenya for DArTseq analysis following the protocol described by Elshire et al. (2011) using 44391 DArTseq codominant markers. All the images from DArTseq platform were analysed using DArTsoft v.7.4.7 (DArT P/L, Canberra, Australia). The DArTseq markers were scored using DArTsoft as binary data zero (0), indicating the presence of the reference allele homozygote, one (1) representing the presence 49 University of Ghana http://ugspace.ug.edu.gh of the SNP allele homozygote or two (2) designating the presence of heterozygote in the genomic representation of each sample as described by DArT Pty Ltd, Australia (https://www.diversityarrays.com) 4.2.4 Filtering of the SNPs The SNP single row file (mapping format) containing DArTseq markers was first converted to a “genlight object” using “dartR” package in “R” for SNP filtering. The initial DArTseq data comprised of 44,391 SNPs for the 152 genotyped orange maize inbred lines. A series of filters was used to select the SNPs for the genetic analysis, in this order: filtering loci on callrate with a threshold of 95%; filtering individuals on call rate with a threshold of 90%; filtering on average repeatability of each locus with a threshold of 100%; filtering out monomorphic loci; filtering out loci with trimmed sequence tags that are too similar (possible paralogues); filtering out all SNPs that did not align with the reference genome. 4.2.5 Summary statistics and cluster analyses The retained markers were subjected to PowerMarker, version 3.25 (Liu and Muse, 2005) for descriptive statistics including number of markers, heterogeneity, gene diversity, minor allele frequency (MAF) and polymorphism information content (PIC) per chromosome and also for Roger’s genetic (Roger, 1972) distance matrix generation. The PIC was calculated according to the following formula: n PIC = 1 − ∑P2i i=1 Where: i = the ith allele of the jth marker, n = the number of alleles at the jth marker and p = allele frequency. 50 University of Ghana http://ugspace.ug.edu.gh A phylogenetic tree was constructed from the genetic distance matrix using the neighbour-joining algorithm with 1000 nonparametric bootstrapping across different loci in PowerMarker. The R- package "Ape" was then used to visualize the genetic relationships among the inbreds. The relative kinship between each pair of lines was calculated using TASSEL software version 5.2.12 (Bradbury et al., 2007). 4.2.6 Population structure and principal component analyses The admixture model-based clustering method was used to infer the population structure of the 152 inbreds using the software package STRUCTURE, version 2.3.4 (Pritchard et al., 2000). To determine the genetic structure of the 152 lines, shared allele frequencies and unlinked loci were assumed. STRUCTURE was run by setting the number of clusters (K) from 1 to 10. Each K was run 10 times with a burn-in period of 10,000 and 100,000 Markov Chain Monte Carlo (MCMC) replications after burn-in. The optimal value of K was estimated using the ad hoc statistic ΔK through online computation, in the Structure Harvester (Evanno et al., 2005). The ΔK is based on the second order rate of change of P(X|K), the posterior probability of the data with respect to a given K and considers that the rate of change of LnP(D) increases as K increases and tends to be maximum at the true value of K. It was computed as follows: 𝑀[|𝐿(𝐾 − 1) − 2𝐿(𝐾) + 𝐿(𝐾 + 1)|] ∆𝐾 = 𝑆[𝐿(𝐾)] where: L(K) is the Kth LnP(D), M is the mean of 10 runs, and S their standard deviation. The probability of membership obtained for the retained K value was used to assign individuals into the different clusters. Individuals with a probability of membership ≥ 70% were assigned to the same group while those with < 70% probability memberships in any single group were assigned to a “mixed” group (Lu et al., 2009; Yang et al., 2011). To confirm the clustering based on the Bayesian statistics, a principal coordinate analysis (PCoA) of the same set of DArTseq markers 51 University of Ghana http://ugspace.ug.edu.gh was performed using GenALEx version 6.5 (Peakall and Smouse, 2006; 2012). The results of the STRUCTURE analysis were used as a priori information for the phylogeny tree and the PCoA. 4.3 Results 4.3.1 Description of DArTseq SNP markers in the Population Four thousand six hundred and twenty (4,620) SNPs (10.41% of the original data) were obtained from the filtering and used in the genetic diversity study. Figure 4.1 presents the summary statistics of the retained 4,620 SNP markers for genetic analysis. The SNPs comprised 26.32% of A, 24.97% of G, 25.54% of T and 23.17% of C. The total transitions among the SNPs occurred at a rate of 59.61% while the transversions occurred at a rate of 40.39%, giving a rate transition/ transversion of 1.48 (Table 4.1). On the average, 10% of the retained markers were distributed on each of the 10 chromosomes, indicating an even distribution of the markers on the chromosomes. The percentage markers on each chromosome varied from 6.1% on chromosome 10 to 15.2% on chromosome 1. The major allele frequency varied from 0.85 to 0.88 while the gene diversity ranged from 0.18 to 0.21 (on chromosomes 2 and 9). The average proportion of the heterozygous inbreds per marker ranged from 0 to 0.87. This proportion varied from 0.03 to 0.04 on average loci per chromosome. The mean polymorphism information content (PIC) across loci per chromosome ranged from 0.15 to 0.18 while, for all the SNPs, it ranged from 0.007 to 0.375 with an average of 0.18. 52 University of Ghana http://ugspace.ug.edu.gh 0.95 16 0.90 0.85 14 0.80 0.75 0.70 12 0.65 0.60 10 0.55 Major.Allele.Frquency 0.50 GeneDiversity 8 0.45 Heterozygosity 0.40 0.35 PIC 6 0.30 %SNP 0.25 4 0.20 0.15 2 0.10 0.05 0.00 0 1 2 3 4 5 6 7 8 9 10 Chromosome Figure 4.1 Summary statistics of the 4620 SNPs used for characterising the 152 orange maize inbred lines from IITA-MIP breeding programs 53 Frequency % SNPs University of Ghana http://ugspace.ug.edu.gh 4.3.2 Genetic purity of the inbred lines The summary of the heterogeneity of the 152 inbred lines using 4620 SNPs evenly distributed on the 10 chromosomes is presented in Figure 4.2. Four inbreds out of 152 (3%) had heterogeneity less than 1%, which represents the expected heterogeneity at six generations of inbreeding. The majority (89%) of the inbred lines showed heterogeneity between 1.1% and 5%. Thirteen inbreds had heterogeneity greater than 5% among which the revealed heterogeneity fell between 5.1% and 12.5% for 9 inbreds while for 4 of them heterogeneity was greater than 12.5%. 90% 89% 80% 70% 60% 50% 40% 30% 20% 10% 6% 3% 3% 0% ≤1% 1.1%-5% 5.1%-12.5% >12.5% Heterogeneity Figure 4.2 Summary of the heterogeneity of 152 inbred lines based on 4,620 polymorphic SNPs 54 Proportion of Inbreds University of Ghana http://ugspace.ug.edu.gh 4.3.3 Genetic distance and relatedness among the inbreds Figure 4.3 presents the distribution of pairwise genetic distance of the inbred lines. Rogers genetic distance between pairwise comparisons of all the 152 inbreds ranged from 0.009 to 0.276 with an average distance of 0.237. Majority of the distances between each pair of inbreds (71.05%) fell between 0.2001 and 0.2500. Distances between 0.2501 and 0.3000 represented 25.39%, giving 96.44% of the pairwise genetic distances falling between 0.2001 and 0.3000. The distribution of the relative kinship coefficients between pairs of inbreds are presented in Figure 4.3 (B). Results show that the relatedness coefficient varied from 0.01 to 0.59 with an overall average of 0.374. However, 55.18 and 37.09% of the values fell in the range of 0.3001 to 0.4000 and 0.4001 to 0.5000 resulting in about 92% of the kinship coefficients falling within 0.3001 and 0.5000. 55 University of Ghana http://ugspace.ug.edu.gh 80% 71.05% A 70% 60% -- 50% 40% 30% 25.39% 20% 10% 1% 1.43% 0.70% 0.58% 0% 0% Genetic distance (Rogers 1972) 60.00% 55.18% B 50.00% -- 40.00% 37.09% 30.00% 20.00% 10.00% 3.39% 1.25% 2.31%0.28% 0.51% 0% 0.00% Relative kinship Figure 4.3 Distribution of pairwise (A) Rogers genetic distance and (B) relative kinship among 152 orange maize inbred lines based on 4,620 SNPs 56 Percentage Percentage University of Ghana http://ugspace.ug.edu.gh 4.3.4 Population structure analysis Figure 4.4 presents results of the second order of likelihood, ΔK, and the clustering of individual inbreds into different groups from K =2 to K =4. Results indicated that there was higher likelihood at K= 4 and sharp decrease of the likelihood at K =5. At K= 6, the likelihood increased again. But the likelihood at K = 4 was higher than that observed at K =6. Therefore, K= 4 was used to partition the inbreds into the respective groups. Each individual is represented by a single vertical line that is partitioned into K coloured segments on the x-axis, with lengths proportional to the estimated probability of membership (y-axis) to each of the K inferred clusters. Thus, a cut-off point of 70% of estimated probability of membership was used to decide on the individuals of group 1 (red colour in the Figure) and group 4 (yellow colour) that showed a shared probability among clusters. All the inbreds with a probability of membership less than 0.7 were assigned arbitrarily to a “mixed” group. A total of 71% of the inbreds were assigned to a group. Group 1 is the largest with 47 inbreds out of 152 followed by group 4 with 30 inbreds. Group 3 (blue, in Figure 4.4) is composed of 21 inbreds and Group 2 (green) is constituted of 10 inbreds. 57 University of Ghana http://ugspace.ug.edu.gh (A) (B) k2 k3 k4 Figure 4.4 Analysis of the population structure of 152 orange maize inbred lines, (A) Estimated Δk over ten repeats of STRUCTURE analysis; (B) Population structure assessed by STRUCTURE for k =2, k =3, and k = 4 58 University of Ghana http://ugspace.ug.edu.gh 4.3.5 Principal coordinate analysis and Distance-based grouping Figure 4.5 shows the percentage variation in the data explained by each of the new axes in the Principal coordinate analysis (PCoA). The results showed that the loci that loaded high on axis 1 and 2 were influential in discrimination among inbreds with a total contribution of 11.5% of the total variation. Results presented in Figure 4.6 indicated a clear disagreement between clusters obtained using the STRUCTURE and the PCoA. Figure 4.6 showed that individuals in groups 1, 3 and “mixed” were clustered together while group 2, which was highly correlated with loci loaded on axis 1, still stands alone as a different group. Group 4 was subdivided into 3 with a portion sharing similar properties with group 3, 1 and “mixed”. On the contrary, genetic distance-based phylogenetic tree revealed almost the same pattern in the population of inbreds compared to the results of the STRUCTURE analysis (Figure 4.7). Members in each of the a priori group were consistently identified by the neighbour-joining method as belonging to the group. In this analysis, two sub-groups were identified in Group 1 with the sub-group (b) containing the majority of members of Group 2. Group 3 while containing a few members of Group 2, has few of its members in Group 3. The “mixed” group obtained from STRUCTURE analysis appeared as a group of members sharing properties with members of other groups and, was singled out as a genetically distant group (Figure 4.7). From the two-dimension plot of PCoA (Figure 4.6), it could be seen that the group 1 obtained from STRUCTURE analysis was completely mixed with the “mixed” group. Furthermore, there were also some subdivisions in Groups 3 and 4. Group 3 could be subdivided into two to obtain subgroup a (3a) and subgroup b (3b) with the subgroup b being mixed with the “mixed” group. Also, the PCoA revealed that Group 4 could be further subdivided into 3 to obtain subgroup a (4a) , b (4b), and c (4c) with subgroup b grouped together with the “mixed” group. 59 University of Ghana http://ugspace.ug.edu.gh Figure 4.5 Contribution of the axis to the total variation explained in the genotypic data set Figure 4.6 Population structure of 152 orange maize inbred lines using 4,620 SNP markers and the a priori information from STRUCTURE analysis in Principal Coordinate Analysis (PCoA) 60 University of Ghana http://ugspace.ug.edu.gh Figure 4.7 Neighbour-joining grouping of 152 orange maize inbred lines based on Rogers genetic distance calculated from 4,620 SNP markers 61 University of Ghana http://ugspace.ug.edu.gh 4.4 Discussions Genetic purity of an inbred line ensures that it possesses genetic properties for which it has been selected. However, during seed regeneration and maintenance (especially, at different places), minor changes in allelic frequencies may occur at different stages of the process (Warburton et al., 2010) without noticeable adverse impact. However, significant changes in the genetic constitution of an inbred line may affect its performance and even lead to the production and distribution of wrong hybrids or varieties (Semagn et al., 2012). The inbreds used in the present study were at S6 generation and therefore, were considered almost pure or fixed. The study clearly showed the presence of some level of genetic heterogeneity in about 8% of the total inbred lines (Heterogeneity >5%) but more importantly, highlighted the homogeneity of the majority, 92%, of the inbreds (Heterogeneity <5%) according to Semagn et al. (2012). Among the homogeneous inbreds, TZEEIOR 54, TZEEIOR 121, TZEEIOR 122 and TZEEIOR 148 reached the expected heterozygosity (< 1%) of S6 lines. On the contrary, Ertiro et al. (2017) reported 30% heterogeneity among 30 inbreds from IITA maize programme and justified it by a probable pollen contamination and seed admixture during inbred line maintenance, as those inbreds were commonly used. The high homogeneity revealed in this study among IITA inbred lines could therefore be explained by the fact that these inbreds are newly developed and the seeds used were from the same source. The study also identified TZEEIOR 75, TZEEIOR 76, TZEEIOR 161 and TZEEIOR 38 as highly heterogeneous inbreds with heterogeneity level greater than 12.5%. This could be explained by seed admixture during processing or lower levels of inbreeding than expected. In the latter, these inbreds might need to go through another three or four generations of inbreeding to be fixed while two generations of selfing may be sufficient to fix those with heterozygosity between 5.1 and 12.5%, assuming the heterogeneity observed was not due to seed admixture. 62 University of Ghana http://ugspace.ug.edu.gh The results of the pairwise genetic distance among the 152 orange maize inbreds revealed that about 96% of the pairs fell within 0.2 and 0.28 suggesting a narrow genetic variation among the lines. This finding supports that of Ertiro et al. (2017) who reported a range of 0.200 and 0.300 for 55% of the pairs among inbreds from IITA maize programme. This could be attributed to the background of the inbreds as all of them were selected for the same main traits. Furthermore, 92% of the inbreds involved in the present study showed relative kinship coefficients ranging from 0.3 to 0.5. These results indicate a moderate relatedness among the inbreds and could be attributed to a shared pedigree. Similarly, Semagn et al. (2012) reported a range of 0.05 to 0.5 for 79% of the kinship coefficients among 450 maize inbred lines from CIMMYT’s eastern and southern Africa maize breeding programmes. The similarity arose from the narrow genetic base of the collection used in the studies. Moreover, even though inbreds in this study were extracted from a genetically broad base population, the selection imposed for the same traits might have led to the isolation of similar genetic constitutions. On the other hand, this finding contradicts those of Ertiro et al. (2017) who reported 59% of the kinship coefficients close to zero for 265 inbred lines; Dao et al. (2014), who reported 61% of the kinship coefficients close to zero for 100 inbred lines, and Wen et al. (2011), who reported that about 60 % of the pairwise kinship coefficients among 359 inbred maize lines were close to zero. It should be emphasized that these authors used in their studies, larger panels of inbreds lines from different sources, thus confirming the argument that the moderate relatedness found among the inbreds in the present study may be attributed to their genetic backgrounds. Variability is needed among individuals subjected to selection for any given trait to ensure high performance of new varieties in plant breeding. It is also well known that crosses between genetically distant parents, in the presence of linkage disequilibrium, increase progeny variance 63 University of Ghana http://ugspace.ug.edu.gh and, thus ensures successful selection of potential parents. Measures such as pairwise genetic distance and relative kinship coefficients are, therefore, often used by plant breeders to assess genotypes before their use in planned crosses. The present study confirms the relatedness and the identity of the inbreds while highlighting the narrow genetic distance among them. But how much is the genetic distance between two parents determining the variability needed in the progeny? In fact, controversial reports about the relationship between genetic distance and progeny genetic variance have been published (Benchimol et al., 2000; Reif et al., 2003; Bertan et al., 2007; Flint- Garcia et al., 2009; Hung et al., 2012). Also, the performance of the inbreds roots in an extra-early broad base population improved both for Striga resistance and elevated levels of carotenoids. Thus, genetic differences could have been expected, assuming a sharp difference among genes for earliness, Striga resistance, and carotenoid accumulation. But the markers used in this study were randomly selected and their relation with the aforementioned traits is unknown and could explain the genetic close relationships among the inbreds. Alternatively, it could be that similar group of genes with similar effects or different group of genes with similar effects were present among the inbreds. The results suggest that specific markers may be needed to explore the true genetic relationship among these inbreds. The maize inbred lines involved in the present study were selected among the top-performers of a population of 253 extra-early inbreds newly developed for Striga resistance and assumed elevated levels of carotenoids due to the colour of their endosperms. They also showed some levels of drought tolerance for two seasons’ evaluations during which high variability was detected for the two stresses. Therefore, one of the major objectives of the current study was to explore whether there could be genetically distinct groups among the selected and if yes, identify quantitative evidence supporting their field performance. In this regard, 4,620 DArTseq SNP 64 University of Ghana http://ugspace.ug.edu.gh markers evenly distributed on the 10 chromosomes were used as genetic variables that were subjected to STRUCTURE, PCoA, and Neighbour-Joining cluster analyses to determine genetic differentiation, population structure and the pattern of relationship among the inbreds. The model-based population structure assigned 71% of the inbreds into four genetically distinct groups with 39% sharing properties of at least two of the defined groups. Using the a priori subdivision of the population in the PCoA and tree-based analysis, results indicated concordance in population partition between STRUCTURE and phylogenetic tree analysis. The PCoA showed a more complex structure with subdivisions within groups 3 and 4 and overlaps of subgroups (“mixed”, group1, group 3 a, and group 4 b). This result is in disagreement with the reports of Yang et al. (2011), Dao et al. (2014), and Wu et al. (2016) that indicated rather a high concordance between model-based structure and the PCoA. However, Yang et al.(2011) highlighted the spurious inference of population structure that often occurs for populations which do not fulfill the algorithm assumptions among which is the assumption of Hardy–Weinberg equilibrium (Falush et al., 2003). The inbreds in this study did not conform to Hardy–Weinberg equilibrium because they were highly selected and thus, the shared pedigree might have caused the low genetic differentiation among some individuals. Many reasons have been given to justify the unreliability of tree-based population structure (Dao et al., 2014). Partly, because a tree-based analysis only classifies individuals to a fixed position while PCoA and STRUCTURE provide the probability for an individual to belong to a group. Yang et al. (2011) also cautioned on the use of phylogenetic trees to explore the genetic relationship among individuals with complex genetic relatedness. Hence, the clustering inferred by the PCoA provided more details about the groups obtained from STRUCTURE and revealed the complexity in grouping the majority of the inbreds due to their relatedness. 65 University of Ghana http://ugspace.ug.edu.gh In conclusion, the study revealed that majority (92%) of the inbreds were genetically pure and fixed. The inbreds were closely related with moderate relatedness. However, 71% of the inbreds were assigned to four different groups while 39% showed admixture using STRUCTURE software. Only 10 inbreds (TZEEIOR6, TZEEIOR8, TZEEIOR9, TZEEIOR10, TZEEIOR12, TZEEIOR13, TZEEIOR15, TZEEIOR19, TZEEIOR39, and TZEEIOR40) were consistently assigned to a distinct genetic group both by STRUCTURE and PCoA. 66 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE 5.0 Estimates of combining abilities and heterosis of extra-early maturing orange inbred lines and performance of their hybrids 5.1 Introduction Drought and Striga constitute the prominent constraints to maize production in savanna agro- ecologies of SSA. Breeding high yielding extra-early maturing hybrids, which combine resistance to Striga and tolerance to drought, will increase the productivity and production of maize in the sub-region. Appropriate breeding strategy in such context, needs to identify parental inbred lines that form superior hybrids under these stresses and also understand the genetic basis for hybrid performance. Therefore, information regarding the nature of combining ability of the parents available as well as the nature of gene action involved in the expression of the desirable traits are of great importance. General combining ability (GCA) and specific combining ability (SCA) effects are indicators of the potential value of inbred lines in hybrid combinations and in grouping materials into heterotic groups. The use of heterotic groups, when aided with good testers in a breeding program can result in the production of high yielding hybrids. However, when there are no appropriate testers to be used, a diallel mating design can be used to identify inbred lines with superior combining ability that may be used as testers in a breeding program, and for identifying superior crosses that may be candidates for single cross hybrids. The study was designed to: i. determine the combining ability of extra-early orange parental lines and the gene action for resistance to Striga and tolerance to drought; ii. identify potential testers for grouping inbred lines into heterotic groups; iii. determine the performance of hybrids and examine the stability of performance in tolerance to Striga and drought. 