Assessing the adaptability and stability of maize hybrids using a Bayesian factor analytic model

Maize (Zea mays L.) is an important crop globally, and obtaining more productive and resistant commercial cultivars is of paramount importance. In this context, adequate analysis of data from multi-environment trials is essential for the accurate modeling of genotype × environment interaction (GEI), thus providing crucial support for decision-making in plant breeding programs. This study uses a Bayesian analytical factorial model (Bayesian factor analytic [BFA]) to analyze the adaptability and stability of grain yield in a collection of 100 maize hybrids evaluated in 14 representative environments of the Southeast region of Brazil. The aim was to highlight and discuss aspects related to the application of the BFA, addressing the advantages and challenges involved. The goal was to explore the interpretations and limitations of the analysis, in order to assist breeders and researchers in the proper use of the employed method. The results allowed us to identify distinct subgroups of genotypes and environments with similar effects, as well as to identify stable genotypes in relation to GEI and to suggest genotype recommendations for specific environments. To achieve this goal, the flexibility of the BFA model was exploited to incorporate inferences to the various parameters, especially bilinear parameters that describe G + GEI in the biplot.