Chemometrics-based approach to modeling quantitative composition-activity relationships for Radix Tinosporae

Quantitative composition-activity relationship (QCAR) study makes it possible to discover active components in traditional Chinese medicine (TCM) and to predict the integral bioactivity by its chemical composition. In the study, 28 samples of Radix Tinosporae were quantitatively analyzed by high performance liquid chromatography, and their analgesic activities were investigated via abdominal writhing tests on mice. Three genetic algorithms (GA) based approaches including partial least square regression, radial basis function neural network, and support vector regression (SVR) were established to construct QCAR models of R. Tinosporae. The result shows that GA-SVR has the best model performance in the bioactivity prediction of R. Tinosporae; seven major components thereof were discovered to have analgesic activities, and the analgesic activities of these components were partly confirmed by subsequent abdominal writhing test. The proposed approach allows discovering active components in TCM and predicting bioactivity by its chemical composition, and is expected to be utilized as a supplementary tool for the quality control and drug discovery of TCM.