Genetic Architecture of Resistance to Maize Lethal Necrosis and Maize Streak Virus in Tropical Germplasm

Maize lethal necrosis (MLN) caused by Maize chlorotic mottle virus and Sugarcane mosaic virus, and their synergistic interactions, and maize streak disease caused by Maize streak virus (MSV), are seriously threatening the maize production in east Africa. MLN and MSV resistances were studied by genome-wide association study (GWAS), linkage mapping and genomic selection (GS). Two association mapping panels comprised of 615 diverse maize inbred lines and three bi-parental populations were used in this study. All genotypes were evaluated in multiple environments and fingerprinted with genotyping-by-sequencing. Phenotypic variation for MLN and MSV resistances was significant and heritability was moderately high. GWAS identified 31 significantly (P<5x10^-5) associated SNPs for MLN resistance and these are located within or adjacent to 29 candidate genes involved in plant disease resistance. Linkage analysis revealed three major QTL on chromosome 3 and 6, and a number of minor QTL. The total phenotypic variance explained by each QTL varied from 4 to 35% and together contributed 45 to 47% of variation for MLN resistance. The promising prediction accuracy for within and across mapping panels encourage the application of GS in MLN resistance breeding. GWAS for MSV resistance identified 19 SNPs on chromosome 1 coincident with the fine mapped region of Msv1. Haplotype identification and trend regression identified a three marker haplotype for MSV resistance. Selection for the resistance haplotype of Msv1 in biparental populations has shown a 25% reduction in disease severity. Significant marker-assisted breeding opportunities exist to combine tolerance to these two important diseases as most MLN tolerant lines identified to date do not carry the favorable allele for Msv1.