Genomic selection on ear height, plant height and grain yield in the primary testing stage of maize hybrids

Genomic selection (GS) is a cutting-edge breeding technology that enables the prediction and early selection of individuals based on genomic estimated breeding values by constructing predictive models. Double haploid (DH) technology has become an efficient method for producing inbred lines in maize, and when combined with GS, it offers significant cost reductions through advanced data and information management. Recent studies have demonstrated the great potential and high expectations of GS in plant breeding, particularly in maize, where the combination of GS and DH has been successfully applied. In this study, 2029 hybrids resulting from crosses were grown in three representative locations, and phenotypic values for three agronomic traits-ear height (EH), plant height (PH), and grain yield (GY)-were measured. Parental genotypes were used alongside genomic predictions to estimate hybrid breeding values, with GY being the primary trait of interest. A combination of traits was then employed as a criterion for advancing hybrids to the primary stage of testing in maize. Predicted breeding estimates showed that the accuracy for EH and PH was approximately 0.75, while for GY, it was 0.43; GY was field validated by including 80% of the top 243 hybrids, measured at about 55%, with moderately high predictive ability. In summary, the study demonstrates a significant reduction in the number of crosses required in the field based on breeding estimates, a decrease in the need for costly multi-site primary field tests, and an increase in breeding efficiency.