不同剂型夏桑菊颗粒HPLC指纹图谱及其模式识别分析(1)
[摘要]建立夏桑菊有糖和无糖颗粒的HPLC指纹图谱,为其鉴别与有效控制质量提供参考。采用高效液相法采集20批无糖型和34批有糖型夏桑菊颗粒指纹图谱,通过模式识别方法(主成分分析,正交最小二乘法判别分析)分类并筛选其主要差异组分;通过对照品对照的方法鉴定其主要的组分。成功建立夏桑菊有糖和无糖颗粒指纹图谱;主成分分析不能完全分类2种颗粒,而正交最小二乘法判别分析可明显的分为2类;2种颗粒之间差异最大的组分主要有6个,其中3个分别为异迷迭香酸苷、木犀草苷和蒙花苷。该研究建立的模式识别方法有助于夏桑菊颗粒整体质量控制,同时为其质量评价提供一种有效手段。
[关键词]HPLC 指纹图谱;模式识别;夏桑菊颗粒;主成分分析;正交最小二乘法判别分析
[Abstract]To establish the fingerprints of Xiasangju granules (with sugar and non-sugar forms) by HPLC, and provide reference for their identification and effective quality control. High performance liquid chromatography (HPLC) method was used to collect the fingerprints of 20 batches of non-sugar Xiasangju granules and 34 batches of sugar type Xiasangju granules. Their main different components were classified and screened by mode identification methods (principal component analysis, PCA, and orthogonal partial least squares discriminate analysis, OPLS-DA). The principal components were identified by comparing with reference standards. The fingerprints of Xiasangju granules (sugar type and non-sugar type) were established. PCA could not fully classify the two types of granules, while OPLS-DA could obviously classify these two different types of Xiasangju granules. Six components showed greatest difference between two types of granules, including salviaflaside, luteoloside and linarin. The developed mode identification method is helpful to control the overall quality of Xiasangju granules, and it provides an effective approach to quality evaluation.
[Key words]HPLC fingerprint; mode identification; Xiasangju granules; principal component analysis; orthogonal partial least squares discriminate analysis
夏桑菊颗粒为纯中药制剂,主要由夏枯草、桑叶和野菊花等3味中药组成,具有清肝明目,疏风散热,除湿痹,解疮毒之功效[1]。用于风热感冒,目赤头痛,高血压,头晕耳鸣,咽喉肿痛,疔疮肿毒等症,并可作清凉饮料。目前,国内外生产夏桑菊颗粒的企业105家,质量标准研究方面主要以蒙花苷、绿原酸、迷迭香酸、异迷迭香酸为指标的含量测定[2-5]。就类型而言,夏桑菊颗粒主要包括有糖型与无糖型等2种类型。同时,两者在工艺上也存在较大的差别,如无糖型以处方组分直接水煮提取进行制备;而有糖型,先乙醇浸渍提取部分野菊花,然后再水提取剩余部分处方进行制备,最后加入野菊花浸渍液体与适量蔗糖制得。因此,2种颗粒在工艺上有较大的差异,再加上中药复方本身成分的复杂性以及靶点的多样性,需要整体考虑其质量控制模式。
模式识别方法已被应用于多个学科的分析工作,包括药物的质量控制、差异标记物的筛选、植物分类等研究[6-8]。尤其在指纹图谱等多维的数据分析中显示出优势,也是其重要的分析手段,已经得到广泛的应用[9-10]。目前常用的方法主要分为2类,一类是无监督的分析方法主要有PCA,ICA,cluster等;另一类被称为有监督的分析方法,主要有PLS-DA,OPLS,DA,KNN,神经网络等[11]。随着数据的不断复杂,一些更先进的机器学习方法被应用,如支持向量机(SVM),随机森林(RF)等[12]。其中PCA,OPLS在多个学科取得了广泛的应用。并体现了其在复杂数据分析中的诸多优势,如较高的预测精度、广泛的适用性、线性数据的分析能力等。因而本文采用PCA,PLS-DA等手段分析有糖型与无糖型夏桑菊颗粒指纹图谱,寻到其主要的差异成分,为夏桑菊颗粒的质量标准和选择夏桑菊颗粒的含量测定指标提供很好的借鉴和参考。 (夏伯候 严东 曹艺 周亚敏 李亚梅 谢嘉驰 柏玉冰 廖端芳 林丽美)
