A Regression-based Approach to Robust Estimation and Inference for Genetic Covariance

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) identify genetic variants linked to complex traits.
• Some variants influence multiple traits, indicating shared genetic architecture.
• Estimating genetic covariance quantifies these shared genetic effects.

Purpose of the Study:

• To develop a unified, robust regression-based method for estimating genetic covariance between general traits.
• To handle nonlinear associations between genetic variants and complex traits.
• To provide robust inference for narrow-sense genetic covariance, even with model mis-specification.

Main Methods:

• A unified regression-based approach for robust estimation and inference of genetic covariance.
• Asymptotic properties of the estimator are derived, demonstrating robustness.
• The method is validated using numerical experiments and applied to mouse GWAS data.

Main Results:

• The proposed method provides robust estimation and inference for genetic covariance of general traits.
• The method is effective even when genetic associations are nonlinear or working models are mis-specified.
• Application to mouse GWAS data revealed significant genetic covariance among developmental traits.

Conclusions:

• The developed regression-based method offers a robust tool for analyzing shared genetic architecture across complex traits.
• This approach enhances understanding of genetic covariance, particularly in the presence of nonlinear genetic effects.
• The findings highlight the utility of the method in uncovering genetic relationships in real-world biological data, such as mouse developmental phenotypes.