Quickstart
Try the API right now — no sign-up required. Click Run on any example below.
Query a cancer gene
Look up EGFR — one of the most studied targets in cancer biology. Hit Run to see live data from the knowledge graph:
curl -s "https://research.usegalen.com/api/v1/entities/EGFR" \
-H "X-API-Key: YOUR_API_KEY" | python3 -m json.toolExplore relationships
EGFR doesn't exist in isolation — it connects to drugs, diseases, pathways, and other genes. See where it leads:
resp = requests.get(
"https://research.usegalen.com/api/v1/entities/EGFR/relationships",
headers={"X-API-Key": "YOUR_API_KEY"},
params={"limit": 5}
)
for rel in resp.json()["items"]:
print(f" {rel['relationship_type']}: {rel['target_name']}"
f" (L{rel['pch_layer']}, {rel['provenance']})")Notice pch_layer — that's the Pearl Causal Hierarchy level. L1 = observed, L2 = interventional, L3 = counterfactual.
See causal inference in action
This is what makes Galen different from PubChem or Open Targets. Instead of “these things are correlated,” Galen answers “if you inhibit EGFR, what happens downstream?” — using a structural causal model, not statistical associations.
Here's what the causal intervention endpoint returns:
{
"target": "EGFR",
"intervention_type": "inhibit",
"downstream_effects": [
{
"entity": "RAS-MAPK pathway",
"direction": "decreased",
"effect_type": "pathway_inhibition",
"confidence": 0.94,
"pch_layer": 3,
"provenance": "chembl_36+cross_evidence_convergence"
},
{
"entity": "PI3K-AKT signaling",
"direction": "decreased",
"effect_type": "pathway_inhibition",
"confidence": 0.91,
"pch_layer": 3,
"provenance": "depmap_crispr+gdsc2"
},
{
"entity": "Erlotinib sensitivity",
"direction": "increased",
"effect_type": "drug_response",
"confidence": 0.88,
"pch_layer": 2,
"provenance": "gdsc2_sensitivity"
},
{
"entity": "Cell proliferation",
"direction": "decreased",
"effect_type": "phenotype",
"confidence": 0.86,
"pch_layer": 2,
"provenance": "depmap_crispr"
}
],
"model": "structural_causal_model",
"evidence_sources": 4
}Causal inference requires a Researcher plan ($49/mo)
The free tier includes knowledge graph queries, all database integrations (cBioPortal, GTEx, STRING, ChEMBL), and hypothesis generation. Causal endpoints (do-calculus, counterfactual reasoning) are available on paid plans. Compare plans →
Get your own API key
The examples above run in sandbox mode with shared rate limits. Create a free account to get your own key with 1,000 requests/day:
- Click below to create an account with your email (magic link, no password)
- Open your Dashboard and click “Generate New Key”
- Copy the key — it's only shown once
Once you have a key, click “Use your own key” on any code example above to switch from sandbox to your account.