Estimativas de parâmetros genéticos e fenotípicos, correlações e diversidade genética em progênies F3 e F4 de soja.

Abstract

In the global market for agricultural commodities, soybean [Glycine max (L.) Merrill] is the main oilseed crop grown in the world. In the 2019/2020 season, Brazil became the world's largest producer and exporter of grains of this species. The study of genetic diversity is a predictive way to choose parents, and obtaining estimates of genetic parameters allows breeders to analyze selection strategies and predict selection gains, consequently helping to identify and select superior genotypes. Thus, the objective of this study was to identify groups of soybean genotypes based on characters from the vegetative stage, in order to select parents for artificial hybridization in soybean and to indicate the relative contribution of the evaluated characters to genetic dissimilarity; estimate genetic and phenotypic parameters, evaluate the association between agronomic traits via phenotypic and genotypic correlations, analyze selection strategies via direct selection and selection indices, and select superior progenies. The experiment was conducted in the field at the Capim Branco experimental farm belonging to the Federal University of Uberlândia (UFU), in a randomized complete block design (RCB), with three replications. Six phenotypic characters were evaluated during the vegetative phase (V2) of the crop: hypocotyl length (CH), epicotyl length (CE), first internode length (CI), unifoliate leaf petiole length (CPFU), petiole length of the first trifoliate leaf (CPFT), plant height at vegetative stage (APV), and twelve agronomic characters: number of days to flowering (NDF), plant height at flowering (APF), number of nodes on the main stem at flowering (NNF), number of days to maturity (NDM), plant height at maturity (APM), number of nodes on the main stem at maturity (NNM), number of productive nodes at maturity (NNPM), number of pods per plant (NVP), first pod insertion height (AIV), lodging (ACM), grain yield per plant (PGP), grain yield in kg ha-1 (PROD). The data obtained at the V2 stage were analyzed using the GENES software, and the genetic diversity among the progenies was estimated using the methods: Tocher optimization, UPGMA and canonical variables. The agronomic traits were evaluated with the aid of the R software, via the mixed models approach (REML/BLUP), in which the genetic and phenotypic parameters, phenotypic and genotypic correlations, selection indices, gain from direct selection, indirectly and through selection indices (Classic, Sum of “Ranks” - M&M and Sum of “Ranks” with no definition of economic weights - M&M2). The PMGS_UFU081 and PMGS_UFU103 genotypes show greater genetic dissimilarity in relation to the others, and PMGS_UFU100 and PMGS_UFU104 are very similar genotypes. On the other hand, the PMGS_UFU081 and PMGS_UFU103 genotypes are quite different from the others, proving to be important genetic resources to explore maximum heterosis in future crosses. Despite the methods of Tocher, UPGMA and Canonical Variables not corroborating the groupings, they are efficient to represent the genetic diversity, with CE being the characteristic that most contributes to the study of genetic diversity in soybean germplasm in the vegetative phase. For the agronomic traits, the existence of genetic variability was detected by the likelihood test, at levels of 0, 01% and 0,001% for all analyzed traits and the heritability estimates ranged from 0,48 to 0,96, evidencing favorable conditions for the selection process. The PGP character presents a positive and high magnitude correlation with the NNM, NNPM and NVP characters, indicating that the selection on these characters can indirectly contribute to the increase of PGP. NDM also showed a positive correlation of high magnitude with the characters NNM, NNPM and NVP, indicating that selection aiming at precocity (reducing NDM) can cause decreases in the average of these characters. The index of the sum of "ranks" with the non-definition of economic weights proves to be the best selection strategy, and the progenies selected by this index were chosen to advance in the breeding program, which are: PMGS_UFU004, PMGS_UFU007, PMGS_UFU009, PMGS_UFU011, PMGS_UFU017, PMGS_UFU 018, PMGS_UFU019, PMGS_UFU021, PMGS_UFU022, PMGS_UFU025, PMGS_ UFU028, PMGS_UFU029, PMGS_UFU076, PMGS_UFU094.