A Genome-Wide Association Analysis for Seed Oil and Protein on 13,000 Soybean Accessions

Genotype-phenotype associations within the soybean germplasm collection could provide valuable information on the frequency and distribution of alleles affecting economically important traits. Our objective was to perform genome-wide association study (GWAS) for seed protein and oil on 13,000 soybean accessions genotyped with 52,041 single nucleotide polymorphisms (SNPs). Existing oil and protein data made available by the USDA Germplasm Resources Information Network was used for the analysis. Within the entire collection, linkage disequilibrium (LD) decays from r²=0.50 to r²=0.40 within 250kb, then decays more slowly to r²=0.20 from 500-1000kb. Rates of LD decay were faster for Glycine soja compared to Glycine max, and differences in rate of LD decay were observed between accessions grouped by country of origin and maturity groups. Both admixture and principal component analyses revealed subpopulation structure, with accessions from common countries of origin clustering together. Using a linear mixed model, a total of 14 association signals were identified for protein on chromosomes 6, 13, 15 and 20. Linked clusters of highly significant markers occurred on chromosomes 15 (3.70-3.96Mb) and 20 (29.39-29.98Mb). GWAS for oil detected 15 significantly associated SNP markers on chromosomes 4, 5, 8, 15 and 20, with the strongest association at 3.82Mb on chr 15. Both oil and protein QTL co-located with previously reported QTLs identified through association and linkage mapping. The levels of significance were much higher in this study and narrower candidate regions were defined, highlighting the potential usefulness of extensive phenotypic data on germplasm collections for mining favorable exotic alleles for soybean improvement.