Purpose To accurately screen and identify key volatile components in cured tobacco leaves, and to provide a scientific basis for precise control of process parameters during tobacco leaf curing stages.
Methods During the three main curing stages (yellowing, color-fixing, and stem-drying) of K326 mature middle leaves, 72 tobacco leaf samples were collected from eight different temperature points, and they were detected by pyrolysis-gas chromatography-mass spectrometry (Py-GC-MS) with a pyrolysis temperature of 600 °C. The Py-GC-MS data were processed using MZmine software, obtaining characteristic peak tables contained peak intensity information, and partial least squares discriminant analysis (PLS-DA) was applied to establish discriminant models for volatile components in tobacco leaves at different curing stages.
Results The PLS-DA visualization results effectively distinguished samples from the three curing stages. Through ten-fold cross-validation, the optimal number of principal components was determined to be five. The training set determination coefficient (R2) was 0.883 with a root-mean-square error of 0.877, and the test set accuracy was 0.800.
Conclusion Based on variable importance in projection (VIP) greater than one, 10 characteristic components were screened, including undec-10-ynoic acid, tridec-2-yn-1-yl ester (VIP=1.5889), N-vinylpyridinium bromide (VIP=1.5497), and oleic acid (VIP=1.5022). These components can serve as key volatile markers for discriminating tobacco leaves at different curing stages. This study provides a data foundation for the development of quality evaluation methods for tobacco leaf curing processes.
DownLoad: