Intelligent characterization multi-components in Yangxinshi tablet by online comprehensive two-dimensional liquid chromatography-quadrupole time-of-flight mass spectrometry combined with deep learning-assisted mass defect filtering classification and multidimensional data annotation strategy.
Journal:
Talanta
PMID:
40020613
Abstract
A comprehensive characterization strategy for the intelligent analysis of multiple chemical components in Yangxinshi Tablet (YXST) was established. The strategy developed the deep learning-assisted mass defect filtering intelligent classification, preferred ions capture list and active exclusion (DLA-MDF-PIL-AE) data acquisition mode by online comprehensive two-dimensional liquid chromatography coupled with quadrupole time-of-flight mass spectrometry (2DLC-Q-TOF-MS/MS). Firstly, the online 2DLC-Q-TOF-MS/MS system was constructed and the orthogonality was evaluated. Secondly, the user interface for deep learning-assisted MDF intelligent classification technology was developed and applied to compounds classification to generate preferred ion capture lists of various types compounds. Finally, molecular networking (MN), associated neutral loss (NL) fragments, and characteristic diagnosis ion (CDI) were utilized for the automatic and manual annotation of compounds, respectively. As a result, a total of 228 compounds including 80 flavanoids, 52 alkaloids, 36 phenolic acids, 15 terpenoids, 17 saponins and 28 others were preliminary identified from YXST and source attribution was assigned to them. Furthermore, 39 compounds were simultaneously quantified by ultra-high performance liquid chromatography coupled with triple quadrupole mass spectrometry (UHPLC-MS/MS) method. Conclusively, the proposed integrated strategy proved to be a powerful method for characterizing multiple components in complex natural products.