Application of Bayesian causal inference and structural equation model to animal breeding

Area of Science:

• Quantitative genetics
• Animal breeding
• Statistical modeling

Background:

• Optimizing breeding goals necessitates understanding trait relationships.
• Traditional multitrait models lack causal inference capabilities.
• Structural Equation Modeling (SEM) offers a framework for analyzing causal effects.

Purpose of the Study:

• To review the application of SEM in animal breeding for inferring causal relationships.
• To highlight the utility of the inductive causation (IC) algorithm in determining causal structures.
• To demonstrate how SEM provides deeper insights than conventional multitrait models.

Main Methods:

• Review of studies applying SEM in quantitative genetics and animal breeding.
• Explanation of the inductive causation (IC) algorithm for causal discovery.
• Comparison of SEM with traditional multitrait models.

Main Results:

• SEM, particularly with the IC algorithm, can infer causal structures without extensive prior knowledge.
• The IC algorithm effectively identifies causal relationships from observed trait associations.
• SEM allows for the inference of intervention effects, which correlations cannot provide.

Conclusions:

• SEM, guided by the IC algorithm, enhances understanding of trait causality in breeding.
• This approach offers superior insights for breeding strategies compared to standard multitrait analyses.
• Utilizing SEM and IC algorithm advances the precision of genetic improvement programs.