enesemretas/hingeprot_fortran
Colab/Jupyter UI wrapper for the HingeProt pipeline to predict protein hinges and rigid parts using elastic network models (GNM/ANM), with chain selection, CA-only support, and 3D visualization.
HingeProt — Colab/Jupyter UI
This repository provides a Google Colab / Jupyter Notebook user interface (UI) for running the classic HingeProt pipeline (rigid parts + hinge prediction) based on Elastic Network Models (GNM/ANM).
It supports:
- Running via PDB code (downloaded from RCSB) or uploading your own PDB/ENT
- Multi-chain detection and chain selection
- CA-only PDB inputs (coarse-grained workflow)
- Interactive 3D preview (3Dmol.js / py3Dmol) and downloadable outputs
🚀 Run on Google Colab
Open the notebook here:
Direct link: https://colab.research.google.com/github/enesemretas/hingeprot_fortran/blob/main/HingeProt.ipynb
Inputs
- PDB Code (e.g.,
4cln)
Downloads from RCSB and prepares the input. - Upload PDB/ENT
Useful for custom structures or modified models. - Chain selection
Detects chain IDs and allows selecting one or more chains. - GNM/ANM cutoffs
Select from a list or enter custom values.
Outputs
- Rigid Parts table (for slowest modes)
- Hinge residues
- Short flexible fragments
- Mode files (mode1/mode2) with download buttons
- Mode trajectory viewer
Outputs are stored under a run folder (default: /content/hingeprot_runs in Colab).
🧩 Requirements
Recommended
- Google Colab (simplest; most dependencies are handled automatically)
Local Jupyter (Linux recommended)
- Python 3.x
- Jupyter Notebook/Lab
pip install ipywidgets requests py3Dmolgit,perl- The compiled HingeProt binaries available in the repository
The UI also includes an automated step to install libg2c.so.0 on Debian/Ubuntu-like environments when needed.
🧠 Notes & limitations
- Single-chain focus (original HingeProt behavior): the classic server/pipeline is designed primarily around single-chain analysis, although this UI allows selecting multiple chains.
- CA-only inputs: supported and recommended for coarse-grained ENM workflows.
- Missing residues: neglected in calculations; results reflect available coordinates.
- Residue limit: HingeProt traditionally limits residue count (commonly cited as 999).
📌 Citation
If you use this work, please cite the original HingeProt paper:
Emekli U, Schneidman-Duhovny D, Wolfson HJ, Nussinov R, Haliloglu T. (2008) HingeProt: Automated Prediction of Hinges in Protein Structures. Proteins, 70(4):1219–1227.