文献类型：期刊文献

中文题名：HPLC结合多变量统计建立野生与人工沉香的识别模型

英文题名：Identification of Wild Agarwood and Cultivated Agarwood by HPLC Coupled with Multivariate Data Analysis

作者：尚丽丽[1] 陈媛[1] 晏婷婷[1] 王茜[1] 李改云[1]

第一作者：尚丽丽

机构：[1]中国林业科学研究院木材工业研究所

年份：2018

卷号：38

期号：6

起止页码：33-41

中文期刊名：林产化学与工业

外文期刊名：Chemistry and Industry of Forest Products

收录：CSTPCD;;Scopus;北大核心:【北大核心2017】；CSCD:【CSCD2017_2018】；

基金：中国林科院中央级公益性科研院所基本科研业务费专项资金(CAFYBB2018SY033)

语种：中文

中文关键词：野生沉香;人工沉香;Fisher线性判别;偏最小二乘判别法(PLS-DA)

外文关键词：wild agarwood;cultivated agarwood;Fisher linear discriminant;partial least square discriminant analysis(PLS-DA)

分类号：TQ35

摘要：为了建立野生、人工沉香的识别方法,采用《中药色谱指纹图谱相似度评价系统》对48批次沉香的HPLC图谱进行分析,并结合多变量统计筛选野生、人工沉香识别特征峰,建立两者的Fisher线性识别模型和偏最小二乘判别法(PLS-DA)识别模型。结果表明:野生、人工沉香HPLC图谱的色酮种类及含量具有较大差异,为识别模型的建立提供了依据;多变量统计筛选出的9个色谱峰变量具有代表性,可用于识别模型的建立;所建立的Fisher识别模型对野生沉香和3年以内人工沉香的判别正确率均为100%,可用于野生与人工沉香的识别; Fisher线性识别模型的线性判别函数:y=-0.193 14x_1-0.086 49x_2+0.073 02x_3+0.136 65x_4+0.053 01x_5-0.058 5x_6+0.097 58x_7+0.040 24x_8-0.130 12x_9,y_1=-1.555 3、y_2=2.641 8、临界值y_0=0.543 3。
In order to establish the identification method for wild and cultivated agarwood,Similarity Evaluation System for Chromatographic Fingerprint of Chinese Materia Medica(version 2012.130723)was used to evaluate the similarity of the HPLC chromatograms of 48 batches of agarwood samples.The identification peaks of wild and cultivated agarwood were screened by HPLC coupled with multivariate data analysis.On this basis,the Fisher linear recognition model and the partial least squares discriminant method(PLS-DA)recognition model were established.The results showed that the HPLC chromatograms of wild and cultivated agarwood were different,which provided the basis for the establishment of the identification model of wild and cultivated agarwood.The nine chromatographic peaks showed good representation in efficiently analyzing the wild and cultivated agarwood by multivariate analysis.Based on Fisher discriminant analysis,the classification accuracies of wild and cultivated agarwood within three years all reached 100%.Therefore,this model could be used for the identification of wild and cultivated agarwood within three years.The Fisher linear recognition model were y=-0.193 14x1-0.086 49x2+0.073 02x3+0.136 65x4+0.053 01x5-0.058 5x6+0.097 58x7+0.040 24x8-0.130 12x9,y1=-1.555 3,y2=2.641 8,y0=0.543 3.