Identification and quality evaluation of Fuyun 6 and Zhongcha 108 tea fresh leaves at different altitudes using non-targeted metabolomics combined with machine learning

Tea cultivar and growing altitude are key factors shaping tea quality. This study systematically investigated their individual and synergistic effects on the quality of fresh leaves of Zhongcha 108 (ZC108) and Fuyun 6 (FY6), which were collected at altitudes from 442 to 945 m. Chemical analysis revealed that ZC108 had higher levels of amino acids, theanine, sugars, flavonoids, and catechins than FY6. Increasing altitude reduced tea polyphenols, flavonoids, caffeine, and catechins in both cultivars. Non-targeted metabolomics identified 50 cultivar and 40 altitude-associated metabolites. Key metabolic pathways, including phenylpropanoid, purine, and flavonoid biosynthesis, were upregulated with increasing altitude. A significant cultivar × altitude interaction was observed for coumarin. Machine learning achieved 100 % accuracy in cultivar classification (RF: 96.7 %, SVM: 93.3 %, KNN: 90 %), highlighting the cultivar factors dominance. Key biomarkers, such as phenylethanol and 6-O-β-D-fructofuranosyl-d-glucose were identified. These findings provide a basis for tea cultivar identification, origin traceability, and quality control.