Diagnostics for Pleiotropy in Mendelian Randomization Studies: Global and Individual Tests for Direct Effects

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Pleiotropy, where a single gene influences multiple traits, can bias Mendelian randomization (MR) analyses.
• The MR-Egger method is commonly used to detect pleiotropy but may lack power.
• Accurate assessment of pleiotropy is crucial for validating MR study findings.

Purpose of the Study:

• To introduce a novel statistical method, global and individual tests for direct effects (GLIDE), for systematically evaluating pleiotropy in MR.
• To compare the performance of GLIDE against existing methods like MR-Egger.
• To demonstrate the utility of GLIDE in refining genetic instrumental variables and improving causal effect estimation.

Main Methods:

• Development and simulation testing of the GLIDE statistical method for detecting pleiotropy.
• Comparison of GLIDE's power and sensitivity against the MR-Egger method through simulations.
• Application of GLIDE as a sensitivity analysis tool to identify and remove pleiotropic single nucleotide polymorphisms (SNPs) in real-world MR studies.

Main Results:

• Simulation experiments indicate GLIDE possesses superior power compared to the MR-Egger method for detecting pleiotropy.
• GLIDE effectively identifies outlier genetic variants exhibiting pleiotropy, enabling the refinement of instrumental variable sets.
• Analysis of body mass index and height associations with colorectal cancer risk revealed pleiotropic SNPs, and their removal attenuated causal effect estimates.

Conclusions:

• The GLIDE method provides a robust approach for diagnosing pleiotropy in MR studies.
• GLIDE enhances the validity of MR analyses by serving as an effective sensitivity analysis tool.
• Utilizing GLIDE can lead to more accurate estimations of causal effects by refining the selection of genetic instrumental variables.