Pro ClinPGx Database

A scientific computing skill for querying ClinPGx — the Clinical Pharmacogenomics Database providing curated information about how genetic variants affect drug response. Pro ClinPGx Database helps you look up gene-drug interactions, dosing recommendations based on genotype, and clinical pharmacogenomics guidelines from CPIC and DPWG.

When to Use This Skill

Choose Pro ClinPGx Database when:

• Looking up pharmacogenomic dosing guidelines for specific gene-drug pairs
• Finding how genetic variants affect drug metabolism or response
• Retrieving CPIC or DPWG clinical recommendations by genotype
• Building pharmacogenomics-aware prescribing support tools

Consider alternatives when:

• You need genetic variant databases broadly (use ClinVar or gnomAD)
• You need drug interaction data without genetics (use DrugBank)
• You need clinical trial information (use ClinicalTrials.gov)
• You need enzyme kinetics data (use BRENDA)

Quick Start

Copy
claude "Look up pharmacogenomics guidelines for CYP2D6 and codeine"

Copy
import requests

# Query ClinPGx for gene-drug interactions
# Example: CYP2D6 metabolizer status and codeine
gene = "CYP2D6"
drug = "codeine"

# CPIC API for guidelines
cpic_url = f"https://api.cpicpgx.org/v1/drug/name/{drug}"
response = requests.get(cpic_url)
drug_data = response.json()

print(f"Drug: {drug_data['name']}")
print(f"CPIC Level: {drug_data.get('cpicLevel', 'N/A')}")

# Get gene-specific recommendations
recs_url = f"https://api.cpicpgx.org/v1/recommendation/drug/{drug}"
recs = requests.get(recs_url).json()

for rec in recs:
    phenotype = rec.get("phenotypes", {}).get(gene, "Unknown")
    print(f"\nPhenotype: {phenotype}")
    print(f"  Recommendation: {rec.get('drugrecommendation', 'N/A')}")
    print(f"  Strength: {rec.get('classification', 'N/A')}")

Core Concepts

Pharmacogenomics Terminology

Term	Definition	Example
Phenotype	Predicted metabolizer status	Poor, Intermediate, Normal, Ultrarapid
Diplotype	Two-allele genotype	CYP2D6 *1/*4
Star Allele	Named haplotype variant	CYP2D6*4 (non-functional)
Activity Score	Quantified enzyme function	AS=0 (poor), AS=1 (intermediate), AS=2 (normal)
CPIC Level	Guideline evidence level	A (guideline available), B (in development)

Metabolizer Phenotypes

Copy
## CYP2D6 Metabolizer Status → Drug Response

| Phenotype | Activity Score | Codeine Effect | Recommendation |
|-----------|---------------|----------------|----------------|
| Ultrarapid | > 2.25 | Excessive morphine → toxicity | Avoid codeine |
| Normal | 1.25-2.25 | Expected analgesic effect | Standard dose |
| Intermediate | > 0, < 1.25 | Reduced efficacy | Consider alternative |
| Poor | 0 | No conversion to morphine | Use alternative analgesic |

Multi-Gene Lookup

Copy
def pharmacogenomics_profile(genes, drug):
    """Get PGx recommendations for multiple genes affecting one drug"""
    profile = {"drug": drug, "genes": {}}

    for gene in genes:
        url = f"https://api.cpicpgx.org/v1/recommendation/drug/{drug}"
        recs = requests.get(url).json()

        gene_recs = [r for r in recs
                     if gene in r.get("phenotypes", {})]
        profile["genes"][gene] = {
            "recommendations": len(gene_recs),
            "phenotypes": list(set(
                r["phenotypes"][gene] for r in gene_recs
            ))
        }

    return profile

# Example: Tamoxifen affected by CYP2D6 and CYP3A4
profile = pharmacogenomics_profile(["CYP2D6", "CYP3A4"], "tamoxifen")

Configuration

Parameter	Description	Default
guideline_source	CPIC, DPWG, or both	CPIC
include_evidence	Include literature references	true
phenotype_system	CPIC standardized or legacy	standardized
allele_notation	Star allele or rsID	star_allele
api_base_url	ClinPGx/CPIC API base URL	https://api.cpicpgx.org/v1

Best Practices

1. Use CPIC standardized phenotype terms. "Poor Metabolizer" is the standard — don't use "slow metabolizer" or "PM" without context. Standardized terms ensure consistent interpretation across clinical systems and publications.

2. Always check the CPIC guideline level. Level A genes have published clinical guidelines with evidence-based recommendations. Level B and below have less evidence — note the evidence level when making clinical recommendations.

3. Consider multi-gene interactions. Some drugs are affected by multiple pharmacogenes. Checking only CYP2D6 for a drug also metabolized by CYP3A4 gives an incomplete picture. Query all relevant genes for comprehensive dosing guidance.

4. Map between star alleles and rsIDs. Clinical labs may report results as star alleles (*1/*4) or rsIDs (rs3892097). Ensure your system can map between both formats, as different labs use different conventions.

5. Update guideline data regularly. CPIC and DPWG publish updated guidelines periodically. Set up quarterly checks for guideline updates and ensure your decision support tools reflect current recommendations.

Common Issues

Star allele not found in database. Rare or novel star alleles may not have clinical annotations yet. Check if the allele has been reclassified or merged with another designation. For novel variants, report as "uncertain function" until characterized.

CPIC and DPWG recommendations differ. These organizations use different evidence evaluation frameworks. CPIC recommendations are generally more conservative. When guidelines conflict, document both and follow institutional policies for which to prioritize.

Activity score calculation doesn't match expected phenotype. Activity scores depend on the specific star alleles reported. Some star alleles have uncertain function (activity score not assignable). In these cases, the phenotype assignment may default to "indeterminate" rather than a specific metabolizer status.