mrrobust

mrrobust performs two-sample summary Mendelian randomization analyses using GWAS summary statistics to estimate and assess causal effects in epidemiologic studies.

Key Features:

• Inverse Variance Weighted Estimation: Provides estimates by weighting each genetic variant's effect size inversely proportional to its variance.
• Median and Modal Estimations: Offers median-based and mode-based estimators that are more robust to outlying instruments.
• MR-Egger Estimation Methods: Includes MR-Egger regression techniques for detecting and adjusting for pleiotropy across genetic variants.
• Visualization Tools: Constructs plots to visualize individual and simultaneous estimates for interpretation.
• Statistical Assessments: Calculates statistics such as IGX2 to assess attenuation bias in causal estimates.

Scientific Applications:

• Causal inference from GWAS summary data: Estimate causal relationships between exposures and outcomes in epidemiologic studies using GWAS summary statistics.
• Pleiotropy detection and adjustment: Identify and adjust for pleiotropy where genetic variants influence multiple traits using MR-Egger methods.
• Robust causal estimation: Obtain causal estimates resilient to outlying instruments using median and modal estimators.

Methodology:

Implements two-sample summary Mendelian randomization via inverse-variance weighted estimation, median and modal estimators, MR-Egger regression, plotting of estimates, and computation of IGX2 to assess attenuation bias.