Geographical origin traceability of traditional Chinese medicine Atractylodes macrocephala Koidz. by using multi-way fluorescence fingerprint and chemometric methods

Atractylodes macrocephala Koidz. (AM) is an important plant of traditional Chinese medicine (TCM), and its status can be comparable with ginseng in China. The efficacy and quality of AM are closely related to the place of origin. Hence, we proposed a simple and fast strategy to classify AM from different geographical origins by using multi-way fluorescence fingerprint combined with chemometric methods. AM samples with different dilution levels have different fluorescence characteristics, resulting from different content of fluorescence components and chemical microenvironment. Therefore, AM samples were diluted 5-fold, 10-fold, and 20-fold with 40% ethanol aqueous solution to obtain excitation-emission matrix data, and multi-way (three-way and four-way) data arrays were constructed. And then, the fluorescence fingerprints of AM samples were characterized by three-way and four-way parallel factor analysis (PARAFAC). In addition, four pattern recognition methods were used to classify AM from different provinces. The results show that the four-way data array can provide more abundant information than three-way data arrays, so it is more conducive to sample classification. According to the results obtained from the analysis of four-way data array, the correct classification rate (CCR) of the cross-validation and prediction set obtained by partial least squares-discrimination analysis (PLS-DA) were 90.5% and 100%, respectively. To sum up, the proposed method can be regarded as a powerful, feasible, convenient, reliable, and universal classification tool for the classification of AM samples from different provinces and can be used as a promising method to realize the geographical origin traceability of other TCMs. (c) 2021 Published by Elsevier B.V.

Automated summary

A strategy using fluorescence fingerprint combined with chemometric methods was proposed to classify AM samples from different geographical origins, showing a high accuracy in sample classification.