Inferring causal networks for genomic trios with high-dimensional confounding variables.

MRGN infers causal networks between molecular phenotypes in genomic trios under PMR while accounting for many confounders, improving inference accuracy and robustness.

• Why it matters: Existing PMR-based causal structure inference struggles with high-dimensional confounding, limiting reliable edge direction and detection when marginal associations are weak.
• What they did: The regression-based MRGN method extends the canonical mediation model, using conditional independence tests given the genetic variant and confounders to infer causal models from individual-level trio data.
• The result: Simulations show improved inference accuracy, robustness, and computational efficiency versus existing approaches, and an application to GTEx data examines cis and trans genetic regulation of gene expression.

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