A Python package for the detection and analysis of protein cavities.
To set up the dependencies for pycasta, you can use either pip (with requirements.txt) or conda (with environment.yml).
Make sure you are in the root directory of the repository (where requirements.txt is located), then run:
pip install -r requirements.txtMake sure you are in the root directory of the repository (where environment.yml is located), then run:
conda env create -f environment.yml
conda activate pycasta-envNote:
Some dependencies such asfreesasaandpymol2may require additional system libraries or specific installation steps, especially on Windows.
If you encounter installation issues, please consult the official documentation of those packages.
After installing the dependencies, you can run the included example analysis with:
cd ./src/pycasta
python run_analysis.pyThis will execute a demo analysis with the default settings.
Check the output and log messages for results or any errors.
A small example dataset is included in the data/ directory for testing and demonstration purposes.
The data/ folder contains three subdirectories:
bounded/– for analysis of molecules in the bound state (paired analysis)unbounded/– for analysis of molecules in the unbound state (paired analysis)tables/– contains Excel files that map the correspondence between each bound and unbound molecule for paired analyses
If you want to perform a paired analysis (comparing bound and unbound forms),
you must provide an Excel file (for example: mapping.xlsx) inside the tables/ directory.
This file should specify the correspondence between each bound and unbound molecule.
The Excel file should contain at least two columns:
| bound_molecule | unbound_molecule |
|---|---|
| 1abc_bound.pdb | 1abc_unbound.pdb |
| 2xyz_bound.pdb | 2xyz_unbound.pdb |
| ... | ... |
Place the mapping file inside the
tables/directory.
For a single-molecule analysis (not paired), simply use a structure where the heteroatom is already present.
You can place such files directly in the appropriate folder.
Feel free to use or modify the example data to test different types of analyses.
For larger or custom datasets, please follow the same folder structure and file naming conventions.