Genomewide prediction to target russet formation in apple

Russet formation in apple (Malus domestica Borkh.) is a superficial skin disorder that detracts from fruit appearance and is likely controlled by many small-effect quantitative trait loci (QTLs). Genomewide prediction has been reported to be an effective breeding approach when targeting highly quantitative traits in apple. Our objective was to investigate the utility of genomewide prediction for russet formation within an apple breeding program. Germplasm included 1,009 unselected offspring from 13 full-sib families derived from 14 breeding parents. 'Honeycrisp' and 'Minneiska', two breeding parents prone to moderate levels of russet, were highly represented. High-quality single nucleotide polymorphism data (947 SNPs) and three years of shoulder and lenticel russet formation data were leveraged in this study. Moderate predictive abilities (r = 0.28−0.35) were observed across training-testing set scenarios and models. In this germplasm, the inclusion of previously detected QTLs as fixed effects in the model did not have significant effects on predictive abilities. Postdiction (retrospective) analyses demonstrated that genomewide predictions and phenotypic observations agreed for 54% of advanced selections. Genomewide prediction is a promising approach when targeting russet formation, a trait that cannot be phenotypically observed in offspring in apple breeding programs until they are past their juvenile phase.