Unified Sequence-Based Association Tests Allowing for Multiple Functional Annotations and Meta-analysis of Noncoding...

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Functional annotation of human genetic variation is crucial for identifying disease-associated variants.
• Existing methods integrating functional annotations can lose power if annotations are not risk-predictive.
• Discovering variants outside protein-coding regions remains challenging.

Purpose of the Study:

• To develop unified statistical tests for integrative association analysis using multiple functional annotations.
• To improve the power of genetic association studies, especially for noncoding variants.
• To enable meta-analysis of sequencing data using summary statistics.

Main Methods:

• Development of unified statistical tests for integrative association analysis.
• Utilizing multiple functional annotations simultaneously with efficient computational techniques.
• Application to meta-analysis of noncoding rare variants in Metabochip data for lipid traits.

Main Results:

• The proposed unified tests significantly improve power when functional annotations predict variant risk status.
• These tests incur minimal power loss when annotations are not predictive and can improve power in some cases.
• A meta-analysis using the unified tests identified significant associations between noncoding rare variants in SLC22A3 and lipid traits, which were missed by standard tests.

Conclusions:

• Unified tests offer a powerful approach for integrative genetic association analysis, leveraging multiple functional annotations.
• These methods enhance the discovery of disease-associated variants, particularly noncoding rare variants.
• The approach facilitates large-scale meta-analyses without requiring individual-level data sharing.