基于致香成分的重庆主要烟草种植区溯源特征研究

Aroma components based traceability characteristics of main tobacco planting areas in Chongqing

【目的】以烤烟中多项致香成分含量为原始数据,采用多元统计分析方法对重庆烟叶产地溯源的可行性进行研究,筛选出判别烤烟产地溯源的有效指标。【方法】利用GC-MS法,检测采自重庆主要烟草种植区(武陵山区、渝中渝西地区、三峡库区和渝东北秦巴山区)4个品种(云烟87、云烟97、K326和贵烟四号)147份烤烟样品中70项致香成分的含量,对试验数据进行方差分析、主成分分析(PCA)、偏最小二乘-判别分析(PLS-DA)、变量重要性因子(VIP)和逐步回归判别分析,逐步筛选出可用于判别烟叶产地溯源的有效指标,建立判别模型,并对模型判别正确率进行检验。【结果】来源于重庆不同地区烟叶样品的致香成分含量具有一定的差异。通过方差分析、PCA、PLSDA、VIP法和逐步回归判别分析,从70项致香成分中筛选出苯乙醇、3-甲基-2-丁烯醛、2-吡啶甲醛、4-吡啶甲醛、糠酸、3-羟基-2-丁酮、面包酮、1-(3-吡啶基)-乙酮、胡薄荷酮、β-紫罗兰酮和亚麻酸甲酯11项,用于有效判别烤烟产地溯源的指标,用11项指标所建立的判别模型对烤烟样品回代检验判别正确率为92.5%,交叉检验判别正确率为90.5%。【结论】将致香成分含量与多元统计分析方法结合可用于烤烟产地溯源。

[Objective]The contents of multiple aroma components in tobacco were used as original data to study the feasibility of understanding traceability of Chongqing tobacco origin by multivariate statistical analysis method and screen out valid indicators for discriminating tobacco origins. [Method] Gas chromatography-mass spectrum （GC-MS） apparatus was applied to detect contents of 70 aroma components in 147 flue-cured tobacco samples of 4 varieties （Yunnan tobacco 87, Yunnan tobacco 97, K326 and Guizhou tobacco No. 4） from main tobacco planting areas in Chongqing （Wuling Mountain Area of Southeast,Mid-dle and West of Chongqing Region, Three Gorges Reservior Region, Oinling-Daba Region of Northeast）. Then, the data were analyzed by different methods including variance analysis, principal component analysis （PCA） ,partial least squares-discriminant analysis （PLS-DA） and stepwise regression discriminant analysis to select valid indicators for discriminating tobacco origin. A discrimination model was built and its accuracy was verified. [Result] The contents of aroma components in tobacco samples from different areas of Chongqing were different. Eleven indicators including phenethyl alcohol, 3-methyl-2-butenal, 2-pyridylalde- hyde, 4-pyridylaldehyde, furoic acid, 3-hydroxide-2-butanone, 2-methyl-tetrahydrofuran-3-ketone, 1- （ 3-pyridine）-ketol,pulegone,]3-ionone and methyl linolenate were screened from 70 aroma components as indicators for effective discriminating tobacco origins by variance analysis, PLS-DA combined with variance importance in the project （VIP）,and stepwise regression discriminant analysis. The return test discrimination accuracy and cross check distinguishing accuracy of established discrimination model using these indicators were 92.5%and 90. 5%, respectively. [Conclusion] The contents of aroma components in combination with multivariate statistical analysis method could be used to discriminate tobacco origins.