Population Genomics Related to Adaptation in Elite Oat Germplasm

Six hundred thirty five oat (Avena sativa L.) lines and 4561 single-nucleotide polymorphism (SNP) loci were used to evaluate population structure, linkage disequilibrium (LD), and genotype–phenotype association with heading date. The first five principal components (PCs) accounted for 25.3% of genetic variation. Neither the eigenvalues of the first 25 PCs nor the cross-validation errors from K = 1 to 20 model-based analyses suggested a structured population. However, the PC and K = 2 model-based analyses supported clustering of lines on spring oat vs. southern United States origin, accounting for 16% of genetic variation (p < 0.0001). Single-locus F-statistic (FST) in the highest 1% of the distribution suggested linkage groups that may be differentiated between the two population subgroups. Population structure and kinship-corrected LD of r2 = 0.10 was observed at an average pairwise distance of 0.44 cM (0.71 and 2.64 cM within spring and southern oat, respectively). On most linkage groups LD decay was slower within southern lines than within the spring lines. A notable exception was found on linkage group Mrg28, where LD decay was substantially slower in the spring subpopulation. It is speculated that this may be caused by a heterogeneous translocation event on this chromosome. Association with heading date was most consistent across location-years on linkage groups Mrg02, Mrg12, Mrg13, and Mrg24