Estimating maternal genetic effects in livestock.

Area of Science:

• Quantitative genetics
• Animal breeding
• Statistical genetics

Background:

• Estimating genetic correlations between direct and maternal effects is challenging.
• Previous models have yielded biased or questionable genetic correlation estimates.
• Environmental covariances between dam and offspring are often implicated in these estimation issues.

Purpose of the Study:

• To evaluate the performance of models implementing Willham's model for genetic (co)variance estimation.
• To investigate the covariance structure of residual terms in Willham's model.
• To develop an unbiased method for estimating direct and maternal genetic (co)variances.

Main Methods:

• Investigated the covariance structure of residual terms arising from Willham's model.
• Employed stochastic simulations to test the proposed model.
• Compared results with models including regression on dam-phenotype.

Main Results:

• A correlation between individual and dam residuals is a consequence of combining Willham's model with standard assumptions.
• Fitting this correlated residual structure yields unbiased estimates of genetic (co)variances.
• Ignoring correlated residuals introduced bias approximately equal to the environmental correlation.

Conclusions:

• Willham's model, accounting for correlated residuals, provides a conceptually appealing and theoretically consistent approach.
• This method yields unbiased genetic (co)variance estimates without the need for regression on dam phenotype.
• Enhancements to existing software for fitting residual variance structures for maternal effects are recommended for animal breeders.