Traditional herbal medicines are gaining an increased interest in the Western world. Their application imposes appropriate quality control tools. An important quality control aspect is the proper identification of the raw materials. One strategy is to identify these materials via comparison with reference samples of (botanically identic) species. Various features can be used for this comparison, which can be visual, chemical or based on other aspects. In our research, chromatographic profiles or fingerprints were used, since they contain information on a multitude of compounds in the medicines. Objective comparison methods were developed using multivariate classification modelling tools. These are built using a representative set of samples from the species that need to be distinguished. Two challenging sample sets were used in this research project. In the first sample set, the number of representative samples was limited, aiming to develop models capable of distinguishing leaf samples from two plant genera and six species. In the second sample set, a distinction had to be made between samples from three groups, formed by the same or very closely related species, and/or different plant parts. In the first study already existing classification techniques were applied, in the second novel approaches were additionally developed. In both challenging situations models were created that successfully could distinguish the considered samples.
- Herbal Medicines
- quality cintrol
- Chromatographic fingerprinting