67 University of Ghana http://ugspace.ug.edu.gh 5.2 Materials and methods 5.2.1 Genetic materials Fifteen promising inbreds, each with at least one positive selection index either under Striga- infested or drought stress environments selected among the 253 inbreds described in Chapter 3 plus two well-known inbred testers (TZdEEI 7 and TZdEEI 12) constituted the parents of the plant material used in this study (Table 5.1). The seventeen extra-early inbred lines were crossed in all possible crosses to generate crosses and reciprocals employing the complete diallel mating design in the IITA breeding nursery, Ibadan in 2016. However, crosses and reciprocals were mixed assuming no maternal effect, giving 136 single cross hybrids generated. The 136 single cross hybrids (Half Diallel) plus 4 experimental hybrid checks were used in the study. 5.2.2 Field Evaluation and data collection A 10 x 14 (10 entries x 14 blocks) alpha lattice design with two replications was used in the present study. Single-row plots each of 3 m long with a spacing of 0.75 m between two adjacent rows and 0.40 m between plants within the row were used. Three seeds were planted per hill, and the seedlings were later thinned to two per hill about 2 weeks after emergence to give a final population density of about 66,667 plants ha-1. Similar to the inbreds trial described in Chapter 3, the hybrid trial was evaluated at three locations (Abuja, Mokwa, and Ikenne), under Striga infestation, induced drought stress, and optimal environments for two years. Evaluation under Striga infestation was carried out in Abuja and Mokwa while those under optimal conditions were conducted at Abuja, Mokwa, and Ikenne, in 2016 and 2017 during June-October rainy seasons. In addition, the same trial was evaluated under managed drought stress during January-May of 2016/2017 and 2017/2018 dry seasons at Ikenne. The management of the trials and data collection 68 University of Ghana http://ugspace.ug.edu.gh under the three conditions (managed drought stress, Striga infestation, and optimal) were done according to the methods described in Chapter 3, section 3.2.2. Table 5.1 Characteristic of the selected 17 extra-early maturing maize inbred lines used in the diallel crosses with their reaction to drought stress and Striga infestation in 2015 Reaction to Entry INBRED Grain colour Pedigree Drought Striga 1 TZEEIOR 12 orange 2009 TZEE-ORI STR S5 3-1/1-1/3-5/6-1/1-1/1 T T 2 TZEEIOR 42 deep orange 2009 TZEE-ORI STR S5 8-3/3-2/3-3/5-2/3-1/1 T T 3 TZEEIOR 53 light orange 2009 TZEE-ORI STR S6 9-1/1-2/2-3/3-1/2-1/1 T S 4 TZEEIOR 76 deep yellow 2009 TZEE-ORI STR S6 12-2/2-1/2-1/1-2/2-2/2 T T 5 TZEEIOR 100 deep orange 2009 TZEE-ORI STR S6 15-2/2-2/2-3/3-1/3-1/3 T T 6 TZEEIOR 113 orange 2009 TZEE-ORI STR S6 20-2/2-1/2-1/2-1/1-1/1 T T 7 TZEEIOR 130 orange 2009 TZEE-ORI STR S6 35-1/2-3/3-1/1-1/3-1/1 T T 8 TZEEIOR 141 orange 2009 TZEE-ORI STR S6 46-1/2-1/3-2/2-2/3-1/2 T T 9 TZEEIOR 145 orange 2009 TZEE-ORI STR S6 46-2/2-1/3-2/2-1/2-1/1 T S 10 TZEEIOR 161 orange 2009 TZEE-ORI STR S6 53-1/2-2/3-2/2-4/4-1/1 T S 11 TZEEIOR 196 deep orange 2009 TZEE-ORI STR S6 68-1/3-2/3-1/2-1/3-1/1 S T 12 TZEEIOR 218 orange 2009 TZEE-ORI STR S5 80-2/3-3/3-1/2-1/6-1/1 T T 13 TZEEIOR 219 orange 2009 TZEE-ORI STR S6 80-2/3-1/1-1/1-3/3-1/1 T T 14 TZEEIOR 222 orange 2009 TZEE-ORI STR S5 80-3/3-3/3-1/2-1/3-1/1 T T 15 TZEEIOR 223 orange 2009 TZEE-ORI STR S6 80-3/3-3/3-2/2-1/6-1/1 T T 16 TZdEEI 7 pale yellow TZEE-Y POP STR 106 S6 189/194-1/1-1/2-4/5-7/9 T T 17 TZdEEI 12 light orange TZEE-Y POP STR 106 S5 2/194-1/1-1/2-1/2-4/4 T S T = tolerant; S = susceptible. 69 University of Ghana http://ugspace.ug.edu.gh 5.2.3 Data Analysis and genetic parameter estimate Before the statistical analysis, the distribution of the data on each measured trait was assessed using the UNIVARIATE procedure. Traits deviating from the normal distribution according to the Shapiro-Wilk test were transformed with either a natural logarithm or square root to obtain a normal distribution. 5.2.3.1 Model description The block model (B) was described as: Environment /Replication /Block /Plot and the genotype model (G) as: genotype (Gen) = GCA1 +GCA2 +SCA. The environment in the model is defined as year and location interaction. The main effects of years and locations were assumed to be negligible and not accounted for. Data analysis was performed using the following general linear model: Y = B + G + B*G + error, where Y is the measured parameter. The linear mixed model was implemented in SAS V9.4 by modifying the macro written by Isik (2009). Data from each trial condition were pooled together and analysed by fitting environments, replicates within environments, and checks as fixed factors while blocks within replicates, GCAs and SCAs and their interactions with the environments were fitted as random effects. The detailed model described in equation 5.1 was used with MIXED procedure. Y = Xβ + ZGCAuGCA + ZSCAvSCA + ZGCA x EnvuGCA x Env + ZSCA x EnvvSCA x Env + e (Equation 5.1) Where: Y is the vector of unadjusted observations. The X-matrix and Z-matrices are incidence matrices belonging to their respective components; β is a vector of fixed effects of environment (Env), replicates within environments, and check; uGCA and vSCA are vectors of general combining ability (GCA) effects and specific combining ability (SCA) effects across the environments. 70 University of Ghana http://ugspace.ug.edu.gh uGCAxEnv, vSCAxEnv, and e are the interaction terms of combining abilities with environments and pooled error. The random effects in the model were assumed to follow a multivariate distribution with means and variances defined by the equations 5.2 and 5.3, respectively. uGCA 0 v 0 SCAu 𝐸 GCA x Env = 0 (Equation 5.2) vSCA x Env 0 [ 𝑒 ] [0] u n 2GCA ⨁ j σGCA 0 0 0 0 v ⨁n SCA σ 2 0 j SCA 0 0 0 Var uGCA x Env = 0 0 IEnv⨂IGenoσ 2 GCA x Env 0 0 vSCA x Env 0 0 0 IEnv⨂IGenoσ 2 SCA x Env0 [ e ] [ 0 0 0 0 R ] (Equation 5.3) Where ⨁ and ⨂ represent the Kronecker direct sum and direct product, respectively (Cullis and Gleeson, 1991). Where 0 is a null matrix; IEnv and IGeno are identity matrices with order equal to the number of environments (Env) and number of genotypes (Geno). 5.2.3.2 Genetic components variance structures Different covariance structures in different parts of the genetic variance were fitted by using multiple RANDOM statements with different options. Thus, banded Toeplitz covariance structure option was used to group parents together and estimate a single variance component across all levels of parents. In addition, GCA and SCA variances were considered heterogeneous across all the environments and, therefore, were fitted using the diagonal matrix as a direct sum of individual matrices with order equals to the number of environments (Equation 5.3). The heterogeneous variance structure was implemented using SUBJECT option in the random statement. 71 University of Ghana http://ugspace.ug.edu.gh The SUBJECT option was removed for the combined analysis for test environments because the loglikelihood failed to converge. Restricted maximum likelihood (REML) in MIXED procedure was used to estimate variances of all random effects. 5.2.3.3 Significance tests The Z probability was used to test the significance of random effects variances while the fixed effects were tested by F probability. Kenward-Roger method was used to approximate the degree of freedom of fixed effects in the analysis of data from each trial. But in the combined analysis of data from all test environments, the residual method was used. 5.2.3.4 Genetic value and heritability estimates The computed best linear unbiased predictions (BLUPs) of the GCA and SCA effects were extracted and used to calculate breeding values of hybrids as the sum of the GCA effects of both parents plus the SCA effects which is the deviation from mean of the estimated GCA values. Broad-sense (H2) and narrow-sense (h2) heritability estimates were calculated on full-sib family mean basis unbiased by variances of all factors in random statement as follow: 2xσ2 + σ2 2xσ2 2 2 GCA SCA GCA + σSCA H = 2 = σP 2 2 2xσ 2 2 2 2 2xσ + σ + GCAxEnv σ + SCAxEnv σBlk σ+ + errorGCA SCA t t b rxt 2xσ2GCA 2xσ 2 2 GCAh = σ2 = 2xσ2 σ2 σ2 σ2P 2xσ2 + σ2GCA SCA + GCAxEnv t + SCAxEnv + Blkt + error b rxt Where, t, b, and r are respectively, number of environments, blocks, and replications and 𝜎2𝑃 , the phenotypic variance. The option ASYCOV of the MIXED procedure was used to produce the variances of variance components (diagonal elements) and the covariances (off-diagonal elements) between them. The 72 University of Ghana http://ugspace.ug.edu.gh standard error of heritabilities were calculated using Dickerson formulae (Isik, 2009), assuming constant variance of heritability as follow: 2 𝑺. 𝑬(ℎ2 2𝑥 𝑉𝑎𝑟(𝜎 ) = √ 𝐺𝐶𝐴 ) 2 ; (𝜎2𝑃) 2 2𝑥 𝑉𝑎𝑟(𝜎 2 𝐺𝐶𝐴)+𝑉𝑎𝑟(𝜎 2 𝑆𝐶𝐴)+2𝐶𝑜𝑣(2𝑥𝜎 2 2 𝑺. 𝑬(𝐻 ) = √ 𝐺𝐶𝐴 ,𝜎𝐺𝐶𝐴). (𝜎2 2 𝑃) Where 𝑉𝑎𝑟(𝜎2𝐺𝐶𝐴), is the variance of GCA component of variance, 𝑉𝑎𝑟(𝜎 2 𝑆𝐶𝐴), is the variance of SCA component of variance, 𝐶𝑜𝑣(𝜎2 2𝐺𝐶𝐴, 𝜎𝐺𝐶𝐴), is the covariance between GCA and SCA variances as given by ASYCOV option. Trial conditions were assumed to represent the different growing conditions in West and Central Africa, which is the target zone of most of the materials developed by IITA-Ibadan. Hence, in each trial type and across growing conditions, the environments were set fixed to account for their effects on genotypic performance prediction. The alternative which could have considered the environments as a random sample of a population of agroecological zones encountered in the sub- region was not dealt with in this research. 5.2.3.5 Relative importance of GCA and SCA The relative importance of GCA and SCA was investigated using two criteria: ➢ the Baker’s (1978) ratio modified by Hung and Holland (2012) and given as follows: 2σ̂2GCA RB = 2σ̂2GCA + σ̂ 2 SCA Where, RB is the value of the ratio; σ̂ 2 GCA is the variance of general combining ability; σ̂ 2 SCA is variance of specific combining ability. When 𝑅 ≈ 1, Additive gene effects are predominant and predictions can be done based on GCA estimates alone. 73 University of Ghana http://ugspace.ug.edu.gh ➢ The GCA/SCA ratio defined as follows: σ̂2GCA R = σ̂2GCA Where, R is the value of the ratio; σ̂2GCA is the variance of general combining ability; σ̂ 2 SCA is variance of specific combining ability. When 𝑅 > 1, Additive gene effects are primary type of gene action; when 𝑅 < 1, non-additive gene effects are primary type of gene action. 5.2.3.6 Heterotic Groupings of parental inbreds Parental inbred lines were assigned to heterotic groups under Striga infestation, drought stress, and across test environments based on GCA of multiple traits (HGCAMT) method proposed by Badu- Apraku et al.(2013, 2015). The GCA effects of the traits with significant GCA variance were scaled (mean of zero and standard deviation of 1) to make variables comparable and the standardised data set were used to compute Euclidean-based distance matrix. The HGCAMT statistical model is defined as follow: 𝑛 (𝑌𝑖 − ?̅?𝑖) 𝑌 = ∑( ) + 𝜀 𝑠 𝑖𝑗 𝑖=1 where Y is HGCAMT, which is the genetic value that measures relationship among genotypes based on the GCA of multiple traits i to n; Yi is the individual GCA effect of genotypes for trait i, ?̅?𝑖 is the mean of GCA effects across genotypes for trait i, s is the standard deviation of the GCA effects of trait i and 𝜀𝑖𝑗 is the residual of the model associated with the combination of inbred i and trait j. Prior to the grouping, clustering tendency or validity of the clustering was assessed using Hopkins statistic which measures the probability that a given data set is generated by a uniform data distribution and a value less than 0.5 indicates that the data contain meaningful clusters. The optimal number of clusters (K) in the data set was determined through Gap statistics method, 74 University of Ghana http://ugspace.ug.edu.gh silhouette approach, and Elbow method. The K suggested by at least two of the three methods was used to perform the hierarchical grouping. Pairs of inbreds were grouped using Ward’s minimum variance (“ward. D2”, implemented in R) linkage method. Ward’s minimum variance method processes by steps in each of which the pair of clusters with minimum between-cluster distance are merged, thus minimising the total within- cluster variance. To assess whether the cophenetic distances (heights in the dendrogram) reflect the original distances accurately, a correlation between the two was calculated and values ≥ 0.75 are considered to reflect good clustering solution. Finally, Baker Gamma correlation coefficient among the three dendrograms generated were computed and used to check if individuals were clustered the same way under different stress environments with the across test grouping. 5.2.3.7 Inbred testers identification The selection of testers was based on the three criteria proposed by Pswarayi and Vivek (2008). According to the authors, an inbred line is considered as a tester if it: i. belongs to a known heterotic group, ii. has a highly significant positive GCA of grain yield across the test environments, iii. has high yield per se. The assessment of the third criteria was done with reference to the results presented in Chapter 3. 5.2.3.8 Heterosis estimates and analysis Gardner & Eberhart (1966) analysis II was used to examine the heterosis for grain yield of the hybrids across test environments and for measured traits involved in the base index in each of the stress environments. The statistical model of the method is given as follows: 1 𝑌𝑖𝑗 = 𝜇𝑣 + (𝑣𝑖 + 𝑣𝑗) + 𝛾ℎ̅ + 𝛾(ℎ𝑖 + ℎ𝑗) + 𝛾𝑆2 𝑖𝑗 75 University of Ghana http://ugspace.ug.edu.gh 0 where i = j With γ = [ ] and ℎ̅ + ℎ𝑖 + ℎ𝑗 + 𝑆𝑖𝑗 = ℎ 1 where i ≠ j 𝑖𝑗 Where: 𝑌𝑖𝑗: observed value of the cross between parents i and j; 𝜇𝑣.: mean of all parental inbreds; 𝑣𝑖 𝑎𝑛𝑑 𝑣𝑗 : variety effects; ℎ̅: average heterosis contributed by the particular set of genotypes used in crosses; ℎ𝑖: the average heterosis contributed by variety I in its crosses measured as a deviation from ℎ̅ (∑𝑖 ℎ𝑖 = 0); 𝑆𝑖𝑗 : specific heterosis that occurs when variety i is mated to variety j; ℎ𝑖𝑗: is the overall heterosis due to differences in gene frequencies in parental inbreds, i and j and to dominance in crosses. The software GENES v.38 (Cruz, 2016) was used to execute the model through two-stage analysis where the error variances and means generated in the analysis of variance (stage I) were used in the model described above (stage II). Mid-parent heterosis (MPH), heterobeltiosis (HPH) for each of the traits were estimated as follows: F1−MP F −HPMPH = ; and HPH = 1 . MP HP Where F1 is the mean performance of the F1 hybrid, MP is the mean of the two inbred parents and HP is the mean of the better parent. 5.2.3.9 Genetic correlation among test environments and stability analysis of hybrid performance META-SAS v4 (Vargas et al., 2013) was used to calculate the heritabilities of the environments and also check the genetic correlation among them before the stability analysis. A graphical assessment of the stability of hybrids performance through GxE interaction analysis was performed in PBTools v1.3 following the GGE Biplot model (Yan et al., 2000; Yan and Kang, 2002): 76 University of Ghana http://ugspace.ug.edu.gh 𝑌𝑖𝑗 − 𝑌?̅? = 𝜆1𝜉𝑖1𝜂𝑗1 + 𝜆2𝜉𝑖2𝜂𝑗2 + 𝜀𝑖𝑗 Where: 𝑌𝑖𝑗, is the average yield of genotype i in environment j; 𝑌?̅?, is the average yield over all genotypes in environment j; 𝜆1and 𝜆2, are the singular values decompositions for the first and second principal components, PC1 and PC2, respectively; 𝜉𝑖1and 𝜉𝑖2, are vector scores of genotype i on PC1 and PC2, respectively; 𝜂𝑗1and 𝜂𝑗2, are vector scores of environment j on PC1 and PC2, respectively; and 𝜀𝑖𝑗, is the residual of the associated model. 5.3 Results 5.3.1 Genetic analysis of performance of extra-early maturing maize inbred lines under contrasting environments Table 5.2 presents variance components and heritability estimates for grain yield and other measured traits under Striga-infested environments. There were significant (𝑝 < 0.01) differences among the 17 inbred parents for GCA component of genetic variance for grain yield, Striga damage at 10 weeks after planting (WAP), number of emerged Striga plants at 8 WAP, husk cover, and ear aspect. Highly significant (𝑝 < 0.001) differences were observed for Striga damage at 10 WAP and number of emerged Striga plants at 10 WAP while GCA variances for anthesis-silking interval (ASI) and ears per plant (EPP) were not significantly different from zero. The SCA component of genetic variance for grain yield, anthesis-silking interval, ears per plant and ear aspect were significantly different from zero (𝑝 < 0.001). Also, no significant differences were detected among the inbreds for number of emerged Striga plants, Striga damage at 8 and 10 WAP, and husk cover. The GCA-environment and SCA-environment interactions effects were highly significant (𝑝 < 0.001) for all the measured traits indicating high differential response among hybrids due to the genetic constitution of inbred parents. Generally, the contribution of GCA variance to the total 77 University of Ghana http://ugspace.ug.edu.gh explained variance were low compared to that of SCA variance for the reported traits. GCA variance contribution ranged from 0.9% for ASI to 3.79% for Striga damage at 8 WAP (SDR8) while the SCA variance contribution ranged from 0.26% for number of emerged Striga plants at 8 WAP (ESP8) to 20.14% for grain yield. 78 University of Ghana http://ugspace.ug.edu.gh Table 5.2 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds under four Striga-infested environments in Nigeria, 2016-2017 Number of Anthesis- Striga damage Husk Ears Ear Num. Grain Emerged Striga Parameters Silking (1-9) Cover per Aspect DF Yield, kg ha-1 plants interval (1-9) plant (1-9) 8 WAP 10WAP 8WAP 10WAP Environment 3 * *** ** * *** *** *** *** NS Replicate (Env) 4 NS NS NS NS NS NS NS NS NS Subject Random effect variances Block (Env x Rep) 412513*** 0.64*** 0.38*** 0.31*** 0.04*** 0.03*** 0.22*** 0.01*** 0.26*** GCA Env 42187** 0.04ns 0.05*** 0.05** 0.01** 0.00*** 0.04** 0.00ns 0.03** SCA Env 340054*** 0.59*** 0.02ns 0.07ns 0.01* 0.00ns 0.03ns 0.01*** 0.19*** Env x GCA 974.38*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** - 0.00*** Env x SCA 17095*** 0.00*** 0.01*** 0.01*** 0.00*** 0.00*** 0.00*** 0.00*** - Residual 875595 2.68 0.94 0.92 0.09 0.07 0.71 0.03 0.63 Narrow sense heritability 0.15 0.07 0.39 0.32 0.15 0.18 0.34 0.08 0.18 S.E. 0.04 0.04 0.08 0.07 0.08 0.07 0.08 0.04 0.05 Broad sense heritability 0.74 0.63 0.46 0.56 0.16 0.22 0.50 0.71 0.72 S.E. 0.11 0.23 0.18 0.17 0.29 0.26 0.15 0.13 0.12 GCA/total variance (%) 2.50 0.90 3.79 3.72 0.98 1.30 3.59 1.31 2.77 SCA/total variance (%) 20.14 14.90 1.44 5.48 0.26 0.59 3.23 19.70 17.36 Baker ratio 0.20 0.11 0.84 0.58 0.88 0.81 0.69 0.12 0.24 GCA/SCA 0.12 0.06 2.63 0.68 3.74 2.20 1.11 0.07 0.16 “*”, “**”, “***” = significance at 𝑝 < 0.05, 𝑝 < 0.01, 𝑝 < 0.001, respectively, “NS” = non-significant. GCA = general combining ability, SCA = specific combining ability, S.E = standard error, Num. DF= numerator degree of freedom, Env = environment, Rep = replication. All “0.00” are not absolute values. All “0” = estimated variance equals zero. 79 University of Ghana http://ugspace.ug.edu.gh The observed relative importance of GCA over SCA was very low for traits such as ASI (0.06) and EPP (0.07) and high for traits such as ESP8 (3.74) and SDR8 (2.63). In addition, Baker ratio for the latter traits as well as the number of emerged Striga plants at 10 WAP (ESP10) were close to 1 with values respectively of 0.88, 0.84, and 0.81. Broad-sense heritability ranged from 0.16 for ESP8 to 0.74 for grain yield. Low (0.07 for ASI) to moderate (0.39 for SDR8) narrow sense heritabilities were observed for all traits. The proportion of additive genetic variance to the phenotypic variance for grain yield was estimated at 0.15±0.04. Table 5.3 presents variance components and heritability estimates for grain yield and other measured traits under drought environments. Genetic variance components were significantly (𝑝 < 0.01) different from zero for most of the reported traits. For grain yield as well as anthesis-silking interval and ear aspect, non-significant differences were found among environments, yet GCA and SCA components of genetic variance for grain yield were significantly (𝑝 < 0.001 and 𝑝 < 0.01, respectively) different from zero. Differences among inbred parents as explained by GCA and SCA components did not follow similar trends for all the traits. The GCA effects were not significantly different from zero for days to 50% anthesis (DYA) and EPP. On the other hand, SCA variance was not significantly different from zero for ASI, days to 50% silking (DYS), ear aspect, and staygreen characteristic (STGC). But, the contribution of GCA component to the total variance, like under Striga infestation, was lower than that of SCA component. The contribution of GCA ranged from 0.24% for EPP to 7.5% for STGC while SCA contribution ranged from 1.96% for STGC to 30.26% for grain yield. 