Meta-analysis of SNP-environment interaction with heterogeneity for overlapping data

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Meta-analysis is crucial for genome-wide association studies (GWAS), but generates heterogeneity.
• Existing methods include random effect model meta-regression (MR) for SNP-environment interactions and fixed effect model overlapping MR for overlapping data.

Purpose of the Study:

• To propose a novel random effect model overlapping MR method.
• To simultaneously address heterogeneity and overlapping data in GWAS meta-analysis.
• To provide a new method for solving determinant calculations in likelihood functions.

Main Methods:

• Development of a random effect model overlapping MR method, integrating existing MR and overlapping MR approaches.
• Implementation of a new technique for calculating the logarithm of the determinant of covariance matrices.
• Application of likelihood ratio statistics for testing SNP-environment interaction and joint effects.

Main Results:

• Simulations confirmed the method's suitability by evaluating null distributions and type I error rates.
• The proposed method demonstrated superior power in detecting SNP-environment interactions.
• The method proved effective for datasets with high heterogeneity and overlapping data.

Conclusions:

• The novel random effect model overlapping MR method is effective for GWAS meta-analysis with complex data structures.
• This approach enhances the power to detect SNP-environment interactions under heterogeneity and data overlap.
• The method offers a robust statistical framework for genetic association studies.