Research on quality homogenization of preparations from multi-part botanical materials based on pseudotargeted metabolomics and machine learning: a case study of Cocculus orbiculatus (L.) DC.

The use of multiple medicinal parts from a single herbal source for disease treatment is a common practice in traditional Chinese medicine. However, variations in the ratio of these parts can lead to inconsistencies in product quality. Therefore, ensuring uniform quality by controlling the ratio of different plant parts is a key challenge for Traditional Chinese Medicine (TCM) products. In this study, UHPLC-Q-TOF-MS was employed to acquire MS data of Cocculus orbiculatus (L) DC. (C. orbiculatus) quality control samples under different collision energies. Ion pairs were screened using the MRM-Ion Pair Finder online platform, and a pseudo-targeted metabolomics method was established based on 741 ion pairs from C. orbiculatus roots and stems using UHPLC-Q-Trap-MS in MRM mode. Chemical profiles of different batches of Ningmitai capsule (NMTC) and mixtures of C. orbiculatus roots and stems in varying ratios were further acquired. Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were applied to identify both differential and non-differential metabolites. Subsequently, the established ratio range was validated using a root-to-stem ratio of 1 : 80. A total of 167 differential metabolites were identified based on VIP > 1, P < 0.05, and FC > 2 or FC < 0.5. Ten key compounds were further selected using the random forest (RF) algorithm as potential quality markers for NMTC. Moreover, machine learning models based on RF, k-nearest neighbors (KNN), and logistic regression were constructed using non-differential markers. Validation results demonstrated that even at an extreme root-to-stem ratio of 1 : 80 encountered in production, the determined mixing ratio range of 70 : 1 to 1 : 290 ensures uniform quality of NMTC. This study not only improves the utilization efficiency of C. orbiculatus but also offers a novel strategy for quality control of multi-part herbal medicines.