80 University of Ghana http://ugspace.ug.edu.gh Table 5.3 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across two drought environments in Nigeria, 2016-2017 Anthesis- Days Days Plant Husk Ear Leaf Num. Grain Ears per Parameters Silking to to aspect cover Aspect Death DF Yield, kg ha-1 plant interval anthesis silking (1-9) (1-9) (1-9) (1-9) Environment 1 NS NS *** *** * *** *** NS * Replicate (Env) 2 NS NS NS NS NS NS NS NS NS Subject Random effect variances Block (Env x Rep) 3868ns 0 0.21ns 0.74* 0.001ns 0 0.001ns 0 0.04ns GCA Env 12478* 0.17* 0.23ns 0.56* 0.027* 0.04* 0.000ns 0.04* 0.08** SCA Env 120139*** 0.30ns 0.88* 1.17ns 0.137*** 0.28*** 0.009** 0.11ns 0.02ns Env x GCA 0 0.00*** 0.01*** 0.01*** 0 0.00*** 0.000*** 0.00*** 0.00*** Env x SCA 1706*** 0.03*** 0.01*** 0 0.001*** 0.02*** 0 0.01*** 0 Residual 258836 5.00 7.47 12.76 0.411 0.68 0.036 1.10 0.89 Narrow-sense heritability 0.12 0.18 0.14 0.20 0.18 0.16 0.01 0.15 0.39 S.E. 0.05 0.07 0.06 0.07 0.06 0.06 0.04 0.06 0.10 Broad-sense heritability 0.69 0.34 0.41 0.41 0.65 0.67 0.50 0.39 0.44 S.E. 0.13 0.19 0.15 0.16 0.13 0.13 0.15 0.17 0.16 GCA/total variance (%) 3.14 3.14 2.63 3.65 4.64 4.18 0.24 2.84 7.50 SCA/total variance (%) 30.26 5.38 9.94 7.70 23.77 27.30 19.68 8.75 1.96 Baker ratio 0.17 0.54 0.35 0.49 0.28 0.23 0.02 0.39 0.88 GCA/SCA 0.10 0.58 0.26 0.47 0.20 0.15 0.01 0.32 3.83 “*”, “**”, “***” = significance at 𝑝 < 0.05, 𝑝 < 0.01, 𝑝 < 0.001, respectively, “NS” = non-significant. GCA = general combining ability, SCA = specific combining ability, S.E = standard error, Num. DF= numerator degree of freedom, Env = environment, Rep = replication. All “0.00” are not absolute values. All “0” = estimated variance equals zero. 81 University of Ghana http://ugspace.ug.edu.gh Moreover, Table 5.3 indicated that GCA was predominant for STGC with the GCA/SCA ratio equal to 3.83 and Baker’s ratio of 0.88, indicating that additive gene effects were the primary type of gene action for STGC. The relative importance of GCA over SCA was 0.10 for grain yield. Almost equal contribution of GCA and SCA to the genetic variance was observed for DYS (0.47) and for ASI (0.58). Broad-sense heritability ranged from moderate to high with the highest values observed for grain yield (0.69), husk cover (0.67), and for plant aspect (0.65). The proportion of additive genetic variance to the total phenotypic variance varied from 0.01 for EPP to 0.39 for STGC. Narrow sense heritability for grain yield and ASI were low under drought stress with values of 12±5% and 18±7%, respectively. Baker’s ratio was 0.88 for STGC. Table 5.4 presents variance components, heritability estimates and combining ability ratios for grain yield and other measured traits under optimal environments. Significant differences were observed among the five optimal environments for all the measured traits. Despite the very high variability among the environments, the contribution of components of genetic variance to the total variance was quite important under optimal conditions than under Striga-infested and managed drought stress environments. The GCA component of variance ranged from 1.87% for EPP to 16.04% for days to 50% anthesis while SCA component of variance varied from 9.08% for husk cover to 31.75% for grain yield. In addition, interactions between environments and components of genetic variance were highly significant for all traits. The relative importance of GCA over SCA as given by the GCA/SCA ratio varied from 0.09 for ears per plant to 0.85 for days to 50% anthesis. In addition to DYA, the predominance of SCA variance was observed in traits such as grain yield (0.22), husk cover (0.26), plant aspect (0.29), and ear aspect (0.39). 82 University of Ghana http://ugspace.ug.edu.gh Table 5.4 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across five optimal environments in Nigeria, 2016-2017 Anthesis- Days Days Plant Husk Ears Ear Num. Grain Root Stalk Parameters -1 Silking to to aspect cover per Aspect DF Yield, kg ha lodging lodging interval anthesis silking (1-9) (1-9) plant (1-9) Environment 4 *** *** *** *** *** *** *** *** *** ** Replicate (Env) 5 * NS ** ** *** * NS NS NS * Subject Random effect variances Block (Env x Rep) 111878*** 0.02* 0.38*** 0.35*** 0.04*** 0.03ns 0.00ns 0.04* 2.8* 8.5** GCA Env 102138*** 0.06*** 0.43*** 0.51*** 0.03*** 0.06* 0.00* 0.06*** 10.4*** 19.0*** SCA Env 462770*** 0.07** 0.51*** 0.66*** 0.10*** 0.24* 0.01*** 0.15*** 14.7*** 23.8*** Env x GCA 6168.21*** 0.00*** 0.00*** 0.00*** 0 0.00*** 0 0.00*** 0 0.5*** Env x SCA 64136*** 0.00*** 0.01*** 0.00*** 0.01*** 0.02*** 0.00*** 0.02*** 1.0*** 4.0*** Residual 710255 0.48 1.34 1.69 0.37 2.26 0.02 0.50 62.0 93.3 Narrow sense heritability 0.27 0.49 0.56 0.55 0.30 0.21 0.12 0.36 0.49 0.52 S.E. 0.05 0.08 0.07 0.07 0.06 0.06 0.04 0.06 0.07 0.08 Broad-sense heritability 0.88 0.78 0.89 0.90 0.80 0.61 0.75 0.82 0.84 0.85 S.E. 0.09 0.12 0.09 0.09 0.11 0.18 0.13 0.11 0.10 0.10 GCA/total variance (%) 7.01 8.93 16.04 15.96 5.53 2.35 1.87 7.63 11.48 12.75 SCA/total variance (%) 31.75 10.47 19.18 20.52 18.99 9.08 20.13 19.62 16.22 15.98 Baker ratio 0.31 0.63 0.63 0.61 0.37 0.34 0.16 0.44 0.59 0.61 GCA/SCA 0.22 0.85 0.84 0.78 0.29 0.26 0.09 0.39 0.71 0.80 “*”, “**”, “***” = significance at 𝑝 < 0.05, 𝑝 < 0.01, 𝑝 < 0.001, respectively, “NS” = non-significant. GCA = general combining ability, SCA = specific combining ability, S.E = standard error, Num. DF= numerator degree of freedom, Env = environment, Rep = replication. All “0.00” are not absolute values. All “0” = estimated variance equals zero. 83 University of Ghana http://ugspace.ug.edu.gh For ASI, DYA, DYS, root and stalk lodging, the values of GCA/SCA ratio was close to one, indicating equal proportions of GCA and SCA variances in determining genetic variability for these traits. Broad-sense heritability under rain-fed conditions was higher compared to the stress environments. It ranged from 0.61 for husk cover to 0.9 for days to silking. But the proportion of the phenotypic variance due to additive genetic effects varied from 0.12 for EPP to 0.56 for DYA. For grain yield, ASI and ear aspect, this proportion was 0.27, 0.49, and 0.36, respectively. Table 5.5 presents variance components, heritability estimates and combining ability ratios for grain yield and other measured traits across test environments. There were significant (𝑝 < 0.001) differences among environments and blocks within replicates nested in environments with the latter justifying the need to recover inter-blocks information. Highly significant (𝑝 < 0.001) differences were also found among inbred parents for interactions between environments and components (GCA and SCA) of genetic variance for all the measured traits. There were significant (𝑝 < 0.05) differences in GCA component of genetic variance for most of the measured traits except for husk cover, ears per plant, and ear aspect. The results indicated, in addition, that there were highly significant (𝑝 < 0.001) differences in SCA estimates among hybrids for grain yield, days to 50% anthesis, and for ear aspect. However, there were not significant differences among crosses for anthesis-silking interval and ear height with regard to SCA estimates. The GCA/SCA ratio, was very low for husk cover (0.04) and ears per plant (0.09), thus indicating the predominance of SCA effect for these traits among this set of inbreds. The highest value of this ratio was observed for stalk lodging, 3.03, indicating the predominance of GCA effect for the trait. Moreover, similar relative importance of GCA over SCA were observed for days to 50% anthesis (1.29), days to 50% silking (1.05), and also for plant and ear height with values of 1.18 84 University of Ghana http://ugspace.ug.edu.gh and 2.55, respectively. The contribution of the additive genetic variance to the phenotypic variance varied from 2% for husk cover to 56% for stalk lodging. This contribution of the additive gene effects to phenotypic variance was 20% for grain yield, 21% for anthesis-silking interval, 11% for ear aspect, and 0.07 for ears per plant with total genetic variance contributing 65, 34, 56 and 43%, respectively, to the total phenotypic variance. 85 University of Ghana http://ugspace.ug.edu.gh Table 5.5 Variance components, heritabilities, and combining ability ratios for grain yield and other traits of extra-early maturing diallel crosses among 17 selected inbreds across 11 test environments in Nigeria, 2016-2017 Grain Anthesis- Days Days Plant Ear Husk Ears Ear Num. Root Stalk Parameters Yield, Silking to to height, height, cover per Aspect DF -1 lodging lodging kg ha interval anthesis Silking cm cm (1-9) plant (1-9) Den DF (Residual) 3054 3051 3054 3051 3051 3051 3054 3054 3054 3054 3054 Environment 10 *** *** *** *** *** *** *** *** *** *** *** Replicate (Env) 11 NS NS NS NS *** *** * NS NS * NS Block (Env x Rep) 198804*** 0.24*** 0.78*** 1.69*** 50.75*** 27.74*** 0.09*** 0.00*** 0.11*** 4.7*** 13.0*** GCA 22296* 0.02* 0.20* 0.28* 4.60* 2.60* 0.00ns 0.00ns 0.01ns 1.4* 6.7* SCA 101296*** 0.03ns 0.15*** 0.27** 3.89* 1.02ns 0.03* 0.00** 0.06*** 0 2.2* Env x GCA 45213*** 0.05** 0.21*** 0.33*** 10.88*** 5.57*** 0.05*** 0.00* 0.04*** 6.2*** 8.8*** Env x SCA 290244*** 0.31*** 0.84*** 1.28*** 28.17*** 13.38*** 0.15** 0.01*** 0.11*** 10.7*** 14.8*** Residual 689121 2.07 3.00 5.26 214.27 101.24 1.42 0.03 0.67 59.2 96.3 Narrow sense heritability 0.20 0.21 0.46 0.54 0.30 0.35 0.02 0.07 0.11 0.35 0.56 S.E. 0.08 0.09 0.14 0.17 0.11 0.12 0.05 0.05 0.07 0.14 0.17 Broad-sense heritability 0.65 0.34 0.64 0.80 0.42 0.41 0.25 0.43 0.56 0.35 0.66 S.E. 0.12 0.12 0.15 0.19 0.13 0.13 0.12 0.12 0.12 0.14 0.17 GCA/total variance (%) 1.66 0.86 3.80 3.08 1.47 1.72 0.06 0.32 0.72 1.69 4.75 SCA/total variance (%) 7.52 1.04 2.94 2.93 1.25 0.67 1.65 3.51 5.75 0.00 1.56 Baker ratio 0.31 0.62 0.72 0.68 0.70 0.84 0.07 0.15 0.20 1.00 0.86 GCA/SCA 0.22 0.83 1.29 1.05 1.18 2.55 0.04 0.09 0.13 - 3.03 “*”, “**”, “***” = significance at 𝑝 < 0.05, 𝑝 < 0.01, 𝑝 < 0.001, respectively, “NS” = non-significant. GCA = general combining ability, SCA = specific combining ability, S.E = standard error, Den. DF = denominator degree of freedom, Num. DF= numerator degree of freedom, Env = environment, Rep = replication. All “0.00” are not absolute values. All “0” = estimated variance equals zero. 86 University of Ghana http://ugspace.ug.edu.gh 5.3.2 Performance in crosses of the 17 inbred parents 5.3.2.1 GCA effects and GCA-based heterotic grouping of the inbred parents Table 5.6 presents the GCA effects of the inbred parents, Hopkins statistic, and the correlation coefficient between the cophenetic distances and original distances. The GCA effects for grain yield ranged from -91 kg ha-1 for TZEEIOR 113 to 142 kg ha-1 for TZdEEI 7. Other inbreds with positive GCA effect for grain yield under drought stress include, in decreasing order of effect, TZEEIOR 42, TZEEIOR 218, TZEEIOR 76, TZEEIOR 141, TZEEIOR 161, TZEEIOR 219, TZdEEI 12, and TZEEIOR 53 with GCA effect of 65, 61, 41, 36, 26,15, 7, and 2 kg ha-1, respectively. Genotypes TZEEIOR 42, TZEEIOR 76, and TZdEEI 7 had negative GCA effects for both STGC and ASI. On the contrary, genotypes TZEEIOR 100 and TZEEIOR 222 expressed significant negative effects for both STGC and ASI, in addition to negative GCA effects for grain yield. The optimal number of heterotic groups among inbred parents under drought environments using HGCAMT method is shown in Figure 5.1. Results indicated that there were two heterotic groups among the inbred parents. The Hopkins value of the clustered data was 0.32 with a validity of 0.74. All the inbreds with positive GCA effect for grain yield plus TZEEIOR 223 were clustered together in group I while others with negative GCA effects were assigned to group II (Figure 5.2). Based on their performance in crosses, TZEEIOR 42 was found closely related to TZdEEI 7 than TZdEEI 7 to TZdEEI 12. Similarly, TZEEIOR 53 was closer related to TZEEIOR 223 than TZEEIOR 223 to TZEEIOR 222. The prediction of GCA effect of TZEEIOR 12 were zero for all the traits. Moreover, it was closely related to TZEEIOR 196 and were clustered together in group II, characterised by negative GCA effects for grain yield. 87 University of Ghana http://ugspace.ug.edu.gh Table 5.6 General combining ability effects of inbred parents for grain yield and other traits across drought stress environments in Nigeria, 2016-2017 Grain Anthesis- Plant Husk Ear Leaf Days to Inbred Yield, kg Silking aspect cover aspect death -1 50% silking ha interval (1-9) (1-9) (1-9) (1-9) TZEEIOR 12 0.000 0.000 0.000 0.000 0.000 0.000 0.000 TZEEIOR 42 64.596 -0.115 -0.457 -0.183 -0.041 -0.066 -0.043 TZEEIOR 53 2.046 -0.019 -0.035 0.024 0.025 0.001 0.137 TZEEIOR 76 41.249 -0.163 0.136 -0.029 -0.024 0.060 -0.458 TZEEIOR 100 -33.363 -0.029 0.350 0.132 0.125 0.163 -0.007 TZEEIOR 113 -90.926 0.013 -0.023 0.078 0.208 -0.043 0.085 TZEEIOR 130 -48.969 0.092 0.240 -0.057 0.125 0.097 0.203 TZEEIOR 141 36.035 0.084 0.120 -0.048 0.108 -0.058 0.172 TZEEIOR 145 -35.232 0.250 -0.054 0.114 0.125 0.104 0.047 TZEEIOR 161 25.555 0.005 0.217 -0.003 0.025 -0.006 0.155 TZEEIOR 196 -69.196 0.562 0.863 0.158 0.000 0.163 -0.130 TZEEIOR 218 61.432 -0.019 -0.873 0.024 -0.083 -0.132 0.027 TZEEIOR 219 15.281 -0.458 -0.379 -0.048 -0.224 -0.073 0.123 TZEEIOR 222 -28.291 -0.043 -0.027 0.051 -0.058 0.016 -0.043 TZEEIOR 223 -50.639 -0.203 -0.247 0.006 -0.008 -0.051 0.105 TZdEEI 7 142.100 -0.362 -0.583 -0.254 -0.182 -0.117 -0.259 TZdEEI 12 6.996 0.060 -0.177 -0.038 -0.074 -0.073 0.029 S.E 88 0.31 0.51 0.12 0.15 0.15 0.17 H 0.3201 R 0.7427 S.E = standard error; H = Hopkins statistic; r = Correlation coefficient between cophenetic distance and the original distance. 88 University of Ghana http://ugspace.ug.edu.gh A B C Figure 5.1 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) under drought environments 89 University of Ghana http://ugspace.ug.edu.gh II II I Figure 5.2 Heterotic grouping of inbred parents under drought environments using HGCAMT-Ward.D2 90 University of Ghana http://ugspace.ug.edu.gh Table 5.7 presents the GCA effects of inbred parents for grain yield, anthesis-silking interval, Striga damage at 8 and 10 WAP, number of emerged Striga plants at 8 and 10 WAP, husk cover, and ear aspect across Striga-infested environments. The inbred TZEEIOR 141 showed the highest GCA effect for grain yield (152) while TZEEIOR 42 had the lowest value of GCA effect (-119) for the trait. Inbreds with positive GCA effects for grain yield ranked in decreasing order of importance were as follow: TZEEIOR 53(140), TZdEEI 7(134), TZEEIOR 145(85), TZEEIOR 12(44), TZEEIOR 219(26), TZEEIOR 113(20), and TZEEIOR 130(16). However, TZdEEI 7 and TZEEIOR 130 showed consistent negative GCA effects for Striga damage and number of emerged Striga plants at 8 and 10 WAP. For the inbred TZEEIOR 12, positive GCA effect (44.8) was associated with consistent positive GCA effects for the other traits. On the contrary, TZEEIOR 196 showed negative GCA effect (-103.2) for grain yield with positive GCA effects for the other traits. Figure 5.3 presents the results of optimal number of heterotic groups in the GCA of multiple traits data under Striga environments. There were three different heterotic groups among the inbred parents (Elbow and Gap statistic methods). Figure 5.4 presents the inbred parents in the three different heterotic groups. The group I was composed of eight inbreds including three inbreds with positive GCA effects for grain yield, TZEEIOR 141, TZEEIOR 53, and TZEEIOR 12. Group III was constituted of the remaining five inbreds with positive GCA effects plus TZEEIOR 76, which was clustered closer to TZEEIOR 113 and TZEEIOR 130 with a cophenetic distance less than 0.5. Inbreds TZEEIOR 222, TZEEIOR 100, and TZdEEI 12 were grouped in the cluster II, which is characterised by negative GCA effects for grain yield and anthesis-silking interval and positive GCA effects for majority of the other traits. The correlation between the cophenetic distances and the GCA-based distances was 0.807. 91 University of Ghana http://ugspace.ug.edu.gh Table 5.7 General combining ability effects of inbred parents for grain yield and other traits across Striga-infested environments in Nigeria, 2016-2017 Striga damage Number of emerged Grain Anthesis- Husk Ear (1-9) Striga plants Inbred Yield, Silking Cover Aspect Kgha-1 Interval 8 WAP 10 WAP 8 WAP 10 WAP (1-9) (1-9) TZEEIOR 12 44.77 0.04 0.05 0.02 0.34 0.49 0.03 0.04 TZEEIOR 42 -119.44 0.04 -0.03 0.01 0.32 0.57 -0.03 0.09 TZEEIOR 53 139.58 0.10 -0.15 -0.18 0.25 0.42 -0.17 -0.11 TZEEIOR 76 -32.85 -0.03 -0.05 -0.03 -0.54 -0.68 -0.01 -0.01 TZEEIOR 100 -92.98 -0.01 0.12 0.09 -0.50 -0.43 0.10 0.14 TZEEIOR 113 20.33 -0.01 0.00 -0.01 -0.15 -0.15 -0.02 0.04 TZEEIOR 130 16.15 0.05 -0.04 -0.06 -0.37 -0.70 -0.03 0.03 TZEEIOR 141 152.21 -0.09 -0.16 -0.18 0.20 0.20 -0.09 -0.12 TZEEIOR 145 85.41 -0.05 0.02 0.03 -0.46 -0.47 -0.01 -0.02 TZEEIOR 161 -13.06 -0.07 -0.01 0.00 0.71 0.64 0.02 -0.07 TZEEIOR 196 -103.22 0.03 0.05 0.06 0.75 1.13 0.06 0.05 TZEEIOR 218 -29.79 -0.01 -0.01 0.02 0.63 0.69 -0.03 -0.03 TZEEIOR 219 25.63 -0.02 -0.04 0.01 -0.57 -0.75 -0.01 -0.01 TZEEIOR 222 -107.95 -0.05 0.19 0.19 0.10 0.00 0.16 0.02 TZEEIOR 223 -28.82 -0.04 0.03 0.01 0.10 0.27 0.00 -0.05 TZdEEI 7 133.92 0.07 -0.10 -0.08 -0.58 -0.93 -0.05 -0.08 TZdEEI 12 -89.88 0.08 0.12 0.10 -0.26 -0.30 0.09 0.08 S.E 159 0.17 0.15 0.15 1.31 1.39 0.13 0.13 H 0.228 r 0.807 S.E = standard error; H = Hopkins statistic; r = Correlation coefficient between cophenetic distance and the original distance. 92 University of Ghana http://ugspace.ug.edu.gh A B C Figure 5.3 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) under Striga environments 93 University of Ghana http://ugspace.ug.edu.gh III II I Figure 5.4 Heterotic grouping of inbred parents under Striga-infested environments using HGCAMT-Ward.D2 94 University of Ghana http://ugspace.ug.edu.gh Table 5.8 presents the GCA effects of grain yield and other measured traits across test environments. The GCA effects of grain yield ranged from -130 for TZEEIOR 100 to 258 for TZEEIOR 141. Inbreds TZEEIOR 141 and TZEEIOR 53 (239) had high positive and significant (𝑝 < 0.01) GCA effects for grain yield. Moderate positive but non-significant GCA effects were found in TZEEIOR 12 (91) and TZdEEI 7 (89). The GCA effect of grain yield in TZEEIOR 141 was associated with significant (𝑝 < 0.001) GCA effects of days to 50% anthesis and silking, respectively. Highly significant (𝑝 < 0.01) GCA effects were observed in TZEEIOR 42 and TZEEIOR 145 for stalk lodging. Also, inbreds TZEEIOR 222 showed significant (𝑝 < 0.05) GCA effect for stalk lodging. Figure 5.5 presents the optimal number of heterotic groups among inbred parents as suggested by the three methods across test environments. There were different heterotic groups suggested by the three methods. Therefore, five heterotic groups suggested by the Elbow method was used. Figure 5.6 shows the clustering of the inbred parents into the five different groups. Genotypes TZEEIOR 141 and TZEEIOR 53, the top best based on GCA effects for grain yield, were singled out in one group (Group II) while the two-top worst, TZEEIOR 100 and TZEEIOR 76, were clustered together with TZdEEI 7 and TZdEEI 12 in the Group III. Group I was composed of six inbreds including TZEEIOR 12, TZEEIOR 42, and TZEEIOR 161, which expressed moderate positive and non-significant GCA effects for grain yield. Members in Group IV were characterised by an average of -47.35 GCA effect for grain yield, negative GCA effects for plant and ear height and for the traits related to Striga resistance. Inbreds TZEEIOR 222 and TZEEIOR 145 composed the Group V, which was characterised by an average of -78.2 GCA effect for grain yield, negative GCA effects for flowering traits and plant and ear heights, but with positive GCA effects for Striga damage rating at 8 and 10WAP. 95 University of Ghana http://ugspace.ug.edu.gh Table 5.8 General combining ability effects of inbred parents for grain yield and other traits across test environments in Nigeria, 2016-2017 Grain Anthesis- Days Plant Ear Husk Ear Days to Ear per Root Stalk Inbred Yield, Silking to height, height, cover Aspect -1 silking plant lodging lodging Kg ha interval anthesis cm cm (1-9) (1-9) TZEEIOR 12 90.90 0.07 0.00 0.08 -0.64 1.65 0.01 0.00 -0.04 -1.02 -1.63 TZEEIOR 42 19.90 0.09 0.04 0.14 3.23* 3.56*** -0.01 0.01 -0.02 0.38 5.94*** TZEEIOR 53 239.39* 0.21* -0.29 -0.02 1.94 -0.20 -0.02 0.01 -0.10 -1.21 -2.05 TZEEIOR 76 -115.33 -0.09 0.05 -0.07 -1.76 -0.74 0.00 -0.01 0.09 0.68 -2.31 TZEEIOR 100 -129.66 -0.05 0.35 0.28 -2.78* -1.50 0.01 -0.01 0.07 -0.18 0.38 TZEEIOR 113 -50.28 0.05 0.20 0.27 -1.08 -0.12 0.01 -0.01 0.08 -1.20 -2.01 TZEEIOR 130 -73.82 0.10 0.37 0.49 -0.16 -0.23 0.00 -0.01 0.06 -0.70 -1.65 TZEEIOR 141 258.39* -0.15 -0.72** -0.91** -1.37 -2.07* 0.00 0.01 -0.08 -0.58 -1.17 TZEEIOR 145 -5.12 -0.09 -0.77 -0.85** -2.78* -0.71 0.00 0.00 0.04 0.27 4.20** TZEEIOR 161 23.57 0.01 -0.45* -0.42 1.77 0.91 0.01 0.00 -0.05 -0.87 -1.75 TZEEIOR 196 -56.03 0.03 0.08 0.12 -0.43 1.27 0.00 0.01 -0.02 0.52 0.88 TZEEIOR 218 -74.61 -0.13 -0.17 -0.33 0.54 -0.89 -0.02 0.00 -0.01 -0.33 -0.73 TZEEIOR 219 -17.95 -0.07 0.50* 0.40 -0.33 -0.07 0.00 0.00 0.00 0.00 -1.00 TZEEIOR 222 -151.13 -0.11 -0.30 -0.43 -0.85 -1.09 0.01 -0.01 0.00 2.09* 2.88* TZEEIOR 223 -51.82 -0.19* 0.15 -0.09 1.49 1.39 -0.01 0.00 -0.02 -0.20 -0.53 TZdEEI 7 88.86 0.13 0.55* 0.71* 0.52 -0.73 0.00 0.00 -0.03 1.43 -0.79 TZdEEI 12 4.75 0.18 0.41* 0.64* 2.68* -0.44 0.01 0.00 0.02 0.92 1.33 S.E 99 0.10 0.21 0.26 1.27 0.91 0.03 0.01 0.07 0.75 1.21 H 0.373 r 0.698 S.E = standard error; H = Hopkins statistic; r = Correlation coefficient between cophenetic distance and the original distance. 96 University of Ghana http://ugspace.ug.edu.gh A B C Figure 5.5 Optimal number of heterotic groups suggested by Elbow method (A), Silhouette method (B), and Gap statistic method (C) across test environments 97 University of Ghana http://ugspace.ug.edu.gh I II V III IV Figure 5.6 Heterotic grouping of inbred parents across test environments using HGCAMT Ward.D2 98 University of Ghana http://ugspace.ug.edu.gh 5.3.2.2 Identification of Testers The GCA effect of grain yield across test environments was positive and significant (𝑝 < 0.01) for TZEEIOR 53 and TZEEIOR 141, thus, making them potential inbred testers. However, the multiple traits GCA-based grouping across test environments clustered TZEEIOR 53 and TZEEIOR 141 in the same group. Inbred TZEEIOR 53 and TZEEIOR 141 yielded 859±243 kg ha-1 and 959±239 kg ha-1, respectively, which represent -8% and 3% yield over the average yield (933±250 kg ha-1) across the pedigree. In summary, inbred TZEEIOR 141 was identified as tester for Group II, because it belongs to a heterotic group, has positive and significant GCA effect for grain yield, has higher than average grain yield per se. In addition, TZEEIOR 53 x TZEEIOR 141 was identified as a single cross tester for group II. However, no inbred lines no single-cross hybrids in the other groups satisfied the criteria for identification of testers. 