GENOMIC SELECTION AS A TOOL FOR MAIZE CULTIVARS DEVELOPMENT

The ability to predict genotypes that have not yet been tested is always a target of plant breeders. Over the last twenty years, many studies presented genomic selection (GS) as a tool contributing to this goal. Currently, many research papers have shown encouraging results in the application of GS. However, there are few examples of long-term, successful applications of GS in plant breeding programs. Furthermore, for breeders and researchers considering the application of GS, there are a series of important considerations on how to adapt a breeding program to maximize the benefit of GS, aiming to reduce the costs and maximize the genetic gains. Under this perspective, we present a review with a general view about applied GS in maize breeding, future perspectives of this technique, and an applied study case of a breeding program using GS. We attempt to provide a brief review of the literature with recent developments, as well as a discussion involving the number of markers required to deploy GS, the different statistical approaches to create GS models, the different ways to define training populations, and the incorporation of non-additive effects and genotype by environment interaction. We end with general recommendations and conclusions about some critical points about adopting GS in maize breeding.