Optimizing genomic selection of agricultural traits using K-wheat core collection

Area of Science:

• Plant genetics and breeding
• Genomics and bioinformatics
• Quantitative trait analysis

Background:

• Complex agricultural traits in wheat breeding are challenging to select due to their quantitative nature.
• Global wheat production requires innovative breeding strategies to sustain demand.
• Genomic tools offer potential solutions for accelerating the improvement of complex traits.

Purpose of the Study:

• To investigate the utility of genome-wide association studies (GWAS) and genome-wide selection (GS) for breeding ten agricultural traits in wheat.
• To identify trait-associated candidate markers using genome-wide single nucleotide polymorphisms (SNPs).
• To evaluate prediction accuracy of various GS models and optimize training population strategies.

Main Methods:

• Genotyping of 567 Korean wheat accessions using an Axiom® 35K wheat DNA chip.
• Phenotypic evaluation of ten agricultural traits including color, length, days to heading, and maturity.
• Application of GWAS for marker identification and GS with six predictive models (G-BLUP, LASSO, BayseA, RKHS, SVM, random forest) and diverse training populations.

Main Results:

• A significant SNP on chr1B was associated with both awn color and ear color, indicating pleiotropy.
• Most GS models (except SVM) achieved prediction accuracies of 0.4 or higher.
• Subgroup-based training populations improved prediction accuracy for several traits, and RKHS model accurately predicted phenotypes in 70% of tested cultivars.

Conclusions:

• GWAS and GS are effective tools for improving complex traits in wheat breeding.
• Genomics-assisted breeding strategies, particularly using optimized training populations, can enhance selection efficiency.
• The findings provide a foundation for developing advanced wheat breeding programs utilizing genomic information.