Using persistent homology and dynamical distances to analyze protein binding.

Area of Science:

• Computational Biology
• Topology
• Biophysics

Background:

• Persistent homology analyzes evolving topological features.
• Barcodes and persistence diagrams are common summary statistics.
• Persistence landscapes offer a functional summary for statistical analysis.

Purpose of the Study:

• Introduce and apply persistence landscapes for biomolecular conformational analysis.
• Detect conformational changes in maltose-binding protein.
• Investigate the utility of persistence landscapes in machine learning and residue analysis.

Main Methods:

• Utilized persistence landscapes as a summary statistic for persistent homology.
• Implemented a permutation test to detect conformational changes.
• Applied support vector machine (SVM) for classification.
• Analyzed the relationship between topological features and protein residues.

Main Results:

• Successfully detected conformational changes between closed and open forms of maltose-binding protein.
• Demonstrated clear separation of conformations using SVM with persistence landscapes.
• Identified active site and allosteric pathway residues near the most persistent topological features.

Conclusions:

• Persistence landscapes provide a powerful tool for analyzing protein dynamics and conformational changes.
• This approach enhances understanding of protein-ligand interactions and allosteric mechanisms.
• The findings offer a novel perspective compared to traditional methods like the anisotropic network model.