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

Area of Science:

• Plant breeding
• Genomics
• Agricultural science

Background:

• Maize (Zea mays L.) is a globally significant staple crop, with breeding advancements driven by experimental design and genome-based approaches.
• Current genomic selection in maize requires optimization for marker dimensionality, line/environment selection, and identifying promising inbred lines for crosses.

Purpose of the Study:

• To reduce high-density single nucleotide polymorphism (SNP) marker data to a low-density format for genomic selection in maize.
• To evaluate genomic selection models for predicting maize hybrid performance in drought and well-watered conditions for grain yield and secondary traits.

Main Methods:

• SNP markers were ranked and selected based on genome-wide association study (GWAS) effects.
• Genomic selection models incorporating general and specific combining abilities (GCA and SCA) with environmental interactions were compared using cross-validation.
• Model performance was assessed for accuracy in predicting traits under different water regimes.

Main Results:

• Reduced-density marker datasets achieved accuracies comparable to complete datasets for all traits in both drought and well-watered conditions.
• A model including GCA, SCA, and their interactions with environments (Model 7) showed superior performance when all environmental data were available.
• A model without interaction effects (Model 6) was more effective when environmental data were incomplete.

Conclusions:

• Optimized marker dimensionality through selection based on GWAS effects is effective for genomic selection in maize.
• The choice of genomic selection model, particularly regarding environmental interactions, impacts prediction accuracy depending on data availability.