Welcome to Journal of Agriculture,

Journal of Agriculture ›› 2024, Vol. 14 ›› Issue (10): 12-18.doi: 10.11923/j.issn.2095-4050.cjas2023-0215

Special Issue: 生物技术 烟草种植与生产

Previous Articles     Next Articles

Stepwise Discriminant Analysis for Tobacco Flavor Based on Organic Acids, Amino Acids, and Aroma Compounds

WANG Wenchao1(), XU Zicheng1, ZHANG Yuan2, QIAN Jiancai2, ZHU Haibo2, XU Jiayang1,3, JIA Wei1, HAO haohao4(), ZHANG Li2()   

  1. 1 College of Tobacco Science, Henan Agricultural University, Zhengzhou 450046, Henan, China
    2 China Tobacco Jiangsu Industry Co., Ltd., Nanjing 210009, Jiangsu, China
    3 College of Resources and Environment, Henan Agricultural University, Zhengzhou 450046, Henan, China
    4 Zhumadian City Company of Henan Province Tobacco Company, Zhumadian 463000, Henan, China
  • Received:2023-09-27 Revised:2024-01-16 Online:2024-10-20 Published:2024-10-17

Abstract:

The research aims to achieve rapid discrimination of five types of flue-cured tobacco flavor profiles, including clean sweet, honey sweet, honey sweet roasted, roasted sweet mellow, and roasted sweet roasted. Fifty-one C3F grade samples of flue-cured tobacco from five provinces were selected. Variance analysis and discriminant analysis were conducted on the organic acids, amino acids, and aromatic substances in the tobacco leaves. A Fisher discriminant function was established. The total amino acid content in clean sweet and honey sweet roasted types was significantly higher than in roasted sweet mellow type. There were no significant differences in total organic acid and aromatic substance content among the five flavor types. However, differences were observed in some higher fatty acids, phenylalanine degradation products, carotenoid degradation products, and quinones. A total of nine indicators, including succinic acid, citric acid, eicosanoic acid, serine, 6-methyl-5-hepten-2-ol, guaiacol, isobutyrophenone, podocarpatrienone-3, and phenylpropanone, were incorporated into the discriminant function. The discriminant function established achieved correct rates of 94.1% and 86.3% through self-validation and cross-validation, respectively. The discriminant function established in this study demonstrated good discrimination performance and high accuracy for the five types of flue-cured tobacco flavor profiles.

Key words: tobacco flavor, discriminant analysis, organic acids, amino acids, aroma compounds