Comparative methods with sampling error and within-species variation: contrasts revisited and revised.

Area of Science:

• Evolutionary biology
• Phylogenetic comparative methods
• Quantitative genetics

Background:

• Comparative methods often assume species phenotypes are true means, neglecting sample size effects on covariance estimation.
• Previous models required known within-species phenotypic covariances, limiting their applicability.

Purpose of the Study:

• To develop a statistical model for estimating evolutionary and within-species covariances when within-species covariances are unknown.
• To correct for bias in covariance estimates arising from finite sample sizes in phylogenetic analyses.

Main Methods:

• A multivariate normal statistical model was applied to species related by a known phylogeny under a Brownian motion model.
• An expectation-maximization algorithm was used to estimate covariances of evolutionary change and within-species phenotypic covariances.
• The method was implemented in the Contrast program of the PHYLIP package.

Main Results:

• The developed model corrects for bias in covariance estimates caused by not accounting for finite sample sizes.
• Computer simulations demonstrated the effectiveness of the proposed method in reducing bias.
• Sampling variation was found to reduce the power of inference for covariation in the evolution of different characters.

Conclusions:

• The new model provides a more accurate estimation of evolutionary covariances by incorporating within-species sample size information.
• This approach enhances the reliability of phylogenetic comparative analyses, particularly for understanding trait evolution.
• The method offers an extension to incorporate additive genetic covariances from genetic experiments.