Identification of protein complexes from multi-relationship protein interaction networks. (July 2016)

• Record Type: Journal Article
• Authors: Li, Xueyong; Wang, Jianxin; Zhao, Bihai; Wu, Fang-Xiang; Pan, Yi
• Is Part Of: Human genomics. Volume 10(2016)Supplement 2
• Page Start: 61
• Page End: 70
• Publication Date: 2016-07
• Subjects: Genomics -- Periodicals; Human genome -- Periodicals; Genetic Research -- Periodicals; Pharmacogenetics -- Periodicals; 611.01816
• DOI: 10.1186/s40246-016-0069-z
• Languages: English
• ISSNs: 1479-7364

Abstract

Abstract Background Protein complexes play an important role in biological processes. Recent developments in experiments have resulted in the publication of many high-quality, large-scale protein-protein interaction (PPI) datasets, which provide abundant data for computational approaches to the prediction of protein complexes. However, the precision of protein complex prediction still needs to be improved due to the incompletion and noise in PPI networks. Results There exist complex and diverse relationships among proteins after integrating multiple sources of biological information. Considering that the influences of different types of interactions are not the same weight for protein complex prediction, we construct a multi-relationship protein interaction network (MPIN) by integrating PPI network topology with gene ontology annotation information. Then, we design a novel algorithm named MINE (identifying protein complexes based on Multi-relationship protein Interaction NEtwork) to predict protein complexes with high cohesion and low coupling from MPIN. Conclusions The experiments on yeast data show that MINE outperforms the current methods in terms of both accuracy and statistical significance.