Selection of core animals in the Algorithm for Proven and Young using a simulation model

Area of Science:

• Animal Breeding and Genetics
• Quantitative Genetics
• Bioinformatics

Background:

• The Algorithm for Proven and Young (APY) facilitates single-step genomic Best Linear Unbiased Prediction (ssGBLUP) in large populations.
• APY separates genotyped animals into core and non-core subsets for computational efficiency in genomic relationship matrix inversion.

Purpose of the Study:

• To investigate the impact of different core subset definitions on the accuracy and bias of genomic evaluations.
• To provide guidance on selecting core animals within the APY framework for large-scale genetic assessments.

Main Methods:

• Simulations involving 95,010 animals across five generations with 50,000 genotypes and 25,500 SNPs.
• Mimicking genotyping errors and missing pedigree data.
• Comparing core animal definitions based on individual generations, equal generational representation, and random selection.

Main Results:

• Core definitions showed similar accuracies and biases for sufficiently large core sizes, even with imperfect genotypes.
• Random and across-generation core definitions significantly improved accuracy and reduced bias when parentage was unknown (p ≤ 0.05).

Conclusions:

• The choice of core definition has minimal impact on ssGBLUP accuracy with adequate core size and good genotype quality.
• Random or across-generation core selection strategies are recommended when dealing with unknown parentage in genomic evaluations.