Applying Genomic Selection to CIMMYT Spring Wheat for End-Use Quality.

Wheat cultivars must pass end-use quality standards to be eligible for varietal release, but breeding for end-use quality is often considered a secondary goal compared to improving yield or disease resistance.  In traditional breeding approaches, bread-making quality cannot be assessed until late in the breeding cycle.  Many materials will be advanced through the breeding pipeline that will not pass end-use standards.  Thus, money, time, and resources are spent on lines which cannot be released.  Genomic selection prediction models will be developed with phenotypes that are normally generated in the current breeding program.  Once model accuracy is determined, genomic selection is applied to lines, and materials can be culled on the basis of quality before expensive multi-location yield and quality tests need to be implemented. CIMMYT advanced lines from the selection cycle 45 and 46 International Bread Wheat Screening Nursery (n=687) were grown over two years under six different environments were used for genomic selection.  High-density genotyping was conducted with genotyping-by-sequencing, and SNPs were emputed.  LSMeans for quality phenotype traits were determined and input into genomic selection models.  Various iterations of training and testing set were performed.  Prediction accuracies for quality traits were recorded.