Protein-protein interaction network comparison based on wavelet and principal component analysis

Abstract

Various Protein-Protein Interaction(PPI) databases effectively promote development of data analysis. But the data in each database are different, even they may come from the same species. So the reliability of assessment on data of protein network becomes an important research topic. This paper first introduces the wavelet and principal component analysis method in this area, and combines them to analyze the similarity of the species PPI network in different databases. The credibility of species PPI Network is established based on the similarity. Numerical experiments illustrates that this method is able to reflect credibility of biological networks, and consistent with the actual level of every special PPI network. And it can be applied to other large networks as well.