As international food trade increases, consumers are becoming increasingly interested in food safety and authenticity, which are linked to geographical origin. Adzuki beans (Vigna angularis) are cultivated worldwide, but there are no tools for accurately discriminating their geographical origin. Thus, our study aims to develop a method for discriminating the geographical origin of adzuki beans through targeted and non-targeted metabolite profiling with gas chromatography time-of-flight mass spectrometry combined with multivariate analysis. Orthogonal partial least squares discriminant analysis showed clear discrimination between adzuki beans cultivated in Korea and China. Non-targeted metabolite profiling showed better separation than targeted profiling. Furthermore, citric acid and malic acid were the most notable metabolites for discriminating adzuki beans cultivated in Korea and China. The geographical discrimination method combining non-targeted metabolite profiling and pareto-scaling showed excellent predictability (Q 2 = 0.812). Therefore, it is a suitable prediction tool for the discrimination of geographical origin and is expected to be applicable to the geographical authentication of adzuki beans.

Perilla and sesame are traditional sources of edible oils in Asian and African countries. In addition, perilla and sesame seeds are rich sources of health-promoting compounds, such as fatty acids, tocopherols, phytosterols and policosanols. Thus, developing a method to determine the geographic origin of these seeds is important for ensuring authenticity, safety and traceability and to prevent cheating. We aimed to develop a discriminatory predictive model for determining the geographic origin of perilla and sesame seeds using comprehensive metabolite profiling coupled with chemometrics. The orthogonal partial least squares-discriminant analysis models were well established with good validation values (Q2 = 0.761 to 0.799). Perilla and sesame seed samples used in this study showed a clear separation between Korea and China as geographic origins in our predictive models. We found that glycolic acid could be a potential biomarker for perilla seeds and proline and glycine for sesame seeds. Our findings provide a comprehensive quality assessment of perilla and sesame seeds. We believe that our models can be used for regional authentication of perilla and sesame seeds cultivated in diverse geographic regions.