Adaptabilidade e Estabilidade de híbridos de milho (Zea mays L.) via modelos mistos e análise gráfica

Abstract

The interaction between hybrids and seasons is one of the obstacles in maize breeding programs. The edaphoclimatic and phytosanitary conditions vary from one season to the next and have a great influence on the phenotypic response of a given genotype at a given location. To explore this type of interaction, the use of sophisticated statistical tools in the study of adaptability and stability are necessary for the accurate choice of superior hybrids based on the productive performance in the same location. In this sense, the present work aimed to use the harmonic mean method relative to the performance of genotypic values (MHPRVG) via REML / BLUP and the Genotype plus Genotype-by-Enviroment (GGE) Biplot model to select superior hybrid genotypes for adaptability and stability based on three harvests in the Lavras-MG region. The experiments were carried out, in alpha-lattice design in the 2018/19 season and in triple lattice in the 2019 and 2019/20 seasons, at the Center for Scientific and Technological Development of the Federal University of Lavras (UFLA). Considering all harvests, 112 hybrids were evaluated in total, five commercial hybrids (controls) and 107 hybrids (genetic treatments) belonging to the UFLA maize breeding program. The following characters were evaluated: grain yield (ton.ha-1), severity of white spot and germination (%). An individual and joint deviance analysis was performed for each character. All characters, in both analyzes, were significant (p <0.05) by the likelihood ratio test (LRT) and the genetic parameters were estimated. In the study of adaptability and stability, hybrids were evaluated by MHPRVG and by graphic analysis via GGE biplot. Both methodologies presented promising and coincident results.