Abstract
Important protein identification methods based on centrality have reached a high level of accuracy. However, there is a need to improve centrality algorithms because they currently consider the nature of protein–protein interaction (PPI) network topology but not protein properties. To improve the centrality algorithm, we introduce the weighted PageRank algorithm, which represents node importance, and the protein interaction combined_score, which represents PPI network edge importance in the STRING database to construct a weighted PPI network. We constructed yeast protein networks for simulations to validate the improved algorithm. Finally, we studied the hepatotoxicity of Pinelliae Rhizoma by applying the PageRank and Edge Clustering (PEC) algorithm. Our study shows that the PageRank and Edge Clustering algorithm can pre-screen important targets and has superior accuracy, sensitivity, and specificity to other centrality algorithms.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
E-mail addresses: zhangaijunzaj{at}126.com (Aj Zhang) slightlzh{at}163.com (Zh Li) qgp-researcher{at}foxmail.com (Gp Qi) drew13{at}163.com (Z XU) 953691839{at}qq.com (X Liu) 1781617074{at}qq.com (Jm Ma) zuojing1006{at}hotmail.com (Zj Li)