Piao Xiang-Lan, Park Jeong Hill, Cui Jian, Kim Dong-Hyun, Yoo Hye Hyun
College of Pharmacy, Seoul National University, Seoul 151-742, Republic of Korea.
J Pharm Biomed Anal. 2007 Sep 3;44(5):1163-7. doi: 10.1016/j.jpba.2007.04.006. Epub 2007 May 4.
This paper describes gas chromatographic/mass spectrometry (GC/MS)-pattern recognition methods for the quality control of Korean Angelica. A total of 57 Angelicae radix samples, including Angelica gigas (Korean origin), A. sinensis (Chinese origin) and A. acutiloba (Japanese origin), were analyzed by GC/MS, with a principal component analysis (PCA) subsequently applied to 10 common peaks selected from each chromatogram. As a result, the samples were clustered according to their origins on the PC score plot. The loading plot revealed that decursin and decursinol angelate were the most contributive principles distinguishing Korean samples from Chinese and Japanese samples, In addition, a discriminant model was developed for classification of the Angelicae radix, using a discriminant analysis (DA), and validated with a training set (three from A. gigas, four from A. sinensis, and three from A. acutiloba). All samples tested were successfully classified according to their species origin.
本文描述了用于韩国当归质量控制的气相色谱/质谱（GC/MS）模式识别方法。通过GC/MS分析了总共57份当归样品，包括当归（韩国产）、当归（中国产）和当归（日本产），随后对从每个色谱图中选出的10个共同峰应用主成分分析（PCA）。结果，样品在PC得分图上根据其产地聚类。载荷图显示，蛇床子素和蛇床子素当归酸酯是区分韩国样品与中国和日本样品的最主要成分。此外，使用判别分析（DA）建立了当归分类的判别模型，并用训练集（三份来自当归，四份来自当归，三份来自当归）进行了验证。所有测试样品均根据其物种来源成功分类。
