The paper by Helgadottir et al highlights the importance and need to develop sensitivity analyses for multivariable MR.10 This is particularly relevant given the recent advances in high‐throughput phenotyping which has led to the introduction of “‐omics” data such as metabolomics, genomics, and proteomics.31 Genome‐wide analyses of high‐dimensional “‐omics” data are becoming more popular,32, 33 yet few MR analyses have been performed using these datasets.21 As summarized data from large consortia become more accessible, the opportunities to use MR on high‐dimensional datasets will only increase. Methods such as multivariable MR‐Egger will be valuable to investigate the causal effects of multiple related phenotypes with shared genetic predictors.
