Data and code for: Causal screening of the plasma proteome and skin transcriptome in keloid: six candidate loci and the limits of Mendelian randomisation for target nomination

# Data and code for "Causal screening of the plasma proteome and skin transcriptome in keloid: six candidate loci and the limits of Mendelian randomisation for target nomination" Version v1.0. This archive accompanies the manuscript above (Li Zhifeng, Li Jingyu; Zhejiang Provincial People's Hospital). It contains the analysis code, documentation and derived results. **It contains no individual-level data and does not redistribute the raw public source data.** ## Files | File | Content ||---|---|| `keloid-xMR_code_and_docs_v1.0.zip` | `scripts/` numbered R/shell pipeline (`config.R` holds all locked thresholds); `docs/` (protocol amendment log in Chinese and English, STROBE-MR checklist in Chinese and English, data manifest, session information); `expression_validation/` (Python scripts and per-dataset result tables for the exploratory expression check, S3 Table / S5 Fig); `README_project_zh.md` (project run-order guide, Chinese) || `keloid-xMR_results_and_figures_v1.0.zip` | `results/` derived summary-level results behind every table and figure (instruments, MR estimates, colocalisation, sensitivity, druggability, PheWAS, main and supplementary tables); `figures/` main and supporting figures (PDF and PNG) || `SHA256SUMS.txt` | Checksums of the two archives | ## Source data (public; not redistributed here) - Keloid GWAS summary statistics: GWAS Catalog GCST90652487 (multi-ancestry), GCST90652488 (European), GCST90652489 (African); Greene et al., Nat Commun 2025;16:7770- Plasma cis-pQTL: ARIC study (Zhang et al.)- Skin and fibroblast cis-eQTL: GTEx v8, and eQTL Catalogue datasets QTD000216, QTD000311, QTD000316- LD reference: 1000 Genomes Project Phase 3 (via MRC-IEU)- Tractability annotation: Open Targets Platform- Expression check: Gene Expression Omnibus GSE163973, GSE158395, GSE218007 ## Reproducing the analysis Scripts are numbered in run order (`00_download_gwas.sh` ... `28_tables_paper.R`; see `README_project_zh.md`). Software: R 4.5.3 (package versions in `docs/package_versions.csv` and `docs/sessionInfo.txt`); plink2 for LD computation; random seed 20260818. Scripts contain absolute paths of the authors' machine (`/Users/hirai/Desktop/2026/keloid-xMR`) that must be edited to reproduce. The plink2 binary is not included. ## Notes on interpretation All analysis thresholds were locked before analysis in `scripts/config.R`. Every deviation, including those made after results were seen, is logged with its timing in `docs/S1_Appendix_protocol_amendments_EN.md`. The expression check (S3 Table) is exploratory and post hoc. Files with the suffix `_VOIDED` in `results/` are analyses withdrawn as invalid and are retained only for audit. ## Licence and funding Creative Commons Attribution 4.0 International. Funding: Medical and Health Science and Technology Project of Zhejiang Provincial Health Commission (General Program, Grant No. 2025HY0098). ## Contact Li Jingyu (corresponding author), [e-mail], ORCID 0009-0008-2211-4021.