ABSTRACT
OBJECTIVE：To identify Panax notoginseng and its processed products . METHODS ：The fingerprint was established by HPLC. Using ginsenoside Rb 1 as reference ，HPLC fingerprints of 15 batches of P. notoginseng and its processed products were drawn and the similarity evaluation was conducted by using the Similarity Evaluation System for TCM Chromatographic Fingerprints（2012 edition）. The common peaks were confirmed by comparing with substance control. SPSS 21.0 and SIMCA 14.1 software were used to perform cluster analysis ，principal component analysis and orthogonal partial least squares-discriminant analysis；taking the variable importance projection （VIP）value greater than 1 as the standard ，the differential marker components causing the quality difference between P. notoginseng and its processed products were screened. IR fingerprints of P. notoginseng and its processed products were established by OMNIC 8.2.0 software，and the spectral similarity was evaluated ；double index sequence analysis was used to analyze absorption peaks of IR fingerprints of 15 batches of P. notoginseng and its processed products. RESULTS ：There were 16 common peaks in the fingerprints of 15 batches of P. notoginseng ， and the similarities were 0.911-1.000；there were 25 common peaks in the fingerprints of processed products ，and the similaritieswere 0.862-1.000. They had 12 identical common peaks ，and wang668@sina.com three of them were ident ified as sanchinoside R 1，ginsenoside Rg1 and ginsenoside Rb 1. Results of cluster analysi s showed that when the distance was 10，15 batches of P. notoginseng could be clustered into two categories ，SW1-SW5 into one category ，SH1-SH5 and SQ 1-SQ5 into one category ，ZW1-ZW5，ZH1-ZH5 and ZQ1-ZQ5 of 15 batches of processed products could be clustered into one category. When the distance was 5，15 batches of P. notoginseng could be clustered into three categories ，SW1-SW5 into one category ，SH2-SH5 and SQ 2 into one category ，SQ1， SQ3-SQ5 and SH 1 into one category. Fifteen batches of processed products could be clustered into two categories ，ZW1-ZW5 into one category ，ZH1-ZH5 and ZQ 1-ZQ5 into one category. The results of principal component analysis showed that the cumulative variance contribution rate of the first two principal components was 80.104％ . The results of orthogonal partial least squares-discriminant analysis showed that the VIP values of the five peaks were greater than 1，which were peak H ，peak G ，peak J，peak F （ginsenoside Rg 1）and peak I. The similarity of IR fingerprints of 15 batches of P. notoginseng and its processed products were 0.889 7-1.000 0 and 0.972 8-1.000 0；the common peak rates were 80％-100％，and the variation peak rates were 0-17.65％ and 0-18.75％，respectively. By comparing the wave numbers of absorption peaks ，it was found that there were differences between P. notoginseng at 3 440 and 1 450 cm－1 and processed products at 1 530 and 575 cm－1. CONCLUSIONS ：Established HPLC fingerprint and IR fingerprint have good similarity ，and could effectively distinguish P. notoginseng and its processed products. P. notoginseng and its processed products from different habitats have high common peak rate and low variation rate ，and their chemical components are different ；peak H ，peak G ，peak J ，ginsenoside Rg 1 and peak I are differential marker components causing the quality difference between P. notoginseng and processed products.