Discrimination of easily confused tea leaves with similar appearance (Gougu tea vs. Gonglao tea) via an integrated method of electronic tongue, HPLC-QTOF-MS-VirtualTaste, electronic nose, electrochemical fingerprinting and machine learning

Gougu tea (GG) and Gonglao tea (GL) were historically misclassified in tea markets for centuries due to their highly similar appearance. To resolve this long-standing challenge, our study focused on two objectives: elucidating the necessity for differentiating them, and constructing an efficient method for their discrimination. For the first time, E-tongue, HPLC-QTOF-MS-VirtualTaste, E-nose, electrochemical fingerprinting, and machine learning were integrated to comprehensively analyze their differences in flavor and composition. E-tongue analysis confirmed bitterness as a shared sensory attribute in GG and GL, while HPLC-QTOF-MS-VirtualTaste revealed their distinct bitter components. Organic acids and triterpenes predominated among the 85 taste components in GG, while alkaloids predominated among the 60 taste components in GL. Quantitative analysis showed that the average chlorogenic acid content (GG's primary bitter component) was 6.4787 mg/g, whereas berberine (GL's main bitter component) reached 17.0383 mg/g. E-nose analysis detected 51 and 38 volatile components in GG and GL, respectively. Eleven common components primarily exhibited fruity and sweet sensory characteristics. Furthermore, electrochemical fingerprinting combined with the random forest algorithm was established, achieving 99.85 % discrimination accuracy. Moreover, this approach possessed the advantages of low cost and simplicity. Our research contributes to addressing the centuries-old challenge of market confusion between GG and GL.