Genomic Prediction and Marker-Assisted Selection for Wheat Resistance to Pre-Harvest Sprouting.

Wheat pre-harvest sprouting (PHS), germination of physiologically matured grains in a spike before harvesting, can cause significant reduction in wheat end-use quality and thus in grain sale price. Evaluation of a large number of wheat lines for PHS resistance in wheat breeding is time and effort consuming, and genetic markers can provide prediction by taking advantages of linkage disequilibrium. In this study, a panel of 185 U.S. cultivars and elite lines was used to compare prediction accuracy between genomic prediction and marker-assisted selection (MAS). This panel was genotyped with the 9K iSelect SNP assays and evaluated for PHS resistance in Manhattan and Hays in both 2013 and 2014. Marker effects were estimated with randomly selected 130 lines from the panel, and PHS resistance prediction was made for the rest of 55 lines in all the four experiments. The ridge regression BLUP (rrBLUP) method showed better prediction accuracies than those of the Bayesian methods on average. However, marker-assisted selection (MAS) made the best prediction using 11 significant SNPs identified in association study with the same population. Therefore, for quantitative traits those are mainly controlled by a few major quantitative trait loci, MAS can be more effective than genomic selection.