PyCoM: a python library for large-scale analysis of residue-residue coevolution data.

Area of Science:

• Computational biology
• Bioinformatics
• Structural biology

Background:

• Computational methods for detecting correlated amino acid positions are crucial for predicting protein contacts, structures, and mutation effects.
• Existing tools for coevolution scoring lack a centralized repository for large-scale studies integrating biological and structural annotations.

Purpose of the Study:

• To introduce PyCoM, a Python library and database for querying and analyzing protein coevolution matrices and sequence alignments.
• To provide a centralized resource for large-scale coevolutionary analyses leveraging UniProtKB/Swiss-Prot annotations.

Main Methods:

• Developed PyCoM, a Python library for accessing a precompiled database (PyCoMdb) of coevolution matrices and sequence alignments for over 457,000 proteins.
• Integrated filters for biological and structural annotations from UniProtKB/Swiss-Prot.
• Supported access via Jupyter Notebooks, Python scripts, and a web API.

Main Results:

• PyCoM enables efficient querying and analysis of coevolution data for a large protein set (≤ 500 residues).
• Facilitates statistical analyses of residue coevolution patterns using rich biological context.
• Provides accessible data for both novice and advanced users.

Conclusions:

• PyCoM serves as a valuable open-source resource for the scientific community.
• It supports the development of data-driven computational models for understanding protein structure, stability, function, and design.
• Enhances large-scale coevolutionary studies by centralizing data and analysis tools.