Synonyms: Weighted mode
Defined as the mode of either an unweighted or inverse variance weighted (IVW) empirical density function of the Wald ratio (or ratio estimates).
This two-sample MR method requires that the largest subset of instrumental variables (IVs) which estimate the same causal effect to be valid IVs. The estimates can be either unweighted or, traditionally, weighted by the inverse variance of the association between the single nucleotide polymorphism (SNP) and the outcome but may be weighted using other methods. This relies on the additional assumption that the largest number of similar individual IV estimates comes from valid IVs, which is known as the "Zero Modal Pleiotropy Assumption (ZEMPA)". See NOME adjustment for more information on these types of weights.
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
- 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
- MR-Path
- Multivariable MR (MVMR) and extensions
- Welch-weighted Egger regression (WWER)