5.3.2.3 Heterosis and SCA effects of inbred parents for grain yield Table 5.9 presents results of Gardner & Eberhart analysis II (1966) for grain yield under Striga infestation, drought stress, and optimal conditions. The mean squares of heterosis were highly significant (𝑝 < 0.001). However, the inbred heterosis, that is the contribution of heterosis by inbred i in its crosses were not significant under drought but were highly significant (𝑝 < 0.001) under Striga, and also significant (𝑝 < 0.05) under optimal conditions. The contribution of the specific heterosis under the different trials managements was also highly significant. Table 5.10 presents the specific heterosis effects, the heterosis, and the heterobeltiosis of the top 20% and 3 worst hybrids obtained under Striga environments. The specific heterosis under Striga environments ranged from -2555 kg ha-1 for the hybrid TZEEIOR 219 x TZEEIOR 222 with significant mid-parent heterosis of -134% to 2869 kg ha-1 for hybrid TZEEIOR 141 x TZEEIOR 99 University of Ghana http://ugspace.ug.edu.gh 222 with mid-parent heterosis of 489% (Table 5.11). Also, the mid-parent heterosis ranged from - 134 to 538 % for TZEEIOR 130 x TZEEIOR 218 while the better parent heterosis varied from - 127 to 483%. The ranking of the genotypes based on the specific heterosis was not consistent with the ranking based on either mid-parent or better-parent heterosis indicating high differences among the per se performances of the inbred parents. Table 5.9 Means squares of 17 extra-early maturing inbred parents with their 136 diallel hybrids for grain yield across environments under Striga infestation, drought stress, and optimal conditions in Nigeria, 2016-2017 Grain yield, Kgha-1Variance across environments components DF Striga Drought Optimal Genotypes 152 2254499.1*** 606845.7*** 8785056.0*** Inbreds 16 3546360.1*** 915732.9*** 7756395.7*** Heterosis 136 2102515.5*** 570506.0*** 8906074.9*** Mean Heterosis (ℎ̅) 1 50001001.6*** 17277684.5*** 738145831.8*** Inbreds Heterosis (ℎ ) 16 1683728.1*** 333185.3ns 1643855.2* 𝑖 Specific Heterosis (𝑆𝑖𝑗) 119 1756314.7 *** 462018.2*** 3754442.6*** Error 305 221295.5 221295.5 810435.4 Mean 2497 1331 4506 CV, % 19 35 20 100 University of Ghana http://ugspace.ug.edu.gh Table 5.10 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across two Striga-infested environments of top 20% and 3 worst extra-early maturing single cross hybrids selected based on the specific heterosis in Abuja in 2016 and Mokwa in 2017 Grain yield, kg ha-1 Hybrid 𝑺𝒊𝒋 H H% Hb Hb% TZEEIOR 141 x TZEEIOR 222 2869.47 5074.45** 488.90 4920.37** 412.78 TZEEIOR 223 x TZdEEI 7 2264.01 3876.30** 271.37 3689.11** 228.34 TZEEIOR 12 x TZEEIOR 222 2144.21 3509.94** 265.00 3069.29** 173.88 TZEEIOR 100 x TZEEIOR 196 2111.38 4771.19** 420.95 4445.13** 304.56 TZEEIOR 196 x TZEEIOR 223 1871.22 4004.03** 260.41 3925.97** 243.00 TZEEIOR 113 x TZEEIOR 219 1550.86 1233.23** 60.21 639.79ns 24.22 TZEEIOR 12 x TZEEIOR 76 1426.78 2162.49** 122.71 2159.57** 122.34 TZEEIOR 130 x TZEEIOR 218 1415.01 3219.42** 537.77 3162.72** 482.60 TZEEIOR 100 x TZEEIOR 113 1395.64 2373.63** 137.64 1456.48** 55.13 TZEEIOR 76 x TZdEEI 7 1331.88 2475.27** 164.99 2216.23** 125.97 TZEEIOR 12 x TZdEEI 12 1307.20 2609.51** 178.43 2306.81** 130.68 TZEEIOR 222 x TZdEEI 7 1223.24 2996.65** 282.03 2817.96** 227.03 TZEEIOR 53 x TZEEIOR 196 1199.41 2459.83** 149.60 2275.10** 124.39 TZEEIOR 76 x TZEEIOR 130 1075.33 2129.05** 176.34 1577.07** 89.64 TZEEIOR 161 x TZEEIOR 219 1026.02 1188.62** 90.33 1049.71* 72.15 TZEEIOR 145 x TZEEIOR 223 1016.90 2096.31** 140.04 1977.59** 122.40 TZEEIOR 42 x TZdEEI 12 1009.73 2904.12** 222.23 2757.06** 189.64 TZEEIOR 53 x TZEEIOR 113 954.50 533.10ns 23.85 126.75ns 4.80 TZEEIOR 53 x TZEEIOR 219 951.93 909.00* 55.36 721.91ns 39.47 TZEEIOR 12 x TZEEIOR 145 937.04 1623.00** 103.26 1429.51** 80.98 TZEEIOR 161 x TZEEIOR 218 934.75 2147.74** 249.89 1830.22** 155.50 TZEEIOR 113 x TZEEIOR 145 828.75 761.06ns 37.86 129.32ns 4.90 TZEEIOR 161 x TZEEIOR 196 827.30 2293.25** 173.96 2151.99** 147.45 TZEEIOR 42 x TZEEIOR 218 827.21 2905.68** 291.17 2449.73** 168.50 TZEEIOR 218 x TZdEEI 12 807.98 2785.53** 327.38 2476.63** 213.55 TZEEIOR 196 x TZEEIOR 219 761.79 2125.98** 145.90 2123.63** 145.50 TZEEIOR 76 x TZEEIOR 141 743.46 2318.41** 157.11 2034.75** 115.66 TZEEIOR 76 x TZEEIOR 223 -1778.28 -649.10ns -38.47 -720.95ns -40.98 TZEEIOR 222 x TZEEIOR 223 -1942.57 -183.37ns -14.67 -549.26ns -34.00 TZEEIOR 219 x TZEEIOR 222 -2554.71 -1564.13** -133.76 -1849.60** -127.14 Minimum -2554.71 -1564.13 -133.76 -1849.60 -127.14 Maximum 2869.47 5074.45 537.77 4920.37 482.60 Mean 0 1286.25 110.34 1002.82 75.72 S.E. 311.15 101 University of Ghana http://ugspace.ug.edu.gh Table 5.11 presents the specific heterosis effects, the heterosis, and the heterobeltiosis of the top best 20% and 3 worst hybrids obtained under drought environments. The specific heterosis ranged from -1121 kg ha-1 for TZEEIOR 12 x TZEEIOR 113 to 1190 kg ha-1 in the cross between TZEEIOR 42 and TZdEEI 7. Also, heterobeltiosis varied from -81.91% in the cross TZEEIOR 76 x TZEEIOR 161 to 666.71% for TZEEIOR 141 x TZEEIOR 222 (hybrid not among the top 20%), which was the best hybrid under Striga infestation based on specific heterosis. Table 5.12 presents the specific heterosis effects, the heterosis, and the heterobeltiosis of the top best 20% and 3 worst hybrids obtained under optimal environments. The specific heterosis ranged from -3414.96 kg ha-1 for TZEEIOR 218 x TZEEIOR 219 to 1977 kg ha-1 for TZEEIOR 113 x TZEEIOR 222. The highest heterosis and heterobeltiosis obtained were 9467.65 kg ha-1 and 3658.11 kg ha-1, respectively and were observed for the hybrid TZEEIOR 113 x TZEEIOR 130 with a specific heterosis of 1363 kg ha-1. 102 University of Ghana http://ugspace.ug.edu.gh Table 5.11 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across two drought stress environments of top 20% and 3 worst extra-early maturing single cross hybrids selected based on the specific heterosis at Ikenne during 2016/2017 and 2017/2018 dry seasons Grain yield, kg ha-1 Hybrid 𝑺𝒊𝒋 H H% Hb Hb% TZEEIOR 42 x TZdEEI 7 1189.99 1442.87** 177.59 1437.36** 175.72 TZEEIOR 196 x TZEEIOR 222 1136.72 2114.39** 587.54 1725.47** 230.44 TZEEIOR 100 x TZEEIOR 130 1100.82 2205.82** 348.78 2161.67** 319.50 TZEEIOR 53 x TZdEEI 12 1056.94 2160.07** 291.28 2072.70** 250.04 TZEEIOR 12 x TZEEIOR 42 995.38 1535.84** 188.73 1529.00** 186.32 TZEEIOR 130 x TZEEIOR 196 946.30 2113.23** 296.52 2077.13** 277.40 TZEEIOR 12 x TZEEIOR 145 904.88 1734.09** 188.62 1635.37** 160.63 TZEEIOR 145 x TZEEIOR 218 715.83 1234.94** 151.49 1032.07* 101.37 TZEEIOR 113 x TZEEIOR 161 698.60 1101.16** 127.35 927.74* 89.37 TZEEIOR 161 x TZEEIOR 222 672.59 1326.77** 262.97 793.18ns 76.41 TZEEIOR 113 x TZEEIOR 145 606.93 1218.57** 142.58 1055.17* 103.64 TZEEIOR 218 x TZdEEI 12 596.39 1375.13** 190.82 1266.82** 152.82 TZEEIOR 219 x TZdEEI 7 557.87 982.09* 120.42 979.70* 119.77 TZEEIOR 12 x TZEEIOR 161 496.21 1116.34** 120.12 1007.60* 97.06 TZEEIOR 12 x TZEEIOR 141 475.74 1589.43** 315.33 1272.85** 155.10 TZEEIOR 42 x TZEEIOR 141 470.02 996.00* 200.32 686.26ns 85.04 TZEEIOR 113 x TZEEIOR 219 465.12 959.34* 127.53 898.38ns 110.48 TZEEIOR 42 x TZEEIOR 219 461.04 585.11ns 72.23 581.99ns 71.57 TZEEIOR 76 x TZEEIOR 130 458.61 1384.03** 172.80 1259.67** 136.14 TZEEIOR 12 x TZEEIOR 218 458.26 1276.33** 178.14 1172.17* 142.84 TZEEIOR 161 x TZEEIOR 223 453.61 1176.75** 227.17 656.64ns 63.25 TZEEIOR 12 x TZEEIOR 219 442.88 1154.67** 141.34 1150.95* 140.25 TZEEIOR 100 x TZEEIOR 113 422.02 1086.13** 169.77 1034.63* 149.67 TZEEIOR 53 x TZEEIOR 223 388.47 1633.94** 501.13 1305.80** 199.60 TZEEIOR 76 x TZEEIOR 113 388.04 872.57* 107.95 755.56ns 81.66 TZEEIOR 76 x TZdEEI 7 366.43 780.96ns 89.60 727.31ns 78.60 TZEEIOR 12 x TZEEIOR 223 355.88 1587.07** 387.78 1175.70* 143.27 TZEEIOR 145 x TZEEIOR 222 -950.04 -86.78ns -17.55 -610.33ns -59.95 TZEEIOR 76 x TZEEIOR 161 -987.91 -793.86* -80.87 -850.27ns -81.91 TZEEIOR 12 x TZEEIOR 113 -1121.40 -210.79ns -27.88 -275.47ns -33.57 Minimum -1121.40 -793.86 -9494.16 -850.27 -81.91 Maximum 1189.99 2205.82 1714.56 2161.67 666.71 Mean 0 756.10 81.68 576.05 69.64 S.E. 311.15 103 University of Ghana http://ugspace.ug.edu.gh Table 5.12 Specific heterosis effect (𝑺𝒊𝒋), heterosis (H), and heterobeltiosis (Hb) for grain yield across three optimal environments of top 20% and 3 worst extra-early maturing single cross hybrids selected based on the specific heterosis at Ikenne in 2016 and 2017 and at Mokwa in 2017 Grain yield, kg ha-1 Hybrid 𝑺𝒊𝒋 H H% Hb Hb% TZEEIOR 113 x TZEEIOR 222 1976.63 5570.51** 985.78 5175.40** 538.99 TZEEIOR 42 x TZEEIOR 218 1861.68 5719.93** 822.80 5690.39** 785.19 TZEEIOR 100 x TZEEIOR 223 1669.83 5581.16** 777.05 5320.58** 543.57 TZEEIOR 130 x TZEEIOR 223 1639.97 6165.03** 2927.19 5917.97** 1293.07 TZEEIOR 223 x TZdEEI 7 1623.75 5317.64** 410.87 4481.06** 210.30 TZEEIOR 145 x TZEEIOR 219 1583.88 5777.33** 881.12 5685.63** 760.74 TZEEIOR 161 x TZEEIOR 218 1560.72 5030.76** 558.68 4795.93** 422.44 TZEEIOR 218 x TZdEEI 7 1540.05 4802.85** 343.49 4070.26** 191.02 TZEEIOR 141 x TZEEIOR 161 1436.51 5541.43** 518.33 5475.23** 482.28 TZEEIOR 76 x TZEEIOR 222 1393.16 4577.10** 477.38 4575.70** 476.53 TZEEIOR 113 x TZEEIOR 130 1362.52 6321.04** 9467.65 6217.83** 3658.11 TZEEIOR 196 x TZEEIOR 223 1320.41 5654.20** 518.95 5022.32** 291.75 TZEEIOR 145 x TZEEIOR 161 1291.90 5870.59** 690.95 5584.94** 491.94 TZEEIOR 161 x TZEEIOR 219 1285.02 4989.69** 530.07 4795.74** 422.42 TZEEIOR 145 x TZEEIOR 218 1278.42 5237.23** 851.84 5186.40** 779.16 TZEEIOR 12 x TZEEIOR 42 1277.79 5178.97** 366.11 4489.09** 213.31 TZEEIOR 196 x TZEEIOR 219 1248.14 5385.48** 436.28 4898.46** 284.56 TZEEIOR 12 x TZEEIOR 141 1223.45 4751.40** 305.81 4200.61** 199.60 TZEEIOR 141 x TZEEIOR 219 1197.48 4917.15** 561.87 4789.39** 477.55 TZEEIOR 76 x TZEEIOR 141 1178.71 5118.35** 522.20 5095.60** 508.08 TZEEIOR 53 x TZEEIOR 196 1120.25 6081.73** 465.95 5665.54** 329.12 TZEEIOR 145 x TZEEIOR 223 1106.44 5496.35** 1075.97 5443.19** 965.13 TZEEIOR 76 x TZdEEI 12 1099.23 4800.27** 678.80 4550.04** 475.25 TZEEIOR 100 x TZEEIOR 218 1096.62 4576.85** 556.64 4420.26** 451.59 TZEEIOR 12 x TZdEEI 12 1042.44 4331.80** 338.23 3508.04** 166.69 TZEEIOR 53 x TZEEIOR 219 1006.64 5150.21** 629.46 5079.39** 571.35 TZEEIOR 141 x TZEEIOR 218 1001.53 4486.57** 537.78 4317.93** 430.54 TZEEIOR 218 x TZEEIOR 223 -2539.25 741.99ns 132.11 638.01ns 95.85 TZEEIOR 222 x TZEEIOR 223 -2900.15 260.28ns 36.71 9.01ns 0.94 TZEEIOR 218 x TZEEIOR 219 -3414.96 -330.18ns -46.73 -371.05ns -49.65 Minimum -3414.96 -330.18 -46.73 -371.05 -49.65 Maximum 1976.63 6321.04 9467.65 6217.83 3658.11 Average 0.00 4035.17 619.60 3702.59 389.42 S.E. 486.19 104 University of Ghana http://ugspace.ug.edu.gh 5.3.3 Performance of hybrids and stability analysis 5.3.3.1 Index-based performance under Striga environments Table 5.13 presents grain yield and other measured traits of the top 20% and 3 worst hybrids using the base index for selection under Striga-infested environments. Grain yield under Striga environments varied from 1134 kg ha-1 for TZEEIOR 53 x TZEEIOR 161 to 4035 kg ha-1 for TZEEIOR 222 x TZdEEI 7 with a mean of 1427 kg ha-1. Striga damage on the plants ranged from mild leaf blotching with some purplish-brown necrotic spots (3) to extensive leaf scorching with mostly grey necrotic spots (6) at eight weeks after planting (WAP). The disease symptoms did not vary much from 8 to 10 WAP (Table 5.14). A similar trend was observed between 8 and 10 WAP for the number of emerged Striga plants which ranged from 4 to 40 at 8 WAP and from 4 to 33 at 10 WAP. Seventy-one out of 140 hybrids evaluated were selected as resistant/tolerant to Striga (positive index value). None of the four experimental hybrid checks used in this study was selected among the top 20% performing hybrids presented in Table 5.14. Hybrids TZEEIOR 222 x TZdEEI 7 (4035 kg ha-1), TZEEIOR 141 x TZEEIOR 218 (3595 kg ha-1), TZEEIOR 76 x TZEEIOR 196 (3326 kg ha-1), TZEEIOR 100 x TZEEIOR 196 (4035 kg ha-1), and TZEEIOR 130 x TZEEIOR 141 (4035 kg ha-1) were the top five performers under Striga. Among the top performing hybrids, TZEEIOR 100 x TZEEIOR 141 and TZEEIOR 100 x TZdEEI 7 supported very few Striga plants with an average of four plants each. They were followed by hybrid TZEEIOR 141 x TZEEIOR 218 which supported on the average 6 Striga plants. Also, these hybrids expressed similar symptoms with an average Striga damage of 4, which was not significantly different from that of the top five performers. Under these conditions, the anthesis-silking interval and the ear aspect ranged from 1 to 5 days and 4 to 7 with an average of 2-5 days, respectively. 105 University of Ghana http://ugspace.ug.edu.gh Table 5.13 Performance of index-based top 20% and 3 worst extra-early maturing single cross hybrids across Striga-infested environments at Abuja and Mokwa in 2016 and 2017 Grain Striga damage Number of emerged Anthesis- Ear Hybrid Yield, (1-9) Striga plants silking Aspect INDEX Kg ha-1 interval (1-9) 8 WAP 10 WAP 8 WAP 10 WAP TZEEIOR 222 x TZdEEI 7 4035 4 4 13 16 2 4 11.60 TZEEIOR 141 x TZEEIOR 218 3595 3 4 6 6 1 4 11.46 TZEEIOR 76 x TZEEIOR 196 3326 3 4 10 12 1 4 10.68 TZEEIOR 100 x TZEEIOR 196 3606 4 4 15 18 2 5 9.11 TZEEIOR 130 x TZEEIOR 141 3674 4 4 16 15 2 4 8.78 TZEEIOR 76 x TZEEIOR 161 3133 4 4 16 13 2 5 7.49 TZEEIOR 223 x TZdEEI 7 3611 4 5 20 18 1 5 6.83 TZEEIOR 145 x TZdEEI 7 3295 4 4 16 17 2 4 6.62 TZEEIOR 42 x TZEEIOR 196 3499 4 4 22 24 2 4 6.10 TZEEIOR 100 x TZEEIOR 141 3161 4 5 4 4 2 5 5.98 TZEEIOR 196 x TZdEEI 7 3596 4 4 16 22 2 5 5.78 TZEEIOR 141 x TZdEEI 12 3688 4 5 12 14 1 5 5.74 TZEEIOR 219 x TZdEEI 7 3198 4 5 14 13 1 5 5.59 TZEEIOR 222 x TZdEEI 12 3075 4 5 14 17 1 5 5.54 TZEEIOR 196 x TZdEEI 12 2914 4 4 11 15 4 5 5.38 TZEEIOR 113 x TZEEIOR 219 3138 4 5 9 10 2 5 5.33 TZEEIOR 76 x TZdEEI 12 3502 4 5 17 18 2 5 5.13 TZEEIOR 113 x TZEEIOR 141 3403 4 5 12 15 2 4 5.11 TZEEIOR 130 x TZEEIOR 196 2853 4 5 10 13 2 4 4.90 TZEEIOR 100 x TZdEEI 7 2725 4 5 5 4 2 5 4.88 TZEEIOR 100 x TZdEEI 12 3561 5 5 11 14 1 5 4.87 TZEEIOR 12 x TZEEIOR 222 3677 4 5 24 25 3 4 4.62 TZEEIOR 196 x TZEEIOR 222 3415 4 5 14 16 3 5 4.58 TZEEIOR 42 x TZdEEI 7 3138 4 5 17 18 2 5 4.41 TZEEIOR 42 x TZEEIOR 222 2885 4 5 10 10 2 5 4.31 TZEEIOR 53 x TZEEIOR 196 3110 4 5 21 23 2 4 3.96 TZEEIOR 100 x TZEEIOR 222 2903 4 5 19 18 2 5 3.85 TZEEIOR 12 x TZEEIOR 130 3397 4 5 12 14 2 5 3.75 TZEEIOR 218 x TZEEIOR 219 1279 5 6 14 16 3 7 -11.29 TZEEIOR 219 x TZEEIOR 222 1134 6 6 17 16 3 7 -12.05 TZEEIOR 53 x TZEEIOR 161 1157 5 6 30 33 4 6 -12.15 Minimum 1134 3 4 4 4 1 4 Maximum 4035 6 6 40 33 5 7 Mean 2621 5 5 17 17 2 5 S.E 458 0 0 1 1 1 0 S.E = standard error. 106 University of Ghana http://ugspace.ug.edu.gh 5.3.3.2 Index-based performance under drought environments Table 5.14 presents grain yield and other measured traits of the top 20% and three worst performing hybrids using the base index for selection under drought environments. There were more days to 50% silking under drought stress than those in Striga infestation with anthesis-silking interval ranging from 1 to 9 days. Ear aspect ranged from 4 to 8 with a mean of 5. Hybrid TZEEIOR 130 x TZEEIOR 196 was the highest yielding with 2881 kg ha-1 while the lowest yielding hybrid was TZEEIOR 42 x TZEEIOR 161 with 505 kg ha-1. On the average, 50-60% of dead leaf area (rate 5) were observed with individual genotype having dead leaf area falling within a range of 20-30% to 60-70%. Seventy-two out of 140 hybrids evaluated had positive base index values, and were therefore considered as tolerant to drought. The index-based ranking of the top five hybrids is as follows: TZEEIOR 100 x TZEEIOR 130 (2496 kg ha-1), TZEEIOR 130 x TZEEIOR 196 (2881 kg ha-1), TZEEIOR 218 x TZdEEI 12 (2082 kg ha-1), TZEEIOR 12 x TZEEIOR 42 (1809 kg ha-1), and TZEEIOR 113 x TZEEIOR 219 (1771 kg ha-1). These hybrids showed, on the average, 20 to 30% dead leaf area with the anthesis-silking interval ranging from 2 to 3 days. Hybrids TZEEIOR 218 x TZdEEI 7 and TZEEIOR 76 x TZEEIOR 130 had anthesis-silking interval of zero and 1, respectively. Hybrid TZEEIOR 218 x TZdEEI 7 was identified as the worst performing hybrid based on the index. Among the five best hybrids selected under drought stress, only TZEEIOR 130 x TZEEIOR 196 and TZEEIOR 113 x TZEEIOR 219 were also selected among the 28 best performing hybrids under Striga infestation. 107 University of Ghana http://ugspace.ug.edu.gh Table 5.14 Index-based performance of top 20% and 3 worst extra-early maturing single cross hybrids across drought-stress environments at Ikenne during 2016/2017 and 2017/2018 dry seasons Grain Ears Anthesis Plant Ear Leaf Hybrid Yield, Per Silking Aspect Aspect death INDEX Kg ha-1 plant interval (1-9) (1-9) (1-9) TZEEIOR 100 X TZEEIOR 130 2496 1 3 5 4 4 13.73 TZEEIOR 130 X TZEEIOR 196 2881 1 3 5 5 4 11.16 TZEEIOR 218 X TZdEEI 12 2082 1 2 5 4 4 10.83 TZEEIOR 12 X TZEEIOR 42 1809 1 3 4 5 4 9.34 TZEEIOR 113 X TZEEIOR 219 1771 1 2 5 4 4 8.10 TZEEIOR 196 X TZEEIOR 222 2316 1 3 4 6 5 8.07 TZEEIOR 130 X TZEEIOR 161 1845 1 3 4 4 5 7.88 TZEEIOR 196 X TZEEIOR 218 1625 1 2 6 4 3 7.81 TZEEIOR 12 X TZEEIOR 145 1882 1 3 5 4 4 7.77 TZEEIOR 76 X TZEEIOR 113 1888 1 2 5 4 5 7.56 TZEEIOR 53 X TZdEEI 12 2225 1 5 5 4 5 7.15 TZEEIOR 130 X TZEEIOR 222 1888 1 3 4 5 4 7.15 TZEEIOR 76 X TZEEIOR 223 2196 1 3 5 4 5 7.08 TZEEIOR 161 X TZEEIOR 222 1903 1 3 5 4 5 6.87 TZEEIOR 12 X TZEEIOR 223 2234 1 5 5 5 5 6.74 TZEEIOR 53 X TZEEIOR 161 1993 1 4 6 5 4 6.35 TZEEIOR 141 X TZEEIOR 196 1832 1 4 5 4 4 6.18 TZEEIOR 76 X TZEEIOR 130 1810 1 1 5 4 5 5.90 TZEEIOR 130 X TZEEIOR 218 1547 1 3 5 5 3 5.50 TZEEIOR 12 X TZEEIOR 222 1875 1 3 5 5 4 5.46 TZEEIOR 12 X TZEEIOR 219 1805 1 3 5 4 5 5.38 TZEEIOR 130 X TZdEEI 7 1813 1 4 5 5 4 5.31 TZEEIOR 12 X TZEEIOR 161 1877 1 3 5 5 5 5.28 TZEEIOR 113 X TZEEIOR 161 1809 1 3 5 4 4 5.03 TZEEIOR 141 X TZdEEI 7 1614 1 4 5 4 5 4.99 TZEEIOR 130 X TZEEIOR 141 1758 1 4 5 5 4 4.97 TZEEIOR 141 X TZEEIOR 219 1603 1 4 6 4 5 4.78 TZEEIOR 76 X TZEEIOR 141 1767 1 4 5 4 4 4.72 TZEEIOR 100 X TZEEIOR 219 684 0 6 6 6 5 -11.46 TZEEIOR 42 X TZEEIOR 161 505 0 9 6 6 6 -14.63 TZEEIOR 218 X TZdEEI 7 516 0 0 7 8 6 -17.80 Minimum 505 0 1 4 4 3 Maximum 2881 1 9 7 8 6 Mean 1427 1 5 6 5 5 S.E 257 0 1 0 0 0 S.E = standard error. 108 University of Ghana http://ugspace.ug.edu.gh 5.3.3.3 Performance of hybrids across test environments. Table 5.15 presents the top 20 and 5 worst hybrids based on the genetic values (GV) across test environments. Forty-eight percent (48%) of the hybrids having positive GV were selected. The top ten performing hybrids were as follow: TZEEIOR 141 x TZEERIOR 219, TZEEIOR 141 x TZEEIOR 223, TZEEIOR 145 x TZdEEI 7, TZEEIOR 12 x TZEEIOR 141, TZEEIOR 161 x TZEEIOR 219, TZEEIOR 53 x TZEEIOR 196, TZEEIOR 12 x TZEEIOR 196, TZEEIOR 53 x TZEEIOR 223, TZEEIOR 141 x TZEEIOR 218, and TZEEIOR 113 x TZEEIOR 141. Out of the top ten performing hybrids identified based on genetic value, inbred TZEEIOR 141 was involved in six crosses. It may be recalled that TZEEIOR 141 was the inbred parent with the highest GCA effect. The best crosses with TZEEIOR 141 involved only two inbreds with positive GCA effects for grain yield, which are: TZEEIOR 12 and TZdEEI 7. Table 5.16 presents the correlation coefficients among the indices used, SCA, and GV. Results showed that grain yield across test environments of the hybrids was significant (𝑝 < 0.01) and positively correlated with their specific combining ability (r = 0.857), their genetic value (r = 0.77), and also with the base index under Striga infestation (0.695). On the contrary, the base index under drought-induced environments was neither correlated to the grain yield of the hybrids across test environments nor any of the other parameters used for selecting the hybrids. High correlation was also observed between SCA and GV. 109 University of Ghana http://ugspace.ug.edu.gh Table 5.15 Performance of the top 20 and 5 worst extra-early maturing single cross hybrids across test environments based on their genetic values across 11 test environments Grain yield kg ha-1 Hybrid GCA_P1 GCA_P2 SCA GV Striga Drought Optimal Across TZEEIOR 141 X TZEEIOR 219 2944 1603 4396 3360 258.39 -17.95 383.57 624.01 TZEEIOR 141 X TZEEIOR 223 3053 1550 4435 3408 258.39 -51.82 393.37 599.94 TZEEIOR 145 X TZdEEI 7 3295 1055 5675 3970 -5.12 88.86 490.36 574.10 TZEEIOR 12 X TZEEIOR 141 2958 1519 4470 3384 90.90 258.39 206.07 555.36 TZEEIOR 161 X TZEEIOR 219 3422 1238 4628 3573 23.57 -17.95 535.01 540.63 TZEEIOR 53 X TZEEIOR 196 3110 1610 4792 3602 239.39 -56.03 285.82 469.18 TZEEIOR 12 X TZEEIOR 196 3184 1472 5411 3885 90.90 -56.03 424.31 459.18 TZEEIOR 53 X TZEEIOR 223 2698 1869 3860 3075 239.39 -51.82 261.17 448.74 TZEEIOR 141 X TZEEIOR 218 3595 798 4052 3294 258.39 -74.61 263.43 447.21 TZEEIOR 113 X TZEEIOR 141 3403 682 3682 3035 -50.28 258.39 223.11 431.22 TZEEIOR 141 X TZEEIOR 222 3207 1425 4115 3296 258.39 -151.13 323.57 430.83 TZEEIOR 53 X TZEEIOR 218 2812 2131 3507 3004 239.39 -74.61 249.09 413.87 TZEEIOR 130 X TZEEIOR 141 3674 1758 3600 3292 -73.82 258.39 195.95 380.52 TZEEIOR 53 X TZEEIOR 141 2161 1838 3332 2635 239.39 258.39 -125.59 372.19 TZEEIOR 222 X TZdEEI 7 4035 1171 4249 3611 -151.13 88.86 395.98 333.71 TZEEIOR 42 X TZEEIOR 53 2477 1534 3744 2882 19.8963 239.39 59.1312 318.42 TZEEIOR 223 X TZdEEI 7 3611 1213 4302 3489 -51.82 88.86 280.69 317.73 TZEEIOR 12 X TZdEEI 12 3108 1387 4160 3273 90.90 4.75 202.1 297.74 TZEEIOR 219 X TZdEEI 7 3198 1231 4331 3355 -17.95 88.86 220.97 291.88 TZEEIOR 141 X TZdEEI 12 3688 1437 3254 3081 258.39 4.75 9.4292 272.56 TZEEIOR 222 X TZEEIOR 223 2512 1402 1586 1889 -151.13 -51.82 -502.14 -705.09 TZEEIOR 218 X TZEEIOR 222 2280 757 1734 1755 -74.61 -151.13 -530.49 -756.23 TZEEIOR 218 X TZEEIOR 223 1856 1213 1625 1634 -74.61 -51.82 -638.99 -765.42 TZEEIOR 218 X TZEEIOR 219 1279 1017 1320 1250 -74.61 -17.95 -814.39 -906.95 TZEEIOR 219 X TZEEIOR 222 1134 954 1461 1250 -17.95 -151.13 -830.44 -999.52 S.E 458 257 496 209 GCA_P1 and GCA_P2, general combining ability of parent 1 and parent 2, respectively; GV: Genetic value of the cross; S.E: standard error. 