Implicações da interação genótipos por ambientes na seleção de progênies de soja

Abstract

Obtaining estimates of genetic and phenotypic parameters are essential for the success of breeding programs. Parameters estimated from only one environment are strongly influenced by the interaction effect, which may lead to errors related to the selection of superior genotypes. The G x E interaction in the soybean crop has been studied mainly using methods to evaluate adaptability and stability. However, in the initial stages of breeding programs, when segregating progenies are available, other tools can be employed to measure the implications of G x E for the classification of the best genotypes. In this context, the achieved heritability, in which the average performance of the superior progenies in different generations is evaluated, is an alternative. The purpose was to study the implication of the G x E interaction (crop years / generation and different environments) in the identification of genetically superior progenies; to obtain estimates of genetic and phenotypic parameters for agronomic traits in soybean progenies; to select genetically superior progenies in order to obtain lins. F3:4 progenies were evaluated in the municipality of Lavras-MG and Itutinga-MG, in the crop season 2016/2017. One line plot of two meters was used, with two replicates, considering the lattice design simple 14 x 14 (194 progenies + 2 parental). F3:5 progenies were evaluated in the summer of 2017/2018 in the municipalities of Lavras-MG, Itutinga-MG and Ijaci-MG. For evaluations, a 10x10 triple lattice (98 progenies + 2 parental) was used, using a three meter line. The traits characters for flowering, full maturity, bottom pod height, the plantheight, lodging score and grain yield were evaluated. Data were analyzed using the mixed model approach. Genetic and phenotypic parameters, expected gain with selection, genetic correlation, correlated response, heritability, and frequency distribution of the adjusted means were estimated. The estimates of the variance components show the existence of variability among the progenies making possible the selection of superior genotypes. The effect of the genotypes x environments interaction was verified for all evaluated traits, and most of them was complex. All evaluated progenies present good agronomic performance, with higher grain yield. From the analysis of genotypic correlation it was possible to observe significant estimates between the traits days to flowering and the full maturity, plant height andlodging score, height and the traits of the 1st legume and lodging score, as well as between the traits insertion of the 1st legume andlodging score. There is influence of the interaction G x E on the estimates of genetic and phenotypic parameters in soybean. Achieved heritability is a tool to study the G x E interaction.