Theoretical accuracy for indirect predictions based on SNP effects from single-step GBLUP

Area of Science:

• Animal breeding and genetics
• Quantitative genetics
• Genomic selection

Background:

• Single-step genomic best linear unbiased prediction (ssGBLUP) allows backsolving SNP effects from genomic estimated breeding values (GEBV).
• Indirect prediction (IP) uses these SNP effects for genomic predictions, especially useful for large genotyped populations or interim evaluations.
• Individual accuracies for IP are of interest but rarely reported.

Purpose of the Study:

• To derive equations for computing individual accuracy of IP based on prediction error covariance (PEC) of SNP effects.
• To evaluate the Algorithm for Proven and Young (APY) in PEC computations.
• To determine the minimum number of genotyped animals required for accurate IP prediction.

Main Methods:

• Developed equations for individual IP accuracy using PEC of SNP effects derived from ssGBLUP GEBV PEC.
• Tested the APY algorithm for PEC computations with large datasets.
• Assessed the impact of reducing genotyped animals in PEC computations on IP accuracy.

Main Results:

• High correlations (≥0.99) were observed between GEBV and IP for validation animals.
• IP accuracies were comparable to GEBV accuracies using direct or APY G inverse.
• Reducing genotyped animals in PEC computations (55,000 to 15,000) maintained high correlations but biased IP accuracies downwards.

Conclusions:

• Theoretical IP accuracy can be reliably computed from ssGBLUP GEBV PEC using direct or APY G inverse.
• Reducing genotyped animals in PEC computations is feasible but may lead to underestimated accuracies.
• Further research is needed to approximate SNP PEC from ssGBLUP to manage computational demands.