KASSAHUN SADESSA BIRATU
GENETIC ANALYSIS OF MAIZE (Zea mays L.) GENOTYPES FOR RESISTANCE TO MAIZE LETHAL NECROSIS DISEASE
ABSTRACT
In sub-Saharan Africa (SSA), maize (Zea mays L.) is a crucial crop for food security and a source of income for smallholder farmers. Maize lethal necrosis (MLN) disease, which can reduce maize yield by up to 100% in environments that are suitable to the pathogens, is a threat to maize production in eastern African countries. The best option for managing MLN disease is the deployment of resistant varieties. Grain yield, diseases resistance, and other relevant agronomic features must be assessed in a wide range of hybrids in order to generate more productive and MLN resistant maize varieties. The breeding efforts aimed at identifying MLN resistant and farmer preferred maize varieties will also benefit from knowledge of the genetic background of breeding materials and identification of quantitative trait loci (QTLs) responsible for MLN resistance. Therefore, the objectives of the present study were to (i) investigate the genomic regions associated with MLN disease resistance and other agronomic traits in a wide array of doubled haploid (DH) populations; (ii) determine the combining ability and performance of testcross hybrids under artificial MLN pressure, managed drought, and optimum management conditions; (iii) determine performances of highland maize hybrids and combining ability of parental inbred lines for agronomic traits and resistance to MLN; (iv) evaluate hybrid performances and efficacy of midaltitude and highland adapted inbred lines converted to MLN resistant versions; (v) determine agronomic performances and grain yield stability of highland maize hybrids across diverse environments. MLN disease pressure, well-watered (WW) and water-stressed (WS) management conditions were used to evaluate a total of 879 DH lines derived from 40 populations. For the genome-wide association study (GWAS), single nucleotide polymorphism (SNP) markers were used to genotype the DH lines. META-R, mixed linear FarmCPU-R package and TASSEL software were used for the data analysis. There were 174 SNP markers found to be associated with MLN resistance, grain yield, and other studied traits under the well-watered (98), water-stressed (54), and MLN (22) management conditions. The discovered SNP markers could be validated, used to develop innovative molecular markers for marker assisted selection, and implemented in the early selection of best elite lines for each targeted trait. In this study, 52 MLN resistant genotypes were identified, among which seven genotypes could be used as resistance donor lines. Eleven DH lines were crossed with 11 single cross testers to generate 115 successful crosses in order to determine the hybrid performance and lines combing ability. Using a 10 x 12 alpha lattice design, all 115 experimental hybrids and five commercial checks were assessed under one managed drought, two MLN artificial infection, and four optimal management situations. Grain yield, agronomic characteristics, and MLN response data were recorded. The analysis of variances and mean values, and combining ability effects were analyzed using META-R and AGD-R statistical software package, respectively. Under all management scenarios, the DH lines L3 and L4 were the best general combiners for grain yield, ear height, and plant height. Lines L6 and L7 were the best general combiners for MLN disease severity, indicating the possibility of developing lines with desirable breeding values for MLN resistance. For grain yield, MLN disease severity, and all other traits, additive gene effects outweighed non-additive gene action, indicating the possibility of improving the studied characteristics by recurrent selection. The best specific combiners for grain yield were the hybrids T8/L10 and T8/L2 under optimal management and managed drought; T4/L6, T5/L7, and T11/L1 under optimal management and MLN infestation; and T7/L9, T5/L10, and T9/L8 under optimal management and controlled drought. Hybrids T9/L4 and T11/L3 had the highest yields under all test environments. The highest yielding hybrids were T10/L3 under both optimal and controlled drought and T9/L1 under MLN infection, showing that these hybrids have the potential to be recommended for commercial production. In order to determine the hybrid performances and the combining ability of the lines, 11 MLN-tolerant highland maize inbred lines were crossed using a diallel mating scheme. Eight optimal and two MLN artificial infestation environments were used to test the resulting 55 hybrids plus five check hybrids. Using 53,538 SNPs, 30 highland MLN resistance maize inbred lines, including those employed in the diallel crosses, were genotyped to determine the genetic relationships and variability among the inbred lines. Grain yield and agronomic data were analyzed using META-R and AGD-R, whereas genotype data were analyzed using DARwin and TASSEL software. High GCA effects for grain yield were seen in lines CELMLN05 and CELMLN06 under optimal management as well as CELMLN01, CELMLN02, CELMLN03, and CELMLN04 under MLN infestation. Similarly, CELMLN02, CELMLN03 and CELMLN04 showed desirable GCA effects for MLN disease severity. Hybrids CELMLN06/CELMLN09 and CELMLN06/CELMLN11 showed higher grain yield under optimum management. Under MLN artificial infestation, the hybrids CELMLN02/CELMLN03, ELMLN01/CELMLN08, CELMLN01/CELMLN03, and CELMLN04/CELMLN11 had high grain yields, whereas the hybrid CELMLN01/CELMLN11 was the best yielding under all management conditions and might be considered and recommended for release. In another experiment, a total of 180 hybrids generated using North Carolina Design II (NC II) crosses of MLN resistant converted highland and mid-altitude adapted inbred lines. Eighty-four hybrids that involve MLN resistant highland inbred lines were evaluated across five locations under optimum management and one location under MLN infestation. Whilst, 96 hybrids generated from MLN resistant converted versions of mid-altitude adapted elite inbred lines were evaluated across eight optimal and one MLN infested environments. Data analysis was performed using the META, TASSEL, and DARwin software. The study identified five highland and three mid-altitude adapted higher yielding and MLN resistant hybrids. Pair comparison of genetic similarity showed that MLN converted versions of 9 highland and 20 mid-altitude adapted lines were 95% genetically similar to the original elite lines, which showed that MLN resistant QTLs were fixed without changing the genetic background of elite lines. The improved lines could be used to constitute higher yielding and MLN disease resistant hybrids. In the last experiment, 28 highland-adapted three-way hybrids with two commercial checks were evaluated under fourteen optimal, four high plant density, and two low soil nitrogen (low N) management strategies to determine the hybrids' agronomic performances and grain yield stability in a range of environments. The data were analyzed using META-R and GEA-R software. The environment Kul13 and Amb21 were the most discriminative and representative for hybrid testing. The best yielding and most stable hybrids found in the test environments were CEH013, CEH022, and CEH009. Hybrids CEH013 and CEH005 were the highest yielders across all management conditions, whereas CEH022 under optimum, CEH017 and CEH009 under high plant density, and CEH001 and CEH010 under low N management, which could be and recommended for release after verification