[关键词]HPLC 指纹图谱;模式识别;夏桑菊颗粒;主成分分析;正交最小二乘法判别分析
[Abstract]To establish the fingerprints of Xiasangju granules (with sugar and non-sugar forms) by HPLC, and provide reference for their identification and effective quality control. High performance liquid chromatography (HPLC) method was used to collect the fingerprints of 20 batches of non-sugar Xiasangju granules and 34 batches of sugar type Xiasangju granules. Their main different components were classified and screened by mode identification methods (principal component analysis, PCA, and orthogonal partial least squares discriminate analysis, OPLS-DA). The principal components were identified by comparing with reference standards. The fingerprints of Xiasangju granules (sugar type and non-sugar type) were established. PCA could not fully classify the two types of granules, while OPLS-DA could obviously classify these two different types of Xiasangju granules. Six components showed greatest difference between two types of granules, including salviaflaside, luteoloside and linarin. The developed mode identification method is helpful to control the overall quality of Xiasangju granules, and it provides an effective approach to quality evaluation.
[Key words]HPLC fingerprint; mode identification; Xiasangju granules; principal component analysis; orthogonal partial least squares discriminate analysis
夏桑菊颗粒为纯中药制剂,主要由夏枯草、桑叶和野菊花等3味中药组成,具有清肝明目,疏风散热,除湿痹,解疮毒之功效[1]。用于风热感冒,目赤头痛,高血压,头晕耳鸣,咽喉肿痛,疔疮肿毒等症,并可作清凉饮料。目前,国内外生产夏桑菊颗粒的企业105家,质量标准研究方面主要以蒙花苷、绿原酸、迷迭香酸、异迷迭香酸为指标的含量测定[2-5]。就类型而言,夏桑菊颗粒主要包括有糖型与无糖型等2种类型。同时,两者在工艺上也存在较大的差别,如无糖型以处方组分直接水煮提取进行制备;而有糖型,先乙醇浸渍提取部分野菊花,然后再水提取剩余部分处方进行制备,最后加入野菊花浸渍液体与适量蔗糖制得。因此,2种颗粒在工艺上有较大的差异,再加上中药复方本身成分的复杂性以及靶点的多样性,需要整体考虑其质量控制模式。
模式识别方法已被应用于多个学科的分析工作,包括药物的质量控制、差异标记物的筛选、植物分类等研究[6-8]。尤其在指纹图谱等多维的数据分析中显示出优势,也是其重要的分析手段,已经得到广泛的应用[9-10]。目前常用的方法主要分为2类,一类是无监督的分析方法主要有PCA,ICA,cluster等;另一类被称为有监督的分析方法,主要有PLS-DA,OPLS,DA,KNN,神经网络等[11]。随着数据的不断复杂,一些更先进的机器学习方法被应用,如支持向量机(SVM),随机森林(RF)等[12]。其中PCA,OPLS在多个学科取得了广泛的应用。并体现了其在复杂数据分析中的诸多优势,如较高的预测精度、广泛的适用性、线性数据的分析能力等。因而本文采用PCA,PLS-DA等手段分析有糖型与无糖型夏桑菊颗粒指纹图谱,寻到其主要的差异成分,为夏桑菊颗粒的质量标准和选择夏桑菊颗粒的含量测定指标提供很好的借鉴和参考。 (夏伯候 严东 曹艺 周亚敏 李亚梅 谢嘉驰 柏玉冰 廖端芳 林丽美)