MIToS.jl: mutual information tools for protein sequence analysis in the Julia language

Area of Science:

• Computational Biology
• Bioinformatics
• Structural Biology

Background:

• Protein sequence and structure analysis is crucial for understanding function.
• Existing tools may lack integrated approaches for mutual information and structural data.
• Efficient management of protein multiple sequence alignments (MSAs) and structures is needed.

Purpose of the Study:

• Introduce MIToS, a Julia environment for mutual information analysis.
• Provide a framework for managing protein MSAs and structures (PDB).
• Streamline the analysis of sequence and structural information for protein contact prediction.

Main Methods:

• Developed MIToS in the Julia language.
• Integrated sequence and structural data using SIFTS.
• Implemented and tested mutual information measures on residue contingency tables.
• Applied a BLOSUM62-based pseudo-count strategy for analysis.

Main Results:

• MIToS facilitates straightforward analysis of Pfam MSAs.
• The framework enables optimization and testing of measures for contact prediction.
• Demonstrated utility with a BLOSUM62-based mutual information analysis example.

Conclusions:

• MIToS offers a unified environment for protein sequence and structure analysis.
• The software simplifies complex mutual information calculations and contact prediction.
• MIToS is a valuable tool for researchers in computational and structural biology.