SkillAgentSearch skills...

clinpgx-database

Query the ClinPGx (formerly PharmGKB) REST API plus the CPIC PostgREST companion API for pharmacogenomic clinical annotations, CPIC/DPWG dosing guidelines, gene-drug pairs, variant-drug associations, FDA/EMA drug labels, and PGx pathways.

Install / Use

npx skills add jaechang-hits/SciAgent-Skills --skill clinpgx-database

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

91/100

Supported Platforms

Universal

Our assessment of clinpgx-database

clinpgx-database scores 91/100 on our quality scale, 242nd of 599 Data & Analytics skills we index (top 41%).

Its SKILL.md is 28 KB long, well organised into 48 sections with 20 code examples: a thorough specification that gives an agent plenty to work with.

It has 367 GitHub stars, a meaningful sign that others use it.

Substance
30/30
Structure
20/20
Description
15/15
Adoption
11/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so clinpgx-database is actively maintained.
  • Our last check on 2026-11-20 found the source still online.
  • No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
  • Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-05. Automated pattern scan on 2026-10-05. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

clinpgx-database compared with similar skills

All 4 of these similar skills score higher than clinpgx-database; compare them before choosing.

SkillScoreStarsUpdatedFormat
clinpgx-database (this skill)by jaechang-hits913673mo agoSKILL.md
Agent-Reachby Panniantong100113.3k12d agoCLAUDE.md
headroomby headroomlabs-ai10080.2ktodayCLAUDE.md
Scraplingby D4Vinci100104.4k2d agoMCP Server
crawl4aiby unclecode10087.6ktodayMCP Server

Frequently asked questions

How do I install clinpgx-database?
Run npx skills add jaechang-hits/SciAgent-Skills --skill clinpgx-database. The install tabs above show the steps for each supported agent.
Which AI agents does clinpgx-database work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is clinpgx-database safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is clinpgx-database still maintained?
The repository was last updated about 3 months ago, so clinpgx-database is actively maintained.

name: "clinpgx-database" description: "Query the ClinPGx (formerly PharmGKB) REST API plus the CPIC PostgREST companion API for pharmacogenomic clinical annotations, CPIC/DPWG dosing guidelines, gene-drug pairs, variant-drug associations, FDA/EMA drug labels, and PGx pathways. Two-host architecture: api.clinpgx.org for annotation records, api.cpicpgx.org for genotype→recommendation lookups. No auth. For germline pathogenicity use clinvar-database; for somatic cancer PGx use cosmic-database or opentargets-database; for drug bioactivity use chembl-database-bioactivity." license: "CC-BY-SA-4.0"

ClinPGx (PharmGKB) Pharmacogenomics Database

Overview

PharmGKB rebranded as ClinPGx in 2024 and the API moved from api.pharmgkb.org to api.clinpgx.org. The old host now returns 404/405; every example here uses the new endpoints. Two complementary APIs are used together:

  • ClinPGx Data API (api.clinpgx.org/v1) — record-style access to genes, drugs, variants, clinical annotations, guideline annotations, drug labels, and pathways. Responses wrap data as {"data": [...], "status": "success"}. Filters use dotted property paths (e.g. relatedChemicals.name=clopidogrel, levelOfEvidence.term=1A).
  • CPIC PostgREST API (api.cpicpgx.org/v1) — relational lookup of genotype → drug recommendation rows. PostgREST filter syntax (column=eq.value, JSON cs.{...} for jsonb containment). Returns flat JSON arrays.

Use ClinPGx for what is known about a gene/drug/variant; use CPIC for how to prescribe given a phenotype. The pattern is ClinPGx for annotations, CPIC for recommendations.

When to Use

  • Retrieving CPIC genotype-specific dosing recommendations for a gene-drug pair (e.g., CYP2C19 + clopidogrel) — use CPIC
  • Looking up all pharmacogenomic clinical annotations for a drug or evidence level — use ClinPGx data/clinicalAnnotation
  • Finding all CPIC/DPWG guideline annotations for a pharmacogene — use ClinPGx data/guidelineAnnotation
  • Resolving a gene symbol, drug name, or rsID to ClinPGx PA identifiers — use data/{gene,drug,variant}
  • Free-text search across all ClinPGx record types (genes, drugs, variants, annotations) — use POST /site/search
  • Retrieving FDA/EMA pharmacogenomic drug label annotations — use ClinPGx data/label
  • Building precision-medicine prescribing workflows that combine annotation evidence with phenotype-specific recommendations
  • For germline disease pathogenicity (not PGx) use clinvar-database
  • For somatic cancer pharmacogenomics use cosmic-database or opentargets-database

Prerequisites

  • Python packages: requests, pandas — both already in standard environments
  • Data requirements: HGNC gene symbols, drug names (lowercase generic), dbSNP rsIDs, or PA identifiers
  • Environment: internet connection; no authentication required for either host
  • Rate limits: the ClinPGx host occasionally returns HTTP 429; insert time.sleep(0.3–0.5) between sequential calls. CPIC is more permissive.

If you are inside a pixi/conda environment that already provides requests and pandas, skip the install — invoke scripts with pixi run python ....

pip install requests pandas

Quick Start

import requests

CLINPGX = "https://api.clinpgx.org/v1"
CPIC    = "https://api.cpicpgx.org/v1"

# CPIC genotype → recommendation: clopidogrel + CYP2C19 Poor Metabolizer
drug = requests.get(f"{CPIC}/drug", params={"name": "eq.clopidogrel"}).json()[0]
recs = requests.get(f"{CPIC}/recommendation",
                    params={"drugid": f"eq.{drug['drugid']}",
                            "phenotypes": 'cs.{"CYP2C19":"Poor Metabolizer"}'}).json()
print(f"clopidogrel CYP2C19=PM: {len(recs)} recommendation(s)")
for rec in recs[:2]:
    print(f"  [{rec['classification']}] {rec['drugrecommendation'][:80]}…")

# ClinPGx side: how many CPIC guideline annotations cover CYP2C19?
glines = requests.get(f"{CLINPGX}/data/guidelineAnnotation",
                      params={"relatedGenes.symbol": "CYP2C19",
                              "source": "CPIC", "view": "base"}).json()["data"]
print(f"CYP2C19 CPIC guidelines: {len(glines)}")

Core API

Module 1: Free-text site search

POST /site/search with a JSON body {"query": "<term>"} is the canonical entry point when you don't know the PA ID. It searches across drugs, genes, variants, clinical annotations, guideline annotations, and labels in one shot.

import requests

CLINPGX = "https://api.clinpgx.org/v1"

r = requests.post(f"{CLINPGX}/site/search",
                  json={"query": "rs4149056"}, timeout=15)
r.raise_for_status()
hits = r.json()["data"]["hits"]
print(f"Total hits: {r.json()['data']['total']}")
for h in hits[:5]:
    print(f"  id={h.get('id')}  name={h.get('name')[:80]}")
# Broader concept search
r = requests.post(f"{CLINPGX}/site/search",
                  json={"query": "TPMT azathioprine"}, timeout=15)
hits = r.json()["data"]["hits"]
print(f"TPMT+azathioprine hits: {len(hits)}")
for h in hits[:5]:
    print(f"  {h.get('id'):>15}  {h.get('name','')[:80]}")

Module 2: Gene, drug, and variant record lookup

The /data/{type} endpoints accept simple property filters. All return {"data": [...], "status": "success"} — use view=base for summary, view=max for full nested objects.

import requests

CLINPGX = "https://api.clinpgx.org/v1"

# Gene by HGNC symbol
gene = requests.get(f"{CLINPGX}/data/gene",
                    params={"symbol": "CYP2D6", "view": "base"}).json()["data"][0]
print(f"{gene['symbol']}  id={gene['id']}  {gene['name']}")

# Drug by name (lowercase generic preferred)
drug = requests.get(f"{CLINPGX}/data/drug",
                    params={"name": "warfarin", "view": "base"}).json()["data"][0]
print(f"{drug['name']}  id={drug['id']}")

# Variant by rsID
var = requests.get(f"{CLINPGX}/data/variant",
                   params={"name": "rs4149056", "view": "base"}).json()["data"][0]
print(f"{var['name']}  id={var['id']}  significance={var.get('clinicalSignificance')}")
# Direct record fetch when you already have a PA ID
r = requests.get(f"{CLINPGX}/data/drug/PA449088", params={"view": "max"}).json()
d = r["data"]
print(f"PA449088 → {d['name']}  (objCls={d['objCls']})")

Module 3: Clinical annotations

data/clinicalAnnotation records associate a variant (location) with one or more drugs (relatedChemicals) and an evidence level (levelOfEvidence.term). The two supported filters are relatedChemicals.name= and levelOfEvidence.term=. There is no working gene= filter on this endpoint — see Module 4 for gene-driven access.

import requests, pandas as pd

CLINPGX = "https://api.clinpgx.org/v1"

# All clinical annotations for clopidogrel
data = requests.get(f"{CLINPGX}/data/clinicalAnnotation",
                    params={"relatedChemicals.name": "clopidogrel",
                            "view": "base"}).json()["data"]
print(f"clopidogrel annotations: {len(data)}")

rows = []
for ann in data[:10]:
    loc = ann.get("location") or {}
    drugs = ", ".join(c.get("name", "") for c in ann.get("relatedChemicals", []))
    rows.append({
        "id": ann["id"],
        "variant": loc.get("displayName"),
        "gene": (loc.get("genes") or [{}])[0].get("symbol"),
        "drug": drugs,
        "level": (ann.get("levelOfEvidence") or {}).get("term"),
        "score": ann.get("score"),
    })
print(pd.DataFrame(rows).to_string(index=False))
# All Level 1A clinical annotations (highest evidence)
data = requests.get(f"{CLINPGX}/data/clinicalAnnotation",
                    params={"levelOfEvidence.term": "1A",
                            "view": "base"}).json()["data"]
print(f"Level 1A annotations: {len(data)}")

drug_to_count = {}
for ann in data:
    for c in ann.get("relatedChemicals") or []:
        drug_to_count[c["name"]] = drug_to_count.get(c["name"], 0) + 1
top = sorted(drug_to_count.items(), key=lambda x: -x[1])[:10]
for d, n in top:
    print(f"  {n:3}  {d}")

