FA-MFMR: a multivariable functional Mendelian randomization method that accounts for correlated longitudinal exposures via factor augmentation.

FA-MFMR accurately estimates time-varying causal effects of multiple correlated longitudinal exposures, reducing estimation error and improving stability in biomedical studies.

• Why it matters: Understanding causal relationships over time with correlated exposures is crucial for biomedical research, but existing MR methods struggle with longitudinal data, multicollinearity, and sparse effects, limiting insight.
• What they did: The authors developed FA-MFMR, an approach that combines genetic instruments, functional exposure trajectories, and a low-dimensional factor structure, tested through simulations and applied to Parkinson's disease biomarkers.
• The result: FA-MFMR demonstrated lower estimation error and more efficient uncertainty quantification than competing methods, revealing age-dependent exposure effects and localized uncertainties in a Parkinson's cohort.