Is it feasible to select for flowering set in soybean?

Abstract

In soybean, potential grain yield is dependent on the number of plants per area, number of pods per plant, number of grains per pod and average grain weight. The contribution of each of these primary production components is variable, however, in some situations the number of pods per plant has been shown to be important. The number of pods depends on the number of flowers produced by the plant and their setting, that is, the proportion of flowers that originate pods. In soybean, normally the number of flowers per plant is much higher than the number of pods, therefore, the floral setting is almost always small. Accordingly, the plant is unable to express all its productive potential, since some stresses (biotic or abiotic) affect flower setting. The objective was to determine the floral setting levels of soybean progenies with different maturity groups; The correlate response for flower set and soybean grain yield; The viability of selection for the floral setting trait. The field experiments were conducted during the 2019/2020 and 2020/2021 crop seasons, in the municipalities of Ijaci, at the Center for Development and Transfer of Technology in Agriculture at UFLA – Palmital Farm, and in the municipality of Lavras, at the Center for Scientific and Technological in Agriculture- Muquém Farm. The F2:4 and F2:5 progenies were evaluated in two experiments: one containing early progenies and checks and another with late progenies and checks, both conducted under a complete randomized block design (RCBD). The plot consisted of two lines with three meters in length and spacing of 0.5 meters. In the central part, six plants were selected in the first row of each plot, a floral receptacle was placed, aiming to collect reproductive structures that would be aborted. The agronomic traits days for full maturity (DTM), yield (YIELD), growth habit (GROWTH) and flower set (FS) were evaluated. Statistical analyzes were performed in the R environment, using mixed linear models (LMM) and generalized linear models (GLMM). The initial stage of data analysis involved fitting both mixed linear models and generalized linear models to the phenotypic data of each character. Only the fixed effects that were significant according to the Wald test and the random effects with non-zero variance estimates were included in the phenotypic model. Growth habit and days to maturity were taken into consideration as covariates in the analysis. Regarding the choice of phenotypic models, the selected model was a First-Order Analytical Factor (FA1) based on the various VCOV matrix structures tested and the criteria used for selection. Estimates of the genetic parameters denote variability among the progenies, making it possible to obtain genotypes with good yield performance. It was concluded that there is a great influence of the G x E interaction on the estimates of genetic and phenotypic parameters associated with Yield and flower set traits in soybean; The correlated response between the FS and Yield traits is negative, making the indirect selection of progenies with high yield and flower setting unfeasible; The selection based on high productivity was satisfactory, highlighting the progenies P72, P58, P30, T169, T136 and T82.