Authentication of herbal medicines from multiple botanical origins with cross-validation mebabolomics, absolute quantification and support vector machine model, a case study of Rhizoma Alismatis

Abstract

Accurate identification of the botanical origins of Rhizoma Alismatis (RA) is pivotal to its precise clinical usage. We herein present a strategy, by integrating untargeted metabolomics, data cross validation, absolute quantification, and vector machine model, for species discrimination and source recognition of Rhizoma Alismatis for the first time. An ultra-high-performance liquid chromatography/LTQ-Orbitrap mass spectrometry with precursor ions list-including data-dependent acquisition approach was developed for metabolite profiling. Holistic, continuous, and pattern recognition chemometrics could make it feasible to unveil forty-one markers, with “multi-duplicated and traceability samples comparison” cross validation to narrow down to twelve robust markers, some of which were further validated by absolute quantification. A support vector machine model was eventually developed to distinguish these two species and predict origins of commercially available varieties. This was the first report on systematic comparison and discrimination of two original species of RA. This integral strategy, in contrast to conventional approaches, renders more convincing data supporting for the discovery of multi-source chemical makers of traditional Chinese medicines (TCM).