Genomic selection of maize test‐cross hybrids leveraged by marker sampling

Maize (Zea mays L.) is a staple crop and the most cultivated cereal worldwide. The expansion of this crop was possible due to efforts in management and breeding. From the breeding standpoint, advances were achieved through field experimental design and analyses, establishing heterotic patterns, and releasing heterotic hybrids. Over the last decade, data analyses have benefited from the surge of genome-based approaches. However, it lacks optimization regarding marker dimensionality, proper selection of tested lines and/or environments, and an indication of promising inbred lines for crosses. This study aimed to convert a high-density single nucleotide polymorphism marker dataset into a low-density dataset and perform genomic selection of maize hybrids tested in drought stress and well-watered environments for grain yield and secondary traits. Single nucleotide polymorphism markers were ranked and selected based on effects from a genome-wide association study. For genomic selection, methods containing general and specific combining abilities (GCA and SCA, respectively) and interaction effects were compared in cross-validation schemes. Accuracies using selected markers were similar to complete marker dataset for all traits under drought nand well-watered conditions. For genomic selection, the model containing the main effects of GCA for inbred lines and testers, SCA for hybrids, and the interaction of GCA and SCA with environments (Model 7) performed better for all traits when information about all environments was included. The model without interaction effects (Model 6) performed better when information about environments was missing.