OptimGS: a dual integrative genomic prediction framework for improving cold stress tolerance in wheat

Area of Science:

• Plant genetics
• Quantitative genetics
• Agricultural science

Background:

• Cold stress tolerance in wheat is a complex quantitative trait with low heritability, challenging conventional breeding.
• Genomic selection (GS) accelerates genetic gain but prediction accuracy depends on model choice and genetic architecture.

Purpose of the Study:

• To propose and evaluate a dual integrative genomic prediction framework to enhance prediction accuracy for complex quantitative traits in wheat.
• To sequentially integrate information across chromosomes and models for improved genomic prediction.

Main Methods:

• Implemented 14 genomic prediction models (Bayesian, BLUP, ML) on a wheat germplasm panel (4269 genotypes).
• Partitioned genome-wide markers chromosome-wise, generating independent predictions per chromosome.
• Combined predictions using a genetic algorithm under two bidirectional strategies: chromosome-first-model-second (CFMS) and model-first-chromosome-second (MFCS).
• Assessed prediction performance using five-fold cross-validation, Pearson's correlation, and mean squared error.

Main Results:

• The CFMS and MFCS strategies consistently outperformed individual models and conventional whole-genome approaches.
• The proposed dual integrative framework demonstrated enhanced prediction accuracy for seedling cold tolerance in wheat.
• The framework proved robust and biologically meaningful for complex quantitative traits.

Conclusions:

• The dual integrative genomic prediction framework offers a significant advancement for predicting complex quantitative traits like cold stress tolerance in wheat.
• This approach holds substantial potential for accelerating genetic gain and improving wheat breeding programs.
• The developed framework provides a robust strategy for enhancing genomic prediction accuracy.