Why Galen Exists
Cancer biology generates more data than any human or team can hold in mind at once. Existing tools address pieces of this complexity in isolation — a mutation database here, a drug screen there, a literature search engine somewhere else. Galen is building the first system that reasons across all of cancer biology from first principles.
The problem
- A single tumor can carry thousands of mutations, each interacting with proteins, pathways, and drug targets in ways that cascade across biological layers.
- Oncologists and researchers currently piece together evidence from dozens of disconnected databases, each with its own schema, nomenclature, and evidence standard.
- General AI can generate plausible-sounding summaries, but cannot compute causal effects, trace claims to specific experiments, or produce deterministic answers.
Galen's answer
- A unified knowledge graph integrating 10+ biomedical databases into a single queryable structure, with provenance on every relationship.
- A formal Structural Causal Model enabling genuine causal inference — do-calculus and counterfactual reasoning — not statistical correlations.
- The Pearl Causal Hierarchy applied to every edge, so you know whether evidence is observational (L1), experimental (L2), or counterfactually validated (L3).
- An autonomous research agent running 24/7, discovering, validating, and promoting causal relationships.
For developers
Galen exposes this system as a REST API. Every response is deterministic, auditable, and provenance-traced. You don't need to understand causal inference theory to use it — but if you do, the full formal machinery is available through the Causal Inference endpoints.
The bigger picture
To learn more about the science behind Galen and the team building it, visit Science and About.
Next: Pearl Causal Hierarchy
Learn about the three-layer causal framework that annotates every relationship in the knowledge graph.
Learn about the Pearl Causal Hierarchy →