110 University of Ghana http://ugspace.ug.edu.gh Table 5.16 Spearman's rho correlation coefficients among the four parameters used for selecting the F1 hybrids and their grain yield across 11 test environments GY across SCA GV Index Striga Index Drought GY across 1 0.857** 0.772** 0.695** 0.129 SCA 0.857** 1 0.771** 0.571** 0.147* GV 0.772** 0.771** 1 0.482** 0.135 Index Striga 0.695** 0.571** 0.482** 1 0.032 Index Drought 0.129 0.147* 0.135 0.032 1 * and **, significance at respectively p< 0.05 and p<0.01 (one-tailed); GV, genetic value of hybrid F1; SCA, specific combining ability; GY across, Grain yield of hybrids F1 across 11 test environments. 5.3.3.4 Genetic correlation and heritability among test environments for grain yield The genetic correlation coefficients among all test environments and their heritability for grain yield are presented in Table 5.17. Results indicated that the two drought stress environments (IK16DT, IK17DT) and Striga-infested environment at Abuja in 2017 (ABJ17STR) were negatively correlated with most of the rest of the environments. However, all of the environments had higher heritability than the threshold (h ≤ 0.20) set for eliminating environment for stability analysis, in this study. Higher genetic correlations were found among optimal environments than those among Striga environments. Drought stress environments and Striga environment at Abuja in 2017 were then removed from the stability analysis despite their relatively high heritability observed for grain yield. 111 University of Ghana http://ugspace.ug.edu.gh Table 5.17 Genetic correlation among test environments and heritability for grain yield Environment Environment 1 2 3 4 5 6 7 8 9 10 11 1 ABJOPT16 1 2 ABJSTR16 0.46 1 3 IKDT16 -0.07 -0.31 1 4 IKOPT16 0.79 0.28 -0.13 1 5 MKOPT16 0.48 0.26 -0.08 0.27 1 6 MKSTR16 0.74 0.47 0.08 0.55 0.15 1 7 ABJSTR17 -0.1 -0.17 -0.04 -0.08 -0.08 0.06 1 8 IKDT17 -0.08 -0.06 0.42 0.16 -0.21 0.16 0.16 1 9 IKOPT17 0.54 0.69 -0.11 0.71 0.3 0.63 -0.16 0.1 1 10 MKOPT17 0.52 0.3 -0.07 0.47 0.21 0.38 -0.14 0 0.66 1 11 MKSTR17 0.26 0.21 0.08 0.3 0.13 0.55 -0.2 -0.02 0.33 0.61 1 Heritability 0.64 0.44 0.53 0.76 0.56 0.49 0.54 0.52 0.63 0.66 0.55 ABJOPT16, ABJSTR16, ABJSTR17 = Optimal trial and Striga-infested trials at Abuja in 2016 and 2017; IKDT16 and IKOPT16, IKDT17 and IKOPT17 = Managed drought and Optimal trials at Ikenne in 2016 and 2017; MKOPT16 and MKSTR16, MKOPT17and MKSTR17= Optimal trials and Striga-infested trials at Mokwa in 2016 and 2017. 5.3.3.5 Stability of the performance of hybrids across test environments Figure 5.7 presents the “which won where” genotype main effect plus genotype x environment biplot of 26 top performing hybrids plus the 4 experimental hybrid checks. The genotype main effects and the genotype x environment interaction (GGE) biplot explained 77.3% of the total variation in grain yield across eight environments with the two axes (PC1 and PC2) constituting the biplot explaining 66.4 and 10.9%, respectively, of the variation captured. The perpendicular lines to the polygon sides divide the biplot into sectors, which are mega-environments suggesting that different genotypes won in different sectors and thus genotype x environment interaction or crossover patterns existed in the data. Environments E8, E10, E1 (Optimal environments) plus E11 (Striga-infested) were grouped as one mega-environment, that is these environments consistently shared the best set of genotypes. Similarly, E7 and E5 (optimal) plus E9 (Striga-infested) shared the same set of best genotypes. However, the Striga-infested environment at Abuja in 2016 (E2) did not share the best genotypes with any of the two mega-environments identified. 112 University of Ghana http://ugspace.ug.edu.gh Figure 5. 7 A ‘which won where’ genotype main effect plus genotype x environment biplot of 30 extra-early maturing maize hybrids evaluated for grain yield across Striga infested and optimal environments in 2016 and 2017 LEGEND Code Hybrid Code Hybrid Code Environment G1 TZEEIOR 12 X TZEEIOR 141 G16 TZEEIOR 145 X TZdEEI 7 E1 Abuja optimal 2016 G2 TZEEIOR 12 X TZEEIOR 196 G17 TZEEIOR 161 X TZEEIOR 219 E2 Abuja Striga-infested 2016 G3 TZEEIOR 12 X TZdEEI 12 G18 TZEEIOR 218 X TZEEIOR 219 E5 Ikenne optimal 2016 G4 TZEEIOR 42 X TZEEIOR 53 G19 TZEEIOR 218 X TZEEIOR 222 E7 Ikenne optimal 2017 G5 TZEEIOR 53 X TZEEIOR 141 G20 TZEEIOR 218 X TZEEIOR 223 E8 Mokwa optimal 2016 G6 TZEEIOR 53 X TZEEIOR 196 G21 TZEEIOR 219 X TZEEIOR 222 E9 Mokwa Striga-infested 2016 G7 TZEEIOR 53 X TZEEIOR 218 G22 TZEEIOR 219 X TZEEIOR 223 E10 Mokwa optimal 2017 G8 TZEEIOR 53 X TZEEIOR 223 G23 TZEEIOR 219 X TZdEEI 7 E11 Mokwa Striga-infested 2017 G9 TZEEIOR 113 X TZEEIOR 141 G24 TZEEIOR 222 X TZEEIOR 223 G10 TZEEIOR 130 X TZEEIOR 141 G25 TZEEIOR 222 X TZdEEI 7 G11 TZEEIOR 141 X TZEEIOR 218 G26 TZEEIOR 223 X TZdEEI 7 G12 TZEEIOR 141 X TZEEIOR 219 G27 TZdEEI 1 x TZdEEI 12 G13 TZEEIOR 141 X TZEEIOR 222 G28 TZEEI 79 x TZEEI 82 G14 TZEEIOR 141 X TZEEIOR 223 G29 (TZEEI 79 x TZEEI82) x TZEEI 95 G15 TZEEIOR 141 X TZdEEI 12 G30 TZEE-Y Pop STR C5 x TZEEI 58 113 University of Ghana http://ugspace.ug.edu.gh The vertex genotype, that is the winning genotype, for E2 was hybrid TZEEIOR 219 x TZdEEI 7. In the mega-environment constituted of E7, E9, and E5, hybrid TZEEIOR 145 x TZdEEI 7 was the winner while TZEEIOR 12 x TZEEIOR 196 was the winner in the mega-environment composed of E1, E8, E10, and E11. Figure 5.8 presents the mean grain yield versus stability of genotype main effect plus genotype x environment biplot of 26 top performing hybrids plus the 4 experimental hybrid checks. The ideal genotype is represented by the head of the arrow on the average environment coordinate (AEC) abscissa (horizontal axis). The arrow shown on the AEC abscissa points in the direction of higher mean performance of genotypes. Results showed that hybrids TZEEIOR 12 x TZEEIOR 196, TZEEIOR 145 x TZdEEI 7, TZEEIOR 222 x TZdEEI 7, TZEEIOR 223 x TZdEEI 7, and TZEEIOR 219 x TZdEEI 7 were the top 5 highest yielding among the hybrids selected for stability analysis. In addition, hybrids TZEEIOR 12 x TZEEIOR 196 and TZEEIOR 222 x TZdEEI 7 had a near-zero projection onto the AEC ordinate (vertical axis), meaning that the rank of these two genotypes was highly consistent across environments. Because these hybrids combined both high grain yield and stability, they were identified as the most stable hybrids. Figure 5.9 presents the discriminating power against the representativeness of the environment in identifying the superior hybrids in mega-environments. Results revealed the optimal environment at Mokwa in 2016 (E8) as the least discriminating of the hybrids with the shortest vector from the biplot origin. All the test environments had similar discriminating power, with the most discriminating environment as the optimal environment at Mokwa in 2017. Moreover, E7 had the smallest angle with the AEC abscissa in the mega-environments composed of E7, E9, and E5. The environment E1 had the smallest angle with the AEC abscissa in the mega-environments composed of E8, E10, E11 and E1. The environments E7 and E1 were, therefore, the most representative of their respective mega-environments. 114 University of Ghana http://ugspace.ug.edu.gh Figure 5. 8 Mean vs. stability view of genotype main effect plus genotype x environment biplot of grain yield of 30 selected extra-early maturing maize hybrids evaluated across Striga-infested and optimal environments in 2016 and 2017 115 University of Ghana http://ugspace.ug.edu.gh Figure 5. 9 Discriminating vs. representativeness view of GGE biplot of 30 selected extra- early maturing hybrid maize for grain yield across Striga-infested and optimum environments in 2016 and 2017 116 University of Ghana http://ugspace.ug.edu.gh 5.4 Discussion and conclusion The observed significant variances for both GCA and SCA interactions with environments, for all the measured traits including grain yield under each trial management type and across test environments, indicated a change in ranking among genotypes in each environment and across test environments. The results also suggest that the environments were discriminating enough to uncover the genetic variability among the hybrids, which is the basis of selection for improvements in the measured traits. Previous studies (Badu-Apraku et al., 2015; Annor and Badu-Apraku, 2016; Talabi et al., 2017) reported genetic variability among hybrids developed in IITA needed for selection of high performing genotypes. However, there were both genetic and environmental variations implying that selection of superior genotypes may be difficult for few evaluations. This result suggested that extensive evaluations may be needed to be able to identify stable genotypes. As indicated by the significant GCA and SCA variances and GCA/SCA ratio, both additive and non-additive gene actions were important in this set of orange maize inbreds for grain yield and other agronomic traits under each of the three groups of environments regarding the management and under test environments. Under Striga infestations, Striga damage at 8 WAP and number of emerged Striga plants at 8 and 10 WAP had GCA/SCA ratio greater than 2 with Baker ratio for each close to unity. This indicated a predominance of GCA over SCA for these traits and implied that additive genetic effects were the primary type of the gene action for these traits, hence improvement for Striga resistance indicator traits in maize could be successfully achieved based on the predictive values of GCA alone. Under drought stress, the results showed that only staygreen characteristic can be improved based solely on GCA predictions of parents. Grain yield under each trial management type and across test environments had a GCA/SCA ratio of less than 0.5, indicating the predominance of non-additive genetic effects in its inheritance. These results disagree with those of Makumbi et al. (2011) and Badu-Apraku and Oyekunle (2012) who reported 117 University of Ghana http://ugspace.ug.edu.gh additive gene action to be predominant in the inheritance of grain yield in early-maturing maize inbred lines. However, it supports the findings of Betran et al. (2003) and Meseka et al. (2006). The difference in gene action for grain yield reported compared to other findings may be explained by the difference in the genetic material used. The important effects of SCA revealed in this study suggests that one should consider not only GCA but also the deviations from predicted values of GCAs in selecting hybrids parents. GCA effects of inbred lines for a given trait are the part of the genetic effect that can be transmitted to progeny. This is particularly important in population development such as synthetics. Results showed that only TZdEEI 7 consistently combined positive GCA effects for grain yield under test environments and negative GCA effects for ASI and STGC, and Striga related traits under drought stress and Striga infestation, respectively. The inbred lines with favourable GCA effects for traits under a specific group of environments could be used as parents to improve such traits in population development for the specific environment (Makumbi et al., 2011, Badu-Apraku et al., 2012b; 2015). For instance, TZEEIOR 42 combined positive GCA effects for grain yield across test environments and negative GCA effects for ASI, and LD while TZEEIOR 141 and TZEEIOR 53 combined positive GCA effects for grain yield across test environments and negative GCA effects for Striga damage rating at 8 and 10 WAP. Narrow sense heritability for a trait is an estimator of the expected genetic variance of a parent in its progeny. It allows measurement of how effective a trait can be inherited. Results of this study showed low to moderate narrow sense heritability for all the measured traits under Striga infestation and across test environments. Under drought stress, low heritability was, on the contrary, observed for all traits except for stay green characteristic with values lower than those reported by Mhike et al. (2011) and Oyekunle and Badu-Apraku (2014). However, the heritability values obtained were relatively higher under optimal conditions. These results highlighted the 118 University of Ghana http://ugspace.ug.edu.gh relative importance of non-additive genetic effects over additives genetic effects discovered for most of the traits, in this set of inbred parents and also the high heterogeneity among error variances. In order to maximise their potential usefulness for the development of productive synthetic varieties and hybrids, the newly developed extra-early orange inbreds needed to be classified into appropriate heterotic groups. This has, also, the advantage of helping to identify the potential extra- early orange testers which are completely lacking in SSA. The inbreds classified into the same heterotic group in the present study may be recombined to form heterotic populations that could be improved through recurrent selection while superior hybrids could be obtained in crosses involving inbreds from different heterotic groups. In this study, TZEEIOR 53 and TZEEIOR 141 had the highest positive and significant GCA effects and appertained to the same heterotic group. Inbred TZEEIOR 141 was identified as inbred tester and, therefore, could be used for grouping new inbreds into different heterotic groups in a line by tester mating scheme. In addition, the cross between TZEEIOR 53 and TZEEIOR 141 was ranked fourteenth among the top fifteen performing hybrids. Therefore, hybrid TZEEIOR 53 x TZEEIOR 141 could be used as a single-cross tester for the development of three-way and double cross hybrids. Maize has been successfully introduced to savannas through extra-early and early varieties. However, currently, only few commercial extra-early maturing orange maize hybrids with significant levels of resistance/tolerance to Striga and drought tolerance are available. Results of the present study indicated that hybrids TZEEIOR 12 x TZEEIOR 96, TZEEIOR 145 x TZdEEI 7, TZEEIOR 222 x TZdEEI 7, TZEEIOR 223 x TZdEEI 7, and TZEEIOR 219 x TZdEEI 7 out- yielded the best check (TZEEI 79 x TZEEI 82) by 35-60% across environments. These outstanding hybrids are also among the top performing hybrids under Striga infestation and could, therefore, 119 University of Ghana http://ugspace.ug.edu.gh be used in these areas to increase the production, incomes, and nutrition of less privileged maize farmers in SSA. Different superior hybrids were identified under different stresses. There is, hence, the need for introgression of new sources of genes responsible for resistance to Striga and tolerance to drought in this set of inbred parents. The high correlation between hybrid performance and the respective SCA and GV is an indication that SCA among parental lines and also GCA and SCA combined can predict hybrid performance, as well as the base index does. In conclusion, only TZEEIOR 53 and TZEEIOR 141 had positive and significant general combining ability effects for grain yield across test environments. Non-additive genetic effects were the primary type of gene action for grain yield and most of the other agronomic traits in all environments except for staygreen characteristic under drought stress and Striga resistance indicator traits, which could be improved based solely on GCA predictions of parental lines. The inbred line TZEEIOR 141 was identified as inbred tester while the cross TZEEIOR 53 x TZEEIOR 141 was identified as hybrid tester. Superior hybrids identified in this study including TZEEIOR 12 x TZEEIOR 196 which was identified as the most stable hybrid could successfully be used to improve maize productivity, production, and well-being of less privileged farmers in sub-Saharan Africa, especially in Striga endemic and drought prone environments. 120 University of Ghana http://ugspace.ug.edu.gh CHAPTER SIX 6.0 Estimate of combining abilities and heterosis for carotenoids in maize kernels of extra-early orange inbreds 6.1. Introduction Breeding for maize cultivars with elevated provitamin A carotenoids is a sustainable and effective way to alleviate provitamin A deficiency in sub-Saharan Africa (SSA). The use of markers for favourable alleles at both lycE and crtRB1 loci, in MAS, allowed increases exceeding the breeding target of 15 μg g−1 such as up to 17.25 μg g−1 (Azmach et al., 2013), from 15 to 20 μg g−1 (Babu et al., 2013) and more recently, up to 22.6 μg g−1 (Menkir et al., 2017) mostly in maize inbred lines. Despite the excellent progress in breeding for higher levels of provitamin A, the current released cultivars contain an average of 6-8 µg g-1 of provitamin A (HarvestPlus, 2017; Menkir et al., 2017). The existence of up to 58 genes involved in the biosynthesis of carotenoids in maize (Owens et al., 2014) makes the trait a complex quantitative trait whose inheritance need to be elucidated. The predominance of additive genetic effects and high heritability were reported (Egesel et al., 2003b) and yet, OPVs with high levels of provitamin A are lacking. On the contrary, Halilu et al. (2016) found non-additive genetic effects to be predominant for all carotenoids and this finding supports that of Burt et al. (2011) who reported that heterosis in carotenoids exist even though rare. This could justify the difficulties in developing hybrids with high levels of provitamin A. The present study aims at adding knowledge to the problem and was designed to: i. determine the combining abilities of selected orange inbreds and the gene action of accumulation of carotenoids in maize kernels; and ii. determine the heterosis and heterotic patterns of carotenoids in single crosses. 121 University of Ghana http://ugspace.ug.edu.gh 6.2. Materials and Methods 6.2.1 Plant material One hundred and thirty-six (136) hybrids from a diallel mating design together with their seventeen parental lines plus three inbred checks and four experimental hybrid checks were planted under optimal growing conditions to produce fresh kernels. Each entry was planted in a single row of one meter (1 m) length with a spacing of 0.2 m between hills, giving a total of five plants per inbred line or per hybrid. The spacing between rows was 0.75 m. All the five plants of each inbred or hybrid in the row were self-pollinated. Each cob was harvested with the husks and kept in the pollination bags to avoid contact with sunlight. The harvested cobs were dried in a glass house for five days. Representative cobs of each genotype were selected, shelled and a sample of hundred kernels was randomly drawn from the bulk to constitute the material for carotenoids analysis. Each sample was packed in a white paper envelop and dispatched for carotenoid analysis at the Institute of Biological Chemistry and Nutritional Science at the University of Hohenheim, Stuttgart, Germany. 6.2.2 Sample preparation Liquid nitrogen was added to the samples, then ground to a fine powder (≈ 0.5 mm) in a Foss CT 1093 Cyclotec sample mill. A drying test was performed to determine the necessary time for the process. In this regard, eight samples were weighed before and after freeze-drying for 8 h, 12 h, and 24 h to generate a curve and determine the appropriate duration of the drying process. The results indicated 12 h to be the appropriate duration. After drying, the samples were transferred into white plastic bags, sealed, and kept at -80 °C from where they were progressively drawn for carotenoid extraction. 122 University of Ghana http://ugspace.ug.edu.gh 6.2.3. Carotenoid extraction and quantification The total carotenoids were extracted from each sample following the Reverse-Phase-HPLC method developed by Wald et al. (2018). The sample size was modified to 200–210 mg per each repetition. Three repetitions were used for each genotype. The HPLC machine described by the same authors with the same setting were used for the chromatographic separation and detection of different compounds of carotenoids at room temperature. The injection volumes were also modified to 10 μL for the standards and 30 μL for the samples to enable clear detection of the cis-isomers of β-carotene. 6.2.4 Parameters measured and data analysis process Quantitative evaluation of standards and samples was performed via peak area using the chromatography software LabSolutions (version 5.71 SP1; Shimadzu Corporation, Kyoto, Japan). A regression line obtained from the eight different concentrations of each carotenoid compound in the standard mixture was used to calculate the carotenoid content in the samples. The concentrations obtained were then transferred to an Excel sheet where the formula below (Equation 6.1) was used to convert the concentrations read by the software to actual concentrations in the samples. 𝐶𝑜𝑛.𝐿𝐴𝐵 ×𝑉 ×100000 𝐼 Equation 6.1 : 𝐶𝐶 = 𝑑 × 𝑠𝑡𝑑 𝑆𝑊 𝐼𝑠𝑎𝑚𝑝𝑙𝑒 Where CC is carotenoid compound concentration (µg per 100g), Con.LAB is concentration of carotenoid compound read by the software (µg per L), Vd is the final dilution volume of the extract, SW is dried sample weight (mg), Istd is injection volume of the standard mixture, Isample is injection volume of the extract. 123 University of Ghana http://ugspace.ug.edu.gh The contents of the carotenoid compounds were used to calculate the total carotenoid content (TC, as the sum of the seven compounds) and the provitamin A content (PVA), as given by the formula below (Equation 6.2). 1 Equation 6.2 𝑝𝑉𝐴 = 𝛽𝐶 + (𝛽𝐶𝑋 + 𝛼𝐶 + 13𝑐𝑖𝑠𝛽𝐶 + 9𝑐𝑖𝑠𝛽𝐶) 2 Where 𝑝𝑉𝐴 is provitamin A carotenoids content, 𝛽𝐶 is beta carotene, 𝛽𝐶𝑋 is beta cryptoxanthin, 𝛼𝐶 is alpha carotene, 13𝑐𝑖𝑠𝛽𝐶 and 9𝑐𝑖𝑠𝛽𝐶 are, the cis isomers of beta carotene. Eight ratios (Venado et al., 2017) described in Table 6.1 were calculated from the compounds and used together for further analysis. Table 6.1 Derived ratios used to assess inbreds and hybrids in 2018 and their significance Ratio code Ratio formula Significance R1 β-branch/α-branch Flux between the carotenoids in the β-branch versus the α-branch R2 βC/βCX Conversion of βC into the next product βCX R3 βC/ (βCX +ZEA) Conversion of βC into the next two products βCX and ZEA R4 βCX/ZEA Conversion of βCX into the next product ZEA R5 (βC+βCX)/ZEA Conversion of βC and βCX into ZEA R6 βC/PVA Contribution of βC to the concentration of PVA R7 βCX/PVA Contribution of βCX to the concentration of PVA R8 βCX/TC Contribution of βCX to the total concentration of carotenoids LUT. = Lutein, ZEA. = Zeaxanthin, βCX = beta cryptoxanthin, 13cis-βC = 13-cis-beta carotene, βC = beta carotene, 9cis-βC = 9-cis-beta carotene. The R software was used to perform a principal component analysis of the raw data of inbreds to determine their carotenoids profiles. On the other hand, the GENES (Genes, 2016) and the SASHAYDIALL macro (Makumbi et al., 2018) were used to perform Gardner and Eberhart (1966) analysis II and Hayman diallel analysis, respectively. Gardner and Eberhart (1966) analysis II provided information on additive and nonadditive genetic effects and heterosis while Hayman 124 University of Ghana http://ugspace.ug.edu.gh analysis decomposed total variance into additive and dominance effects of genes, average degree of dominance, proportion of dominance, direction of dominance, distribution of genes, number of groups of genes that control a trait and exhibit dominance, ratio of dominant to recessive alleles in all the parents, and broad-sense (H2) and narrow-sense (h2) heritability estimates (Hayman 1954). 