Genomewide predictions from maize single-cross data.

Area of Science:

• Plant breeding
• Genetics
• Agricultural science

Background:

• Maize (Zea mays L.) breeders assess numerous single-cross hybrids annually across diverse environments.
• Accurate prediction of hybrid performance is crucial for efficient breeding programs.

Purpose of the Study:

• To evaluate the utility of genomewide predictions for identifying superior untested single crosses.
• To assess the effectiveness of these predictions for selecting elite inbreds within biparental crosses.

Main Methods:

• Utilized 479 experimental maize single crosses and genotypic data from 669 single nucleotide polymorphism (SNP) markers.
• Employed best linear unbiased prediction (BLUP) and ridge-regression BLUP (RR-BLUP) models for performance prediction.
• Assessed prediction accuracy (r(MG)) for grain yield, moisture, and lodging traits.

Main Results:

• BLUP achieved high prediction accuracies (0.84-0.90) for grain yield and moisture.
• RR-BLUP models did not significantly differ from BLUP in predicting single-cross performance.
• RR-BLUP predictions for testcross performance within biparental populations yielded low and often negative accuracies.

Conclusions:

• Genomewide marker effects from single-cross data offer no advantage over BLUP for predicting single-cross performance.
• The utility of these predictions for identifying top inbreds in biparental crosses is unreliable.
• Further research may be needed to improve prediction models for biparental testcrosses.