Genomic Prediction of Baking Quality in Winter Wheat

Phenotyping of wheat baking quality parameters is expensive and time-consuming, so improving baking quality often have lower priority rather than increasing the yield of new varieties in breeding programs. Using dense DNA markers to predict the baking quality of new lines requires fewer resources than phenotyping and enables selection of lines at early stages of breeding programs.

In this study 672 F₆ winter wheat lines from two breeding cycles of the plant breeding company Nordic Seed were phenotyped for the quality parameters grain protein content, test weight, Zeleny sedimentation, thousand-kernel weight, and falling number. A total of 10,763 SNP markers were used for GWAS to identify SNPs in regions with large effects on quality parameters, and for genomic predictions using GBLUP models to predict the parameters more accurately based on all the SNPs. Accuracy of genomic predictions were evaluated using several kinds of cross-validations.

SNPs significantly associated with Zeleny sedimentation were identified on chromosome 1B, 1D, and 5D. For the remaining traits, SNPs close to the significance threshold were found on several chromosomes. The GWAS indicate that the quality parameters are controlled by many QTLs with small effects, so genomic prediction based on all SNPs appears to be a promising strategy. Correlations between observed and predicted phenotypes corrected for fixed effects ranged from 0.5 for protein to 0.79 for Zeleny sedimentation based on Leave-One-Out cross-validations, showing good potential for developing genomic predictions for quality traits in wheat.