[Bioinformatics] PyCoM: a python library for large-scale analysis of residue–residue coevolution data https://doi.org/10.1093/bioinformatics/btae166
Python Library (both local and remote!) & REST API for accessing the PyCom Database.
PyCom is an open-source, queryable database of Protein Residue-Residue Contacts (Coevolution Matrices) and annotation data derived from UniProt (Swiss-Prot to be precise!) using HHBlits/HHFilter/CCmpred.
More information can be found at: https://pycom.brunel.ac.uk/
Install the PyCom library with:
pip3 install git+https://github.com/scdantu/pycom
Requirements:
- Python 3.8+
- numpy / pandas / h5py / requests, installed automatically
Get the pycom.db and pycom.mat files (115 GB total) from:
https://pycom.brunel.ac.uk/downloads/
Not necessary for remote use.
Examples can be found in 00_Getting_Started_Locally.ipynb.
This requires the pycom.db and pycom.mat files.
The library can be used remotely without downloading the pycom.db and pycom.mat files.
A tutorial can be found in 00_Getting_Started_Remotely.ipynb.
It is also possible to call the REST API directly. Documentation (Swagger): https://pycom.brunel.ac.uk/api/spec/ Endpoint: https://pycom.brunel.ac.uk/api/ Mini Tutorial: 00_WebAPI.ipynb
We also provide a webservice to download the alignment data used to generate the coevolution matrices.
See: https://pycom.brunel.ac.uk/alignments/
If our work was helpful, you can cite us!
@article{bibik2024pycom,
author = {Bibik, Philipp and Alibai, Sabriyeh and Pandini, Alessandro and Dantu, Sarath Chandra},
title = "{PyCoM: a python library for large-scale analysis of residue–residue coevolution data}",
journal = {Bioinformatics},
volume = {40},
number = {4},
pages = {btae166},
year = {2024},
url = {https://doi.org/10.1093/bioinformatics/btae166},
}The PyComDB database creation and HPC deployment was largely done by Philipp Bibik under scientific advisement from Sarath Dantu.
The Python interface and REST API (this repo) was primary written by Philipp Bibik, with contributions from Sarath Dantu.
Tutorials and documentation were created by Sarath Dantu and Philipp Bibik.
Sarath Dantu served as the PI for this project. Without additional support from Alessandro Pandini, this project would not have been possible.