HPLC fingerprint of flavonoids in Rosa davurica by principal component analysis and cluster analysis

Different varieties and different origins fingerprints of flavonoids in 30 batchs of Rosa davurica were established by HPLC method, principal component analysis and cluster analysis were carried out. The athena-C18 column was selected for separation with gradient elution and the mobile phases was acetonitrile-0.1% ( v/v) phosphoric acid aqueous solution.The flow rate was 1.0 mL/min, the detection wave length was 360 nm. The results showed that the similarities of 30 batches of Rosa davurica were between 0.772 and 0.995, among which 22 batchs of Rosa xanthine Lindl were more than 0.958, and the RSD of relative retention time was less than 1.0%.15 common characteristic peaks were selected by total pattern method, and 3 common peaks were identified, including rutin, hyperoside and quercetin. The peak area RSD of precision, repeatability and stability of the method were less than 2.9%, 4.1% and 4.9%, respectively.The recoveries of rutin, hyperoside and quercetin were 97.3% ±1.2%, 96.2% ± 1.9% and 98.2% ± 1.1% respectively, and the RSD were 3.2%, 2.4% and 2.6% respectively, which were in accordance with the fingerprint test requirements. The fingerprints were identified with the models of principal component analysis and cluster analysis carried out by SPSS 22.0. The results indicated that 30 batches of samples were divided into 2 categories by principal component analysis, and the cumulative variance contribution rate of the two principal components was 80.3%.The samples were divided into 3 categories by cluster analysis, and the two classification methods were consistent with the similarity evaluation results.The method had the advantages of good precision, repeatability and stability.In combination with similarity evaluation, principal component analysis and cluster analysis, the methods can be used for rapid identification of different varieties of Rosa davurica and provide a effective evaluation method for their quality control.