Practical implementation of cost-effective genomic selection in commercial pig breeding using imputation

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Bioinformatics

Background:

• Genomic selection (GS) offers advancements in livestock breeding.
• Genotype imputation can reduce the cost of GS implementation.
• Assessing imputation accuracy and its impact on genomic estimated breeding values (gEBV) is crucial for practical application.

Purpose of the Study:

• To investigate the accuracy of genotype imputation in pigs using various low-density panels.
• To evaluate the impact of imputation accuracy on the reliability of genomic breeding values (gEBV).
• To determine optimal strategies for low-density panel design and genotyping to maximize imputation and gEBV accuracy.

Main Methods:

• Construction of three low-density SNP panels (450, 3,071, 5,963 SNPs) from high-density data.
• Imputation of low-density genotypes using AlphaImpute under different genotyping scenarios.
• Calculation of gEBV using single-step genomic evaluation and assessment of retained information via correlation with true gEBV.

Main Results:

• High imputation accuracy was achieved across scenarios, decreasing with fewer SNPs and reduced genotyping of ancestors.
• The largest decrease in imputation accuracy was observed for the lowest density panel (L450) and when dams/granddams were not genotyped.
• Excluding non-ancestor genotypes minimally impacted imputation accuracy but significantly reduced gEBV accuracy.

Conclusions:

• Genomic breeding value accuracy from imputed genotypes is sensitive to the genotyping level of close relatives and the overall dataset size.
• Fewer high-density genotyped individuals are required for accurate imputation than for accurate gEBV.
• Strategic development of low-density panels is essential for improving both imputation and gEBV accuracy in pig breeding programs.