Pro ClinVar Workspace

A scientific computing skill for querying ClinVar — NCBI's public archive of human genetic variants and their clinical significance. Pro ClinVar Workspace helps you search for variant-disease associations, retrieve pathogenicity classifications, and integrate clinical variant data into genomics research workflows.

When to Use This Skill

Choose Pro ClinVar Workspace when:

• Looking up the clinical significance of human genetic variants
• Searching for pathogenic variants associated with specific diseases
• Retrieving variant classifications from multiple submitting labs
• Building variant interpretation pipelines with ClinVar annotations

Consider alternatives when:

• You need population allele frequencies (use gnomAD)
• You need cancer-specific somatic variants (use COSMIC)
• You need pharmacogenomic variants (use ClinPGx/PharmVar)
• You need protein structure impact predictions (use AlphaFold + variant tools)

Quick Start

Copy
claude "Search ClinVar for pathogenic BRCA1 variants"

Copy
from Bio import Entrez
import json

Entrez.email = "[email protected]"

# Search ClinVar for BRCA1 pathogenic variants
handle = Entrez.esearch(
    db="clinvar",
    term="BRCA1[gene] AND pathogenic[clinical significance]",
    retmax=10
)
results = Entrez.read(handle)
print(f"Total pathogenic BRCA1 variants: {results['Count']}")

# Fetch variant details
for var_id in results["IdList"]:
    handle = Entrez.efetch(db="clinvar", id=var_id, rettype="vcv",
                           is_variationid="true", from_esearch="true")
    record = handle.read()
    print(f"Variant ID: {var_id}")

Copy
# Using ClinVar API directly
import requests

# Search via NCBI E-utilities
url = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi"
params = {
    "db": "clinvar",
    "term": "BRCA1[gene] AND pathogenic[clinsig]",
    "retmode": "json",
    "retmax": 20
}
response = requests.get(url, params=params)
data = response.json()
ids = data["esearchresult"]["idlist"]
print(f"Found {len(ids)} variants")

Core Concepts

ClinVar Classification System

Classification	Meaning	Action
Pathogenic	Causes disease	Report, clinical action
Likely Pathogenic	Probably causes disease	Report, consider clinical action
Uncertain Significance (VUS)	Insufficient evidence	Report, no clinical action
Likely Benign	Probably harmless	May omit from report
Benign	Harmless	Omit from report

Variant Identifiers

Copy
# ClinVar uses multiple identifier systems
variant_ids = {
    "variation_id": 12345,          # ClinVar internal ID
    "rcv": "RCV000012345",         # Reference ClinVar accession
    "scv": "SCV000012345",         # Submitter ClinVar accession
    "rs_id": "rs28897696",         # dbSNP rsID
    "hgvs_c": "NM_007294.4:c.5266dupC",  # Coding DNA HGVS
    "hgvs_p": "NP_009225.1:p.Gln1756ProfsTer74",  # Protein HGVS
}

Batch Variant Lookup

Copy
def batch_clinvar_lookup(variant_list):
    """Look up multiple variants in ClinVar"""
    results = []
    for variant in variant_list:
        handle = Entrez.esearch(
            db="clinvar",
            term=f"{variant}[variant name]"
        )
        search = Entrez.read(handle)

        if search["IdList"]:
            handle = Entrez.esummary(
                db="clinvar",
                id=search["IdList"][0]
            )
            summary = Entrez.read(handle)
            doc = summary["DocumentSummarySet"]["DocumentSummary"][0]
            results.append({
                "variant": variant,
                "clinical_significance": doc.get("clinical_significance", {}).get("description", "N/A"),
                "gene": doc.get("genes", [{}])[0].get("symbol", "N/A"),
                "conditions": [t.get("trait_name", "") for t in doc.get("trait_set", [])]
            })
    return results

variants = ["NM_007294.4:c.5266dupC", "NM_000546.5:c.215C>G"]
results = batch_clinvar_lookup(variants)

Configuration

Parameter	Description	Default
Entrez.email	Required email for NCBI API	Required
Entrez.api_key	NCBI API key for higher limits	None
significance_filter	Filter by clinical significance	None (all)
review_status_min	Minimum review star rating	None
assembly	Reference genome (GRCh37 or GRCh38)	GRCh38

Best Practices

1. Filter by review status for clinical use. ClinVar assigns star ratings (0-4) based on the level of review. For clinical reporting, use variants with ≥ 2 stars (criteria provided, multiple submitters). Single-submitter classifications may not be validated.

2. Check for conflicting interpretations. Different labs may classify the same variant differently. ClinVar flags these conflicts. Always review the individual submitter classifications (SCV records) when a variant has conflicting interpretations.

3. Use HGVS nomenclature for precise variant identification. rsIDs can be ambiguous (one rsID may map to multiple variants). HGVS notation (e.g., NM_007294.4:c.5266dupC) is unambiguous and includes the reference transcript version.

4. Cross-reference with population databases. A variant classified as pathogenic in ClinVar should be rare in the general population. Cross-check allele frequency in gnomAD — if a "pathogenic" variant has >1% frequency, the classification may warrant review.

5. Subscribe to variant reclassification updates. ClinVar classifications change as new evidence emerges. For clinically important variants, set up alerts for reclassifications, especially for VUS that may be upgraded or downgraded.

Common Issues

Search returns no results for a known variant. The variant may be described using different nomenclature. Try searching by rsID, HGVS notation, gene name + genomic position, or the protein change. ClinVar indexes variants in multiple ways.

Variant shows "conflicting interpretations." This means different submitting labs disagree on clinical significance. Review each submitter's evidence (SCV records), check their submission dates, and consider the most recent expert panel review if available.

Coordinate mapping errors between genome builds. ClinVar reports positions on both GRCh37 and GRCh38. Ensure your query coordinates match the expected assembly. Use LiftOver tools to convert between assemblies if needed.