A two-stage testing strategy for detecting genes×environment interactions in association studies

Area of Science:

• Statistical Genetics
• Genomics
• Computational Biology

Background:

• Identifying gene-environment interactions (G×E) is crucial for understanding complex diseases.
• Detecting G×E interactions involving rare variants remains a significant challenge in genome-wide association studies.
• Current statistical methods for G×E interaction detection have limitations, especially with rare variants.

Purpose of the Study:

• To develop a novel statistical method for detecting G×E interactions, specifically addressing the challenge of rare variants.
• To extend the adaptive combination of P-values (ADA) method to create an improved strategy for G×E interaction analysis.
• To introduce a two-stage testing approach for comprehensive G×E interaction analysis across genomic regions.

Main Methods:

• Proposed a novel two-stage testing strategy named iSADA.
• Utilized score statistics in the first stage to obtain preliminary P-values for trait value and gene-environment interaction terms.
• Constructed a full test statistic in the second stage by adaptively combining P-values from the first stage, inspired by the ADA method.
• Evaluated the method's performance using simulation studies and the GAW17 dataset.

Main Results:

• The iSADA method demonstrated higher statistical power compared to existing methods across various scenarios.
• The simulation studies confirmed the effectiveness and improved performance of iSADA.
• The GAW17 dataset analysis illustrated the practical applicability of the iSADA method.

Conclusions:

• The iSADA method provides a powerful and effective approach for detecting G×E interactions, especially when rare variants are involved.
• This novel strategy advances the field of statistical genetics by improving the analysis of complex disease etiology.
• iSADA offers a valuable tool for genomic research, enhancing the ability to identify G×E effects in large-scale studies.