Implications of using an average relationship matrix in genetic evaluation for a population using multiple-sire...

Area of Science:

• Animal Breeding and Genetics
• Quantitative Genetics
• Statistical Genomics

Background:

• Accurate genetic evaluation is crucial for effective animal breeding programs.
• Ambiguous paternity, where sires are not definitively known, presents a challenge in genetic analysis.
• The average numerator relationship matrix (average A) is a method to incorporate unknown sires into genetic models.

Purpose of the Study:

• To evaluate the impact of using the average numerator relationship matrix (average A) in mixed model equations (MME) for populations with ambiguous paternity.
• To assess the accuracy of estimated breeding values (EBVs) and genetic trend predictions when using the average A.
• To quantify the overestimation of genetic trend caused by the average A substitution.

Main Methods:

• Simulated a population of 40 females and 2 males over 8 breeding cycles.
• Applied both random mating and sequential selection scenarios.
• Calculated EBVs using an animal model and the average A within MME.
• Computed variances of EBVs and prediction errors.

Main Results:

• The average A matrix incorrectly estimates the variance of family sizes among sires.
• It underestimates the variance reduction caused by selection.
• These inaccuracies lead to a significant overestimation of the genetic trend.

Conclusions:

• The average A matrix is a viable approach for handling ambiguous paternity in MME.
• However, its use leads to an overestimation of genetic trend due to misperceptions of variance components.
• Researchers and breeders should be aware of this bias when interpreting genetic evaluations in such scenarios.