Objective To propose a new method for mining compatibility rules of traditional Chinese medicine (TCM) formulation from the perspective of weighted projection of bipartite networks, and to predict the combination of new drugs to provide a basis for guiding the clinical treatment of acute nasopharyngitis.
Methods Using the acute nasopharyngitis prescription data in the Traditional Chinese Medicine Integrated Database (TCMID) as the data source, a bipartite network is constructed by extracting the prescription and drug nodes.A drug network projection map was then obtained using weighted projection.Social network analysis was performed combined with weighted projection of bipartite networks, and compatibility rules of "Jun-Chen-Zuo-Shi" in TCM via hierachical cluster analysis based on Pearson correlation.Besides, link prediction was used for core drug prediction.
Results The combination of bipartite network weighted projection and Pearson correlation for systematic clustering analysis played a significant role in the study of the compatibility rules of TCM prescriptions.In link prediction, 11 link prediction indicators were selected, and weighted and unweighted algorithms were distinguished.The final calculation showed that the area under the curve (AUC) of the weighted indicators were generally higher than that of the unweighted network.Among the weighted indicators, the indicator with the highest AUC index was the network resource allocation Index.A total of 7 groups of drug combinations were predicted, including Baitouweng-Maokezi, Anxixiang-Shijiaocao, Baihuacha-Fuzi, etc.
Conclusion The bipartite network weighted projection method is practical and effective in revealing the compatibility rules of TCM and predicting drug combination.
DownLoad: