Application of Improved K-mean Clustering in Predicting Protein-Protein Interactions

With the increasing availability of complete genome sequences, there are more excellent opportunity for further research and development of tools for functional studies in proteomics. Phylogenetic profile is a basic non-homology method in functional proteomics. This paper describes the theory of phylogenetic profile, focuses on the improvement of K-mean clustering algorithm in analyzing the protein phylogenetic profile; also compare the method with the traditional K-mean clustering algorithm. The experimental results show that the improved K-mean clustering algorithm is fast and effective; it can quickly converge to the approximate optimal solution.