A biobank-scale method for learning modulators of gene-environment interaction underlying human complex traits from...

Area of Science:

• Genetics
• Environmental Health
• Computational Biology

Background:

• Genetic effects on complex traits are influenced by environmental context.
• Large-scale biobanks offer opportunities to study gene-environment (G×E) interactions.
• Current methods struggle to model polygenic G×E with numerous exposures.

Purpose of the Study:

• To develop a method for accurately modeling polygenic G×E interactions with multiple environmental exposures.
• To distinguish true G×E effects from environmental noise.
• To enable efficient G×E analysis at biobank scale.

Main Methods:

• Developed Efficient multi-eNvironmental Gene-environment Interaction iNference Estimator (ENGINE), a supervised variance-component framework.
• ENGINE learns an environmental embedding combining multiple exposures.
• The method efficiently processes large genotype data for biobank-scale inference.

Main Results:

• Simulations demonstrate ENGINE's control of type I error rates and high statistical power.
• ENGINE accurately recovers environmental embeddings and is computationally efficient.
• Application to UK Biobank data revealed significantly larger G×E variance compared to single exposures or principal components.

Conclusions:

• ENGINE provides a powerful and efficient framework for analyzing complex G×E interactions in large biobanks.
• The method enhances the ability to identify genetic influences on complex traits shaped by multiple environmental factors.
• This approach advances our understanding of the interplay between genes and environment in human health.