Models for Estimating Genetic Parameters of Milk Production Traits Using Random Regression Models in Korean Holstein...

Area of Science:

• Animal Genetics and Breeding
• Dairy Cattle Production
• Quantitative Genetics

Background:

• Accurate estimation of genetic parameters is crucial for effective genetic selection in dairy cattle.
• Traditional lactation models may not fully capture the dynamic changes in milk production traits throughout lactation.
• Random regression models (RRMs) offer a more flexible approach to modeling trait expression over time.

Purpose of the Study:

• To estimate genetic parameters for milk, fat, protein, and solids-not-fat yield in Holstein cows using RRMs.
• To compare the goodness of fit of various RRMs with homogeneous and heterogeneous residual variances.
• To determine the optimal model for genetic evaluation of milk production traits in Korean Holstein cattle.

Main Methods:

• Utilized 126,980 test-day milk production records from Korean Holstein cows (2007-2014).
• Applied RRMs with Legendre polynomials (3rd-5th order) for genetic and permanent environmental effects.
• Compared models with homogeneous (HOM) and heterogeneous (HET15, HET60) residual variances using AIC and BIC criteria.

Main Results:

• The best-fitting models were L5-HET15 for milk, protein, and SNF yields, and L4-HET15 for fat yield, based on BIC.
• Heterogeneous residual variance models (HET15) generally outperformed homogeneous models.
• Heritability estimates varied across lactation stages, with genetic variances showing dynamic changes.

Conclusions:

• Heterogeneity of residual variances is important and should be considered in test-day analyses.
• RRMs provide more accurate genetic parameter estimates than traditional lactation models.
• Recommended using RRMs for national dairy cattle genetic evaluations of milk production traits in Korea.