Tobacco leaf is highly complex, containing hundreds of different components that range in molecular weight and polarity, and these components can be used to characterise and differentiate leaf types using advanced analytical techniques and data processing.

Metabolomics is a powerful tool used for the characterisation of biological samples in life science research and has become a ‘Hot Topic’ for the analysis of complex samples, making it an ideal process to be used with tobacco leaf. In this study, a metabolomics approach was applied to the characterisation and comparison of tobacco leaf type (flue-cured, Burley and Oriental).

Tobacco leaves were extracted using different solvents (hexane, water / MeOH / MeCN), and the extracts were analysed by GC-MS. The GC-MS data was processed using deconvolution software (AMDIS, Agilent Technologies) prior to analysis in multivariate software (MPP, Agilent Technologies). Using MPP, various manipulations were performed, such as filtering by frequency, T-test, Volcano and PCA plot construction along with the identification of unique entities.

By comparing the PCA plots from the different extraction conditions, it was observed that the results were complementary to one another. The PCA plot obtained from the hexane extracts showed good discrimination between the three leaf types. On the other hand, the polar extracts (water / MeOH / MeCN extracts) were only able to distinguish Burley from the flue-cured and Oriental. Lastly, using the information obtained from the Volcano Plots, feature identification of the hexane extracts was performed and 18 potential marker components were identified.