Using Genomewide Selection to Increase the Efficiency of Apple Breeding

Apple (Malus domestica) is an important horticultural crop, both economically and in terms of contributions to human health. Breeders must meet the demands of the consumers by exceeding flavor, texture, and appearance expectations. Selecting for multiple traits adds complexity and requires development and assessment of thousands of seedlings each year. Many of the traits being selected on are complex traits, controlled by many small effect quantitative trait loci (QTL). However, current genetic tests only allow for the selection of a few traits that are controlled by larger effect QTL. Genomewide selection allows breeders to forgo tests of significance and draws on a large, random set of marker data as the basis of selection. This approach is already widely used in maize (Zea mays), but has been more slowly adopted in highly heterozygous crops like apple. In an effort to assess whether accounting for heterozygosity improves accuracy of prediction, models with and without a heterozygous parameter have been used to predict the traits of crispness, firmness, and juiciness. Additionally, analysis of marker-trait associations has been conducted. Marker-trait associations representing major QTL have the potential to be treated as fixed effects in prediction models as a means of increasing prediction accuracy.