Traditional and Genomic Methods for Improving Fusarium Ear Rot Resistance in Maize

Fusarium ear rot of maize is a problem in maize-growing regions worldwide, impacting both food availability and quality. The deployment of maize hybrids possessing genetic resistance to the disease is an effective strategy for controlling Fusarium ear rot and reducing the prevalence of mycotoxins in livestock feed and human foodstuffs. However, the highly polygenic nature of resistance, in combination with a large environmental component, has made it a challenge to improve resistance levels of adapted maize breeding pools. A key step in improving Fusarium ear rot resistance is the demonstration that resistance alleles from unadapted donor sources can be introgressed into adapted germplasm and confer resistance without having an undue effect on agronomic performance. In this study, I evaluated a set of BC4F3:4 and BC4F4:5 lines derived from backcrosses of the highly resistant donor source GE440 into the elite propriety stiff stalk line FR1064. These lines were evaluated for disease resistance across three years, and topcrosses to NC478 and LH283xLH287 were evaluated for both disease resistance and agronomic potential for three years. Two inbred lines, coded as NC301 and NC303, were identified as having superior disease resistance compared to FR1064 but are not statistically different for many agronomic traits including yield. Based on the improved disease resistance and favorable agronomic characters of NC301 and NC303, the USDA-ARS Plant Science Research Unit and North Carolina State University propose to release these two lines given their usefulness for improving disease resistance in elite temperate breeding pools. Identifying and utilizing novel resistance allele variants is critical in making continued genetic gains in Fusarium ear rot resistance. Traditional QTL mapping approaches have had limited success in identifying useful allele variants as most Fusarium ear rot QTL have small effects and are not consistent between populations. Alternatively, genome-wide association studies (GWAS) offer finer-scale mapping combined with the ability to mine diverse breeding material for novel resistance alleles. I conducted a GWAS of Fusarium ear rot resistance on a commonly-used maize core diversity panel of 279 inbred lines using phenotypic data collected from North Carolina and Galicia, Spain. Although no significantly associated markers were identified in Galicia, three markers significantly associated with improved resistance were identified in the North Carolina data set, each with a relatively small additive effect. Targeted allele selection for these variants may be useful for some improvement of resistance levels in breeding programs. In addition, a large amount of additive background polygenic variance was observed in the analysis, indicating the potential usefulness of genomic selection in follow-up studies. Within the last year, a tremendous amount of genotypic data has become available on the USDA-ARS national maize inbred seed bank, consisting of over 2,800 diverse inbred lines from public and private breeding programs worldwide. I used a set of 200,978 SNP markers to conduct a GWAS of Fusarium ear rot resistance in a sample of 1,689 inbred lines from this collection. Seven SNPs located on three chromosomes were identified as significantly associated with ear rot resistance. The three most significant SNPs were then included as fixed linear effects in a genomic best linear unbiased prediction (G-BLUP) model and compared to a traditional G-BLUP model. Although the three SNPs had small effects, their inclusion in the genomic prediction model significantly improved prediction accuracies from 11% to 15%. These results suggest a relatively simple and straight-forward method for improving genomic selection models by accounting for known GWAS hits in addition to background polygenic variation.