The Causal Inference Engine for Cancer Biology.

import requests

API_KEY = "gk_live_your_key_here"
BASE = "https://research.usegalen.com/api/v1"

# Query: What happens downstream when you inhibit EGFR?
# This uses BFS over a Structural Causal Model — not
# statistical correlations. Every effect is traceable.
causal = requests.post(
    f"{BASE}/causal/intervention",
    headers={"X-API-Key": API_KEY},
    json={
        "target": "EGFR",
        "intervention_type": "inhibit"
    }
)

# Each downstream effect includes:
# - The affected entity and mechanism
# - Effect size and confidence score
# - Pearl Causal Hierarchy layer (L1/L2/L3)
# - Original data source (e.g., "ChEMBL", "cBioPortal")
for effect in causal.json()["downstream_effects"]:
    print(f"{effect['entity']}: {effect['effect_type']}"
          f" (confidence: {effect['confidence']:.2f})")
    print(f"  Source: {effect['provenance']}"
          f" | Pearl Level: L{effect['pch_layer']}")

# Example output:
# EGFR inhibition → suppresses MAPK signaling (confidence: 0.94)
#   Source: ChEMBL bioactivity | Pearl Level: L2 (experimental)
# EGFR inhibition → reduces cell proliferation (confidence: 0.87)
#   Source: cBioPortal clinical genomics | Pearl Level: L2 (experimental)
# EGFR inhibition → may restore apoptosis via BIM (confidence: 0.71)
#   Source: SCM counterfactual validation | Pearl Level: L3 (counterfactual)