Co-expression-wide association studies link genetically regulated interactions with complex traits

Area of Science:

• Genetics and Systems Biology
• Computational Biology
• Disease Risk Prediction

Background:

• Transcriptome- and proteome-wide association studies (TWAS/PWAS) identify genes/proteins linked to disease risk.
• Existing methods overlook co-expression and interaction effects between genes or proteins.
• Understanding molecular networks is crucial for complex trait genetics.

Purpose of the Study:

• Introduce the co-expression-wide association study (COWAS) method.
• Identify gene/protein pairs with genetically regulated co-expression associated with complex traits.
• Investigate the role of protein-protein interactions in disease etiology.

Main Methods:

• COWAS trains models to predict expression and co-expression from genetic variation.
• Tests association between imputed co-expression and traits, accounting for direct effects.
• Applies the method to UK Biobank plasma proteomic data.

Main Results:

• Identified dozens of interacting protein pairs associated with cholesterol levels, Alzheimer's, and Parkinson's diseases.
• Demonstrated that protein co-expression can influence complex traits independently of individual protein effects.
• Showcased COWAS's ability to disentangle direct and interaction effects in molecular networks.

Conclusions:

• COWAS provides a novel approach to uncover genetically regulated co-expression patterns associated with complex traits.
• Protein-protein interactions revealed by COWAS offer new insights into disease mechanisms.
• The method enhances understanding of molecular networks mediating genetic effects on disease.