6.3 Results 6.3.1 Carotenoids profile of the selected parental lines Table 6.2 presents the results of the analysis of variance of carotenoid compounds for the inbred lines. There were highly significant differences p(α< 0.001) in all the measured parameters among the inbreds. Table 6.3 presents the content of carotenoid compounds and different ratio values in the inbreds. The total carotenoid content ranged from 11.19 μg g−1 (TZdEEI 7) to 42.12 μg g−1 (TZEEIOR 42) with an average of 26.77 μg g−1. On the other hand, the total provitamin A carotenoids content ranged from 0.97 μg g−1 (TZEEI 79, yellow inbred check), to 10.53 μg g−1 (TZEEIOR 100) with an average of 5.54 μg g−1. The inbred lines in this set showed different levels of various compounds within each of the two branches. However, α-carotene was not found in almost all the 18 inbreds with an average of 0.05 μg g−1. Zeaxanthin was the most abundant carotenoid compound and represented, on average, more than twice the content of lutein (5.94 vs 13.08 μg g−1). Levels of β- cryptoxanthin ranged from 0.27 μg g−1 (TZdEEI 7) to 6.97 μg g−1 (TZEEIOR 141) whereas those of β-carotene varied from 0.55 μg g−1 (TZEEI 79) to 6.93 μg g−1 (TZEEIOR 100). 125 University of Ghana http://ugspace.ug.edu.gh Table 6.2 Mean squares of different compounds of carotenoids in 18 tropical inbreds lines Source of DF LUT. ZEA. βCX. 13βC αC. βC. αC. PVA TC variation Rep 2 0.86ns 4.66ns 0.33ns 0.02ns 0.00ns 0.12ns 0.01ns 0.56* 16.90* Entry 19 20.00*** 41.50*** 10.88*** 0.12*** 0.05*** 7.78*** 0.34*** 17.93*** 144.15*** Error 38 0.37 1.67 0.12 0.01 0.00 0.04 0.00 0.17 5.10 R-Square 0.96 0.93 0.98 0.92 0.96 0.99 0.98 0.98 0.93 CV, % 10.51 9.97 9.36 14.67 48.54 6.65 7.41 7.42 8.44 ‘ns’ non-significance; “*”, ‘***’ significance at 𝑝(𝛼 <0.05) and 𝑝(𝛼 <0.001); DF, degree of freedom; LUT, lutein; ZEA, zeaxanthin; βCX., β-cryptoxanthin; 13βC, 13-cis-β-Carotene; α-Carotene; β-Carotene; PVA, pro vitamin A carotenoids; TC; total carotenoids. 126 University of Ghana http://ugspace.ug.edu.gh Table 6.3 Profile of the 17 parental inbred lines and one checks in carotenoid compounds and derived parameters and ratios Entry LUT ZEA β-CX 13βC αC βC 9βC PVA TC R1 R2 R3 R4 R5 R6 R7 R8 TZEEIOR 12 3.27 11.17 2.21 0.33 0.01 1.98 0.25 3.37 19.21 0.20 0.90 0.15 0.20 0.38 0.59 0.66 0.12 TZEEIOR 42 13.83 18.62 3.87 0.67 0.40 4.07 0.67 6.87 42.12 0.51 1.05 0.18 0.21 0.43 0.59 0.56 0.09 TZEEIOR 53 5.00 12.08 2.54 0.47 0.00 2.44 0.68 4.28 23.21 0.28 0.96 0.17 0.21 0.41 0.57 0.59 0.11 TZEEIOR 76 4.54 13.66 2.29 0.58 0.00 2.58 1.03 4.53 24.68 0.23 1.13 0.16 0.17 0.36 0.57 0.51 0.09 TZEEIOR 100 2.86 13.40 5.30 0.76 0.00 6.93 1.14 10.53 30.38 0.10 1.31 0.37 0.40 0.91 0.66 0.50 0.17 TZEEIOR 113 4.47 13.68 2.69 0.37 0.00 1.99 0.61 3.83 23.82 0.23 0.74 0.12 0.20 0.34 0.52 0.70 0.11 TZEEIOR 130 6.40 10.51 1.66 0.38 0.00 2.15 0.59 3.47 21.69 0.42 1.30 0.18 0.16 0.36 0.62 0.48 0.08 TZEEIOR 141 6.18 16.52 6.97 0.78 0.04 3.43 1.19 7.92 35.10 0.22 0.49 0.15 0.42 0.63 0.43 0.88 0.20 TZEEIOR 145 8.28 19.02 4.43 0.40 0.13 2.19 0.66 5.00 35.10 0.32 0.49 0.09 0.24 0.35 0.44 0.89 0.13 TZEEIOR 161 7.82 14.05 2.17 0.28 0.00 1.89 1.02 3.63 27.23 0.40 0.87 0.12 0.16 0.29 0.52 0.60 0.08 TZEEIOR 196 7.13 16.69 3.03 0.41 0.00 3.00 0.81 5.12 31.06 0.30 0.99 0.15 0.18 0.36 0.59 0.59 0.10 TZEEIOR 218 5.40 12.59 5.18 0.49 0.21 2.55 0.49 5.74 26.91 0.26 0.49 0.14 0.41 0.61 0.45 0.90 0.19 TZEEIOR 219 4.92 13.98 4.97 0.51 0.08 2.73 0.49 5.75 27.67 0.22 0.55 0.14 0.36 0.55 0.47 0.86 0.18 TZEEIOR 222 4.41 13.45 6.03 0.57 0.00 3.43 0.79 7.13 28.69 0.18 0.57 0.18 0.45 0.71 0.48 0.85 0.21 TZEEIOR 223 4.51 14.77 6.63 0.67 0.00 3.50 0.82 7.56 30.89 0.17 0.53 0.16 0.45 0.69 0.46 0.88 0.22 TZdEEI 7 4.08 4.13 0.27 0.48 0.00 1.62 0.61 2.30 11.19 0.57 6.02 0.37 0.07 0.46 0.70 0.12 0.02 TZdEEI 12 4.70 11.42 2.98 0.57 0.00 2.63 0.89 4.84 23.18 0.25 0.88 0.18 0.26 0.49 0.54 0.62 0.13 TZEEI 79 9.08 5.69 0.65 0.07 0.00 0.55 0.14 0.97 16.17 1.28 0.84 0.09 0.11 0.21 0.56 0.66 0.04 Mean 5.94 13.08 3.55 0.49 0.05 2.76 0.72 5.16 26.57 Minimum 2.86 4.13 0.27 0.07 0.00 0.55 0.14 0.97 11.19 Maximum 13.83 19.02 6.97 0.78 0.39 6.93 1.19 10.53 42.12 S.E.M 0.33 0.49 0.24 0.03 0.01 0.21 0.05 0.32 0.91 “*”, “**”, “***”, = significance at p(𝛼 <0.05), p(𝛼 <0.01), and p(𝛼 <0.001) respectively; “ns” = non-significant. LUT. = Lutein, ZEA. = Zeaxanthin, βCX = beta cryptoxanthin, 13cis-βC = 13-cis-beta carotene, βC = beta carotene, 9cis-βC = 9-cis-beta carotene; S.E.M, standard error of mean. R1 = Flux between the carotenoids in the β-branch versus the α-branch, R2 = Conversion of βC into the next product βCX, R3 = Conversion of βC into the next two products βCX and ZEA, R4 = Conversion of βCX into the next product ZEA, R5 = Conversion of βC and βCX into ZEA, R6 = Contribution of βC to the concentration of PVA, R7 = Contribution of βCX to the concentration of PVA, R8 = Contribution of βCX to the total concentration of carotenoids. 127 University of Ghana http://ugspace.ug.edu.gh Figure 6.1 describes the inbred lines regarding the content of carotenoid compounds and the importance of different ratios in determining the variation observed among the lines. The principal component (PCA) analysis biplot indicated that lutein and α-carotene contributed less to the 72.1% variation explained by the principal components 1 and 2, from the total variation among the inbreds. β-cryptoxanthin and total provitamin A carotenoids contributed most to the variation explained by PCA1 while lutein content and the conversion of β-carotene to β-cryptoxanthin and zeaxanthin were most determinant of the PCA2. Inbreds TZEEIOR 141, TZEEIOR 222, and TZEEIOR 223 were characterized by high levels of β-cryptoxanthin, high contribution of β- cryptoxanthin to total provitamin carotenoids, and low conversion of β-cryptoxanthin to zeaxanthin. On the other hand, inbreds TZEEIOR 42, TZEEIOR 145, TZEEIOR 218, and TZEEIOR 219 were characterized by high levels of total carotenoids and zeaxanthin. TZEEIOR 42 had, in addition, the highest level of lutein with 13.83 μg g−1. Figure 6.2 presents the inbred lines in the different groups based on their profiles. Three different groups were suggested by Elbow method. Inbred TZEEI 79 and TZdEEI 7, characterised by low levels in all carotenoid compounds in the β-branch of the carotenoid biosynthetic pathway, were grouped together in Group I. The inbreds TZEEIOR 42, TZEEIOR 100, TZEEIOR 141, TZEEIOR 145, TZEEIOR 218, TZEEIOR 219, TZEEIOR 222, and TZEEIOR 223 constituted the Group II. This group was composed of the inbreds with high total carotenoids with moderate to high levels of carotenoids in the β-branch. This group was also characterised by equal contribution of β- cryptoxanthin and β-carotene to the total provitamin A carotenoids level. The remaining inbreds were clustered together in Group III and were characterized by low to moderate levels of total carotenoids and variable content of carotenoids in the β-branch. 128 University of Ghana http://ugspace.ug.edu.gh Figure 6.1 Carotenoids profile of 18 maize inbred lines using Principal Components Analysis (PCA) biplot Legend Code Name Code Name 1 TZEEIOR 12 10 TZEEIOR 161 2 TZEEIOR 42 11 TZEEIOR 196 3 TZEEIOR 53 12 TZEEIOR 218 4 TZEEIOR 76 13 TZEEIOR 219 5 TZEEIOR 100 14 TZEEIOR 222 6 TZEEIOR 113 15 TZEEIOR 223 7 TZEEIOR 130 16 TZdEEI 7 8 TZEEIOR 141 17 TZdEEI 12 9 TZEEIOR 145 18 TZEEI 79 129 University of Ghana http://ugspace.ug.edu.gh A I B III II Figure 6.2 Optimal number of clusters suggested by Elbow method (A) and carotenoids profile-based grouping of 18 maize inbred lines (B) 130 University of Ghana http://ugspace.ug.edu.gh 6.3.2 Heterosis for carotenoid accumulation in maize kernels Table 6.4 presents the results of Gardner & Eberhart analysis II and Hayman analysis of diallel of carotenoid traits among extra-early maize inbreds. There were highly significant (p< 0.001) differences among inbreds effects and effects of their derived crosses for all the carotenoid compounds, total provitamin A, and total carotenoids content. Average heterosis was significant (p< 0.001) for lutein and zeaxanthin, on one hand, and highly significant (p< 0.001) for beta- carotene and its cis isomers, on the other hand. The contribution of heterosis by each inbred to its crosses was also highly significant for all the measured parameters except for lutein while the specific heterotic effect in each cross was highly significant for all parameters. Table 6.5 presents the summary statistics for the mid-parent heterosis, the better parent heterosis, and the worst parent heterosis. There were small and positive heterotic effects for all the traits for heterosis and heterobeltiosis. The average heterotic effects were 35.98%, 5.68%, 2.78%, 1.3%, 0.86%, and 0.09% for 9-cis-β-carotene, 13-cis-β-carotene, β-cryptoxanthin, PVA, zeaxanthin, and total carotenoids, respectively. Table 6.6 presents the heterosis and the heterobeltiosis of β-cryptoxanthin for the 136 single crosses. The highest heterosis effect for β-cryptoxanthin was obtained for TZEEIOR 113 x TZEEIOR 130 with value of 67.4%. The highest heterosis obtained among high-level parents for β-cryptoxanthin was 20.37% between TZEEIOR 218 and TZEEIOR 219. The Inbred TZEEIOR 113 showed a consistent positive and significant heterosis effect in all the crosses except its crosses with TZEEIOR 141 and TZEEIOR 223. A similar trend was observed for the better parent heterosis with the highest being 49.69% between TZEEIOR 53 and TZEEIOR 113. The cross TZEEIOR 100 x TZEEIOR 218 exhibited 24.75% heterobeltiosis, the highest among high-level parents for β-cryptoxanthin. 131 University of Ghana http://ugspace.ug.edu.gh Table 6.4 Analysis of variance for carotenoid compounds, provitamin A (PVA), and total carotenoids (TC) of selected extra- early maturing orange maize Source of Gardner & Eberhart (1966) mean squares of Carotenoids compounds and derived parameters variation DF LUT. ZEA. 𝛃𝐂𝐗 𝟏𝟑𝛃𝐂 𝛃𝐂 𝟗𝛃𝐂 PVA TC Replication 2 0.02ns 2.28* 0.15ns 0.00ns 0.09ns 0.01ns 0.29ns 5.06ns Genotype 152 11.4*** 22.25*** 6.24*** 0.06*** 1.82*** 0.16*** 6.31*** 95.53*** Inbreds 16 72.2*** 166.82*** 50.67*** 0.34*** 13.44*** 0.68*** 49.45*** 722.43*** Hybrids 136 4.2*** 5.24*** 1.01*** 0.03*** 0.46*** 0.10*** 1.24*** 21.77*** H.Mean 1 3.8** 0.29ns 0.15ns 0.02** 1.12*** 2.49*** 0.001ns 0.16ns H.Inbreds 16 4.8 9.41*** 1.98*** 0.05*** 0.87*** 0.16*** 2.67*** 39.26*** H.Specific 119 4.1*** 4.72*** 0.88*** 0.03*** 0.40*** 0.07*** 1.06*** 19.60*** E rror 3 04 0.4 0.66 0.1 0.002 0.04 0.01 0.14 2.79 Hayman (1954) mean squares of Carotenoids compounds and derived parameters Replication 2 0.02ns 2.28* 0.15ns 0.00ns 0.09ns 0.01ns 0.29ns 5.06ns Genotype 152 11.36*** 22.25*** 6.24*** 0.06*** 1.82*** 0.16*** 6.31*** 95.53*** a 16 30.51*** 32.17*** 8.61*** 0.14*** 4.05*** 0.36*** 11.76*** 179.36*** b 136 9.10*** 21.08*** 5.96*** 0.05*** 1.56*** 0.14*** 5.67*** 85.66*** b1 1 1.14ns 2.03* 8.08** 0.01ns 0.24* 0.02ns 4.42* 16.62* b2 16 3.75*** 22.66*** 7.34*** 0.09*** 2.51*** 0.22*** 9.41*** 104.28*** b3 119 9.89*** 21.03*** 5.75*** 0.05*** 1.45*** 0.13*** 5.18*** 83.74*** Residual 304 0.36 0.66 0.1 0.0 0.04 0.01 0.14 2.79 a x Rep 32 0.38 0.8 0.13 0 0.06 0.01 0.21 3.63 b1 x Rep 2 0.17 0.03 0.01 0 0.01 0 0.03 0.13 b2 x Rep 32 0.39 0.43 0.05 0 0.03 0.01 0.11 2.29 b3 x Rep 238 0.35 0.68 0.1 0 0.04 0.01 0.14 2.77 b x Rep 272 0.35 0.64 0.09 0 0.04 0.01 0.14 2.69 Total x Rep 304 0.36 0.66 0.1 0 0.04 0.01 0.14 2.79 Mean 5.5 13.6 3.8 0.53 2.7 1 5.4 27.1 CV % 10.9 6 8.3 8.6 7.6 10.4 7 6.2 “**”, “***”, = significance at p(𝛼 =0.01), and p(𝛼 =0.001) respectively; “ns” = non-significant. LUT. = Lutein, ZEA. = Zeaxanthin, βCX = beta cryptoxanthin, 13cis-βC = 13-cis-beta carotene, βC = beta carotene, 9cis-βC = 9-cis-beta carotene. a, additive effect; b, dominance effect; b1, measure of directional dominance; b2, measure of ambi-directional dominance; b3, residual dominance. 132 University of Ghana http://ugspace.ug.edu.gh Table 6.5 Minimum, maximum, and average heterosis (H) and heterobeltiosis (Hb) for carotenoids compounds, provitamin A (PVA), and total carotenoids (TC) of selected extra- early maturing orange maize H% Hb(>P) % Hb(P) = heterobeltiosis higher parent; Hb( ?̂?2). Also, there was not equal distribution of alleles with positive and negative effects in the parents (?̂?2⁄ < 0.5) at loci 4?̂?1 exhibiting dominance. The proportion of genes with positive and negative effects was less than 0.2 for all the traits in this set of inbreds. The results showed that two genes or block of genes which exhibited total dominance for 9-cis-β- carotene were present in the parents (ℎ̂2⁄ = 1.88). For the rest of the measured traits, there were ?̂?2 no genes or blocks of genes with complete dominance, at the loci exhibiting dominance. Table 6.9 presents the general combining ability estimates of inbred parents for carotenoid compounds. Expected small GCA effects were observed for all measured traits but differences among inbreds were highly significant. Only the inbred TZEEIOR 42 expressed positive and highly significant effect for all the traits. The inbred TZEEIOR 100 showed a positive and significant effect for carotenoids in the β-branch. The highest additive effect for β-cryptoxanthin was observed for TZEEIOR 218, TZEEIOR 223, and TZEEIOR 100 with 1.18, 1.09, and 1.06 μg g−1, respectively. For β-carotene, the highest effect was found in TZEEIOR 100 with 1.40 μg g−1, followed by TZEEIOR 42 with a far lower effect, 0.8 μg g−1. 136 University of Ghana http://ugspace.ug.edu.gh Table 6.8 Genetic parameter estimates for carotenoid compounds, provitamin A carotenoids (PVA), and total carotenoids (TC) of 17 selected extra-early maturing orange maize inbreds and their 136 hybrids Carotenoid compounds Parameter PVA TC LUT ZEA. 𝛃𝐂𝐗 𝟏𝟑𝛃𝐂 𝛃𝐂 𝟗𝛃𝐂 ?̂? 4.22 ± 0.16 48.38 ± 2.59 6.42 ± 0.57 11.50 ± 0.73 3.55 ± 0.26 0.02 ± 0.0 1.52 ± 0.09 0.06 ± 0.01 ?̂?1 2.21 ± 0.28 36.52 ± 4.61 6.48 ± 1.01 8.78 ± 1.3 1.77 ± 0.46 0.05 ± 0.01 0.79 ± 0.16 0.15 ± 0.02 ?̂?2 1.37 ± 0.22 24.44 ± 3.66 5.02 ± 0.8 5.89 ± 1.03 1.14 ± 0.36 0.03 ± 0.01 0.52 ± 0.12 0.10 ± 0.01 ?̂?2 0.00 ± 0.15 0.00 ± 2.44 0.27 ± 0.54 0.00 ± 0.69 0.00 ± 0.24 0.00 ± 0.0 0.08 ± 0.08 0.19 ± 0.01 ?̂? 0.05 ± 0.04 0.94 ± 0.61 0.12 ± 0.13 0.22 ± 0.17 0.03 ± 0.06 0.00 ± 0.0 0.01 ± 0.02 0.00 ± 0.0 ?̂? 1.59 ± 0.34 8.35 ± 5.69 2.85 ± 1.25 2.40 ± 1.61 0.56 ± 0.56 0.01 ± 0.01 0.89 ± 0.19 0.06 ± 0.02 √?̂?𝟏⁄ 0.72 0.87 1 0.87 0.7 1.51 0.72 1.61 ?̂? ?̂?𝟐⁄ 0.16 0.17 0.19 0.17 0.16 0.17 0.16 0.17 𝟒?̂?𝟏 𝑲?̂?⁄ 1.71 1.22 1.57 1.27 1.25 1.37 2.36 1.84 𝑲?̂? ?̂?𝟐⁄ 0 0 0.05 0 0 0.05 0.16 1.88 ?̂?𝟐 r 0.69 056 0.86 0.35 0.94 -0.24 0.65 -0.15 𝑯𝟐𝒏 0.82 0.79 0.65 0.78 0.85 0.58 0.76 0.47 𝑯𝟐𝒃 0.98 0.97 0.97 0.97 0.98 0.97 0.98 0.94 ?̂?, component of variation due to additive effect of genes; ?̂?1, component of variation due to dominance effects of genes; ?̂?2, dominance component indicating asymmetry of positive and negative effects of genes; ℎ̂2, overall mean dominance effect of heterozygous loci; ?̂?, relative frequency of dominant and recessive alleles in the parents; ?̂?, replicate variation. √?̂?1 𝐾⁄ = Mean Degree of Dominance, ?̂?2⁄ = Proportion of genes with positive and negative effects in the parents, ?̂?⁄ = Proportion of dominant & recessive genes in parents, ?̂? 4?̂?1 𝐾?̂? ℎ̂2⁄ = Number of groups of genes which control the character and exhibit dominance, 𝐻2𝑛 = Narrow-sense Heritability, 𝐻 2 𝑏= Broad-sense Heritability. LUT. = Lutein, ZEA. = ?̂?2 Zeaxanthin, βCX = beta cryptoxanthin, 13cis-βC = 13-cis-beta carotene, βC = beta carotene, 9cis-βC = 9-cis-beta carotene. 137 University of Ghana http://ugspace.ug.edu.gh Table 6.9 General combining ability (GCA) estimates of selected extra-early maturing orange maize for different carotenoid compounds and derived traits GCA effect Inbred LUT ZEA βCX 13βC αC βC 9βC PVA TC TZEEIOR 12 -0.84*** -1.42*** -0.94*** -0.13*** -0.03*** -0.28*** -0.23*** -0.94*** -3.87*** TZEEIOR 42 3.40*** 3.18*** 0.56*** 0.18*** 0.11*** 0.80*** 0.16*** 1.30*** 8.38*** TZEEIOR 53 0.40*** -0.19ns -0.52*** 0.07*** -0.02*** -0.11** 0.03* -0.32*** -0.32ns TZEEIOR 76 -1.40*** -0.40* -0.60*** -0.03*** -0.04*** -0.11*** 0.06*** -0.42*** -2.52*** TZEEIOR 100 -1.10*** 1.53*** 1.06*** 0.12*** -0.03*** 1.40*** 0.24*** 2.10*** 3.23*** TZEEIOR 113 -0.24* 0.53*** 0.21*** 0.01* 0.00* -0.06* 0.00ns 0.05ns 0.45ns TZEEIOR 130 0.70*** -0.76*** -1.17*** -0.03** -0.02*** -0.18*** 0.04** -0.77*** -1.42*** TZEEIOR 141 -0.32** 0.83*** 0.90*** 0.03*** -0.03*** -0.11** -0.04* 0.32*** 1.26*** TZEEIOR 145 0.15ns 0.76*** -0.02ns -0.06*** 0.00* -0.50*** -0.08*** -0.58*** 0.26ns TZEEIOR 161 0.49*** -0.05ns -0.78*** -0.09*** 0.02*** -0.51*** 0.05*** -0.91*** -0.87** TZEEIOR 196 1.19*** 0.86*** -0.41*** -0.05*** -0.03*** 0.07* 0.02ns -0.17** 1.65*** TZEEIOR 218 0.48*** -0.13ns 1.18*** 0.01ns 0.10*** 0.07* -0.06** 0.68*** 1.65*** TZEEIOR 219 -0.15ns 0.65*** 0.75*** -0.02* 0.02*** -0.07* -0.10*** 0.26*** 1.09*** TZEEIOR 222 -0.32*** 0.07ns 0.95*** 0.01ns -0.02*** 0.10** -0.01ns 0.56*** 0.78* TZEEIOR 223 -0.69*** 0.71*** 1.09*** 0.05*** -0.03*** 0.27*** 0.01ns 0.83*** 1.41*** TZdEEI 7 -0.95*** -5.25*** -2.05*** -0.06*** 0.00* -0.64*** -0.15*** -1.77*** -9.09*** TZdEEI 12 -0.79*** -0.92*** -0.21*** -0.01ns -0.03*** -0.14*** 0.03* -0.25*** -2.06*** Baker ratio 0.66 0.79 0.85 0.57 0.54 0.77 0.44 0.83 0.8 “*”, “**”, “***”, = significance at p(𝛼 <0.05), p(𝛼 <0.01), and p(𝛼 <0.001) respectively; “ns” = non-significant. LUT. = Lutein, ZEA. = Zeaxanthin, βCX = beta cryptoxanthin, 13cis-βC = 13-cis-beta carotene, βC = beta carotene, 9cis-βC = 9-cis-beta carotene. 138 University of Ghana http://ugspace.ug.edu.gh 6.4 Discussion and conclusion The observed range of levels for all the carotenoid compounds among the orange maize inbred lines in this study are within the ranges reported by previous studies involving yellow or orange endosperm colour maize (Egesel et al., 2003b; Menkir et al., 2008). These inbreds are characterised by significantly improved levels of all the β-branch carotenoids with zeaxanthin and β-cryptoxanthin being the most abundant. This is in line with the method used for their development. Indeed, these inbreds were developed from a converted yellow population to orange using only visual scores of the grain colour for carotenoids improvement. Egesel et al. (2003b) found zeaxanthin to be a major pigment in endosperm colour. Thus, the results of this study support the idea that grain colour can be used as a secondary trait for indirect selection for zeaxanthin and β-cryptoxanthin (Chandler et al., 2013). The moderate levels of β-carotene in this set of inbreds suggest that there was not much improvement of the trait during the selection process. Moreover, the orange colour source, STR-34-1-1-1-1-2-1-B*5/NC354/SYN-Y-STR-34-1-1-1-1-2-1-B*5 (OR1), used to convert 2004 TZEE-Y STR C4 to an orange population from which the inbreds were derived was, later on, discovered to carry none of the functional alleles of the crtRB1 gene (Azmach et al., 2013). In a different study that involved the inbred parents in this research, it was observed that only TZEEIOR 196 and TZdEEI 7 contained the favourable allele at the 3’TE of lycE loci while all of the other inbreds contained the unfavourable allele at 5’TE of crtRB1 loci. It is clear that the 3’TE functional allele found in TZEEIOR 196 and TZdEEI 7 is not coming from the donor (OR1). The presence of this functional allele in these genotypes is a validation of the finding of Yan et al. (2010), who reported 3’TE favourable allele to be present in tropical germplasm at 4.6%. Even though TZEEIOR 196 possesses the 3’TE functional allele, which has proved to be correlated with increased levels of β-carotene, its level in β-carotene is three times 139 University of Ghana http://ugspace.ug.edu.gh lower than that of TZEEIOR 100 and two times lower than that of TZEEIOR 42. Discovering genotypes with no known functional allele of the most used major genes in marker-assisted breeding (PSY1, lcyE, and crtRB1) having good levels of carotenoids is not a surprising event. Indeed, Azmach (2013) identified a line (entry 50) having unfavourable alleles at lcyE SNP, crtRB1 5’TE, crtRB1 InDel4, and crtRB1 3’TE (T, 1, 0, 3), which exceeded by 23.48 μg g−1 that of the average total carotenoid of those genotypes carrying the favourable alleles. The existence of up to 58 genes involved in the carotenoids biosynthesis (Owens et al., 2014b) may explain the performance of some genotypes whose genome has been explored for the well-known candidate genes used up to date. Beside, Yan et al. (2010) found the most pronounced effect leading to higher βC concentrations could be attributed to the 206-bp insertion allele of 5'TE, a rare allele detected only in temperate germplasm at a frequency of 2.9% , although additive genetic effects were also reported for combinations of favourable alleles at 5’TE, InDel4, and 3’TE and for combinations of favourable alleles of crtRB1 and lcyE genes. Thus, the presence of only 3’TE favourable alleles in TZEEIOR 196 and TZdEEI 7 was not enough to result in higher increase of β-carotene. The variability discovered, even in this small set of inbred lines, indicates the potential to select parental lines for genetic improvement of the compounds for which they perform. Principal component biplot on the carotenoid composition of the inbred lines revealed that some inbred lines had higher concentrations of provitamin A carotenoids, as well as the xanthophylls (lutein, and zeaxanthin). The inbred TZEEIOR 42 can be used as a parent for improvement of both lutein and β-branch provitamin A carotenoids while all the inbreds in group II can be used purposely for all β-branch carotenoids. But, the fact that most of the inbreds in this study did not possess favourable alleles of the known candidate genes despite their performance makes them the best candidates for gene introgression. 