Symposium review: Single-step genomic evaluations in dairy cattle

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Dairy Cattle Breeding

Background:

• Genomic selection has revolutionized dairy cattle breeding, significantly reducing sire age and increasing genetic progress.
• Traditional genetic evaluations are becoming biased due to the shift from progeny-tested to genomically selected young sires.
• Existing multi-step genomic evaluations can also be distorted when using biased traditional evaluations as input.

Purpose of the Study:

• To highlight the necessity of incorporating genomic information into evaluations for unbiased genetic progress estimation.
• To discuss the challenges and advancements in implementing single-step genomic evaluation models in large-scale national dairy evaluations.
• To address current computational and convergence issues in single-step evaluation models.

Main Methods:

• Review of single-step evaluation models, including ssGBLUP and single-step marker models.
• Discussion of computational strategies like sparse matrix representations (APY-G, ssGTBLUP) to overcome initial limitations.
• Exploration of methods to improve convergence in iterative solving, such as preconditioned conjugate gradient iteration and accounting for genetic groups.

Main Results:

• Single-step evaluation models, though introduced in 2010, are not yet widely implemented in national dairy evaluations.
• Computational costs hindering single-step evaluations have been largely mitigated by algorithmic and hardware advancements.
• Current research focuses on improving convergence for iterative solvers, especially for low-heritability traits and complex models, and addressing genomic evaluation inflation.

Conclusions:

• Single-step genomic evaluation is the only viable long-term solution to maintain unbiased genetic progress estimates in dairy cattle.
• Ongoing research is crucial to refine single-step models, addressing convergence issues and potential overvaluation of genomic information.
• Further implementation of single-step evaluations is essential for accurate and reliable genetic improvement programs in dairy cattle.