Cellence is an AI-powered biomedical pipeline that combines deep learning, cheminformatics, and bioinformatics to accelerate early-stage drug discovery.
It generates, validates, and ranks novel, drug-like molecules tailored to a given protein target — all served through a clean, modular Flask-based API.
✅ Generate valid, novel SMILES molecules conditioned on a chemical prompt (via MolGPT)
✅ Validate, clean, and deduplicate molecules using RDKit
✅ Evaluate drug-likeness (Lipinski’s Rule of 5)
✅ Predict ADMET properties for pharmacokinetic assessment
✅ Predict binding affinity of molecules to a protein sequence (DeepPurpose)
✅ Output results as CSV, including molecular structure images
✅ REST API endpoints for easy integration
User → /generate → [MolGPT → RDKit] → Valid SMILES
→ /score → [Lipinski → ADMET → DeepPurpose] → Ranked molecules + CSV + Images
Results also include Lipinski & ADMET evaluation, molecular images, and a downloadable CSV.
Cellence.ipynib
git clone https://github.com/JabezJesudasonJena/Cellence-Drug-Discovery-model.git
| Component | Role |
|---|---|
| MolGPT | Generate novel SMILES strings |
| RDKit | Validate, canonicalize, Lipinski evaluation |
| DeepPurpose | Predict binding affinity |
| ADMET predictor | Estimate ADMET properties |
| Python | Core implementation |
✨ Interactive AI chatbot advisor
✨ 3D protein–ligand visualization
✨ Diversity clustering & synthetic accessibility scoring
✨ Multi-objective optimization
MIT License — feel free to use, modify, and contribute!
Contributions, issues, and feature requests are welcome!
Please open an issue to discuss changes before submitting a PR.
If you’re interested in AI for healthcare, computational biology, or drug design — feel free to connect with me on LinkedIn or open a discussion here.
Cellence — bridging AI and biomedical science for smarter drug discovery.