constrained IVs

Method to construct valid instrumental variables (IVs) and obtain unbiased and adjusted causal effect estimates when pleiotropy exists and when pleiotropic phenotypes are available.

Whilst aiming to mathematically maximise IV strength, the constrained IV method aims to construct a set of valid IVs that are strongly related to the exposure of interest, where the association between the IVs and phenotypes that are independently associated with the outcome (i.e., pleiotropic phenotypes) is zero. This requires the ability to estimate the relationship between each IV and both the exposure and all possible pleiotropic phenotypes. Once the constrained IV is created, conventional MR methods can be applied.

References

• Jiang L, Oualkacha K, Didelez V, Ciampi A, Rosa-Neto P, Benedet AL, et al. Constrained instruments and their application to Mendelian randomization with pleiotropy. Genetic Epidemiology 2019; 43: 373–401

Other terms in 'Pleiotropy-robust one-sample MR methods':

• Adaptive Lasso
• MR for gene-environment interactions (MR-GxE)
• MR using G-Estimation under No Interaction with Unmeasured Selection (MR-GENIUS)
• Over-identification tests
• Pleiotropy-robust MR (PRMR)
• Some Invalid Some Valid Instrumental Variable Estimation (SISVIVE)
• Two-stage hard thresholding (TSHT) with voting