An interactive cheminformatics web app to explore the drug-likeness of small molecules using RDKit and Streamlit.
This tool allows users to upload or search for molecular structures, compute molecular descriptors, visualize drug-likeness, and download results — all in a clean web interface.
- SMILES Input: Upload molecules via CSV or search by compound name using ChEMBL.
- Drug-Likeness Filtering: Apply Lipinski's Rule of Five to filter drug-like compounds.
- Descriptor Calculation:
- Molecular Weight
- LogP
- Hydrogen Bond Donors / Acceptors
- Rotatable Bonds
- Molecule Visualization: Display molecule structures in a grid.
- Descriptor Distributions: Bar chart of key descriptors.
- Downloadable Results: Export filtered data as CSV.
| Tool | Purpose |
|---|---|
Streamlit |
Web app framework |
RDKit |
Cheminformatics toolkit |
Pandas |
Data manipulation and table handling |
Plotly |
Descriptor visualization (optional) |
ChEMBL API |
Molecular search via compound names |
drug_pipeline_project/ ├── app.py # Main Streamlit app ├── environment.yml # Conda-based environment (used by Streamlit Cloud) ├── .streamlit/ │ └── config.toml # Streamlit Cloud settings