MR data challenge

This Data Challenge was developed for “The Future of Mendelian Randomization Studies 2021” workshop.

It aims to highlight some of the methodological challenges of inferring causal effects from real-world data using Mendelian Randomization (MR) and provide a concrete anchor for discussions of these complexities among conference attendees. In particular, we encourage participants to consider issues of selecting potential instruments, clearly defining causal estimands, and timing.

On the corresponding GitHub page, you’ll find the following files needed to complete the challenge:

• Instructions and prompts for the data challenge
• Two datasets: one corresponding to Part 2 of the data challenge and one corresponding to Part 3 of the data challenge
• Data dictionary describing the variables in each of the datasets