This directory contains Jupyter notebooks that can be directly launched on Google Colab.
To run it locally on your system see the instructions below (does require a working python3 (3.10 to 3.13) installation).
| Notebooks | Description | Colab |
|---|---|---|
| HADDOCK3-antibody-antigen.ipynb | antibogy-antigen tutorial (based on our online tutorial) | Launch Colab |
# create a directory
cd $HOME
mkdir haddock3-tutorial
cd haddock3-tutorial
# setup python
python3.13 -m venv .venv
source .venv/bin/activate
# install jupyter
pip install notebook
# download the notebook
wget https://raw.githubusercontent.com/haddocking/haddock3/refs/heads/main/notebooks/HADDOCK3-antibody-antigen.ipynb
# run the jypyter notebook server
jupyter notebookgo to http://localhost:8888/notebooks/HADDOCK3-antibody-antigen.ipynb - click Run All
Marimo notebooks are reactive Python notebooks that run locally. Unlike Jupyter, every cell updates automatically when its inputs change, making them well suited for interactive analysis workflows.
| Notebook | Description |
|---|---|
HADDOCK3-interface-analysis.py |
Energy minimization → contact map (chord chart + heatmap) with optional alanine scanning, all driven from a single configuration panel; includes a HADDOCK scoring table with energy terms (Evdw, Eelec, Edesolv, BSA) and per-interface scores for multi-chain complexes |
HADDOCK3-scoring.py |
Score and cluster a set of PDB models (emscoring → clustfcc → caprieval → contactmap); displays per-model and per-cluster statistics tables with traceback to the original input model, interactive 3D viewer and download button for selected models, contact-map chord charts; optional reference structure upload enables full CAPRI metrics (irmsd, fnat, lrmsd, DockQ) |
Prerequisites: HADDOCK3 installed in development mode (pip install -e '.[dev]') and CNS available on your PATH (see docs/pages/CNS.md).
# install marimo (if not already present)
pip install marimo
# launch the notebook from the repository root
marimo run notebooks/HADDOCK3-interface-analysis.pyAlternatively you can start it in edit mode to see the Python code and be able to edit it.
marimo edit notebooks/HADDOCK3-interface-analysis.py
The notebook opens in your browser at http://localhost:2718. Upload a multi-chain PDB file, adjust the configuration panel, and click Run HADDOCK3 Workflow.
marimo run notebooks/HADDOCK3-scoring.py
# or in edit mode:
marimo edit notebooks/HADDOCK3-scoring.pyUpload one or more PDB complex files (each file is one model of the complex), optionally upload a reference structure to enable CAPRI quality metrics, adjust the clustering and contact-map settings, and click Run HADDOCK3 Scoring Workflow.
Results include:
- Sortable per-cluster and per-model statistics tables; an input model column traces each scored model back to the original uploaded file via
haddock3-traceback - Interactive 3D viewer and a download button for any selected model (requires internet access for the 3Dmol.js CDN)
- Chord charts and heatmaps from the contact-map analysis
Free for non-profit, but does require registration with the EGI SSO.
Once logged in and the session is active, start a Jypyter notebook and import the notebook from an URL specifying the same https address as above to the local execution.