Discovering protein complexes in protein interaction networks via exploring the weak ties effect

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Protein complexes are vital for understanding biological processes and structure-function relationships.
• Existing protein-protein interaction network analysis often relies on subgraph density, overlooking crucial edge topological information.
• A need exists to explore edge roles for improved protein complex discovery.

Purpose of the Study:

• To investigate the significance of edge topology in protein-protein interaction networks for protein complex identification.
• To develop a novel method for protein complex detection that leverages edge roles and topological information.

Main Methods:

• Analysis of weak ties and their relationship with topological similarity in protein-protein interaction networks.
• Construction of a virtual network using bridges to identify complex cores as maximal cliques.
• Development of a core-attachment based method for detecting protein complex cores and attachments.

Main Results:

• Identified a 'weak ties' phenomenon where edges connecting less similar vertices are crucial for global network connectivity.
• Demonstrated a negative correlation between weak tie strength and topological similarity.
• The novel core-attachment method outperformed existing algorithms in detecting protein complexes on the yeast PPI network.

Conclusions:

• The weak tie effect is present in protein-protein interaction networks.
• Network density alone is insufficient for characterizing protein complex topology.
• The proposed edge-centric method offers superior performance and biologically significant results for protein complex discovery.