Likelihood-based Mendelian randomization analysis with automated instrument selection and horizontal pleiotropic modeling

Abstract

Mendelian randomization (MR) is a common tool for identifying causal risk factors underlying diseases. Here, we present a method, MR with automated instrument determination (MRAID), for effective MR analysis. MRAID borrows ideas from fine-mapping analysis to model an initial set of candidate single-nucleotide polymorphisms that are in potentially high linkage disequilibrium with each other and automatically selects among them the suitable instruments for causal inference. MRAID also explicitly models both uncorrelated and correlated horizontal pleiotropic effects that are widespread for complex trait analysis. MRAID achieves both tasks through a joint likelihood framework and relies on a scalable sampling–based algorithm to compute calibrated P values. Comprehensive and realistic simulations show that MRAID can provide calibrated type I error control and reduce false positives while being more powerful than existing approaches. We illustrate the benefits of MRAID for an MR screening analysis across 645 trait pairs in U.K. Biobank, identifying multiple lifestyle causal risk factors of cardiovascular disease–related traits.

Statistical methods for Mendelian randomization in genome-wide association studies: A review

“…2 ). These methods include the inverse variance weighted MR [39] , MR-Egger [43] , weighted median estimation [60] , Bayesian weighted MR [61] , robust adjusted profile score [62] , MRMix [63] , CAUSE [64] , and MRAID [65] . These methods differ in their approaches to three considerations that include instrument selection (from among the available SNPs), modeling and controlling for horizontal pleiotropy, and statistical inference procedure.…”

“…In the short time since publication of [46] , researchers have recognized two types of horizontal pleiotropy [65] . The first occurs via exposure-independent paths.…”

“…It induces correlation between horizontal pleiotropic effects and SNP-exposure effects. Both types of horizontal pleiotropy violate standard MR modeling assumptions and can bias causal effect estimates and can increase false discoveries [65] . Early MR analyses avoided confounding from horizontal pleiotropy by discarding instrumental SNPs that might be associated with the outcome [46] , [47] .…”

“…Early MR analyses avoided confounding from horizontal pleiotropy by discarding instrumental SNPs that might be associated with the outcome [46] , [47] . More recent methods have attempted to model horizontal pleiotropy [64] , [65] , [85] . CAUSE [64] and MRMix [63] both use a mixture of normal distributions to control for both types of horizontal pleiotropy.…”

“…3 C). CAUSE and MRAID are two methods that aim to model correlated horizontal pleiotropy [64] , [65] . Their approaches to modeling correlated horizontal pleiotropy differ in important ways that ultimately affect performance.…”