Powerful multi-marker association tests: unifying genomic distance-based regression and logistic regression

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Case-control studies utilize statistical tests to detect genetic associations with complex diseases.
• Multi-marker association tests offer increased power over single-marker tests, especially with linkage disequilibrium (LD).
• Existing methods often focus on limited distributional aspects, necessitating new approaches for comprehensive analysis.

Purpose of the Study:

• To reformulate genomic distance-based regression (GDBR) as a logistic regression model.
• To develop a unified framework integrating GDBR with existing statistical methods.
• To enhance the power and flexibility of genetic association testing.

Main Methods:

• Reformulation of GDBR within a logistic regression framework.
• Utilizing asymptotic distributions for P-value calculation, replacing permutations.
• Incorporating covariates such as gene-gene interactions.

Main Results:

• GDBR is successfully integrated into logistic regression, creating a unified analytical framework.
• The reformulated approach allows for efficient P-value derivation and covariate inclusion.
• Fisher's P-value combining method boosts power by integrating allele frequencies, LD patterns, and higher-order interactions.

Conclusions:

• The reformulation of GDBR offers a powerful and flexible approach for genetic association studies.
• This unified framework overcomes limitations of GDBR and logistic regression, enabling combined analyses.
• The method enhances statistical power by leveraging diverse genetic information for disease detection.