Each scaffold hop used to mean — a week of manual profiling
Without running variants through a separate ADMET tool, then a separate compliance check, then a separate comparison. Scaffold hopping and AI optimization generate variants with ADMET, QED, LogP, and compliance inline — in one call.
“Optimize this lead for better oral bioavailability and lower hERG liability while maintaining EGFR selectivity.”
How it works
Submit a lead and objectives
Provide a SMILES and optional property targets — QED, LogP, molecular weight, similarity threshold. Choose scaffold hopping for structural diversity or MolMIM for property-directed fine-tuning.
Variants generated and enriched
Scaffold hopping swaps ring systems (benzene↔pyridine, cyclohexane↔piperidine). MolMIM generates AI-guided variants targeting your profile. Every variant auto-enriched with ADMET predictions and FAVES compliance.
Ranked variants with full profiles
You receive variants sorted by property match, each with Tanimoto similarity to seed, patent risk assessment, compliance status, and complete ADMET profile. Ready for docking.
Proof
Two paths: lead_optimization (RDKit scaffold hopping, 30+ ring pairs) and optimize_molecule (NVIDIA MolMIM, property-directed).
Post-optimization enrichment: chem-props (SA/properties) + addie-models (41 ADMET models) run in parallel. FAVES auto-screens every variant.
Patent risk via Tanimoto similarity to Pinecone patent index. Each variant returns compliance_status, tanimoto_to_seed, and patent_risk.
Optimize without hand-offs
Scaffold hopping + AI optimization. ADMET inline. Sign up and optimize your first lead.