Comparison of methodologies (REML/BLUP vs. GGE Biplot) for soybean selection in the Brazilian Cerrado

Abstract The selection of adapted and stable soybean genotypes for the Brazilian Cerrado is hindered by the complex genotype x environment (G x E) interaction. This study aimed to compare the effectiveness of the REML/BLUP and GGE Biplot. Yield data from 33 genotypes in 44 environments (2019-2020 and 2020-2021 growing seasons) were analyzed. The results indicated a significant G x E interaction and high selective accuracy (0.86), validating the selection precision. Both methodologies proved to be effective and concordant in identifying a group of elite genotypes (commercial checks G3 (BASF), G6 (BASF), G4 (BASF), and the experimental lines G17 (BASF) and G18 (BASF)). REML/BLUP provided the quantitative basis for selection, whereas the GGE Biplot elucidated the formation of mega-environments and the positioning of genotypes relative to the ideotype. It is concluded that the methodologies are complementary, enhancing decision-making and maximizing genetic gains in soybean breeding programs for the Cerrado.