Agent skill

bio-causal-genomics-proteome-mr-drug-target

Runs cis-pQTL Mendelian randomization for drug-target validation using UKB-PPP (Olink), deCODE (SomaScan), Fenland, INTERVAL, ARIC, and FinnGen-PPP proteomes plus colocalization triangulation, phenome-wide on-target adverse-effect scans, cross-platform Olink/SomaScan replication, and PAV (protein-altering variant) sensitivity. Use when nominating or de-risking a drug target from plasma-proteome GWAS, mimicking pharmacological inhibition via cis-pQTL instruments, separating shared-causal from LD-confounded signal under the Schmidt 2020 cis-MR framework, screening on-target adverse phenotypes ph

BioTender-maxgithub.com/BioTender-maxGitHub ↗
claude-codeNOASSERTION
Install
npx skills add BioTender-max/awesome-bio-agent-skills --skill proteome-mr-drug-target --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

Facts
Files in the skill folder: 4
SKILL.md size: 38 KB
Bundled scripts: none
Path: skills/bioskills/proteome-mr-drug-target/SKILL.md
Open the folder on GitHub →
Where it comes from
Stars: 135
Language: Python

Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.

Review
written from the skill's own SKILL.md · Aug 5, 2026

What it does

Runs cis-pQTL Mendelian randomization for drug-target validation using multiple proteomes, plus colocalization triangulation, phenome-wide on-target adverse-effect scans, cross-platform replication, and PAV sensitivity. Intended for nominating or de-risking a drug target from plasma-proteome GWAS, mimicking pharmacological inhibition via cis-pQTL instruments, and separating shared-causal from LD-confounded signal under the Schmidt 2020 cis-MR framework. It supports screening on-target adverse phenotypes and producing publication-grade STROBE-MR plus PP.H4 evidence for a target gene.

How it works

  • Uses cis-pQTLs in the gene window as instruments under the Schmidt 2020 framework. Performs triangulation with colocalization (3-tier PP.H4 ladder) and cross-platform replication (Olink vs SomaScan), and flags protein-altering-variant (PAV) confounding.
  • Primary inference uses TwoSampleMR for cis-IVW and related methods; supports Cis-IVW with correlation (correl=TRUE) using an LD matrix; includes various MR sensitivity methods (Egger, weighted median, MR-PRESSO) as described.
  • Performs colocalization analyses (coloc.abf or coloc.susie) within the same cis window; conducts pheWAS via ieugwasr::associations against OpenGWAS; annotates cis-pQTLs with VEP for PAV signaling via VEP CLI (annotate every cis-pQTL with vep --species homo_sapiens --canonical --check_existing).
  • Implements a decision rule: a drug-target nomination requires PP.H4 >= 0.8 minimum; industry-grade claims require PP.H4 >= 0.95 with full triangulation and cross-platform agreement, and PAV-excluded sensitivity.
  • Guides version checks for software packages (TwoSampleMR, MendelianRandomization, MR-PRESSO, coloc, susieR) and data source verification across UKB-PPP, deCODE, Fenland, INTERVAL, ARIC, FinnGen-PPP, and others.

When to use it

  • Use when nominating or de-risking a drug target from plasma-proteome GWAS.
  • Use for screening on-target adverse phenotypes (pheWAS-style) and to require cross-platform replication before clinical claims.
  • Use to produce publication-grade evidence with MR + coloc + cross-platform agreement + PAV-excluded sensitivity.

What it can touch

  • Data inputs: cis-pQTL instruments within the gene window; summary statistics from UKB-PPP (Olink), deCODE (SomaScan), Fenland, INTERVAL, ARIC, FinnGen-PPP; OpenGWAS phenotypes for pheWAS.
  • Analysis tools: TwoSampleMR, MendelianRandomization, MR-PRESSO, coloc, susieR, ieugwasr, VEP CLI; R and CLI commands as specified inside the workflow.

