Multichannel chemical fingerprinting and exploratory chemometric analysis for Taiwanese tea traceability

This study explored a multichannel analytical framework for characterizing chemical differentiation among Taiwanese oolong teas. HPLC-UV, electrochemical detection (ECD), high-resolution ESI-MS/MS, and HILIC-based amino acid profiling were integrated to provide complementary chemical information. ECD contributed selective responses to electroactive constituents, particularly catechin-related compounds, complementing UV and mass-spectrometric signals. Thirty-six secondary metabolite- and amino acid-related features were characterized. Principal component analysis (PCA) revealed exploratory group-associated patterns, while quantitative K-means validation supported three clusters. Partial least squares-discriminant analysis (PLS-DA) and Random Forest (RF) were evaluated using Tea-ID-grouped validation to prevent information leakage between duplicate measurements; both correctly classified all nine independent teas in the held-out test subset. Feature-importance analyses indicated contributions from amino acid-, polyphenol-, UV-, and electrochemical-response features. These findings demonstrate the feasibility of integrating detector-specific chemical information for tea characterization, while larger independent cohorts are required to establish its applicability to authentication and traceability.