Hit to lead

Best fit: you have a hit compound and a target structure, and you want the whole hit-to-lead read — binding, safety, and whether you can actually make and source it.

The prompt

Compare imatinib and nilotinib head-to-head on the ABL1 DFG-out conformation
(PDB: 1IEP) — verify each binding mode in the pocket, rescore the poses for
consensus, and tell me how their CYP3A4 metabolism and hERG cardiotoxicity
stack up. For the cleaner compound, plan how I'd actually synthesize it at
scale and check vendor availability for the starting materials, then recommend
which drug has both the better recapitulated binding mode and the more
procurement-ready route.

What it does

The orchestrator carries the hit from binding all the way to procurement:

1. DiffDock docks each compound into the target pocket and returns poses.
2. MM-GBSA rescores the poses for a physics-based consensus.
3. ADMET-AI screens metabolism and cardiotoxicity.
4. AiZynthFinder plans retrosynthetic routes.
5. ZINC checks which starting materials are purchasable and lists their vendor catalog codes.

Watch for

• Rescore before you trust a pose. DiffDock confidence and MM-GBSA energy are different signals — agreement between them is the useful result.
• Procurement is a real filter. A route with fewer steps but single-source building blocks may be riskier at scale than a longer route with in-stock materials.
• Hazard flags. Ask the route planner to flag cryogenic or heavy-metal steps, as the prompt does — they change feasibility.

Read next

• Fold, find pockets, dock, triage — get to a docked hit from a sequence first.
• Which tool? — generative and synthesis alternatives.
• Lead optimization — the guided version: multi-parameter refinement step by step.
• Synthesis planning and Compound procurement — the route and sourcing steps in depth.