140 University of Ghana http://ugspace.ug.edu.gh Both additive and non-additive genetic variances were observed to be significantly different from zero revealing the complexity in the inheritance of the studied traits. The values of Baker’s ratio (>0.5) indicated the predominance of additive genetic variance for all the traits except for 9-cis- β-carotene (0.44). However, the predictability of selecting superior progeny based on estimates of GCA alone can only be possible for β-cryptoxanthin for which Baker’s ratio was close to unity. As the levels of PVA in this set of inbreds were mainly determined by their levels in β- cryptoxanthin, selecting superior progeny for PVA is also found to be predictable based on estimates of GCA alone. High narrow sense heritabilities were also observed for the majority of the traits. This report on the predominance of additive genetic effects and high heritability, as also reported in earlier studies (Egesel et al., 2003a; Halilu et al., 2016; Menkir et al., 2014), suggests that progress in breeding maize for increased carotenoid content especially, provitamin A carotenoids to be achievable. But, from the discovery of favourable alleles of lcyE (Harjes et al., 2008) and that of crtRB1 (Yan et al., 2010) to their use in marker-assisted selection for elevated levels of provitamin A carotenoids (Babu et al., 2013; Menkir et al., 2017), high levels of PVA exceeding the actual breeding target (15 μg g−1) were only reported in inbreds. Released PVA hybrids contained a range of 6 to 8 μg g−1 of PVA. Given the availability of inbreds with much higher levels suggest some constraints to a realized heterosis in hybrid development. This may possibly be due to the presence of significant non-additive genetic effects as found in this study and also reported by Halilu et al. (2016). Dominance genetic effects were found to significantly contribute to the inheritance of most carotenoid traits in the present study. There were more dominant genes than recessive genes in the parents with the ratio of the dominant to recessive being greater than 2 (2.36) for β-carotene. Hayman diallel analysis revealed that at loci exhibiting dominance, recessive alleles were mostly 141 University of Ghana http://ugspace.ug.edu.gh positive for β-cryptoxanthin, lutein, and to some extent for β-carotene while dominant genes carried both effects for the rest of the traits. This supports the importance of the ambi-directional dominance detected. The presence of more dominant alleles with negative effects for β- cryptoxanthin and β-carotene could explain the negative to low heterotic effects observed for these traits in this set of inbreds. The negative effects of dominant genes might not have only nullified the effect of the recessive homozygotes in the parents but also could have significantly reduced the effects of additive gene effects over loci. This finding, consequently, suggests that effort in hybrid development for elevated levels of provitamin A can be hampered by the predominant negative genetic effects of dominant genes and could explain the difficulties in developing such hybrids needed to compensate for post-harvest degradation of carotenoids, degradation during processing, and storage. Using molecular markers, Menkir et al. (2017) developed inbred lines with provitamin A levels ranging from 0.2 to 22.6 μg g−1, from 2013 to 2016. At the same time, best hybrids developed from these lines have provitamin A levels ranging from 10.1 to 11.1 μg g−1. This depicts the unavoidable need of the usage of markers to support selection based on colour rating, as recommended by Chander et al. (2008). In conclusion, this study showed a similar trend in carotenoid levels in both parental lines and crosses derived from them. The genetic analysis revealed presence of negative dominance effects across loci which reduces the effects of the predominantly additive effects discovered for all carotenoid traits. Consequently, heterosis and heterobeltiosis were found to be mostly negative with the highest values obtained from crosses involving parents with low-level of the considered trait. 142 University of Ghana http://ugspace.ug.edu.gh Finally, genotypic selection based on gain colour rating alone can be effective for increased levels of β-cryptoxanthin and zeaxanthin in maize kernels. However, effective breeding for much higher carotenoids content may require the use of molecular markers to support and guide selection based on endosperm colour. Validation of introgressed favourable alleles of the major genes could also be considered in hybrids before their evaluation to determine the overall genetic effect of favourable alleles over other genes with possible epistatic effects. Development of hybrids with increased levels of carotenoids should consider the finding of this study by opting for a progressive gene pyramiding or simultaneous targeting of major genes and other genes with epistatic effects in a genomic selection programme. The latter option could be achieved only if maximum genes or QTLs directly or indirectly influencing the accumulation of different carotenoids are discovered and well known. In this regard, QTL analysis of the genetic pool from which the inbreds involved in this study are derived will bring clarification on the genomic regions that are linked to their performance for β-cryptoxanthin. 143 University of Ghana http://ugspace.ug.edu.gh CHAPTER SEVEN 7.0 CONCLUSIONS AND RECOMMENDATIONS Conclusion The main goal of this research was to develop extra-early maturing maize hybrids that combine high yield, resistance to Striga, tolerance to drought and elevated levels of carotenoids, especially provitamin A carotenoids for Striga endemic and drought-prone areas of the savanna agro- ecologies of SSA in order to alleviate food shortage in the sub-region. Towards this end, four experiments were conducted. Adequate genetic variability among the selected 152 orange inbreds with respect to each trial management (Striga infestation, drought stress and optimal) was observed. The residual variability measured was due to different levels of resistance to each of the stresses. The predicted breeding values (Pedigree-based BLUPs) under Striga infestation were almost thrice greater than that under drought stress for all the inbreds, indicating that there was a higher level of resistance/tolerance to Striga than the level of tolerance to drought. However, the higher additive genetic effects’ contribution to the total genetic variability observed confirmed the genetic basis of the resistance levels to each of the stresses and the fact that the traits are highly heritable. Across all test environments, the breeding values of the inbreds were similar to those under Striga infestation indicating that their performance across test environments may be driven by the performance under Striga. Predictions over pedigree allowed identification of several inbreds not evaluated with breeding values that fell among the top best selected and even greater than those of some previously selected inbreds. Molecular characterisation of the 152 orange maize inbred lines using 4620 SNPs revealed high homogeneity in 92% of the inbreds. However, TZEEIOR 75, TZEEIOR 76, TZEEIOR 161 and 144 University of Ghana http://ugspace.ug.edu.gh TZEEIOR 38 were highly heterogeneous with heterozygosity level greater than 12.5%. Also, the study revealed that the inbreds were closely related regarding the genetic distance and the kinship coefficient with the latter confirming their origin from a shared pedigree. The unknown relationship between the markers used in this study and the traits for which the inbreds were selected could explain the results obtained. This suggests that specific markers may be needed to explore the true genetic relationship among these inbreds. However, model-based population structure and neighbour-joining clustering concorded in distinguishing four groups in 71% of the inbreds while 39% shared properties of different groups, probably due to the low genetic distances observed among the inbreds. Distant inbreds could be used in planned crosses to maximise the performance in their progenies. In the evaluation of the seventeen selected inbreds for their performance in crosses, there were significant variations among environments, components of genetic variance (GCA and SCA), and environments by variance components interactions among inbred parents for grain yield and most of the traits. This indicated the presence of adequate variability among inbred parents and also among hybrids for breeding purposes and also the need for testing in several environments for selecting performing inbred parents and hybrids. Additive and non-additive genetic effects were both important in controlling grain yield and most of the measured traits in each trial and across environments. However, for leaf senescence and Striga related traits, additive genetic effects were the primary type of gene action suggesting that selection for these traits could easily be done based on predictions of GCA alone. Again, GCA values under Striga infestation were greater than those under drought stress indicating that inbred parents contributed more to the hybrid performance under Striga than under drought. The performance of the hybrids across test environments was highly correlated to the SCA and to their genetic values. This highlighted not only the importance 145 University of Ghana http://ugspace.ug.edu.gh of SCA in predicting hybrid performance but also the contribution of GCA to the performance. That is, a hybrid from two parents with high GCA and SCA will yield higher than a hybrid from two parents with low GCA effects compared to the first parents but the same SCA value as the first hybrid. Among the selected outstanding hybrids across test environments, TZEEIOR 12 x TZEEIOR 196 and TZEEIOR 222 x TZdEEI 7 were the most stable and yielded respectively, 3885 kg ha-1 and 3611 kg ha-1 across environments and 5411 kg ha-1 and 4249 kg ha-1 under optimal conditions. They could thus be proposed for on-farm evaluations or included in mother-baby trials to test their adaptability to farmers conditions and acceptance. Analysis of inbred parents and their hybrids for carotenoid levels revealed three different profiles in the inbreds: inbreds with high levels of total carotenoids, inbreds with high levels of beta-branch carotenoids, and inbreds with almost equal levels of β-carotene and β-cryptoxanthin. Inbred TZEEIOR 42 had both high levels of β-cryptoxanthin and xanthophylls. The levels of provitamin A in this set of inbreds was mainly due to the equal contribution of β-carotene and β-cryptoxanthin. Dominance genetic effects were found to significantly contribute to the inheritance of most carotenoid traits. Also, there were more dominant genes than recessive genes in the parents with the ratio of dominant to recessive being greater than 2 (2.36) for β-carotene. Dominant alleles were found to have negative effects on most of the traits, especially for β-cryptoxanthin and β-carotene. That explains the negative to low heterotic effects observed for these traits in this set of inbreds. The negative effects of dominant genes in the parents might have reduced the effects of additive gene effects over loci, thus yielding low levels in most of the hybrids. Thus, breeding for elevated levels of carotenoids in maize using colour rating improves significantly the level of zeaxanthin and β-cryptoxanthin but not in β-carotene. Success in breeding heterosis for provitamin A might need to consider the genetic background of inbred parents with one important step being the 146 University of Ghana http://ugspace.ug.edu.gh reduction of dominant genes with negative effects by increasing the number of favourable alleles through introgression. Recommendations to: ❖ IITA-MIP 1. The best inbreds not evaluated but selected by pedigree-based BLUPs should be evaluated to confirm their performance; 2. The level of tolerance to drought of inbreds should be increased through introgression of novel resistance alleles from different sources; 3. Specific markers such as validated drought tolerance and resistance to Striga should be used to further characterize the inbreds; 4. Request for the readily available sources of 5’TE favourable alleles from CIMMYT to be introgressed into the inbreds which have acceptable levels of β-branch carotenoids; ❖ IITA-MIP and other research institutes 5. Adopt a more progressive programme for gene pyramiding or simultaneous selection of major genes in a genomic selection programme for elevated levels of carotenoids in maize. ❖ IITA collaborators 6. Hybrids TZEEIOR 12 x TZEEIOR 196, TZEEIOR 145 x TZdEEI 7, and TZEEIOR 222 x TZdEEI 7 should be tested in regional trials and if suitable, released for adoption and commercialization in the sub-region; ❖ Further researche 7. Study the interaction between genes for tolerance to drought or resistance to Striga and genes for increased levels of β-branch carotenoids. 147 University of Ghana http://ugspace.ug.edu.gh REFERENCES AATF, 2006. Empowering African farmers to eradicate Striga from maize croplands. The African Agricultural Technology Foundation. Nairobi, Kenya. 17 pp. Akanvou, L., Doku E.V., and Kling J., 1997. Estimates of genetic variances and interrelationships of traits associated with Striga resistance in maize. Africa Crop Science Journal 5: 1–8. Almeida Rios, S. de, Dias Paes, M. C., Cardoso, W. S., Borém, A., and Teixeira, F. F., 2014. Color of Corn Grains and Carotenoid Profile of Importance for Human Health. American Journal of Plant Sciences, 05 (06), 857–862. Annor, B. and Badu-Apraku, B., 2016. Gene action controlling grain yield and other agronomic traits of extra-early quality protein maize under stress and non-stress conditions. Euphytica, 212 (2), 213–228. Atera, E. A., Ishii, T., Onyango, J. C., Itoh, K., and Azuma, T., 2013. Striga Infestation in Kenya: Status, Distribution and Management Options. Sustainable Agriculture Research, 2 (2), 99. Atera, E. A., Itoh, K., Azuma, T., and Ishii, T., 2012. Farmers’ perspectives on the biotic constraint of Striga hermonthica and its control in western Kenya. Weed biology and management, 12 (1), 53–62. Atera, E. A., Itoh, K., and Onyango, J. C., 2011. Evaluation of ecologies and severity of Striga weed on rice in sub-Saharan Africa. Agriculture and biology journal of North America, 2 (5), 752–760. Atlin GN, Palacios N, Babu R, Das B, Twumasi-Afriyie S, Friesen DK, De Groote H, Vivek B, Pixley KV., 2011. Quality protein maize: progress and prospects. Plant Breeding Reviews doi: 10.1002/9780470880579.ch3 Azmach, G., Gedil, M., Menkir, A., and Spillane, C., 2013. Marker-trait association analysis of functional gene markers for provitamin A levels across diverse tropical yellow maize inbred lines. BMC plant biology, 13 (1), 227. Babu, R., Rojas, N. P., Gao, S., Yan, J., and Pixley, K., 2013. Validation of the effects of molecular marker polymorphisms in LcyE and CrtRB1 on provitamin A concentrations for 26 tropical maize populations. Theoretical and Applied Genetics, 126 (2), 389–399. 148 University of Ghana http://ugspace.ug.edu.gh Badu-Apraku, B., Akinwale, R. O., Ajala, S. O., Menkir, A., Fakorede, M. A. B., and Oyekunle, M., 2011. Relationships among traits of tropical early maize cultivars in contrasting environments. Agronomy Journal, 103 (3), 717–729. Badu-Apraku, B. and Fakorede, M. A. B., 2017. Advances in Genetic Enhancement of Early and Extra-Early Maize for Sub-Saharan Africa [online]. Cham: Springer International Publishing. Available from: http://link.springer.com/10.1007/978-3-319-64852-1 [Accessed 21 Sep 2018]. Badu-Apraku, B., Fakorede, M. A. B., Gedil, M., Talabi, A. O., Annor, B., Oyekunle, M., Akinwale, R. O., Fasanmade, T. Y., Akaogu, I. C., and Aderounmu, M., 2015. Heterotic responses among crosses of IITA and CIMMYT early white maize inbred lines under multiple stress environments. Euphytica, 206 (1), 245–262. Badu-Apraku, B., Fakorede, M. A. B., and Lum, A. F., 2007. Evaluation of experimental varieties from recurrent selection for Striga resistance in two extra-early maize populations in the savannas of West and Central Africa. Experimental Agriculture, 43 (2), 183–200. Badu-Apraku, B., Fakorede, M. A., Oyekunle, M., and Akinwale, R. O., 2012a. Selection of extra- early maize inbreds under low N and drought at flowering and grain-filling for hybrid production. Maydica, 56 (2). Badu-Apraku, B., Fakorede, M. A., Oyekunle, M., and Akinwale, R. O., 2012b. Selection of extra- early maize inbreds under low N and drought at flowering and grain-filling for hybrid production. Maydica [online], 56 (2). Available from: http://cra- journals.cineca.it/index.php/maydica/article/view/686 [Accessed 15 Feb 2017]. Badu-Apraku, B. and Oyekunle, M., 2012. Genetic analysis of grain yield and other traits of extra- early yellow maize inbreds and hybrid performance under contrasting environments. Field Crops Research, 129, 99–110. Badu-Apraku, B., Oyekunle, M., Akinwale, R. O., and Aderounmu, M., 2013. Combining Ability and Genetic Diversity of Extra-Early White Maize Inbreds under Stress and Nonstress Environments. Crop Science, 53 (1), 9. Badu-Apraku, B., Oyekunle, M., Akinwale, R. O., and Lum, A. F., 2011. Combining ability of early-maturing white maize inbreds under stress and nonstress environments. Agronomy journal, 103 (2), 544–557. 149 University of Ghana http://ugspace.ug.edu.gh Badu-Apraku, B., Oyekunle, M., Fakorede, M. A. B., Vroh, I., O Akinwale, R., and Aderounmu, M., 2013. Combining ability, heterotic patterns and genetic diversity of extra-early yellow inbreds under contrasting environments. Euphytica, 192 (3), 413–433. Baker, R. J., 1978. Issues in diallel analysis. Crop Science, 18 (4), 533–536. Barker, D. H., Stark, L. R., Zimpfer, J. F., Mcletchie, N. D., and Smith, S. D., 2005. Evidence of drought-induced stress on biotic crust moss in the Mojave Desert. Plant, Cell & Environment, 28 (7), 939–947. Benchimol, L. L., de Souza Jr, C. L., Garcia, A. A. F., Kono, P. M. S., Mangolin, C. A., Barbosa, A. M. M., Coelho, A. S. G., and De Souza, A. P., 2000. Genetic diversity in tropical maize inbred lines: heterotic group assignment and hybrid performance determined by RFLP markers. Plant Breeding, 119 (6), 491–496. Berner, D. K., Kling, J. G., and Singh, B. B., 1995. Striga research and control. A perspective from Africa. Plant Disease, 79 (7), 652–660. Bertan, I., Carvalho, F. de, and Oliveira, A. de, 2007. Parental selection strategies in plant breeding programs. Journal of Crop Science and Biotechnology, 10 (4), 211–222. Betrán, F. J., Ribaut, J. M., Beck, D., and de León, D. G., 2003. Genetic Diversity, Specific Combining Ability, and Heterosis in Tropical Maize under Stress and Nonstress Environments. Crop Science, 43 (3), 797–806. Bolaños, J. and Edmeades, G. O., 1996. The importance of the anthesis-silking interval in breeding for drought tolerance in tropical maize. Field Crops Research, 48 (1), 65–80. Bouis, H. E., Hotz, C., McClafferty, B., Meenakshi, J. V., and Pfeiffer, W. H., 2011. Biofortification: a new tool to reduce micronutrient malnutrition. Food and nutrition bulletin, 32 (1_suppl1), S31–S40. Bradbury, P. J., Zhang, Z., Kroon, D. E., Casstevens, T. M., Ramdoss, Y., and Buckler, E. S., 2007. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics, 23 (19), 2633–2635. Bromley, C. M., Van Vleck, L. D., Johnson, B. E., and Smith, O. S., 2000. Estimation of Genetic Variance in Corn from F 1 Performance with and without Pedigree Relationships among Inbred Lines. Crop Science, 40 (3), 651–655. Burt, A. J., Grainger, C. M., Shelp, B. J., and Lee, E. A., 2011. Heterosis for carotenoid concentration and profile in maize hybrids. Genome, 54 (12), 993–1004. 150 University of Ghana http://ugspace.ug.edu.gh Butler, D. G., Cullis, B. R., Gilmour, A. R., and Gogel, B. J., 2009. ASReml-R reference manual. The State of Queensland, Department of Primary Industries and Fisheries, Brisbane. Chander, S., Guo, Y. Q., Yang, X. H., Zhang, J., Lu, X. Q., Yan, J. B., Song, T. M., Rocheford, T. R., and Li, J. S., 2008. Using molecular markers to identify two major loci controlling carotenoid contents in maize grain. Theoretical and Applied Genetics, 116 (2), 223–233. Chandler, K., Lipka, A. E., Owens, B. F., Li, H., Buckler, E. S., Rocheford, T., and Gore, M. A., 2013. Genetic Analysis of Visually Scored Orange Kernel Color in Maize. Crop Science, 53 (1), 189. Chapman, S. C. and Edmeades, G. O., 1999. Selection improves drought tolerance in tropical maize populations: II. Direct and correlated responses among secondary traits. Crop Science, 39 (5), 1315–1324. Charnikhova, T. V., Gaus, K., Lumbroso, A., Sanders, M., Vincken, J.-P., De Mesmaeker, A., Ruyter-Spira, C. P., Screpanti, C., and Bouwmeester, H. J., 2017. Zealactones. Novel natural strigolactones from maize. Phytochemistry, 137, 123–131. Chen, J., Zavala, C., Ortega, N., Petroli, C., Franco, J., Burgueño, J., Costich, D. E., and Hearne, S. J., 2016. The Development of Quality Control Genotyping Approaches: A Case Study Using Elite Maize Lines. PLOS ONE, 11 (6), e0157236. Cruz, C. D., 2016. Genes Software-extended and integrated with the R, Matlab and Selegen. Acta Scientiarum. Agronomy, 38 (4), 547–552. Cullis, B. R. and Gleeson, A. C., 1991. Spatial analysis of field experiments-an extension to two dimensions. Biometrics, 1449–1460. Dao, A., Sanou, J., Mitchell, S. E., Gracen, V., and Danquah, E. Y., 2014. Genetic diversity among INERA maize inbred lines with single nucleotide polymorphism (SNP) markers and their relationship with CIMMYT, IITA, and temperate lines. BMC genetics, 15 (1), 127. De Groote, H., Wangare, L., Kanampiu, F., Odendo, M., Diallo, A., Karaya, H., and Friesen, D., 2008. The potential of a herbicide resistant maize technology for Striga control in Africa. Agricultural Systems, 97 (1–2), 83–94. Denmead, O. T. and Shaw, R. H., 1960. The Effects of Soil Moisture Stress at Different Stages of Growth on the Development and Yield of Corn 1. Agronomy Journal, 52 (5), 272–274. Dörr, I., 1997. How Striga parasitizes its host: a TEM and SEM study. Annals of Botany, 79 (5), 463–472. 151 University of Ghana http://ugspace.ug.edu.gh Edmeades, G., Bänziger, M., Campos, H., and Schussler, J., 2006. Improving tolerance to abiotic stresses in staple crops: a random or planned Process? In: Plant Breeding: The Arnel R. Hallauer International Symposium. Wiley Online Library, 293–309. Edmeades, G. O., n.d. Progress in Achieving and Delivering Drought Tolerance in Maize - An Update, 44. Edmeades, G. O., Bänziger, M., Elings, A., Chapman, S. C., and Ribaut, J.