Spatial analysis: an approach to optimize the selection of wheat families

Wheat production in Brazil is expanding in non-traditional regions, such as the Cerrado; this trend demands breeding tools that improve selection efficiency despite the strong environmental influence on field trials. This study evaluated the impact of spatial analysis on the selection of 400 F_3:4 wheat families and 10 parents as controls, in an augmented block design at Federal University of Viçosa, in 2024. Data for days to heading, plant height, ear size, wheat blast severity, grain yield, and thousand-grain weight were analyzed using the SpATS package to generate raw (S1) and spatially adjusted (S2) models, followed by deviance analysis in Genes software. A combined index was also applied to raw and adjusted data, generating S3 and S4, and selection gains were estimated using the MGIDI index. S2 reduced residual variance, increased heritability, and lowered AIC values compared to S1, particularly for grain yield, whose heritability rose from 65.71% to 81.28%. Selection gains were consistently higher with spatial adjustment, and the S4, which combines spatial correction with the genetic index, achieved the greatest gains across all traits. These results confirm that spatial analysis, especially when integrated with genetic value prediction, substantially improves selection accuracy and efficiency in wheat breeding programs.