Protein complex prediction in large ontology attributed protein-protein interaction networks

Area of Science:

• Computational biology
• Systems biology
• Bioinformatics

Background:

• Protein complexes are crucial for cellular organization and function.
• Existing computational methods for protein complex prediction primarily rely on protein-protein interaction (PPI) network topology.
• Gene Ontology (GO) annotation information is often overlooked in current prediction approaches.

Purpose of the Study:

• To develop a novel computational approach for predicting protein complexes.
• To integrate protein-protein interaction network structure with gene ontology (GO) annotation information.
• To improve the accuracy and effectiveness of protein complex prediction.

Main Methods:

• Constructed ontology-attributed PPI networks using PPI data and GO resources.
• Proposed a novel approach named CSO (Clustering based on Network Structure and Ontology attribute similarity).
• Utilized the complementary nature of structural and GO attribute information for prediction.

Main Results:

• CSO effectively leverages the correlation between frequent GO annotation sets and dense subgraphs.
• The approach was applied to four yeast PPI datasets, successfully predicting numerous known protein complexes.
• Experimental results demonstrated CSO's value and state-of-the-art performance in protein complex prediction.

Conclusions:

• Integrating network structure and GO annotations enhances protein complex prediction.
• CSO offers a valuable and effective method for identifying protein complexes.
• The approach achieves superior performance compared to existing methods.