Caveats

  • Data-versioning caveats: UKB-PPP, deCODE, Fenland layout changes; verify column headers and panel memberships before formatting data.
  • Platform discordance between Olink (antibody-based) and SomaScan (aptamer-based) can yield discordant cis-pQTL signals; if irreconcilable, avoid clinical claims.
  • PAV confounding can inflate or distort cis-MR results; require PAV-excluded concordance with defined criteria to publish a drug-target claim.
  • Trans-pQTLs are discouraged as instruments due to horizontal pleiotropy; typically should be avoided.
  • Version compatibility notes: ensure matching versions for TwoSampleMR, MendelianRandomization, MR-PRESSO, coloc, susieR, ieugwasr, and PLINK2; verify with packageVersion and CLI checks as described.
From the SKILL.md

## Version Compatibility Reference examples tested with: TwoSampleMR 0.5.11+, MendelianRandomization 0.10+, MR-PRESSO 1.0+, coloc 5.2.3+, susieR 0.12.35+, ieugwasr 1.0+, plink2 2.00a5+, R 4.4+. Before using code patterns, verify installed versions match. If versions differ: - R: `packageVersion('<pkg>')` then `?function_name` to verify parameters - CLI: `plink2 --version`; VEP `vep --help` If code throws OAuth or rate-limit errors from OpenGWAS, or a missing `dataset$N` from coloc, introspect the installed API and adapt the example rather than retrying. UKB-PPP, deCODE, and Fenland summary statistics changed file layouts between 2023 and 2025; verify column headers before passing into `format_data()`. # Proteome-Wide Drug-Target Mendelian Randomization **"Does genetically lowering plasma protein X cause a change in disease Y, mimicking a drug?"** -> Use cis-pQTLs in the gene window for protein X as instruments under the Schmidt 2020 framework (Nat Commun 11:3255), restrict the exclusion-restriction violation to the geometric neighbourhood of the encoding gene, triangulate with colocalization (3-tier PP.H4 ladder below) and cross-platform replication (Olink vs SomaScan), and flag pr

What's inside
Steps it walks through
  1. Version Compatibility
  2. PP.H4 Three-Tier Threshold Ladder
  3. Data Source Taxonomy
  4. Platform and Cohort Versioning (2024-2026)
  5. Cis-MR Methodological Taxonomy
  6. Decision Tree by Scenario
  7. Per-Method Failure Modes
  8. Olink vs SomaScan platform discordance
  9. LD-based pleiotropy in the cis-window
  10. Reverse causation from disease state on plasma protein
  11. PAV (protein-altering-variant) confound
  12. Trans-pQTL pleiotropy if used as instrument
  13. Sample overlap when both ends are UKB
  14. Triangulation Requirement (Operational Postdoc Rule)
Ships with 3 files
  • examples/cis_pqtl_mr.R
  • examples/phewas_drug_target_mr.R
  • usage-guide.md
Commands it runs
echo -e "chr1\t55039548\t.\tG\tT" > cis_pqtls.vcf
vep --species homo_sapiens --assembly GRCh38 --canonical --check_existing \
Ensembl VEP for PAV annotation
conda install -c bioconda ensembl-vep
vep_install -a cf -s homo_sapiens -y GRCh38 -c $HOME/.vep
Prebuilt at https://mrcieu.github.io/ieugwasr/
More from awesome-bio-agent-skills
All skills →
report-templatePublication-quality PDF report generation using Typst templates. Produces professional scientific…Code Review & Qualitysec-reportSEC (size-exclusion chromatography) analysis with peak detection, oligomer classification, and…Code Review & Qualitybio-atac-seq-consensus-peaksetBuild a differential-ready consensus peakset from per-replicate ATAC-seq peaks using iterative…Otherbio-toolsBiology research tools reference. Always available inside agent containers.AI & Agentsbio-remote-homologyDetect distant homologs using profile and structure-aware methods that go beyond standard BLAST.…Otherbio-variant-callingCall SNPs and indels from aligned reads using bcftools mpileup and call. Use when detecting…Other
About this skill
What does the bio-causal-genomics-proteome-mr-drug-target skill do?

Runs cis-pQTL Mendelian randomization for drug-target validation using UKB-PPP (Olink), deCODE (SomaScan), Fenland, INTERVAL, ARIC, and FinnGen-PPP proteomes plus colocalization triangulation, phenome-wide on-target adverse-effect scans, cross-platform Olink/SomaScan replication, and PAV (protein-altering variant) sensitivity. Use when nominating or de-risking a drug target from plasma-proteome GWAS, mimicking pharmacological inhibition via cis-pQTL instruments, separating shared-causal from LD-confounded signal under the Schmidt 2020 cis-MR framework, screening on-target adverse phenotypes ph

How do I install it?

Run `npx skills add BioTender-max/awesome-bio-agent-skills --skill proteome-mr-drug-target --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.

Where does this skill come from?

From BioTender-max/awesome-bio-agent-skills, a repository with 135 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.

Is a popular skill a good skill?

Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.

Keep going
All skills →
The full catalog, by category and by tool.
Open Source Radar →
The repositories these skills come from, watched daily.
The people behind them →
Makers and builders across the ecosystem.