A Symphony of Genes Make a Maize Plant

under a specific cross-validation scheme are commonly used to evaluate their relative performances. Hadasch et al. compare the predictive abilities of contending models for MAS under several contrasting cross-validation schemes using empirical datasets on two quantitative traits with different heritabilities collected from multienvironment trials. The predictive abilities of models for MAS were strongly influenced by the presence of marker loci with large effects and thus indirectly by the heritability of the target trait. The predictive abilities of models for MAS also varied across the different cross-validation schemes, but none of the schemes consistently produced the highest predictive abilities across all models or traits. Consequently, it is recommended to evaluate the predictive abilities of models for MAS under multiple cross-validation schemes that are based on sampling genotypes, environments, or both. 