Development of simultaneous interaction prediction approach (SiPA) for the expansion of interaction network of traditional Chinese medicine

摘要: Due to the lack of enough interaction data among compositions, targets and diseases, it is difficult construct a complete network Traditional Chinese Medicine (TCM) that comprehensively reflects active compositions their synergistic in terms specific diseases. Therefore, mapping full spectrum between compounds central importance when we use pharmacology approach explore therapeutic potential TCM. To address this challenge, developed large-scale simultaneous prediction (SiPA) integrated one based simple inference model (SIM), focusing on ‘logical relevance’ compounds, proteins or another compound-target correlation space (CTCS-IPM) was built basis canonical analysis (CCA) estimate position (or targets) compound-protein correlated space. Then SiPA applied discover reliable multiple interactions for expansion TCM, compound Salvia miltiorrhiza. By means analysis, related synergy underlying cardiovascular diseases were evaluated expanded original networks. Part new validated with existing experimental evidence molecular docking. As known test dataset, established combination proved make highly accurate prediction, showing well performance SIM high recall rate 85.2% CTCS-IPM. 710 pairs interactions, 24 compound-cardiovascular disease 294 disease-protein predicted Results suggested could dramatically improve completeness effectiveness network. Validation results literature docking manifested inferred had good reliability. We provided practical efficient way TCM ingredients, which not limited by negative samples, sample size target 3D structures. help researchers more accurately prioritize effective completely treating indicating effectively identifying candidate target) drug discovery.