Untargeted volatile fingerprinting of sesame oils using HS-SPME–GC×GC/HRMS for geographical origin discrimination

Sesame oil is widely consumed across Asia, yet reliable analytical approaches for geographical origin authentication remain limited, particularly after roasting-induced chemical transformation. In this study, untargeted volatile fingerprinting using headspace solid-phase microextraction coupled with comprehensive two-dimensional gas chromatography/high resolution mass spectrometry (HS-SPME–GC×GC/HRMS) was applied to discriminate roasted sesame oils from Korea, China, and India. A total of 205 volatile compounds were identified, and roasting markedly increased volatile chemical complexity through Maillard reaction pathways. Multivariate statistical analyses, including principal component analysis (PCA) and hierarchical clustering analysis (HCA), revealed clear origin-dependent clustering patterns despite extensive thermal transformation. ClassyFire-based chemical ontology analysis further demonstrated significant differences in volatile chemical superclass distributions among origins, particularly in hydrocarbons and organic acid-related compounds. In addition, an integrated marker prioritization strategy combining FDR-corrected ANOVA significance and PLS-DA variable importance in projection (VIP), supported by PCA loading magnitudes, identified multiple candidate volatile compounds associated with geographical origin discrimination. Several compounds showed strong association with Korean or Chinese sesame oils, whereas discrimination of Indian-origin oils appeared to rely more on multivariate compositional patterns than on individual marker compounds. These findings highlight the potential of untargeted GC×GC/HRMS volatile fingerprinting for sesame oil authentication and geographical origin traceability.