Integration of Genomic Selection in the Nebraska Wheat Breeding Program

The Nebraska wheat breeding program has released ~36 varieties to date, and has a vital role in feeding millions of people. The goal of this study is to utilize genomic selection (GS) to make improved selection decisions, and accelerate shuttling of advanced lines as parents in the crossing block, to ultimately develop new and improved wheat cultivars. This work comprised 1,100 entries from four independent F_3:6 nurseries evaluated during 2012-2015 at 27 environments (year x location combinations). Each year the F_3:6 nursery was composed of ~267 entries and three checks, which were grown in a single replicate augmented design at five to eight locations in Nebraska. Yield (kg/h) was analyzed using mixed models accounting for global-trends (experimental design), local-trends (spatial variation within the trial), or both. For 22 of the 27 environments models adjusting for spatial variation provided better fit to the data. SNP calling was performed by combining genotype-by-sequencing data of ~3,300 genotypes sequenced so far. The average accuracy of SNP calling was >95%. Nearly 206,622 SNPs were identified and further stringent filtering provided 41,913 high-quality SNPs across the 1,100 entries. The key questions currently being addressed are: 1) can GS improve selections in a new F_3:6 nursery? 2) can the prediction accuracy be increased by incorporating environment information 3) can multi-year, multi-location evaluation of F_3:6 nurseries assist in selecting entries suitable for advanced multi-location yield trials from the F_3:5 nursery, which has ~1,800 entries and is tested at a single location?