Selecting causal risk factors from high-throughput experiments using multivariable Mendelian randomization

Study on multivariate Mendelian Randomization (MR) based on Bayesian model that scales to high-throughput and can select biomarkers as risk factors for disease.

In principle, multivariate MR makes the assumption that genetic variants influence a set of multiple risk factors and thus accounts for measured pleiotrophy. The main goal of this approach is risk factor selection, not the precise estimation of causal effects.