Effects of culling on sire evaluations by mixed models

Area of Science:

• Animal Science
• Dairy Cattle Genetics
• Quantitative Genetics

Background:

• Accurate genetic evaluation of dairy sires is crucial for genetic improvement.
• Selection bias can affect the accuracy of genetic predictions, particularly when using multiple records.

Purpose of the Study:

• To assess the impact of culling low-yield second lactation records on Holstein sire genetic evaluations.
• To evaluate the effectiveness of adjustment methods for selection bias in sire evaluations.

Main Methods:

• Utilized milk and fat yield records from 677,800 Holstein daughters of 200 sires.
• Applied Best Linear Unbiased Prediction (BLUP) for sire evaluations.
• Simulated culling of 10%, 20%, and 30% of lowest second lactation milk yield records.

Main Results:

• Culling low-yield second records reduced milk evaluation standard deviations and correlations between first and second records.
• Evaluations using both records showed minimal impact from culling, with high correlations.
• Adjustment for selection effectively mitigated the effects of culling on sire evaluations.

Conclusions:

• Culling low-yield second records has a limited negative impact on sire genetic evaluations when using both records.
• Adjustment for selection is effective in removing bias and maintaining the accuracy of genetic predictions.
• BLUP and Modified Contemporary Comparison procedures yield similar results regarding culling and adjustment effects.