Combining stable C, N, O, H isotope and multi-element with chemometrics for identifying the geographical origins of Codonopsis pilosula

Codonopsis pilosula (C. pilosula) is a renowned traditional Chinese medicine, and its geographical origin plays a crucial role in determining its quality. Therefore, advanced techniques are required to determine the geographical origin of C. pilosula. In this study, we collected a total of 210 samples of C. pilosula from seven prominent production areas in China. Four stable isotope ratios (δ13C, δ15N, δ18O and δ2H) and 42 elements were analyzed together for origin traceability and farming authentication purposes. Analysis of variance (ANOVA) and principal component analysis (PCA) were used to compare the stable isotope ratios and elements content, and statistically significant differences were found among C. pilosula samples from 7 geographical regions (p < 0.05). Linear discriminant analysis (LDA), k-nearest neighbor (KNN), random forest (RF), and support vector machine (SVM) were implemented to construct models for C. pilosula authentication. Among these models, the SVM model was suggested as the optimal option for identifying the origin of C. pilosula based on its superior discriminant accuracy rate (100%) and predictive accuracy rate (100%). Thus, this approach could serve as a crucial alternative method for ensuring authenticity and combatting issues of origin mislabeling and fraudulent activities associated with C. pilosula products.