Combining functional and topological properties to identify core modules in protein interaction networks

Area of Science:

• Proteomics
• Systems Biology
• Bioinformatics

Background:

• Large-scale proteomics technologies generate extensive Protein Interaction Networks (PINs).
• Existing methods for identifying sub-graphs in PINs rely solely on graph theory, neglecting functional context.
• There is a need to integrate domain knowledge with network topology for identifying functionally relevant sub-graphs.

Purpose of the Study:

• To develop and evaluate novel network measures that combine functional information with topological properties for PIN analysis.
• To introduce SWEMODE (Semantic WEights for MODule Elucidation), an algorithm for identifying functional modules in PINs.
• To demonstrate the effectiveness of the proposed methods in identifying functionally coherent protein clusters.

Main Methods:

• Introduced two weighted network measures: weighted clustering coefficient and weighted average nearest-neighbors degree.
• Calculated interaction weights based on semantic similarity of proteins using Gene Ontology terms.
• Developed the SWEMODE algorithm, which ranks proteins by weighted neighborhood cohesiveness to seed module identification.

Main Results:

• Systematically compared weighted measures with topological counterparts in the yeast PIN.
• Demonstrated that SWEMODE successfully identifies dense sub-graphs containing functionally similar proteins.
• Many identified modules correspond to known protein complexes or their subunits.

Conclusions:

• The developed weighted network measures effectively integrate functional and topological information in PINs.
• SWEMODE provides a robust approach for discovering functionally relevant modules within large protein interaction networks.
• This method enhances the biological interpretability of PINs by identifying functionally cohesive protein groups.