Module 4: Guideline annotations (gene-driven access)

data/guidelineAnnotation supports both relatedGenes.symbol= and relatedChemicals.name=, plus source= (CPIC, DPWG, CPNDS, RNPGx). This is the canonical way to get gene→guideline coverage.

import requests

CLINPGX = "https://api.clinpgx.org/v1"

# All CPIC guidelines mentioning CYP2C19
data = requests.get(f"{CLINPGX}/data/guidelineAnnotation",
                    params={"relatedGenes.symbol": "CYP2C19",
                            "source": "CPIC",
                            "view": "base"}).json()["data"]
print(f"CYP2C19 CPIC guidelines: {len(data)}")
for g in data[:5]:
    print(f"  PA{g['id']}: {g['name'][:80]}")
# Guidelines for a specific drug across all bodies (CPIC, DPWG, …)
data = requests.get(f"{CLINPGX}/data/guidelineAnnotation",
                    params={"relatedChemicals.name": "clopidogrel",
                            "view": "base"}).json()["data"]
by_source = {}
for g in data:
    for s in (g.get("crossReferences") or []):
        by_source.setdefault(s.get("resource", "?"), 0)
        by_source[s["resource"]] = by_source.get(s["resource"], 0) + 1
print(f"clopidogrel guidelines: {len(data)} ({list({g.get('source') for g in data})})")

Module 5: Regulatory drug labels (FDA / EMA)

data/label records are PharmGKB-curated annotations of FDA/EMA pharmacogenomic labeling. Filter by relatedChemicals.name= and source= (FDA, EMA, HCSC, PMDA, Swissmedic).

import requests, pandas as pd

CLINPGX = "https://api.clinpgx.org/v1"

data = requests.get(f"{CLINPGX}/data/label",
                    params={"relatedChemicals.name": "warfarin",
                            "source": "FDA",
                            "view": "base"}).json()["data"]
print(f"warfarin FDA labels: {len(data)}")

rows = [{
    "name": d["name"][:60],
    "biomarker_status": d.get("biomarkerStatus"),
    "testing_required": d.get("testingRequired"),
    "alternate_drug": d.get("alternateDrugAvailable"),
} for d in data]
print(pd.DataFrame(rows).to_string(index=False))

Module 6: CPIC genotype → recommendation chain

CPIC's PostgREST API uses column=eq.value for equality and column=cs.{...} for JSONB containment. The standard lookup chain is drug → drugid → recommendation, optionally filtered by phenotype.

import requests

CPIC = "https://api.cpicpgx.org/v1"

# Resolve drug name to drugid (RxNorm-prefixed)
drug = requests.get(f"{CPIC}/drug",
                    params={"name": "eq.clopidogrel"}).json()[0]
print(f"clopidogrel drugid: {drug['drugid']}")

# All phenotype-specific recommendations for clopidogrel
recs = requests.get(f"{CPIC}/recommendation",
                    params={"drugid": f"eq.{drug['drugid']}"}).json()
print(f"Total recommendations: {len(recs)}")
for rec in recs[:3]:
    print(f"  {rec['phenotypes']}  [{rec['classification']}]")
    print(f"    {rec['drugrecommendation'][:90]}…")
# Phenotype filter via jsonb containment (cs.{...})
# The phenotypes column is a jsonb dict; cs. checks that the query is a subset.
recs = requests.get(f"{CPIC}/recommendation",
                    params={"drugid": f"eq.{drug['drugid']}",
                            "phenotypes": 'cs.{"CYP2C19":"Poor Metabolizer"}'}
                    ).json()
for rec in recs:
    print(f"  [{rec['classification']}] {rec['drugrecommendation'][:90]}…")

# Gene-driven: list every drug with a CPIC pair for CYP2C19
pairs = requests.get(f"{CPIC}/pair",
                     params={"genesymbol": "eq.CYP2C19"}).json()
print(f"\nCYP2C19 CPIC pairs: {len(pairs)}")
drug_ids = sorted({p["drugid"] for p in pairs})
print(f"Sample drug IDs: {drug_ids[:5]}")

Key Concepts

Two-host architecture

| Question | Use | Why

Truncated for display — read the full file on GitHub.

Related Skills

View on GitHub
GitHub Stars367
CategoryData
Updated2mo ago
Forks36

Languages

Python

Trust signals

88/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 medium