Estratificação ambiental em milho por meio de redes de similaridade

Abstract

The complexity of genotype by environment interactions poses challenges to identifying environments conducive to consistent performance of maize hybrids (Zea mays L.). This aspect is crucial for the success of breeding programs, which rely on the ability to provide hybrids with optimized yields in diverse environmental conditions. In this context, the central hypothesis of this study suggests that maize hybrids exhibit significant performance variations in different regions, although it is possible to identify similar patterns. This identification is of great relevance, reducing the need for extensive trials and optimizing investments and human resources. Therefore, the research objective is to identify environmental strata using the GGE biplot technique integrated with similarity networks. The aim is to identify regions that provide similar performance for maize hybrids, aiming not only to reduce the need for extensive trials but also to optimize the efficient allocation of resources. The methodology involved evaluating production data from 759 hybrids in 114 experiments in the state of Minas Gerais over three harvests. A randomized block design and incomplete block design were used, integrating the GGE biplot methodology and similarity networks in five distinct steps. The results of the analysis of variance indicated the existence of different mega-environments in the studied region, highlighting the importance of environmental stratification. The use of similarity networks allowed observing environmental patterns. Environmental stratification, based on graphical analysis integrating the GGE biplot and similarity networks, proved effective in identifying environmental strata. This approach provided valuable insights into environmental diversity and the formation of homogeneous groups of environments, enabling more precise decision-making in selecting representative locations for hybrid testing. By considering similarity patterns, the analysis allowed detecting strata in which different sets of hybrids exhibit similar production responses. This approach stands out as a promising tool to optimize resource allocation in maize breeding programs.