Fixed-effect variance and the estimation of repeatabilities and heritabilities: issues and solutions.

Area of Science:

• Quantitative genetics
• Evolutionary biology
• Statistical modeling

Background:

• Linear mixed-effects models are standard for estimating heritability and repeatability.
• Current methods often exclude variance from fixed effects, potentially biasing heritability and repeatability estimates.
• Accurate phenotypic variance estimation is crucial for biologically meaningful genetic parameter interpretation.

Purpose of the Study:

• To propose an alternative method for estimating heritability and repeatability using mixed-effects models.
• To ensure phenotypic variance accurately reflects the relevant ecological setting by including fixed effects.
• To address challenges in estimating heritability with heritable covariates and in nonlinear/generalized mixed models.

Main Methods:

• Fit linear mixed-effects models incorporating all relevant effects, including variance explained by fixed effects.
• Include fixed effects variance in the phenotypic variance estimation for unbiased heritability and repeatability.
• Discuss integration over fixed effects for nonlinear and generalized mixed models.

Main Results:

• The proposed method allows for optimizing phenotypic variance estimation by including biologically relevant variation.
• This approach avoids upward bias in heritability and repeatability estimates often caused by omitting fixed effects.
• Provides a framework for interpreting heritability with heritable covariates and in complex mixed models.

Conclusions:

• Including variance explained by fixed effects in mixed-effects models provides more accurate heritability and repeatability estimates.
• This method enhances the biological relevance of quantitative genetic parameters by using a more complete phenotypic variance.
• The approach is adaptable for complex models, including nonlinear and generalized mixed models.