Mendelian Randomization using Public Data from Genetic Consortia.

Area of Science:

• Epidemiology
• Statistical Genetics
• Biostatistics

Background:

• Mendelian randomization (MR) is a robust instrumental variable method for causal inference in observational studies.
• Publicly available genome-wide association study (GWAS) meta-analysis data enables large-scale MR analyses.

Purpose of the Study:

• To investigate the properties of Mendelian randomization using publicly available GWAS data.
• To evaluate the causal effect of birth weight on adult glucose levels using MR.
• To clarify the distinction between using estimated versus true coefficients for genetic risk scores in MR.

Main Methods:

• Utilized publicly available genome-wide association study meta-analysis results.
• Employed simulation studies to examine the statistical properties of the MR approach.
• Applied MR to assess the relationship between birth weight and adult glucose levels.

Main Results:

• Demonstrated the utility of MR for causal inference with public genetic data.
• Highlighted the importance of using true underlying coefficients for genetic risk scores in MR.
• Showed that methods based on estimated coefficients can underestimate variance and require careful correction to avoid bias.

Conclusions:

• Mendelian randomization using public data is a feasible and accessible method for causal inference.
• Researchers must exercise caution to prevent false precision and bias, particularly when correcting for estimated coefficients.
• Accurate estimation of variance is crucial for reliable causal effect estimates in MR.