指纹图谱结合模式识别、一测多评的连翘质量评价

Quality Evaluation of Forsythiae Fructus Based on Combinative Method of Fingerprint, Pattern Recognition and Quantitative Analysis of Multi-components by Single Marker

关键词:  连翘  指纹图谱  模式识别  一测多评  聚类分析  主成分分析

Key words:  Forsythiae Fructus  fingerprint  pattern recognition  quantitative analysis of multi-components by single marker(QAMS)  cluster analysis(CA)  principal component analysis(PCA)

摘要:

目的 建立连翘药材的指纹图谱并结合模式识别、一测多评技术对其进行系统、全面的质量评价。方法 采用HPLC测定10批次连翘样品的指纹图谱，建立指纹图谱共有模式，通过相似度分析，结合聚类分析、主成分分析等模式识别技术对其进行定性评价。以连翘苷为内参物，建立连翘酯苷A、（+）松脂醇-β-D-吡喃葡萄糖苷、连翘酯素与内参物的相对校正因子，并进行含量计算，实现一测多评的定量评价；同时采用外标法测定10批连翘中4个成分的含量，比较计算值与实测值的差异，验证一测多评法的准确性。结果 建立的指纹图谱共标定13个共有峰，并对其中4个进行了归属，分别是连翘酯苷A、（+）松脂醇-β-D-吡喃葡萄糖苷、连翘苷、连翘酯素，10批次连翘样品指纹图谱与对照指纹图谱的相似度均>0.98，通过聚类分析及主成分分析将其分为3类。建立的相对校正因子重现性良好，采用校正因子计算的含量值与实测值之间相关系数均>0.99，相对误差均<3%。结论 本研究建立的分析方法直观、准确、简便、可行。指纹图谱结合模式识别、一测多评技术进行定性定量分析可以为连翘的质量控制提供科学依据。

Abstract:

OBJECTIVE To establish the fingerprint of Forsythiae Fructus and conduct a systemic and comprehensive quality evaluation using pattern recognition and quantitative analysis of multi-components by single marker(QAMS). METHODS Ten batches of Forsythiae Fructus were determined by HPLC and a common mode of fingerprint was established. The similarities between fingerprints of 10 batches samples and common mode were evaluated. The further qualitative analysis of the drug was carried out by similarity evaluation, cluster analysis(CA), principal component analysis(PCA). QAMS was performed. Forsythin was selected as internal reference and the relative correction factors(RCF) of forsythiaside A, (+)pinoresinol-β-D-glucoside and phillygenin to it were calculated. At the same time, the content of 4 components in 10 batches of forsythia was measured by external standard method, and the difference between the calculated value and the measured value was compared to verify the accuracy of QAMS. RESULTS There were 13 common peaks in fingerprints of 10 batches of Forsythiae Fructus, four of which were identified. The similarities of fingerprints were >0.98. The batches of the drug were divided into 3 categories by CA and PCA. The established RCF had a good reproducibility. The correlation coefficients between the content values calculated using the correction factors and the measured values were all >0.99, and the relative errors were all <3%. CONCLUSION The developed method is accurate, simple and feasible. The quality of Forsythiae Fructus could be evaluated qualitatively and quantitatively using a fingerprint combined with pattern recognition technique and QAMS technique. Moreover, it will also lay a solid scientific basis for the further quality control of Forsythiae Fructus.