Latent mixture MR model that exploits effect heterogeneity and clusters instrumental variables (IVs) together based on the similarity of their causal effect estimates using summary-level data in a two-sample MR setting.
Effect heterogeneity is the phenomenon whereby individual genetic variants being used as IVs have a differential influence on the exposure. Such heterogeneity can be due to the validity of MR assumptions (e.g., violations of the independence or exclusion retriction assumptions) or due to genuinely different biological mechanisms by which the genetic variant on the exposure. MR-Path uses a Monte-Carlo expectation-maximization algorithm to fit a transparent mixture model that maximizes a likelihood function to capture mechanistic heterogeneity. This model assumes that each genetic variant being used as an IV has a specific causal effect and the genetic variants on the same biological pathway have similar causal effects and form clusters. The mean of each cluster corresponds to the Wald ratio estimand of that pathway. See also MR-Clust, as similar methodology.
References
Other terms in 'Pleiotropy-robust two-sample MR methods':
- Bayes MR
- Bayesian implementation of the MR-Egger Estimator (BMRE)
- Bayesian multi-instrument Mendelian randomization (MIMR)
- Bayesian network analysis
- Causal Analysis Using Summary Effect estimates (MR-CAUSE)
- Contamination mixture models
- Generalized Summary MR (GSMR)
- Genetic Instrumental Variable (GIV)
- Hierarchical joint Analysis of Marginal summary statistics (hJAM)
- Inverse variance weighted (IVW) random effects model
- Iterative Mendelian Randomization and Pleiotropy (IMRP)
- Leave-one-out analysis
- Median-based estimate
- Mode-based estimate
- MR accounting for Correlated and Idiosyncratic Pleiotropy (MRCIP)
- MR accounting for Linkage Disequilibrium and Pleiotropy (MR-LDP)
- MR Lasso
- MR Mixture (MRMix)
- MR using Robust regression (MR Robust)
- MR with penalized weights
- MR with regularization
- MR-Clust
- MR-Egger regression and extensions
- MR-Link
- Multivariable MR (MVMR) and extensions
- Welch-weighted Egger regression (WWER)