Generalized correlation-based dynamical network analysis: a new high-performance approach for identifying allosteric...

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Molecular interactions govern cellular processes, including protein complex formation and enzyme activity.
• Identifying critical residues is vital for understanding biochemical pathways and for therapeutic or engineering applications.
• Dynamical network analysis (DNA) is established for identifying molecular pathways but is computationally limited for large systems.

Purpose of the Study:

• To evolve the dynamical network analysis method for scalability to large biomolecular systems.
• To develop an intuitive and interactive interface for analyzing molecular interactions.
• To enable the identification of essential residues and information pathways in complex biological molecules.

Main Methods:

• Developed an enhanced dynamical network analysis protocol integrated with Jupyter notebooks.
• Implemented an optimized, parallel generalized correlation calculation linear with system size.
• Incorporated automatic detection of solvent/ion residues, community clustering, and betweenness centrality calculations.
• Utilized visual molecular dynamics (VMD) for network visualization on biomolecular structures.

Main Results:

• The enhanced DNA method successfully analyzed large systems up to 2.5 million atoms.
• Investigated OMP-decarboxylase, leucyl-tRNA synthetase, and respiratory complex I.
• Demonstrated efficient and scalable identification of molecular pathways and key residues.

Conclusions:

• The updated protocol offers an intuitive and interactive interface for large macromolecular complex analysis.
• This advancement overcomes previous computational limitations of DNA for large biomolecular systems.
• Provides a valuable tool for researchers in drug discovery and bioengineering.