Table of Contents Author Guidelines Submit a Manuscript
BioMed Research International
Volume 2015 (2015), Article ID 165186, 7 pages
Research Article

Protein Complex Discovery by Interaction Filtering from Protein Interaction Networks Using Mutual Rank Coexpression and Sequence Similarity

1Institute of Biochemistry and Biophysics, University of Tehran, Enghelab Avenue, P.O. Box 13145-1384, Tehran, Iran
2Science College, University of Tehran, Tehran, Iran

Received 19 August 2014; Revised 19 November 2014; Accepted 1 December 2014

Academic Editor: Zhirong Sun

Copyright © 2015 Ali Kazemi-Pour et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


The evaluation of the biological networks is considered the essential key to understanding the complex biological systems. Meanwhile, the graph clustering algorithms are mostly used in the protein-protein interaction (PPI) network analysis. The complexes introduced by the clustering algorithms include noise proteins. The error rate of the noise proteins in the PPI network researches is about 40–90%. However, only 30–40% of the existing interactions in the PPI databases depend on the specific biological function. It is essential to eliminate the noise proteins and the interactions from the complexes created via clustering methods. We have introduced new methods of weighting interactions in protein clusters and the splicing of noise interactions and proteins-based interactions on their weights. The coexpression and the sequence similarity of each pair of proteins are considered the edge weight of the proteins in the network. The results showed that the edge filtering based on the amount of coexpression acts similar to the node filtering via graph-based characteristics. Regarding the removal of the noise edges, the edge filtering has a significant advantage over the graph-based method. The edge filtering based on the amount of sequence similarity has the ability to remove the noise proteins and the noise interactions.