A stochastic simulation study on using different models for prediction of breeding values while changing the breeding...

Area of Science:

• Animal breeding and genetics
• Quantitative genetics
• Dairy cattle improvement

Background:

• Effective breeding programs require accurate prediction of breeding values.
• Breeding goals evolve to incorporate multiple traits, including health and production.
• Stochastic simulation allows for robust evaluation of complex genetic scenarios.

Purpose of the Study:

• To investigate the impact of simultaneous changes in breeding goal and prediction model on genetic response.
• To compare the efficacy of univariate, approximate multitrait, and full multitrait models.
• To assess genetic trends and inbreeding under different breeding scenarios.

Main Methods:

• Simulated a population of 100,000 cows over 30 years in two 15-year steps.
• Applied univariate and trivariate models for initial selection.
• Evaluated six scenarios combining two breeding goals and three prediction models over 15 years with 20 replicates.

Main Results:

• Approximate and full multitrait models yielded similar genetic responses.
• Multitrait models significantly improved total merit response compared to univariate models under a broad breeding goal.
• Narrow breeding goals showed no significant difference in response between models.
• Multitrait models redistributed genetic progress towards low heritability traits with high economic value.

Conclusions:

• Multitrait prediction models are beneficial for broad breeding goals in dairy cattle.
• The choice of breeding goal significantly influences the effectiveness of different prediction models.
• Exploiting available data with multitrait models can enhance genetic gain but may increase inbreeding.