Geographical origin classification of tobacco by stable isotope and multi-elemental analysis in combination with chemometric methods

Abstract

Tobacco is an important economic crop widely grown around the world. Its quality and features are closely interrelated to its geographical origin. With a view to identify the geographical origin of tobacco leaves, stable isotope ratios (δ13C, δ15N and δ2H) and 33 element contents in 260 tobacco samples collected from six provinces in China were determined by stable isotope ratio mass spectrometry (IRMS) and inductively coupled plasma mass spectrometry (ICP-MS), respectively, and analyzed with the aid of chemometric methods such as Orthogonal Partial Least-Squares Discriminant Analysis (OPLS-DA) and Random Forest (RF). Experimental results revealed that δ2H, Zn and Tl were significantly different in samples from different regions, which indicated their potential to be used in geographical origin classification of tobacco samples. Both OPLS-DA and RF models displayed satisfactory overall performance for the discrimination of geographical origins of tobacco samples from six cultivation areas. Particularly, the RF model built on the fused data of IRMS and ICP-MS had an exquisite sensitivity of 98.5% in comparison with the RF models on either the data set produced by IRMS or ICP-MS alone, whose sensitivity values were 92.3% and 95.4%, respectively. It was found that the top ten influential indicators for identifying the geographic origin of tobacco were δ2H, Cr, δ15N, 87/86Sr, Li, Ba, Sb, Pb, Tl, and Zn. This work may open up an avenue for geographical origin classification of tobacco.