Selection Strategies in Wheat Populations

Abstract The selection of segregating populations is a key point in plant breeding programs. These should gather favorable phenotypes for multiple target characters, which makes it difficult to identify populations with high potential. Thus, this research aims to select potential wheat populations with precocity and grain yield by multivariate and multigeneration approaches. To achieve these objectives, 54 wheat populations in generations F2 and F3 were tested in two environments and evaluated for the traits days for heading and grain yield. Four analytic strategies were imposed: Strategy I: univariate model for each generation; Strategy II: univariate model and multigeneration; Strategy III: multivariate model for each generation; Strategy IV: multivariate model and multigeneration. In this scenario, the strategies that involved the multi-generation model (environments) provided greater gains, strategies I and III. Nevertheless, strategy I, involving a univariate model, provided the greatest gain. Within the strategies, the 1D, 1G, 1H, 2D, 2E, 2F, 2G and 3H populations were selected by different strategies, being the most promising for derivation of early cycle and with grain productivity.