Rxivist logo

A robust and efficient method for Mendelian randomization with hundreds of genetic variants: unravelling mechanisms linking HDL-cholesterol and coronary heart disease

By Stephen Burgess, Christopher N Foley, Elias Allara, James R. Staley, Joanna MM Howson

Posted 04 Mar 2019
bioRxiv DOI: 10.1101/566851

Mendelian randomization uses genetic variants as instrumental variables to make causal claims based on observational data. When genetic variants are not all valid instrumental variables, the validity of findings from a Mendelian randomization investigation is not guaranteed. This problem is becoming more and more common as the number of genetic variants discovered in genome-wide association studies increases. We here introduce the contamination mixture method as a robust and efficient method for Mendelian randomization. Compared to other robust methods, it had the best performance in a simulation study in terms of mean squared error across a range of realistic scenarios, and gave estimates with low bias and low Type 1 error rate inflation when up to 40% of the genetic variants were invalid instruments. The method has linear computational time in the number of genetic variants, and so can perform analysis with hundreds of variants in a fraction of a second. The method can also identify when there are multiple groups of genetic variants with similar causal effect estimates, which may represent different mechanisms by which the risk factor influences the outcome. We demonstrated the presence of two such groups in a Mendelian randomization analysis for high-density lipoprotein (HDL) cholesterol and coronary heart disease (CHD) risk, and showed that 11 variants associated with increased HDL-cholesterol, increased triglyceride levels, and decreased CHD risk had the same pattern of associations with platelet distribution width and other blood cell traits, suggesting a shared mechanism linking lipids and CHD risk relating to platelet aggregation.

Download data

  • Downloaded 1,022 times
  • Download rankings, all-time:
    • Site-wide: 21,137
    • In genetics: 1,009
  • Year to date:
    • Site-wide: 75,374
  • Since beginning of last month:
    • Site-wide: 69,449

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide