RESUMEN
Objective: To accelerate the development of nation medicine, ultra-high performance liquid chromatography (UPLC) combined with content analysis of index compound, cluster analysis (HCA), and principal component analysis (PCA) was used to evaluate and discriminate three kinds of national folk medicinal plants, Swertia davidi, S. punicea, and S. angustifolia. Methods: The chromatograms of 27 samples from three kinds of Swertia L. were collected. The chromatograms were imported in Similarity Evaluation System for Chromatographic Fingerprint of Chinese Materia Medica 2004A to obtain data retention time and peak area of samples. Three sets of samples of similarity were analyzed. The peak area data were imported in SPSS and SIMCA-P+ software for cluster analysis and principal component analysis, dendrogram and principal component scores were obtained, and the clustering effect was observed. Results: The line relationship of this way was good (R2 > 0.9993), with high precision instrument (RSD of 0.19%-2.45%), the method had good reproducibility (RSD of 0-1.77%), recovery was between 95.8% and 102.3% (RSD of 0.45%-2.81%). Three kinds of Swertia L. medicinal plants were different in the contents of index compounds, and the contents of index compounds in the same species were different. Swertiamarin: S. davidi (0.54-10.05 mg/g), S. angustifolia (undetectable), S. punicea (1.76-15.62 mg/g); gentiopicroside: S. davidi (4.27-11.18 mg/g), S. angustifolia (0.26-3.58 mg/g), S. punicea (1.13-16.11 mg/g); sweroside: S. davidi (0.02-0.28 mg/g), S. angustifolia (0.02-0.14 mg/g), S. punicea (0.11-44.52 mg/g); mangiferin: S. davidi (0.14-0.55 mg/g), S. angustifolia (0.03-0.04 mg/g), S. punicea (0.01-13.49 mg/g). The accuracy of dendrogram classification was 85.2%. The contents of index compounds of three false anomalies were less than the remaining nine samples of S. punicea. The effect of discrimination by PCA scores plot of three Swertia L. plants was good, and S. punicea with dispersive distribution indicated that the individuals were significant difference. Conclusion: The method to evaluate and discriminate three Swertia L. medicinal plants is feasible by UPLC combined with analysis of index compound HCA and PCA. In this article, S. punicea is more valuable compared S. davidi and S. angustifolia.