Bellerophon is a computational tool that automatically decomposes heterobifunctional degraders (PROTACs) into their three structural components:
Warhead (target-binding ligand), Linker, and E3 Ligase Ligand.
Accurately splitting PROTACs into their constituent parts is a complex task. Existing resources provide valuable manually curated information but lack automated, flexible parsing.
Bellerophon overcomes this limitation with a structure-based algorithm that identifies and extracts these fragments directly from a PROTAC’s molecular structure.
By standardizing the decomposition process, Bellerophon enables:
- High-throughput annotation and data curation
- Comparison and recombination of validated building blocks
- Structure-driven de novo PROTAC design workflows
Bellerophon uses a multi-step, structure-based comparison procedure:
- The input PROTAC is matched against a curated warhead library.
- Once the warhead is identified and removed, the remaining structure is compared against the E3 ligase library.
- The linker is inferred from the remaining molecular fragment.
- The process is repeated in reverse (starting from the E3 ligand) to ensure consistency.
- Only decompositions yielding identical linkers in both directions are retained.
To minimize false positives, the algorithm applies multiple structural integrity checks, ensuring the identified components fully reconstruct the original PROTAC (heavy atom count, ring count, and connectivity).
- Language: Python
- Core library: RDKit
- Interface: Streamlit
- Source code and default libraries: Available in this repository
- Web app: 🌐 Bellerophon Streamlit App
- Automatic identification of warhead, linker, and E3 ligand
- Consistency validation via heavy-atom and ring-count checks
- Built-in curated libraries (warheads and E3 ligases)
- Option to upload custom
.sdflibraries - Batch processing of multiple PROTACs
- Downloadable results in
.csvor.txtformat - Internal error logging and reporting
👉 Try it instantly on Streamlit:
https://bellerophon-protac-decomposing-tool.streamlit.app/
# Clone the repository
git clone https://github.com/yourusername/bellerophon.git
cd bellerophon
# Install dependencies
pip install -r requirements.txt
# Launch Streamlit app
streamlit run sprotac.py-
Paste Mode:
Directly paste PROTAC names and SMILES strings. -
File Upload:
Upload.txt,.csv, or.sdffiles containing the columns Name and SMILES.
The tool returns a table listing all valid decompositions, including:
- Compound Name
- PROTAC SMILES
- Warhead SMILES
- Linker SMILES
- E3 Ligand SMILES
All results can be downloaded as .csv or .txt files.
A separate file containing any errors encountered during processing is also available.
Curated collections of warheads and E3 ligands are included in the repository.
These were compiled from degraders reported in clinical trials and recent literature.
Users can also upload custom .sdf libraries to tailor the decomposition process to specific datasets or design goals.
Developed by the MedChemBeyond Group,
Department of Molecular Biotechnology and Health Sciences,
University of Turin, in collaboration with Alvascience.
📧 Contact: giulia.apprato@unito.it
This repository is distributed under the MIT License (see LICENSE).
Use is permitted for non-commercial purposes only.
If you use Bellerophon in your research, please cite:
Bellerophon: A structure-based tool for automatic PROTAC decomposition
Web app: https://bellerophon-protac-decomposing-tool.streamlit.app/
Special thanks to contributors from the University of Turin, MedChemBeyond members, Alvascience
for supporting the development of Bellerophon.