The investigation into genotype-by-environment interaction (G × E) has gained prominence owing to its theoretical and practical importance in understanding the genetic architecture of complex traits. Utilizing data from diverse environments to perform genome-wide association studies (GWAS) provides a convenient means to study G × E.

However, conducting GWAS with multi-environment data is challenging, mainly due to the absence of a systematic framework. In this study, we developed a GWAS pipeline that efficiently tests the main effects of markers across environments, the specific effects of markers in each environment, and the interaction effects between markers and environments. In addition, it is capable of handling both additive and dominance effects.

As an application, the established model was used to dissect the pattern of marker-by-environment interactions of yellow rust (YR) resistance in wheat with a dataset comprised of 5,243 hybrids and their 597 parents from 3 different experimental series and 19 environments.

We found that markers with additive effects showed stronger G × E interaction and explained high proportion of phenotypic variance. Additionally, QTLs with dominant effects were detected across more environments.