Adjusting for medication status in genome-wide association studies.

GEMRA reduces bias and increases statistical power in genome-wide association studies by empirically adjusting phenotypes of medicated participants, outperforming common medication-handling strategies in simulations.

• Why it matters: Medication use can bias genetic effect estimates in GWAS, and standard fixes may leave residual distortion or lose power.
• What they did: GEMRA applies a heuristic search for an empirical medication adjustment to phenotypes, and simulations compare it against covariate adjustment, unmedicated-only restriction, and no adjustment.
• The result: In UK Biobank and the Uganda Genome Resource, relative adjustments fit LDL cholesterol, SBP, and DBP best, with consistent reductions of 40%, 10%, and 10%, and misspecified effect scales can create spurious cross-population differences.

AI summary of the abstract · may contain errors · not medical advice · read the paper.