Comparison of repeatability and multiple trait threshold models for litter size in sheep using observed and simulated...

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Reproductive Biology

Background:

• Estimating genetic parameters for litter size across multiple parities is crucial for sheep breeding programs.
• The choice between repeatability models (RM) and multiple trait threshold models (MTM) impacts the accuracy of genetic evaluations.

Purpose of the Study:

• To compare the appropriateness of RM and MTM for estimating genetic parameters of litter size in Mule ewes.
• To evaluate the impact of model choice on the accuracy of breeding value estimation and the risk of model misspecification.

Main Methods:

• Bayesian analyses were employed to estimate genetic parameters using 5580 litter size records from 1758 Mule ewes.
• Both RM and MTM were fitted to observed data, and their performance was compared using simulated data based on estimated parameters.
• Model selection was also assessed using the Bayesian Information Criterion (BIC).

Main Results:

• MTM yielded higher heritability estimates (0.12-0.18) compared to RM (0.08).
• Positive genetic correlations (0.24-0.71) were found between litter sizes across parities.
• Simulations indicated a substantial loss in accuracy and increase in mean squared error (MSE) when using RM if MTM parameters were true.
• BIC favored RM due to a penalty for the larger number of parameters in MTM.

Conclusions:

• MTM is recommended when accuracy of breeding values and reduction of MSE for litter size across parities are primary concerns.
• RM is suggested if minimizing the risk of using an incorrect model is the main objective, as indicated by BIC.
• The findings highlight a trade-off between model complexity, accuracy, and risk assessment in genetic parameter estimation for sheep litter size.