Métodos AMMI, GGE BIPLOT, REML/BLUP e análise de fatores na estabilidade e estratificação de ambientes de safrinha para seleção de híbridos de milho

Abstract

The present study evaluated the effects of genotype by environment interaction, adaptability, stability, selection of the best maize hybrids, and to stratify the most suitable environments for selection of genotypes adapted to second harvest in Central Brazil, by the methods AMMI (Additive Main Effects and Multiplicative Interaction Analysis), GGEbiplot (Genotype and Genotypes by Environment Interaction), AF (Factor Analysis) and REML/BLUP. Data from the yield evaluations from Dow AgroSciences, for 25 maize hybrids in 13 locations distributed in the States of Mato Grosso do Sul, Mato Grosso, Goiás and Distrito Federal, in two second harvests, 2011 and 2012 were evaluated .Analyses of variance were performed, followed adaptability and stability analyses. GGE biplot and AMMI analyses associated with the predicted genetic values estimated by REML/BLUP methodology allowed the indication of the best hybrids for cultivation for second harvest. The best hybrids for planting for the second harvest in Central Brazil considering the adaptability and stability were: H10, H17, H07, H05 and H09. The REML/BLUP method provides results that are interpreted directly as genotypic values and presented high coincidence with the average yield data; however, it does not identify specific interactions between genotypes x environments such as in AMMI and GGE biplot. The AMMI, GGE biplot and Factor analyses allowed stratification of environments based on altitude. The locations that have the best potential for selection of superior genotypes for second harvest were Jataí - GO, Montividiu - GO, Campo Novo do Parecis - MT and Campo Verde - MT. The GGE biplot method was superior to the models AMMI1, AMMI2 and Factors Analysis, explaining greater proportion of genotype x environment interactions and those associated with the Critical Difference Percentage (CDP) allowed the reduction of the number of environments for future trials in 23%.