Pathway-based genetic association analysis for overdispersed count data.

Area of Science:

• Genetics
• Biostatistics
• Bioinformatics

Background:

• Overdispersion is prevalent in genetic count data, like gene expression.
• Current pathway analysis methods are unsuitable for overdispersed count data.
• Investigating gene expression associations with genetic variants in pathways is crucial.

Purpose of the Study:

• To propose a novel hierarchical approach for analyzing overdispersed count data in genetic association studies.
• To develop methods for assessing associations between gene expression and low-frequency genetic variants within pathways.

Main Methods:

• Utilized negative binomial regression for overdispersed count responses.
• Derived score-type test statistics for fixed and random effects of genetic variants.
• Introduced a procedure for efficiently combining statistics for global testing.

Main Results:

• Simulation studies showed the proposed method is more powerful than existing approaches.
• The method effectively identified associations in a colorectal cancer study.
• Demonstrated improved power across various scenarios for genetic association analysis.

Conclusions:

• The proposed hierarchical approach offers a powerful tool for analyzing overdispersed genetic count data.
• This method enhances the identification of gene expression-pathway associations.
• Applicable to real-world genetic association studies, including cancer research.