Multi-line ssGBLUP evaluation using preselected markers from whole-genome sequence data in pigs.

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Genomic Prediction

Background:

• Genomic evaluations in pigs can be enhanced by integrating multi-line data with whole-genome sequencing (WGS).
• Large-scale data are necessary to capture population variability for effective genomic evaluations.

Purpose of the Study:

• To investigate strategies for combining large-scale data from different terminal pig lines in multi-line genomic evaluation (MLE).
• To assess the utility of preselected variants from WGS data within single-step GBLUP (ssGBLUP) models.

Main Methods:

• Employed single-step GBLUP (ssGBLUP) models for multi-line genomic evaluations (MLE) across five traits in three terminal pig lines.
• Explored Unknown Parent Groups (UPG) and metafounders (MF) to manage genetic differences among lines.
• Preselected sequence variants using multi-line genome-wide association studies (GWAS) or linkage disequilibrium (LD) pruning.

Main Results:

• Multi-line genomic evaluation (MLE) using UPG and MF showed minimal to no gain in prediction accuracy compared to single-line evaluations (SLE).
• Incorporating preselected variants from GWAS into commercial SNP chips yielded a maximum accuracy increase of 0.02 for average daily feed intake in specific lines.
• No benefits were observed from using preselected sequence variants in MLE, and BayesR weights did not improve ssGBLUP performance.

Conclusions:

• Limited benefits were found in using preselected whole-genome sequence variants for MLE in pigs, even with substantial imputed sequence data.
• Accurate accounting for line differences is essential for MLE to achieve predictions comparable to SLE.
• The primary benefit of MLE is enabling comparable predictions across different pig lines.