A Combined Study of Headspace Volatiles using Human Sensory, Mass Spectrometry and Chemometrics

K G Mcadam,C Liu,J Cote,L Bishop,S Lubbe,J Tetteh,H Digard

doi:10.1038/s41598-020-64491-6

Abstract

Smokeless tobacco products (STPs) are widely used in certain parts of the world, yet there is limited understanding of how they are consumed, particularly the impact of chemosensory characteristics on their use. In order to develop an understanding of the drivers of STP use and product acceptability we conducted both human sensory panel testing and chemical analyses on a range of STPs. Free-sorting paired odour testing using sensory panellists identified similarities and clear differences between eleven different STPs. Headspace volatiles, analysed by headspace solid-phase microextraction gas chromatography mass spectrometry (HS-SPME-GC-MS), identified 20 to 70 components depending upon the STP. Key differences in headspace volatiles were found between STPs. For example, the headspace of Skoal Bandits Wintergreen was dominated by methyl salicylate, while Marlboro Spice consists of a more complex profile including pinene, nicotine, eugenol and cymene. Chemometric Target Factor Analysis (TFA) and Hierarchical Cluster Analysis (HCA) of chemistry and sensory data was used to deduce chemical drivers of sensory perceptions. The chemometric strategy used showed that headspace analysis is a complementary screening tool to sensory analysis in classification studies. This study is generic with applications across various product sectors that require routine human sensory panel evaluation.

Highlights

This is followed by results relating to Target Factor Analysis (TFA) extraction of the core volatile chemical constituents associated with the sensory effects
du Maurier Original (DMO), MSS, Pall Mall Original Portion (PMOP), Pall Mall White Portion (PMWP), LSCnf, Sensory 3D-scatter-Principal Component Analysis (PCA) grouping based on Fig. 4
The results obtained have shown that complex data from headspace GC-MS and sensory evaluations of smokeless tobaccos can be combined and analyzed using chemometrics to derive insights into the flavour characteristics of smokeless tobacco products (STPs)

Summary

Materials and Methods

We describe choice of STPs, sample preparation methods, the sensory panel methodology, the procedure for headspace volatile analysis and chemometric data analysis strategies. Where peak identification was regarded as tentative, samples of these compounds were analysed by matrix addition to establish retention times and confirm spectra in the analytical system The use of both retention times and library matched mass spectra provided a 2-fold confirmation of identities across the whole data set. 205 unique chemical entities were identified resulting in a final data matrix of 205 compound rows (×) 22 columns of samples. The cross product (similarity) matrix for both sensory and headspace data for each sample were calculated and used for the PCA59–61. 3-d scatter plots were used to visualize similarities between samples This involved (a) conducting principal component analysis (PCA) (b) selecting the first n significant principal components as input for the number of clusters to be modelled (c) plot the identified clusters in x, y, z space. The Matlab “clusterdata” and “scatter3” functions were used in this study[58]

Results and Discussion

Clusters

Summary of Results and Discussion

Conclusions