Genomic Prediction Based on SNP Functional Annotation Using Imputed Whole-Genome Sequence Data in Korean Hanwoo...

Area of Science:

• Animal Genomics
• Quantitative Genetics
• Livestock Breeding

Background:

• Whole-genome sequence (WGS) data offers higher predictive ability in genomic predictions by including causal mutations.
• Existing SNP panels may not capture all relevant genetic variation for complex traits.
• Improving genomic prediction accuracy is crucial for efficient livestock breeding programs.

Purpose of the Study:

• To enhance the predictive performance of the Hanwoo 50k SNP panel for four carcass traits.
• To investigate the utility of adding pre-selected variants from WGS data to a customized SNP panel.
• To evaluate the impact of different functional annotations on predictive power.

Main Methods:

• Utilized WGS imputed genotypes and phenotypes from 16,892 Hanwoo cattle for four carcass traits.
• Performed genome-wide association studies (GWAS) for variant pre-selection.
• Added sets of pre-selected SNPs (1000-10,000) to the 50k panel and assessed predictive performance using GBLUP.

Main Results:

• Adding 3000 pre-selected variants per trait to the 50k panel improved prediction accuracies by up to 9.9%.
• The inclusion of 12,000 variants (3000 per trait) increased prediction accuracies for backfat thickness, carcass weight, longissimus muscle area, and marbling score.
• Pre-selected SNPs from the intergenic region (IGR) showed the highest improvement, comparable to using whole-genome data.

Conclusions:

• The predictive performance of the Hanwoo 50k SNP panel can be effectively improved by incorporating selected variants from WGS data.
• The optimal number of added variants appears to be trait-dependent, with 3000 per trait being sufficient for significant improvement.
• While IGR variants showed promise, WGS data generally provided higher accuracy than specific genomic regions.