Discrimination of Dried Ginger and Its Different Processed Products: Untargeted Metabolomics, Network Pharmacology, and Machine Learning.

Dried ginger (Gan-Jiang (GJ)) is the dried rhizome ofZingiber officinale Rosc., which has two different processed products, baked ginger (Pao-Jiang (PJ)) and ginger charcoal (Jiang-Tan (JT)). However, PJ and JT are still confused with each other in the market due to the absence of quality markers. In this study, metabolomics and chemometrics were used to detect metabolites and screen featured metabolites in GJ, PJ, and JT. Network pharmacology was employed to evaluate their activity strength. Machine learning algorithms were applied to distinguish GJ, PJ, and JT. Our findings illustrated that a total of 171 metabolites were identified. Among them, 20 featured metabolites were selected, including gingerenone C, methyl-6-gingerdiol, and gingerenone A. Significant changes in the content of featured metabolites were attributed to the conversion of gingerols and their degradation under high temperatures. Network pharmacology analysis suggested that featured metabolites mainly regulate the calcium signaling pathway, PI3K-Akt signaling pathway, and NF-κB signaling pathway, exerting different activity strengths. Furthermore, all 10 machine learning models exhibited a relatively good predictive performance in training sets (100%) and testing sets (97.83%). Overall, this study provides a theoretical and application basis for identifying and controlling the quality of GJ and its different processed products.