Integrating Genome-Wide Association Study (GWAS) and Marker-Assisted Selection for Enhanced Predictive Performance of Soybean Cold Tolerance.

Soybean (Glycine max (L.) Merr.), as a crucial source of oil and protein globally, is widely cultivated in many countries. Low-temperature stress has become one of the major environmental factors affecting soybean production, especially in colder regions, making the improvement of cold tolerance traits in soybean a key breeding objective. This study integrates Genome-Wide Association Studies (GWAS) and Marker-Assisted Selection (MAS) to enhance the predictive performance of soybean cold tolerance traits. First, three GWAS methods-Fast3VmrMLM, fastGWA, and FarmCPU-were used to analyze soybean cold tolerance traits, and significant SNP markers were identified. Principal Component Analysis (PCA) was employed to reveal genetic differences among various soybean germplasm. Then, based on the identified SNP markers, multiple Genomic Selection (GS) models, such as GBLUP, BayesA, BayesB, BayesC, BL, and BRR, were used for prediction to evaluate the contribution of genetic effects to phenotypic variation. The results showed that the markers selected through GWAS significantly improved the prediction accuracy of genomic selection, especially with the Fast3VmrMLM and FarmCPU methods in larger datasets. Finally, Gene Ontology (GO) analysis was performed to further identify candidate genes associated with cold tolerance traits and their biological functions, providing theoretical support for molecular breeding of cold-tolerant soybean varieties.