Bread wheat for human consumption makes a great contribution to health and nutrition of the growing global population, but competition for human digestible grains that are fed to livestock has concerning implications for sustainability and food security. Specific quality requirements must be met for wheat crops to be suitable for bread making and these are both genetically and environmentally controlled. In the UK, breeders have historically focussed more on increasing yield rather than quality due to the difficulty in selecting for multiple low throughput and high cost quality traits. This thesis aims to redress this imbalance by taking a quantitative genetics approach to enable enhanced selection for milling and baking quality in collaboration with the DSV UK breeding programme.

A wheat multi-parent advanced generation intercross population was used to investigate the genetic control of multiple wheat quality and micronutrient traits. This analysis identified multiple quantitative trait loci (QTL) with co-locating pleiotropic effects that could explain much of the complementary and antagonistic relationships among these traits. Of note, a QTL that co-located with the awn length inhibitor locus on chromosome 5A was found to consistently increase grain calcium content while not decreasing grain specific weight, despite the established negative correlation between these two traits. Principal component based multi-trait analysis increased the power to detect novel QTL that have effects that contradict overall trait correlations so may be useful to optimise antagonistic trait trade-offs.

Genomic prediction of quality and loaf baking quality traits were then investigated in a panel of released high quality wheat varieties and recent breeding lines. Historical trends in these traits and changes in frequency of QTL alleles identified through genome wide association analysis revealed evidence for breeders’ selection for decreased protein content but increased loaf baking quality for the Chorley Wood Baking Process. QTL identified here have direct application for improvement of quality traits, such as grain specific weight and Hagberg falling number, for which little improvement has been made in recent decades of breeding. However, most QTL identified here could only explain a small proportion of the heritability of most complex quality traits. Genomic selection was shown to be highly applicable for prediction of costly loaf baking quality traits and offer increased prediction accuracy at reduced costs in comparison to phenotypic selection based on early-stage predictive traits.

Stability of wheat quality traits across environmental variation is also an important target for selection. A large dataset of long-term field trials in the UK was analysed with genetic marker and pedigree data to characterise genotypes and weather and soil covariates to characterise environments. Cross validation of untested genotypes in untested years demonstrated that prediction models were able to successfully predict environmental and genotype by environment interaction effects. Predictions into future environments simulated from climate projection models enabled prediction of climate change impacts on UK wheat quality and the potential for breeding to mitigate these impacts.

This project provides several quantitative genetics tools and resources for enhanced selection of wheat milling and baking quality with direct relevance to a UK breeding programme.