Combining different functions to describe milk, fat, and protein yield in goats using Bayesian multiple-trait random...

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Dairy Science

Background:

• Accurate genetic evaluation of dairy goats is crucial for breeding programs.
• Previous studies often used single-trait models or multiple-trait models with a single function for all traits.

Purpose of the Study:

• To develop and evaluate multiple-trait random regression models (MTRRM) combining different functions for dairy goat genetic evaluation.
• To compare the fit of MTRRM with single-trait random regression models (STRRM).

Main Methods:

• Employed Bayesian inference for genetic evaluation.
• Tested various STRRM including Legendre polynomials, B-splines, Ali and Schaeffer, and Wilmink functions.
• Combined best-performing STRRM into MTRRM and selected the optimal model based on DIC and PMP.

Main Results:

• Combined MTRRM showed superior fit (lower DIC, higher PMP) compared to STRRM.
• The Ali_Leg5_Ali model, using Ali for milk yield (MY) and protein percentage (PP) and Leg5 for fat percentage (FP), provided the best fit.
• Heritability estimates for MY, FP, and PP ranged from 0.25-0.54, 0.27-0.48, and 0.35-0.51, respectively.
• Genetic correlations between traits varied, indicating complex genetic relationships.

Conclusions:

• Combining different functions within MTRRM is a viable approach for joint genetic evaluation of milk yield and constituents in goats.
• The Ali_Leg5_Ali model offers an improved framework for dairy goat genetic analysis.