FLOW-MR Published in Nature Communications

Our paper, Causal mediation analysis for time-varying heritable risk factors with Mendelian Randomization, has been published in Nature Communications:

👉 Nature Communications doi: 10.1038/s41467-025-61648-7
👉 Original preprint: bioRxiv 2024.02.10.579129

The paper introduces FLOW-MR, a computational framework for estimating causal structural equations for temporally ordered traits using only GWAS summary statistics. FLOW-MR enables decomposition of total genetic effects into:

• direct effects
• indirect mediation effects
• pathway-specific causal contributions

Key methodological innovations include:

• handling correlated longitudinal risk factors with limited GWAS sample sizes,
• improved stability and efficiency under strong polygenicity and weak instruments through a spike-and-slab prior, and
• robust inference even in noisy settings.

Using FLOW-MR, we identify a childhood-specific protective effect of BMI on breast cancer, and analyze the evolving causal impacts of BMI, systolic blood pressure, and cholesterol on stroke risk over the life course.