Detection of Genetic Diversity and Footprints of Selection in an East African Wheat (Triticum aestivum L.) Panel Designed for Association Mapping

Wheat productivity in the East African region has perennially suffered losses due to diseases. The recent emergence of the highly virulent race TTKS of stem rust has imposed effort to affirm the genetic diversity of local germplasm and to discover resistance genes/QTLs controlling variability in traits of agronomic and adaptive importance. Towards this goal, we have assembled an association mapping panel of 300 lines, 90% of which are past and present East African cultivars plus a few landraces and breeder lines. A number of historically important Minnesota cultivars are included. Cognizant that the resolution of the panel to dissect intended traits functionally depends on the nature of past successive bottlenecks, migrations, adaptations and human selection, we here report scans of the genetic diversity, signatures of selection and population structure in the present mapping set. The germplasm was genotyped using the 9,000 genomewide SNP assay leading to about 6000 high quality SNPs. Both model-based clustering and discriminant analysis of principal components successfully assigned lines up to 6 distinct subpopulations, with the East African lines forming the largest 2 clusters. Analysis of molecular variance (AMOVA) revealed higher genetic differentiation (88%) among individuals within subpopulations than among subpopulations (12%).  The average polymorphism information content (PIC) among subpopulations ranged from 0.14 to 0.27 while that for estimated expected heterozygosity was 0.17 to 0.34. Both coalescent-based simulation and Bayesian-based tests were effective in estimating Fst and identifying outlier SNPs, suggestive of directional selection. At least 16 such highly significant SNPs and source ESTs are reported.