Universality in protein residue networks

Area of Science:

• Structural bioinformatics
• Network science
• Computational biology

Background:

• Protein residue networks exhibit complex topological features.
• Understanding these network properties is crucial for deciphering protein structure-function relationships.
• Deviations from universal characteristics in small proteins require further investigation.

Purpose of the Study:

• To investigate the universal topological characteristics of protein residue networks.
• To explain deviations from universality in small proteins based on their domain structure.
• To explore the relationship between topological cavities and protein binding sites.
• To analyze the impact of cutoff values on residue network topology.
• To compare the topological properties of residue networks with various network growth models.

Main Methods:

• Analysis of residue networks from 595 nonhomologous proteins.
• Characterization of network topology, focusing on modularity and cavities.
• Investigation of the effect of different cutoff distance values on network construction.
• Comparison of empirical residue network properties with Erdös-Rényi, Barabási-Albert, and Watts-Strogatz models.

Main Results:

• Protein residue networks universally display modular topology with interconnected clusters and cavities.
• Small proteins (<200 residues) show deviations attributed to their domain structure.
• Topological cavities align well with experimentally identified protein binding sites.
• Cutoff value significantly influences detected cavities and overall network homogeneity.
• Watts-Strogatz model effectively reproduces the topological class and properties of residue networks, unlike Erdös-Rényi or Barabási-Albert models.

Conclusions:

• Protein residue networks possess a universal modular architecture, with variations linked to protein size and domain organization.
• Topological cavities serve as reliable indicators of protein binding sites.
• The Watts-Strogatz model provides a suitable framework for simulating protein residue network topology, with potential for biologically relevant modifications.