Incorporating prior biologic information for high-dimensional rare variant association studies

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Rare variant association studies are growing in scale, necessitating advanced analytical methods.
• High-dimensional genetic data presents challenges for traditional association analyses.

Purpose of the Study:

• To develop a novel Bayesian risk index method for high-dimensional rare variant association studies.
• To integrate multiple data sources and allow for region-specific risk indices.

Main Methods:

• The method extends existing Bayesian risk indices by incorporating external biological variant-specific covariates.
• A second level of uncertainty is included to identify associated regions.

Main Results:

• Simulations demonstrate increased power to detect true associations compared to existing methods.
• The approach enables multi-level inference at pathway, region, and variant levels.

Conclusions:

• The method effectively integrates diverse data types for high-dimensional genetic studies.
• It is applicable to candidate gene and pathway studies, as shown in breast and colon cancer analyses.