Fine-tuning genomic prediction in peach using additive and non-additive genetic effects for phenology and disease tolerance

International audience

Inglés. Genomic predictions rely on modelling genome-wide marker effects and are especially powerful to predict complex traits. Like for other perennial fruit crops, the implementation of genomic selection could have a big impact on breeding programs by shortening the breeding schemes and/or gaining in selection intensity and reliability. Despite vast available genomic resources in public institutions, both in terms of plant material and genomic information, only few studies have assessed prediction accuracy in peach. In this study, we explore the potential of genomic selection by calculating genomic estimated breeding values for traits with contrasted genetic architecture. As it is known that non-additive effects and directional dominance can strongly affect trait variation, and because these effects can be easily fixed in clonally propagated crops, we also proposed a model comparison from the simplest, purely additive model, to a more complete model accounting for inbreeding, dominance and epistatic effects. Our training set is a core-collection of 192 accessions planted in three sites in different climatic regions of South-East France, and phenotyped for green tip emergence, flowering date (three years of observation), and for leaf curl symptoms observed under zero pesticide (one year of observation). Using linear mixed models, we will present the impact of inbreeding and non-additive effects on trait mean and variance in peach. Genomic prediction will be tested with classical rrBLUP as well as GBLUP approaches using a total of six different models accounting for different combinations of genetic variance and inbreeding parameters. Across-site predictions for phenological traits over years and environments will also be tested. In a translational perspective, it is planned to use the same approach on apricot, a crop closely related to peach, to compare results in the light of crop biology and trait architecture.