AGGrEGATOr: A Gene-based GEne-Gene interActTiOn test for case-control association studies.

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Gene-gene interactions are crucial for understanding complex diseases.
• Existing gene-based methods in genome-wide association studies (GWAS) have limitations in detecting cumulative SNP-SNP signals.
• There is a need for powerful and interpretable methods to identify gene interactions.

Purpose of the Study:

• To develop a novel gene-based statistical method, AGGrEGATOr, for identifying gene-gene interactions in GWAS.
• To evaluate the statistical power, type-I error control, and robustness of AGGrEGATOr compared to existing methods.
• To apply AGGrEGATOr to identify and replicate gene pairs associated with rheumatoid arthritis (RA).

Main Methods:

• AGGrEGATOr employs a minP procedure, testing single nucleotide polymorphism (SNP)-SNP interactions first, then aggregating p-values for gene-level testing.
• Simulations were conducted to assess type-I error control and statistical power across various disease models.
• The method was applied to the GSE39428 dataset for RA and validated using the Wellcome Trust Case Control Consortium dataset.

Main Results:

• AGGrEGATOr demonstrated robust type-I error control and superior statistical power in simulations.
• The analysis of the GSE39428 dataset identified 13 potential gene-gene interactions for RA, with one replicated.
• AGGrEGATOr successfully replicated seven previously reported gene pairs associated with RA, Crohn's disease, or coronary artery disease in the validation dataset.

Conclusions:

• AGGrEGATOr offers an effective approach for detecting gene-gene interactions in GWAS, outperforming existing methods.
• The method provides enhanced statistical power and biological interpretability for complex disease genetics.
• AGGrEGATOr holds promise for advancing the discovery of genetic architectures underlying complex diseases.