Analysis code for: Genetic and multi-omic causal triage identifies triglyceride metabolism as an exploratory axis outside current treatment classes for heavy menstrual bleeding

Published: 23 June 2026| Version 1 | DOI: 10.17632/w86tfgnb22.1
Contributors:
,
,

Description

This repository contains the analysis code accompanying the manuscript "Genetic and multi-omic causal triage identifies triglyceride metabolism as an exploratory axis outside current treatment classes for heavy menstrual bleeding" submitted to eBioMedicine. SCOPE The code reproduces a prespecified two-sample Mendelian randomization (MR) causal-triage pipeline with bounded-null calibration across ten systemic exposures against two heavy menstrual bleeding (HMB) outcomes (FinnGen R12 menorrhagia, n_case = 28,642; GCST90483501, n_case = 31,309), followed by drug-target cis-MR with colocalization and safety phenome scanning at five triglyceride-network targets (APOC3, ANGPTL3, LPL, ANGPTL4, APOB), F2/prothrombin, and a calibration panel of six clinically-validated HMB targets (PLG, PGR, PTGS1, PTGS2, ESR1, GNRHR). Five multi-omic layers are integrated for tissue-of-action mapping: GTEx v8, Tabula Sapiens, GSE23339, STRING v12, and AutoDock Vina docking. CONTENTS - Scripts for harmonisation, IVW / MR-Egger / weighted-median MR, LCV, MVMR, colocalization, multi-omic integration, and figure generation. - A manifest mapping each script to the manuscript figure, table, or supplementary item it produces. - A README with environment specification (R / Python versions and package versions) and end-to-end run instructions. - SHA256 checksums for every shipped output (reference_outputs/) so downstream users can verify byte-identical reproduction. DATA NOT INCLUDED Raw GWAS summary statistics are NOT redistributed; each is cited by its original accession (FinnGen R12, GWAS Catalog GCST*, GTEx v8, eQTLGen, deCODE, UK Biobank-PPP) and downloaded from the source by the user via the provided fetch scripts. This complies with the data redistribution terms of each source consortium. USAGE Clone or download the archive, follow README.md to install the conda environment and R package set, run fetch_data.sh, then run run_pipeline.sh. Expected wall-clock on a 16-core workstation is approximately 6-8 hours end-to-end. CITATION If you use this code, please cite the accompanying manuscript (DOI to be assigned upon acceptance) and this dataset. CORRESPONDENCE Wei Yuan — rita.w.yuan@gmail.com, Fei Yang — finnyang@yangtzeu.edu.cn

Files

Institutions

Categories

Clinical Genetics, Bioinformatics, Clinical Endocrinology, Mendelian Randomization

Licence