Application of single-step genomic evaluation using multiple-trait random regression test-day models in dairy cattle

Area of Science:

• Animal Genetics and Breeding
• Dairy Cattle Genomics
• Quantitative Genetics

Background:

• Test-day traits are crucial for genetic evaluation in dairy cattle.
• Multiple-trait random regression models (RRM) offer improved modeling for these traits.
• Genomic estimated breeding values (GEBV) enhance genetic selection accuracy.

Purpose of the Study:

• To evaluate the reliability and bias of GEBV predicted by multiple-trait RRM via single-step genomic best linear unbiased prediction (ssGBLUP).
• To compare ssGBLUP with traditional BLUP for genetic evaluation in Canadian dairy breeds.
• To investigate the impact of scaling factors and genotype inclusion strategies on GEBV accuracy.

Main Methods:

• Utilized two multiple-trait RRM for milk, fat, protein yields, and somatic cell score across three lactations.
• Applied ssGBLUP to predict individual additive genomic random regression coefficients.
• Compared daily GEBV with traditional daily parent averages from BLUP, testing various scaling factors for relationship matrices and genotype inclusion criteria.

Main Results:

• ssGBLUP demonstrated considerably higher validation reliabilities than BLUP without genomic information.
• Scaling factors for combining genomic (G^-1) and pedigree (A^-1₂₂) matrices had minor effects on reliability but influenced GEBV inflation.
• Optimal scaling factors reduced GEBV inflation compared to traditional BLUP parent averages; results were consistent regardless of genotype inclusion strategy.

Conclusions:

• ssGBLUP combined with multiple-trait RRM significantly enhances the reliability of GEBV for dairy cattle.
• This approach effectively reduces bias in breeding value predictions for young animals across Ayrshire, Holstein, and Jersey breeds.
• The findings support the adoption of ssGBLUP with RRM for more accurate genetic evaluations in Canadian dairy populations.