Skip to content

usage-guide.md

Latest commit

 

History

History
75 lines (49 loc) · 2.76 KB

usage-guide.md

File metadata and controls

75 lines (49 loc) · 2.76 KB

Pleiotropy Detection - Usage Guide

Overview

Detect and correct for horizontal pleiotropy in Mendelian randomization analyses. Implements MR-PRESSO for outlier removal, MR-Egger regression for directional pleiotropy detection, and Steiger filtering for instrument directionality.

Prerequisites

install.packages('remotes')
remotes::install_github('MRCIEU/TwoSampleMR')
remotes::install_github('rondolab/MR-PRESSO')

# For MR-RAPS (archived from CRAN March 2025)
install.packages('MendelianRandomization')

Quick Start

Tell your AI agent what you want to do:

  • "Run sensitivity analyses on my MR results to check for pleiotropy"
  • "Use MR-PRESSO to detect and remove outlier instruments"
  • "Check if the MR-Egger intercept is significant for my analysis"

Example Prompts

MR-PRESSO

"Run MR-PRESSO on my harmonized MR data and identify outlier SNPs"

"Remove pleiotropic outliers and compare the corrected estimate to the original"

MR-Egger

"Test for directional pleiotropy using the Egger intercept"

"What is the I-squared for my MR-Egger analysis -- is the NOME assumption satisfied?"

Comprehensive Sensitivity

"Run all standard MR sensitivity analyses: IVW, Egger, weighted median, MR-PRESSO, and leave-one-out"

"Generate a summary table comparing all MR methods and pleiotropy diagnostics"

Steiger Filtering

"Filter my instruments using the Steiger test and re-run MR"

"Are any of my instruments acting through a reverse causal pathway?"

What the Agent Will Do

  1. Run MR-PRESSO global test for horizontal pleiotropy
  2. Identify and remove outlier instruments
  3. Test MR-Egger intercept for directional pleiotropy
  4. Compute I-squared for NOME assumption assessment
  5. Apply Steiger filtering for instrument directionality
  6. Compare causal estimates across robust methods
  7. Report comprehensive sensitivity analysis summary

Tips

  • MR-PRESSO sims - Use at least 5000 distributions for stable p-values; 1000 is a minimum
  • Egger intercept - Low power with fewer than 10 instruments; non-significant does not mean no pleiotropy
  • I-squared - Below 0.9 indicates NOME violation; Egger estimate may be biased toward null
  • Method agreement - Consistent results across IVW, Egger, median, and MR-PRESSO strengthen causal claims
  • MR-RAPS - The CRAN package was archived March 2025; use the MendelianRandomization package instead
  • STROBE-MR - Report all sensitivity analyses, not just favorable results
  • Outlier removal - After removing MR-PRESSO outliers, re-check heterogeneity

Related Skills

  • causal-genomics/mendelian-randomization - Primary MR analysis workflow
  • causal-genomics/fine-mapping - Prioritize causal variants at instrument loci
  • population-genetics/association-testing - GWAS for instrument selection