Integration of UPLC-Q-TOF-MS/MS, chemometrics and network pharmacology to discovery potential quality markers in Sinomenii Caulis

Rationale: There are significant differences in Sinomenii Caulis (SC) obtained from different geographical regions and medicinal plant parts. This study aims to explore potential quality markers that are correlated with clinical efficacy in SC by a comprehensive strategy that integrates chemical profiling, chemometrics, and network pharmacology. Methods: First, an alkaloid database was created through the utilization of the UNIFI system to qualitatively analyze of alkaloids in SC. Then, differential compounds in SC collected from various geographic regions were screened by applying multivariate data analysis. Subsequently, the support vector machine (SVM) and random forest (RF) algorithms are adopted to calculate the grouping accuracy of different components. Finally, network pharmacology was conducted to analyze the pharmacological properties and potential associations of these target compounds. Results: A total of 81 alkaloids were identified from SC samples, including 13 aporphine alkaloids, 18 protoberberine alkaloids, 32 morphine alkaloids, 10 benzylisoquinoline alkaloids, and 8 other types of alkaloids. Notably, palmatine, sinoracutine, and magnoflorine are active ingredients with the ability to differentiate the different regions of SC samples. And thus should be prioritized when selecting quality markers. Additionally, it was observed that the RF algorithms demonstrated higher classification accuracy than the SVM model. Conclusion: This comprehensive strategy may prove to be a powerful technique for screening the quality markers components, which could be used for the quality control of SC, and can serve as reference for design of quality control of other herbs.