Analysis of Genetic Loci Associated with Agronomic Performance in Previously Plant-Variety-Protected Elite Commercial Maize Germplasm
The low cost of genotyping coupled with the availability of high performance computers has enhanced the ability of plant breeders to maintain and potentially increase the rate of genetic gain through genotype-based selection methods. Genotypic selection works well when single-nucleotide polymorphisms (SNPs) are identified which are linked to quantitative trait loci (QTL) that control agronomic performance traits. An effective way to identify SNPs correlated with superior crop performance is a genome-wide association study (GWAS) within a genetically diverse population of former elite commercial inbreds. This study considers a population of 413 maize (Zea mays subsp. mays) inbreds from three groups: (a) 283 inbreds with expired Plant Variety Protection certificates (known hereafter as ex-PVP inbreds); (b) 66 public inbreds, composed of the main historical contributors to contemporary North-American commercial germplasm; and (c) 64 Dow AgroSciences proprietary inbreds. Inbred genotypes were obtained via genotyping-by-sequencing (GBS) and filtered down to 77,314 high-quality SNPs. Population genetic analyses showed that stratification is consistent with the three heterotic group divisions of maize: Stiff Stalk, Non-Stiff Stalk, and Iodent. Fixation index analysis revealed specific SNPs and genomic regions responsible for genetic differences between heterotic groups. Linkage disequilibrium decays to an average of r2 = 0.2 at approximately 2 to 3 Kb. The phenotypic data was collected on hybrids which were produced by crossing each individual inbred to five testers: two Stiff Stalk (SS) testers, two Non-Stiff Stalk (NSS) testers, and one Iodent tester. Testcrosses were grown in standard yield trial format in six environments. Measurements for six performance-related traits were recorded: grain yield, test weight, percent moisture, GDU at 50% pollen shed, plant height, and ear height. Best linear unbiased predictors (BLUPs) were produced for each of the six traits within each of the ve tester groups. In addition, BLUPs were calculated for inbred general combining ability (GCA) by assembling the 12 SS tester environments into one group, the 18 NSS and Iodent environments into a second group, and considering the testers as treatments and multiple testcross groups within each environment as unique replications. Results from 42 unique genome-wide association analyses across all six traits identified 144 significant SNP-trait associations, described in three categories: 20 QTL linked to known genes; 106 QTL linked to described gene models; and 18 QTL with no known genes or gene models. In addition, at 11 of these 144 QTL, a significant SNP-trait association occurred in more than one GWAS (i.e. tester-trait combination). Furthermore, major and minor alleles and their effects on phenotypic traits were identified at each QTL. By identifying alleles associated with superior agronomic performance, these results present an important resource that can be applied to inbred or hybrid development in both public and private breeding programs.
Rocheford, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our