Abstract—Protein-protein interactions play a key role in the completion of cellular functions and usually correlate to each other in the form of a protein-protein interaction network. In this paper, we propose an algorithm to study the proteinprotein interactions using the properties of protein-protein interaction networks. First, the algorithm constructs a proteinprotein network according to the two query proteins. All the neighbors of the two query proteins obtained from the online protein databases are also included in the network. Second, the improved network partition algorithm was used to split the network into sub networks. Finally, a scoring function was proposed based on network clusters to predict the proteinprotein interactions. The experimental results show that the scoring function based on the PPI network predicts the protein-protein interactions accurately.
Index Terms—Protein-protein interaction; graph theory; score function.
The authors are with the Department of Systems and Computer Science, Howard University, Washington, DC 20059, USA.(e-mail: hli@scs.howard.edu; chunmei@scs.howard.edu; blegand@scs.howard.edu).
Cite:Hui Li, Chunmei Liu, and Legand Burge, "Predicting Protein-Protein Interactions Based on PPI Networks," International Journal of Machine Learning and Computing vol.2, no. 6, pp. 794-797, 2012.