Translating Genomic Research into Cultivar Development in the Nebraska Wheat Program

Extreme weather events during a growing season can reduce phenotypic selection accuracy impeding breeding progress. We present the use of genomic-assisted breeding tools, genomic selection (GS) and marker-assisted selection (MAS) to mitigate this challenge during early and intermediate phases of testing in a breeding program. The datasets used in this study comprise three nurseries, F_3:5 (2015-2016; ~4,000 lines; tested at 1 location), F_3:6 (2012-2016; 1,340 lines; 8-10 locations) and F_3:7 (2013-2016; 228 lines; 6-9 locations) that represent early, intermediate and advanced testing stages, respectively. Genotyping-by-sequencing (GBS) provided ~100,000 high-quality SNPs for GS and MAS efforts. MAS was performed by identifying GBS-derived SNPs in strong linkage disequilibrium with markers tagging known loci for disease resistance, height and grain quality. This was critical in 2015 when grain quality could not be measured due to heavy rain and disease after flowering. We found that using GS in early and intermediate stages to predict grain yield (1) made selections more precise by using the variation from multiple growing-years under different environmental conditions; (2) identified elite lines sooner for crossing; (3) selected best lines after a hail event destroyed half of the single location F_3:5 trial grown in 2016; and (4) identified locations resulting in higher or lower predictive ability. Performance of lines advanced from F_3:6 to F_3:7 indicated GS would have outperformed phenotypic selection during two extreme weather years and emerged ~50% successful in a normal year. In summary, GS and MAS have increased selection accuracy and accelerated cultivar development in challenging weather conditions.