Modeling Genotype By Environment Interaction with Large Genomic Studies in Oats

Since the advent of agriculture, plant breeding has successfully improved plants for human benefit. Modern plant breeding activities consists in evaluating the genetic merit of lines discerning genetic from environment and noise components. Plant breeding activities could therefore be grouped into three categories: traditional, marker assisted, and genomic selection. Furthermore, most of the relevant traits for agriculture like yield, grain quality and diseases are quantitative traits that are also affected by the environment. Therefore, good phenotyping is crucial in any plant breeding program. This creates two challenges. First, field trials and experimental designs for large number of genotypes should be carefully chosen to reduce spatial heterogeneity and experimental error and to increase heritability. Second, genotype by environment interaction (GEI) is widespread in plants, and affect especially quantitative traits that are of main importance for plant breeding. Mixed models have been used for modeling GEI. Our objective was to present some strategies that are successfully being used in plant breeding programs. We discuss the use of genomic information for genomic selection, and then we discuss how to handle GEI with and without genomic information. Finally, we present some ideas for an oat proposal to address these issues.