Eric S. Kawaguchi

The paper introduces pathway-based polygenic risk scores (pPRS) to detect gene-environment interactions, finding a significant interaction between NSAIDs and the TGF-β/GRHR pathway in colorectal cancer.

The paper introduces a regularized hierarchical Cox model that integrates meta-features to improve prediction and feature selection for time-to-event outcomes in omic studies.

The paper improves two-step tests for gene-environment interactions by addressing displacement issues and multiple testing, showing increased power in simulations and a GxSex interaction in colorectal cancer.