Evaluating Genomic Selection in the First Two Cycles of a Winter Barley Breeding Program

Genomic selection (GS), which takes advantage of inexpensive genome-wide marker data,  may accelerate rates of gain from selection because it allows breeders to evaluate lines in earlier generations and at multiple times per year.  For three breeding cycles, we selected lines that were predicted by marker data to have the best low temperature tolerance, malt extract, height, heading date, and grain yield. The lines were evaluated for these traits as well as Fusarium head blight severity (not under selection) in two or three locations in 2014. Genome-wide markers used to generate genomic predictions also allowed us to assess changes in genotypic variation across selection cycles.  The objectives of this study were to 1) assess accuracy of predictions based on progeny performance, 2) assess rate of gain from GS, and 3) determine the change in genotype frequencies across cycles of selection for markers associated with winter hardiness. Low temperature tolerance improved significantly after two cycles of GS. However, Fusarium head blight severity, heading date, and grain yield did not change significantly. Markers for several loci known to affect winter hardiness moved toward fixation of the favorable genotype after the first cycle of GS and became nearly fixed after the second cycle. This implies that GS can rapidly shift the frequency of large effect genes that have previously been the target of traditional marker assisted selection. This empirical evaluation of GS will inform plant breeders who are considering whether to adopt this breeding strategy.