-M., 1997. Recent advances in breeding for drought tolerance in maize. In: Applications of systems approaches at the field level. Springer, 63–78. Edmeades, G. O., Bola n ˜ os, J., Chapman, S. C., Lafitte, H. R., and B a ¨ nziger, M., 1999. Selection Improves Drought Tolerance in Tropical Maize Populations: I. Gains in Biomass, Grain Yield, and Harvest Index. Crop Science, 39 (5), 1306–1315. Edmeades, G. O., Trevisan, W., Prasanna, B. M., and Campos, H., 2017. Tropical Maize (Zea mays L.). In: Genetic Improvement of Tropical Crops [online]. Cham: Springer International Publishing, 57–109. Available from: http://link.springer.com/10.1007/978-3- 319-59819-2_3 [Accessed 2 Dec 2018]. Egesel, C. O., Wong, J. C., Lambert, R. J., and Rocheford, T. R., 2003a. Combining Ability of Maize Inbreds for Carotenoids and Tocopherols. Crop science, 43, 6. Egesel, C. O., Wong, J. C., Lambert, R. J., and Rocheford, T. R., 2003b. Gene dosage effects on carotenoid concentration in maize grain. Maydica, 48 (3), 183–190. Ejeta, G. and Gressel, J., 2007. Integrating new technologies for striga control: towards ending the withch-hunt. Elshire, R. J., Glaubitz, J. C., Sun, Q., Poland, J. A., Kawamoto, K., Buckler, E. S., and Mitchell, S. E., 2011. A robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. PloS one, 6 (5), e19379. Ertiro, B. T., Semagn, K., Das, B., Olsen, M., Labuschagne, M., Worku, M., Wegary, D., Azmach, G., Ogugo, V., Keno, T., Abebe, B., Chibsa, T., and Menkir, A., 2017. Genetic variation and population structure of maize inbred lines adapted to the mid-altitude sub-humid maize agro-ecology of Ethiopia using single nucleotide polymorphic (SNP) markers. BMC Genomics [online], 18 (1). Available from: http://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-017-4173-9 [Accessed 29 Nov 2017]. 152 University of Ghana http://ugspace.ug.edu.gh Evanno, G., Regnaut, S., and Goudet, J., 2005. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular ecology, 14 (8), 2611– 2620. Falush, D., Stephens, M., and Pritchard, J. K., 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics, 164 (4), 1567–1587. Fan, X. M., Chen, H. M., Tan, J., Xu, C. X., Zhang, Y. M., Huang, Y. X., and Kang, M. S., 2008. A New Maize Heterotic Pattern between Temperate and Tropical Germplasms. Agronomy Journal, 100 (4), 917. Farooq, M., Wahid, A., Kobayashi, N., Fujita, D., and Basra, S. M. A., 2009. Plant drought stress: effects, mechanisms and management. In: Sustainable agriculture. Springer, 153–188. Flint-Garcia, S. A., Buckler, E. S., Tiffin, P., Ersoz, E., and Springer, N. M., 2009. Heterosis is prevalent for multiple traits in diverse maize germplasm. PloS one, 4 (10), e7433. Fraser, P. D. and Bramley, P. M., 2004. The biosynthesis and nutritional uses of carotenoids. Progress in lipid research, 43 (3), 228–265. Gallagher, C. E., Matthews, P. D., Li, F., and Wurtzel, E. T., 2004. Gene duplication in the carotenoid biosynthetic pathway preceded evolution of the grasses. Plant Physiology, 135 (3), 1776–1783. Gardner, C. O. and Eberhart, S. A., 1966. Analysis and Interpretation of the Variety Cross Diallel and Related Populations. Biometrics, 22 (3), 439. Gethi, J. G. and Smith, M. E., 2004. Genetic Responses of Single Crosses of Maize to Striga hermonthica (Del.) Benth. and Striga asiatica (L.) Kuntze. Crop Science, 44 (6), 2068– 2077. Gilmour, A. R., Gogel, B. J., Cullis, B. R., Welham, Sj., and Thompson, R., 2015. ASReml user guide release 4.1 structural specification. Hemel hempstead: VSN international ltd. Gobena, D., Shimels, M., Rich, P. J., Ruyter-Spira, C., Bouwmeester, H., Kanuganti, S., Mengiste, T., and Ejeta, G., 2017. Mutation in sorghum LOW GERMINATION STIMULANT 1 alters strigolactones and causes Striga resistance. Proceedings of the National Academy of Sciences, 114 (17), 4471–4476. 153 University of Ghana http://ugspace.ug.edu.gh Graham, R. D., Welch, R. M., and Bouis, H. E., 2001. Addressing micronutrient malnutrition through enhancing the nutritional quality of staple foods: principles, perspectives and knowledge gaps. Halilu, A. D., Ado, S. G., Aba, D. A., and Usman, I. S., 2016. Genetics of carotenoids for provitamin A biofortification in tropical-adapted maize. The Crop Journal, 4 (4), 313–322. Hallauer, A. R., 1990. Methods used in developing maize inbreds. Maydica, 35 (1), 1–16. Harjes, C. E., Rocheford, T. R., Bai, L., Brutnell, T. P., Kandianis, C. B., Sowinski, S. G., Stapleton, A. E., Vallabhaneni, R., Williams, M., Wurtzel, E. T., Yan, J., and Buckler, E. S., 2008. Natural Genetic Variation in Lycopene Epsilon Cyclase Tapped for Maize Biofortification. Science, 319 (5861), 330–333. HarvestPlus, 2017. Progress update: Crop development of biofortified staple food crops under HarvestPlus. African Journal of Food, Agriculture, Nutrition and Development, 17 (02), 11905–11935. Hayman, B. I., 1954. The analysis of variance of diallel tables. Biometrics, 10 (2), 235–244. Hung, H.-Y., Browne, C., Guill, K., Coles, N., Eller, M., Garcia, A., Lepak, N., Melia-Hancock, S., Oropeza-Rosas, M., Salvo, S., Upadyayula, N., Buckler, E. S., Flint-Garcia, S., McMullen, M. D., Rocheford, T. R., and Holland, J. B., 2012. The relationship between parental genetic or phenotypic divergence and progeny variation in the maize nested association mapping population. Heredity, 108 (5), 490–499. Hung, H.-Y. and Holland, J. B., 2012. Diallel analysis of resistance to Fusarium ear rot and fumonisin contamination in maize. Crop Science, 52 (5), 2173–2181. Isik, F., 2009. Analysis of Diallel Mating Designs, 34. Johnson, I. J. and Miller, E. S., 1938. Variation in carotenoid pigment concentration among inbred and crossbred strains of corn. Cereal Chemistry, 15, 345–330. Kamali, B., Abbaspour, K. C., Wehrli, B., and Yang, H., 2018. Drought vulnerability assessment of maize in Sub-Saharan Africa: Insights from physical and social perspectives. Global and Planetary Change, 162, 266–274. Kandianis, C., 2010. Genetic dissection of carotenoid concentration and compositional traits in maize grain. [online]. University of Illinois at Urbana-Champaign. Available from: http://www.ideals.illinois.edu/handle/2142/14699 [Accessed 11 Mar 2016]. 154 University of Ghana http://ugspace.ug.edu.gh Kassam, A. H., Kowal, J., Dagg, M., and Harrison, M. N., 1975. Maize in West Africa: and its potential in savanna areas. World Crops (UK). Kassam, A. H. and Kowal, J. M., 1973. Productivity of Crops in the Savannah and Rain Forest Zone in Nigeria [by]. Ahmadu Bello University. Kim, S. K., 1991. Breeding for tolerance and general resistance in maize: a novel approach to combating Striga in Africa. In: Improving Striga management in Africa: 2nd general workshop of the Pan-African Striga Network (PASCON), Nairobi, Kenya. 168–176. Kim, S. K. and Winslow, M. D., 1991. Progress in breeding maize for Striga tolerance/resistance at IITA. In: 5. International Symposium of Parasitic Weeds, Nairobi (Kenya), 24-30 Jun 1991. CIMMYT. Kim, S. K., 1994. Genetics of maize tolerance of Striga hermonthica. Crop Science, 34 (4), 900– 907. Li, F., Vallabhaneni, R., and Wurtzel, E. T., 2008. PSY3, a new member of the phytoene synthase gene family conserved in the Poaceae and regulator of abiotic stress-induced root carotenogenesis. Plant Physiology, 146 (3), 1333–1345. Liu, K. and Muse, S. V., 2005. PowerMarker: an integrated analysis environment for genetic marker analysis. Bioinformatics, 21 (9), 2128–2129. Lobell, D. B., Bänziger, M., Magorokosho, C., and Vivek, B., 2011. Nonlinear heat effects on African maize as evidenced by historical yield trials. Nature climate change, 1 (1), 42. Lu, Y., Yan, J., Guimarães, C. T., Taba, S., Hao, Z., Gao, S., Chen, S., Li, J., Zhang, S., Vivek, B. S., Magorokosho, C., Mugo, S., Makumbi, D., Parentoni, S. N., Shah, T., Rong, T., Crouch, J. H., and Xu, Y., 2009. Molecular characterization of global maize breeding germplasm based on genome-wide single nucleotide polymorphisms. Theoretical and Applied Genetics, 120 (1), 93–115. Ludlow, M. M. and Muchow, R. C., 1990. A critical evaluation of traits for improving crop yields in water-limited environments. In: Advances in agronomy. Elsevier, 107–153. Makumbi, D., Alvarado, G., Crossa, J., and Burgueño, J., 2018. SASHAYDIALL: A SAS Program for Hayman’s Diallel Analysis. Crop Science, 58 (4), 1605. Makumbi, D., Betrán, J. F., Bänziger, M., and Ribaut, J.-M., 2011. Combining ability, heterosis and genetic diversity in tropical maize (Zea mays L.) under stress and non-stress conditions. Euphytica, 180 (2), 143–162. 155 University of Ghana http://ugspace.ug.edu.gh Mbuvi, D. A., Masiga, C. W., Kuria, E. K., Masanga, J., Wamalwa, M., Mohamed, A., Odeny, D., Hamza, N., Timko, M. P., and Runo, S. M., 2017. Novel sources of witchweed (Striga) resistance from wild sorghum accessions. Frontiers in Plant Science, 8, 116. Meenakshi, J. V., Banerji, A., Manyong, V., Tomlins, K., Mittal, N., and Hamukwala, P., 2012. Using a discrete choice experiment to elicit the demand for a nutritious food: Willingness- to-pay for orange maize in rural Zambia. Journal of Health Economics, 31 (1), 62–71. Menkir, A., 2006. Assessment of reactions of diverse maize inbred lines to Striga hermonthica (Del.) Benth. Plant Breeding, 125 (2), 131–139. Menkir, A., Gedil, M., Tanumihardjo, S., Adepoju, A., and Bossey, B., 2014. Carotenoid accumulation and agronomic performance of maize hybrids involving parental combinations from different marker-based groups. Food Chemistry, 148, 131–137. Menkir, A. and Kling, J. G., 2007. Response to recurrent selection for resistance to Striga hermonthica (Del.) Benth in a tropical maize population. Crop Science, 47 (2), 674–682. Menkir, A., Liu, W., White, W. S., Maziya-Dixon, B., and Rocheford, T., 2008. Carotenoid diversity in tropical-adapted yellow maize inbred lines. Food Chemistry, 109 (3), 521–529. Menkir, A., Maziya-Dixon, B., Mengesha, W., Rocheford, T., and Alamu, E. O., 2017. Accruing genetic gain in pro-vitamin A enrichment from harnessing diverse maize germplasm. Euphytica [online], 213 (5). Available from: http://link.springer.com/10.1007/s10681-017- 1890-8 [Accessed 19 Jun 2018]. Meseka, S. K., Menkir, A., Ibrahim, A. E. S., and Ajala, S. O., 2006. Genetic analysis of performance of maize inbred lines selected for tolerance to drought under low nitrogen. Maydica, 51 (3), 487. Messias, R. da S., Galli, V., Silva, S. D. D. A. E., Schirmer, M. A., and Rombaldi, C. V., 2015. Micronutrient and Functional Compounds Biofortification of Maize Grains. Critical Reviews in Food Science and Nutrition, 55 (1), 123–139. Mhike, X., Lungu, D. M., and Vivek, B., 2011. Combining ability studies amongst AREX and CIMMYT maize (Zea mays L.) inbred lines under stress and non-stress conditions. African Journal of Agricultural Research, 6 (8), 1952–1957. Modanesi, S., Massari, C., Camici, S., Brocca, L., and Amarnath, G., 2018. ESA-CCI soil moisture for agricultural drought: validation by using crop data. Plinius Conference Abstracts, Plinius16-52-2. 156 University of Ghana http://ugspace.ug.edu.gh Mohamed, K. I., Papes, M., Williams, R., Benz, B. W., and Peterson, A. T., 2006. Global invasive potential of 10 parasitic witchweeds and related Orobanchaceae. AMBIO: A Journal of the Human Environment, 35 (6), 281–288. Mora, J. O., 2003. Proposed vitamin A fortification levels. The Journal of Nutrition, 133 (9), 2990S–2993S. Musyoki, M. K., Cadisch, G., Enowashu, E., Zimmermann, J., Muema, E., Beed, F., and Rasche, F., 2015. Promoting effect of Fusarium oxysporum [f. sp. strigae] on abundance of nitrifying prokaryotes in a maize rhizosphere across soil types. Biological Control, 83, 37– 45. Muzhingi, T., Langyintuo, A. S., Malaba, L. C., and Banziger, M., 2008. Consumer acceptability of yellow maize products in Zimbabwe. Food Policy, 33 (4), 352–361. NeSmith, D. S. and Ritchie, J. T., 1992. Effects of soil water-deficits during tassel emergence on development and yield component of maize (Zea mays). Field Crops Research, 28 (3), 251–256. Ortiz-Monasterio, J. I., Palacios-Rojas, N., Meng, E., Pixley, K., Trethowan, R., and Pena, R. J., 2007. Enhancing the mineral and vitamin content of wheat and maize through plant breeding. Journal of Cereal Science, 46 (3), 293–307. Owens, B. F., Lipka, A. E., Magallanes-Lundback, M., Tiede, T., Diepenbrock, C. H., Kandianis, C. B., Kim, E., Cepela, J., Mateos-Hernandez, M., Buell, C. R., Buckler, E. S., DellaPenna, D., Gore, M. A., and Rocheford, T., 2014a. A Foundation for Provitamin A Biofortification of Maize: Genome-Wide Association and Genomic Prediction Models of Carotenoid Levels. Genetics, 198 (4), 1699–1716. Owens, B. F., Lipka, A. E., Magallanes-Lundback, M., Tiede, T., Diepenbrock, C. H., Kandianis, C. B., Kim, E., Cepela, J., Mateos-Hernandez, M., Buell, C. R., Buckler, E. S., DellaPenna, D., Gore, M. A., and Rocheford, T., 2014b. A Foundation for Provitamin A Biofortification of Maize: Genome-Wide Association and Genomic Prediction Models of Carotenoid Levels. Genetics, 198 (4), 1699–1716. Oyekunle, M. and Badu-Apraku, B., 2014. Genetic Analysis of Grain Yield and Other Traits of Early-Maturing Maize Inbreds under Drought and Well-Watered Conditions. Journal of Agronomy and Crop Science, 200 (2), 92–107. 157 University of Ghana http://ugspace.ug.edu.gh Palaisa, K. A., Morgante, M., Williams, M., and Rafalski, A., 2003. Contrasting effects of selection on sequence diversity and linkage disequilibrium at two phytoene synthase loci. The Plant Cell, 15 (8), 1795–1806. Parker, C. and Riches, C. R., 1993. Parasitic weeds of the world: biology and control. CAB International. Passioura, J.B. Grain yield, harvest index, and water use of wheat. Journal of the Australian Institute of Agricultural Science 43: 117-120. 1977. Peakall, R. O. D. and Smouse, P. E., 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular ecology notes, 6 (1), 288–295. Peakall, R. and Smouse, P. E., 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research: an update. Bioinformatics 28, 2537e2539. Piepho, H. P., Möhring, J., Melchinger, A. E., and Büchse, A., 2008. BLUP for phenotypic selection in plant breeding and variety testing. Euphytica, 161 (1–2), 209–228. Pillay, K., Derera, J., Siwela, M., and Veldman, F. J., 2011. Consumer acceptance of yellow, provitamin A-biofortified maize in KwaZulu-Natal. South African Journal of Clinical Nutrition, 24 (4), 186–191. Pritchard, J. K., Stephens, M., and Donnelly, P., 2000. Inference of population structure using multilocus genotype data. Genetics, 155 (2), 945–959. Pswarayi, A. and Vivek, B. S., 2008. Combining ability amongst CIMMYT’s early maturing maize (Zea mays L.) germplasm under stress and non-stress conditions and identification of testers. Euphytica, 162 (3), 353–362. Reif, J. C., Melchinger, A. E., Xia, X. C., Warburton, M. L., Hoisington, D. A., Vasal, S. K., Srinivasan, G., Bohn, M., and Frisch, M., 2003. Genetic distance based on simple sequence repeats and heterosis in tropical maize populations. Crop Science, 43 (4), 1275–1282. Rice, A. L., West Jr, K. P., and Black, R. E., 2004. Vitamin A deficiency. Comparative quantification of health risks: global and regional burden of disease attributes to selected major risk factors. Geneva: World Health Organization, 211–56. Roger, J. S., 1972. Measure of genetic similarity and genetic distance. Studies in genetics VII. University of Texas publication, 7213, 145–153. Rosen, M. A. and Scott, D. S., 1992. Exergy analysis of hydrogen production from heat and water by electrolysis. International Journal of Hydrogen Energy, 17 (3), 199–204. 158 University of Ghana http://ugspace.ug.edu.gh Safa, S. B., Jones, B. M. G., and Musselman, L. J., 1984. Mechanisms favouring outbreeding in Striga hermonthica [Scrophulariaceae]. New Phytologist, 96 (2), 299–305. Sansaloni, C., Petroli, C., Jaccoud, D., Carling, J., Detering, F., Grattapaglia, D., and Kilian, A., 2011. Diversity Arrays Technology (DArT) and next-generation sequencing combined: genome-wide, high throughput, highly informative genotyping for molecular breeding of Eucalyptus. In: BMC proceedings. BioMed Central, P54. Sauerborn, J., 1991. The economic importance of the phytoparasites Orobanche and Striga. In: Proceedings of the 5th international symposium of parasitic weeds, Nairobi, Kenya, 24-30 June 1991. CIMMYT (International Maize and Wheat Improvement Center), 137–143. Schmidt-Vogt, H., Dyrenkov, S. A., Evers, F. H., Gussone, H. A., Man’ko, J. I., Pardé, J., and Thomasius, H., 1991. The spruce. Volume II/3. Silviculture, ecosystems, virgin forests, commercial forests, nutrition, fertilizer treatment, prospects. Verlag Paul Parey. Schussler, J. R. and Westgate, M. E., 1995. Assimilate flux determines kernel set at low water potential in maize. Crop Science, 35 (4), 1074–1080. Semagn, K., Beyene, Y., Makumbi, D., Mugo, S., Prasanna, B. M., Magorokosho, C., and Atlin, G., 2012. Quality control genotyping for assessment of genetic identity and purity in diverse tropical maize inbred lines. Theoretical and Applied Genetics, 125 (7), 1487–1501. Sommer, A., 1995. Vitamin A deficiency and its consequences: a field guide to detection and control. World Health Organization, Geneva. 88 pages. Spallek, T., Mutuku, M., and Shirasu, K., 2013. The genus Striga : a witch profile: Profile of Striga , the witchweed. Molecular Plant Pathology, 14 (9), 861–869. Stevens, G. A., Bennett, J. E., Hennocq, Q., Lu, Y., De-Regil, L. M., Rogers, L., Danaei, G., Li, G., White, R. A., Flaxman, S. R., Oehrle, S.-P., Finucane, M. M., Guerrero, R., Bhutta, Z. A., Then-Paulino, A., Fawzi, W., Black, R. E., and Ezzati, M., 2015. Trends and mortality effects of vitamin A deficiency in children in 138 low-income and middle-income countries between 1991 and 2013: a pooled analysis of population-based surveys. The Lancet Global Health, 3 (9), e528–e536. Stevens, R. and Winter-Nelson, A., 2008. Consumer acceptance of provitamin A-biofortified maize in Maputo, Mozambique. Food Policy, 33 (4), 341–351. Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., 2013. IPCC, 2013: Climate Change 2013: The Physical 159 University of Ghana http://ugspace.ug.edu.gh Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, 1535 pp. Cambridge Univ. Press, Cambridge, UK, and New York. Talabi, A. O., Badu-Apraku, B., and Fakorede, M. A. B., 2017. Genetic Variances and Relationship among Traits of an Early Maturing Maize Population under Drought-stress and Low Nitrogen Environments. Crop Science, 57 (2), 681. Tanumihardjo, S. A., 2011. Vitamin A: biomarkers of nutrition for development. The American Journal of Clinical Nutrition, 94 (2), 658S-665S. Vargas, M., Combs, E., Alvarado, G., Atlin, G., Mathews, K., and Crossa, J., 2013. META: A Suite of SAS Programs to Analyze Multienvironment Breeding Trials. Agronomy Journal, 105 (1), 11. Venado, R. E., Owens, B. F., Ortiz, D., Lawson, T., Mateos-Hernandez, M., Ferruzzi, M. G., and Rocheford, T. R., 2017. Genetic analysis of provitamin A carotenoid β-cryptoxanthin concentration and relationship with other carotenoids in maize grain (Zea mays L.). Molecular Breeding [online], 37 (10). Available from: http://link.springer.com/10.1007/s11032-017-0723-8 [Accessed 19 Jun 2018]. Wald, J. P., Nohr, D., and Biesalski, H. K., 2018. Rapid and easy carotenoid quantification in Ghanaian starchy staples using RP-HPLC-PDA. Journal of Food Composition and Analysis, 67, 119–127. Warburton, M. L., Setimela, P., Franco, J., Cordova, H., Pixley, K., Bänziger, M., Dreisigacker, S., Bedoya, C., and MacRobert, J., 2010. Toward a Cost-Effective Fingerprinting Methodology to Distinguish Maize Open-Pollinated Varieties. Crop Science, 50 (2), 467– 477. Weber, E. J., 1987. Carotenoids and tocols of corn grain determined by HPLC. Journal of the American Oil Chemists’ Society, 64 (8), 1129–1134. West Jr, K. P., 2002. Extent of vitamin A deficiency among preschool children and women of reproductive age. The Journal of nutrition, 132 (9), 2857S–2866S. West, K. P. and Darnton-Hill, I., 2008. Vitamin A deficiency. In: Nutrition and health in developing countries. Springer, 377–433. 160 University of Ghana http://ugspace.ug.edu.gh Westgate, M. E., Forcella, F., Reicosky, D. C., and Somsen, J., 1997. Rapid canopy closure for maize production in the northern US corn belt: radiation-use efficiency and grain yield. Field Crops Research, 49 (2–3), 249–258. World Health Organization (WHO), 2009. Global prevalence of vitamin A deficiency in populations at risk 1995-2005: WHO global database on vitamin A deficiency. [online]. Available from: http://apps.who.int/iris/handle/10665/44110 [Accessed 15 Feb 2017]. Wu, Y., San Vicente, F., Huang, K., Dhliwayo, T., Costich, D. E., Semagn, K., Sudha, N., Olsen, M., Prasanna, B. M., Zhang, X., and Babu, R., 2016. Molecular characterization of CIMMYT maize inbred lines with genotyping-by-sequencing SNPs. Theoretical and Applied Genetics, 129 (4), 753–765. Wurtzel, E. T., Cuttriss, A., and Vallabhaneni, R., 2012. Maize Provitamin A Carotenoids, Current Resources, and Future Metabolic Engineering Challenges. Frontiers in Plant Science [online], 3. Available from: http://journal.frontiersin.org/article/10.3389/fpls.2012.00029/abstract [Accessed 19 Jun 2018]. Yan, J., Kandianis, C. B., Harjes, C. E., Bai, L., Kim, E.-H., Yang, X., Skinner, D. J., Fu, Z., Mitchell, S., Li, Q., and others, 2010. Rare genetic variation at Zea mays crtRB1 increases [beta]-carotene in maize grain. Nature Genetics, 42 (4), 322–327. Yan, W., Hunt, L. A., Sheng, Q., and Szlavnics, Z., 2000. Cultivar evaluation and mega- environment investigation based on the GGE biplot. Crop Science, 40 (3), 597–605. Yan, W. and Kang, M. S., 2002. GGE biplot analysis: A graphical tool for breeders, geneticists, and agronomists. CRC press. Yang, X., Gao, S., Xu, S., Zhang, Z., Prasanna, B. M., Li, L., Li, J., and Yan, J., 2011. Characterization of a global germplasm collection and its potential utilization for analysis of complex quantitative traits in maize. Molecular Breeding, 28 (4), 511–526. Yoneyama, K., Arakawa, R., Ishimoto, K., Kim, H. I., Kisugi, T., Xie, X., Nomura, T., Kanampiu, F., Yokota, T., and Ezawa, T., 2015. Difference in Striga-susceptibility is reflected in strigolactone secretion profile, but not in compatibility and host preference in arbuscular mycorrhizal symbiosis in two maize cultivars. New Phytologist, 206 (3), 983–989. 161