Polygenic approaches to detect gene-environment interactions when external information is unavailable

Area of Science:

• Genetics and Bioinformatics
• Epidemiology
• Statistical Genomics

Background:

• Gene-environment interactions (G × E) are crucial for understanding disease etiology and developing personalized prevention strategies.
• Current methods often rely on external genome-wide association study (GWAS) data to construct genetic risk scores (GRS), which may not be available for all populations, particularly non-Caucasian ethnicities.
• The performance of GRS in detecting G × E without external information remains unclear.

Purpose of the Study:

• To explore and compare the power of different polygenic approaches for detecting gene-environment interactions (G × E) in the absence of external GWAS data.
• To evaluate the efficacy of the adaptive combination of Bayes factors method (ADABF) against GRS based on marginal SNP effects (GRS-M) and GRS based on SNP × E interactions (GRS-I).
• To provide guidance on selecting appropriate polygenic methods for G × E analysis when external information is limited.

Main Methods:

• Comparison of three polygenic methods: ADABF, GRS-M, and GRS-I, for testing G × E.
• Simulation studies to assess the power of each method under different scenarios, including the presence or absence of single-nucleotide polymorphism (SNP) main effects.
• Application of the methods to real-world data from the Taiwan Biobank to identify G × E on blood pressure (BP).

Main Results:

• ADABF demonstrated the highest power for detecting G × E in the absence of SNP main effects.
• GRS-M was generally the most powerful method when SNP main effects were present.
• GRS-I exhibited the lowest power due to its data-splitting strategy.
• Analysis of Taiwan Biobank data revealed significant gene × alcohol and gene × smoking interactions affecting blood pressure, where BP-increasing alleles had a greater impact in individuals who consumed alcohol or smoked.

Conclusions:

• The choice of polygenic method for detecting G × E without external information depends on the presence or absence of SNP main effects.
• ADABF is a powerful alternative when external GWAS data is unavailable, especially when SNP main effects are minimal.
• The findings highlight the importance of considering G × E in blood pressure regulation and offer practical guidance for genetic association studies.