Bayesian GGE model for heteroscedastic multienvironmental trials

The dissection of genotype × environment interaction (GEI) is a crucial aspect the final stages plant breeding pipelines and recommendation cultivars. Linear-bilinear models used to analyze this interaction, such as additive main effects multiplicative (AMMI) plus GEI (GGE), often assume homogeneity residual variances across environments which affects estimates therefore, interpretations conclusions. Our objective was propose GGE model that considers heteroscedasticity using Bayesian inference evaluate its implications in interpretation real simulated data. assuming common variance also fitted for comparison purposes. great flexibility transferred biplots, allowing construction credible regions genotypic environmental scores. on stability adaptability genotypes might change when ignored. When data are used, different patterns correlations between affect representativeness discrimination target environment. modeling allowed clustering into subgroups, with similar GEI. proposed more adequate realistic deal scenarios heterogeneous multienvironment trials, can be useful exploiting