Assessing the stability and adaptability of corn hybrid yield with the Bayesian AMMI model

Abstract

The objective of this work was to evaluate the stability and adaptability of commercial and experimental corn hybrids tested in three locations in Mato Grosso do Sul (MS) in the 2021 and 2022 agricultural years. We identified superior genotypes that can be recommended broadly and/or restrictedly to optimize productivity. The randomized complete block design (RCBD) was used to evaluate the grain yield per plot (ton/ha) with the additive main effects model and multiplicative interaction via Bayesian inference (BAMMI). It was observed from the results that the genotype by environment interaction (G x E) did not behave in a complex way and that the genotypes, in general, do not have significant contributions to the interaction. Furthermore, it was found that stable and productive hybrids can be selected regardless of genetic structure. The G26 genotype, which is the simple commercial hybrid B2702, stood out as one of the best cultivars. However, it was observed that there are experimental hybrids as good as B2702, such as G9, which is a double hybrid. This result is relevant since double hybrid cultivation is more viable in low investment contexts, as occurs in family farming. Finally, the BAMMI model proved to be a versatile tool with flexibility to overcome the limitations inherent to classical and frequentist analyses of linear-bilinear models.