TwoStepCisMR

TwoStepCisMR implements a two-step cis-Mendelian randomization approach to attenuate bias from linkage disequilibrium in causal inference of genetic variants' effects on outcomes.

Key Features:

• R package: Implemented as an R package.
• Bias Attenuation: Reduces confounding bias induced by linkage disequilibrium when genetic variants within gene regions are used as instrumental variables.
• Integration of Methods: Combines Mendelian randomization with the product and difference method for mediation analysis.
• Addresses multivariable MR limitations: Targets scenarios where multivariable Mendelian randomization is limited by high correlation among genetic variants within gene regions.
• Simulation and Application: Developed and evaluated through simulation experiments and applied examples, including assessing interleukin 6 receptor signaling on coronary artery disease risk.

Scientific Applications:

• Drug target validation: Provides quasi-experimental evidence for on-target drug effects by estimating causal effects of genetic variants on drug target proteins.
• Pharmacogenomics: Strengthens causal inference in pharmacogenomics and related fields by mitigating linkage disequilibrium–induced bias in cis-MR.
• cis-MR studies: Enhances the reliability of causal estimates and conclusions drawn from cis-Mendelian randomization analyses.

Methodology:

Uses genetic variants within a gene region as instrumental variables, integrates Mendelian randomization with the product and difference method for mediation analysis to address confounding from linkage disequilibrium, and was developed and evaluated via simulation experiments and applied examples such as interleukin 6 receptor signaling on coronary artery disease risk.