Rapid identification model based on decision tree algorithm coupling with 1H NMR and feature analysis by UHPLC-QTOFMS spectrometry for sandalwood
| Autor: | Mingxin Guo, Juanxia Wang, Youzhen Tan, Yifan Feng, Xia Wu, Biying Chen, Kexing Shi, Cui Ren |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Sandalwood
Chromatography biology Chemistry Decision tree learning 010401 analytical chemistry Clinical Biochemistry Decision tree Cell Biology General Medicine Nuclear magnetic resonance spectroscopy Carbon-13 NMR biology.organism_classification Mass spectrometry 030226 pharmacology & pharmacy 01 natural sciences Biochemistry 0104 chemical sciences Analytical Chemistry 03 medical and health sciences 0302 clinical medicine Feature (machine learning) Proton NMR Biological system |
| Zdroj: | Journal of Chromatography B. 1161:122449 |
| ISSN: | 1570-0232 |
| DOI: | 10.1016/j.jchromb.2020.122449 |
| Popis: | Sandalwood is one of the most valuable woods in the world. However, today's counterfeits are widespread, it is difficult to distinguish authenticity. In this paper, similar genus (Dalbergia and Pterocarpus) and confused species (Gluta sp.) of sandalwood were quickly and efficiently identified. Rapid identification model based on 1H NMR and decision tree (DT) algorithm was firstly developed for the identification of sandalwood, and the accuracy was improved by introducing the AdaBoost algorithm. The accuracy of the final model was above 95%. And the feature components between different species of sandalwood were further explored using UHPLC-QTOFMS and NMR spectrometry. The results showed that 183 compounds were identified, among which 99 were known components, 84 were unknown components. The 1H NMR and 13C NMR signals of 505 samples were assigned, among them, 14 compounds were attributed, characteristic chemical shift intervals with great differences in the model were analysed. Furthermore, the fragmentation pattern of different compounds from sandalwood, in both positive and negative ion ESI modes, was summarized. The results showed a potential and rapid tool based on DT, NMR spectroscopy and UHPLC-QTOFMS, which had performed great potential for rapid identification and feature analysis of sandalwood. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect