Automatic authentication and distinction of Epimedium koreanum and Epimedium wushanense with HPLC fingerprint analysis assisted by pattern recognition techniques

This study aims to identify and discriminate between two commonly confused traditional Chinese medicines, Epimedium wushanense and Epimedium koreanum, using pattern recognition aided fingerprint analysis of their secondary metabolites. Samples of the two species were collected during different stages of their growth period. The HPLC generated chromatographic data were analyzed using principal components analysis (PCA) and hierarchical cluster analysis (HCA). Two major clusters were formed, each consisting of a single species. The entire dataset was then divided into two: a training set and a test set. Supervised pattern recognition techniques, soft independent modeling by class analogy (SIMCA) and back propagation artificial neural network (BP-ANN), were performed. SIMCA failed to predict one sample, whereas BP-ANN precisely predicted the whole test set. In conclusion, fingerprint analysis assisted by pattern recognition techniques is a potential strategy for the authentication and differentiation of species used in herbal medicines.

Graphical abstractFigure optionsDownload full-size imageDownload as PowerPoint slideHighlights► Supervised pattern recognition was applied to distinguish two Epimedium species. ► The HPLC fingerprint profile was taken as input to pattern recognition techniques. ► Significant variation in the composition of two species was visualized by PCA.