Chemometrics-driven discrimination of flue-cured tobacco aroma types via GC-MS/MS and multivariate analysis

Abstract

In this work, a novel classification model for flue-cured tobacco aroma types is presented by integrating chemometric modeling with quantitative aroma component analysis. Three representative types of flue-cured tobacco samples were selected for their distinct flavor profiles and commercial importance. Sensory characteristics were quantified by descriptive analysis of a trained panel. Gas chromatography-triple quadrupole tandem mass spectrometry (GC-MS/MS) was employed to rapidly identify the aroma components. The aroma types of flue-cured tobacco were studied using correlation analysis, hierarchical clustering, principal component analysis (PCA), and discriminant analysis. In total, 31 aroma components of flue-cured tobacco were identified by GC-MS/MS. Each flue-cured tobacco sample was first assigned an aroma style based on geographical origin and subsequently corroborated by the descriptive panel. Correlation analysis successfully identified compounds related to aroma substances and the descriptive analysis indices of flue-cured tobacco. Cluster analysis cleanly segregated the samples into the three predefined aroma types. Six principal components were extracted from the PCA to construct the discriminant model. Internal and cross-validation both confirmed the discriminant model's reliability and accuracy. This study evaluated the potential of using tobacco aroma components to distinguish and classify flue-cured tobacco aroma types.