A strategy for identifying effective and risk compounds of botanical drugs with LC-QTOF-MS and network analysis: A case study of Ginkgo biloba preparation

Abstract

Botanical drugs have unique advantages in the treatment of complex diseases. In order to ensure the efficacy and safety of botanical drugs, ascertaining the effective and risk compounds is quite necessary. However, the conventional identification method is laborious, time-consuming, and inefficient. In this work, a 3-steps strategy was presented to rapidly identify the effective and risk compounds of botanical drugs, and a Ginkgo biloba preparation, Shu-Xue-Ning injection (SXNI), was taken as a case study. Firstly, mass spectral molecular networking was used to rapidly identify the compounds of SXNI. Secondly, three networks (i.e. the compound-target network, the indication-related biomolecule network, and the adverse drug reaction-related biomolecule network) are constructed. Finally, a novel network analysis algorithm was used to predict the effective and risk compounds in SXNI. By this strategy, a total of 138 compounds were identified including the firstly reported terpenoid glycosides and lignan glycosides. Among them 71 compounds were predicted as effective ones, and 42 compounds as risk ones. Especially, 31 compounds relevant to both efficacy and safety should be scientifically controlled during manufacturing. In addition, ten pathways were enriched to preliminarily explain the action mechanism of SXNI. This strategy for MS data analysis can be applied to provide important basis for the manufacturing and quality control, as well as valuable points for research on the pharmacological mechanisms of botanical drugs.