Compound (ChEMBL bioactivity targets) × Disease (Open Targets associations) → overlap + graph. Back to MolGen.
Input
Target-matching between ChEMBL and Open Targets is approximate name
string matching, not a proper ID crosswalk — treat overlap as a lead,
verify manually before acting on it. No PPI expansion or pathway
enrichment stats yet (Phase 2). Disease name resolution (search query)
wasn't independently verified live this session — if it errors,
report the exact message back and it's a quick schema fix.