Multi-environment Genomic Selection in Rice Elite Breeding Lines.

Area of Science:

• Plant breeding
• Genomics
• Agricultural science

Background:

• Elite line performance assessment is crucial for breeding programs.
• Genotype by environment interactions pose a significant challenge.
• Genomic prediction models offer a solution by integrating multi-environment trial data and environmental covariates.

Purpose of the Study:

• To assess the predictive ability of different genomic prediction models.
• To optimize the utilization of multi-environment trial data.
• To evaluate model performance for predicting elite rice line performance in untested environments.

Main Methods:

• Utilized 111 elite rice breeding lines from the International Rice Research Institute program.
• Evaluated three traits (days to flowering, plant height, grain yield) across 15 environments in Asia and Africa.
• Employed seven multi-environment genomic prediction models and three cross-validation scenarios with 882 SNP markers.

Main Results:

• Elite lines belonged to the indica group, specifically indica-1B subgroup.
• High phenotypic correlations were observed for days to flowering and plant height, but low for grain yield.
• Models with main effects (genotype, environment, covariates) performed comparably to those including genotype-by-environment interactions; using all environments improved predictions.

Conclusions:

• Multi-environment genomic prediction models focusing on main effects are adequate for accurate phenotypic prediction in target environments.
• These findings can refine testing strategies and enhance genomic prediction models.
• Simpler models are sufficient, optimizing resource allocation in breeding programs.