Preprocessing and exploratory analysis of chromatographic profiles of plant extracts

详细信息    查看全文

• 作者：Hendriks ; Margriet M.W.B. ; Cruz-Juarez ; Leyre ; Bont ; Dries De ; Hall ; Robert. D.
• 关键词：Metabolic profiling ; Metabolomics ; Preprocessing ; Alignment ; Weighted principal component analysis ; Liquid chromatography ; Plant extracts ; Willow (Salix sp.)
• 刊名：Analytica Chimica Acta
• 出版年：2005
• 出版时间：July 22, 2005
• 年：2005
• 卷：545
• 期：1
• 页码：53-64
• 全文大小：642 K

文摘

The characterization of herbal extracts to compare samples from different origin is important for robust production and quality control strategies. This characterization is now mainly performed by analysis of selected marker compounds. Metabolic fingerprinting of full metabolite profiles of plant extracts aims at a more rapid and thorough screening or classification of plant material. We will show that HPLC is an appropriate technique for metabolic fingerprinting of secondary metabolites, given that adequate preprocessing of raw profiles is performed. Additional variation, which results from sample preparation and changing measurement conditions, usually obscures the information of interest in these raw profiles. This paper illustrates the importance of preprocessing of chromatographic fingerprinting data. Different alignment methods are discussed as well as the influence of normalization. Weighted principal component analysis is introduced as a valuable alternative to autoscaling of data. LC–UV data on Willow (Salix sp.) extracts is used to evaluate these preprocessing methods and their influence on exploratory data analysis.