A novel framework with automated horizontal pleiotropy adjustment in mendelian randomization.

Area of Science:

• Genetics
• Epidemiology
• Statistical Genetics

Background:

• Horizontal pleiotropy, where genetic variants affect an outcome through multiple pathways, poses a significant challenge in Mendelian randomization (MR) studies.
• Existing MR methods often struggle to fully account for and mitigate the bias introduced by horizontal pleiotropy.

Purpose of the Study:

• To propose and evaluate a novel two-stage framework, Conditional MR (CMR), designed to address horizontal pleiotropy in MR analyses.
• To demonstrate CMR's ability to reduce pleiotropic bias and enhance the accuracy of causal effect estimates.

Main Methods:

• The CMR framework integrates a conditional analysis of multiple genetic variants to remove linkage disequilibrium-induced pleiotropy.
• Subsequent application of robust MR methods to model the adjusted genetic effect estimates.

Main Results:

• Extensive simulations across diverse horizontal pleiotropy scenarios show CMR outperforms standard MR methods that model marginal genetic effects.
• CMR effectively reduces horizontal pleiotropy, leading to improved performance of existing MR techniques.

Conclusions:

• The proposed CMR framework offers a more reliable approach for causal inference in the presence of horizontal pleiotropy.
• CMR demonstrates potential for delivering robust results in real-world genetic epidemiology applications, such as investigating the causal role of body mass index in diseases.