An integrated analysis for determining the geographical origin of medicinal herbs using ICP-AES/ICP-MS and 1H NMR analysis

ICP-MS and 1H NMR are commonly used to determine the geographical origin of food and crops. In this study, data from multielemental analysis performed by ICP-AES/ICP-MS and metabolomic data obtained from 1H NMR were integrated to improve the reliability of determining the geographical origin of medicinal herbs. Astragalus membranaceus and Paeonia albiflora with different origins in Korea and China were analysed by 1H NMR and ICP-AES/ICP-MS, and an integrated multivariate analysis was performed to characterise the differences between their origins. Four classification methods were applied: linear discriminant analysis (LDA), k-nearest neighbour classification (KNN), support vector machines (SVM), and partial least squares-discriminant analysis (PLS-DA). Results were compared using leave-one-out cross-validation and external validation. The integration of multielemental and metabolomic data was more suitable for determining geographical origin than the use of each individual data set alone. The integration of the two analytical techniques allowed diverse environmental factors such as climate and geology, to be considered. Our study suggests that an appropriate integration of different types of analytical data is useful for determining the geographical origin of food and crops with a high degree of reliability.