MRBEE: A novel bias-corrected multivariable Mendelian Randomization method.

Area of Science:

• Genetics and Epidemiology
• Statistical Genetics
• Biostatistics

Background:

• Mendelian Randomization (MR) is a key method for inferring causality between exposures and outcomes using genetic variants.
• Existing MR approaches are susceptible to biases such as weak instruments, sample overlap, and measurement error.
• The increasing availability of genome-wide association study (GWAS) summary statistics necessitates robust methods for causal inference.

Approach:

• Introduce MRBEE, a computationally efficient multivariable MR method designed to correct for multiple known biases concurrently.
• Validate MRBEE through theoretical derivations, simulations, and real-world data analyses.
• Compare MRBEE's performance against existing MR methods to highlight its advantages.

Key Points:

• MRBEE simultaneously corrects for weak instruments, sample overlap, and measurement error in multivariable MR analyses.
• Analysis revealed that Body Mass Index (BMI) influences coronary artery disease risk exclusively through blood pressure.
• MRBEE identified a significantly stronger causal effect of cannabis use disorder on schizophrenia risk compared to traditional MR methods.

Conclusions:

• MRBEE provides a robust and unbiased approach for causal inference in the context of multiple risk factors and disease outcomes.
• The method demonstrates superior accuracy in estimating causal effects, particularly in complex genetic epidemiological studies.
• MRBEE is a valuable tool for leveraging large-scale GWAS data to understand disease etiology.