Computational refinement for structural identification of catechins and oxidation products in Chinese black tea

Characterization of natural products and their metabolites in foods is hindered by isomeric ambiguity, limited authentic standards, and compounds absent from libraries. Here, this study presents a confidence-stratified identification workflow adapted from Metabolomics Standards Initiative (MSI) Levels 1–4, spanning standard-confirmed identification, putative annotation, tentative structural or class assignment, and pathway-supported characterization of unknowns. Using UHPLC-QTOF, the workflow was applied to catechins and their oxidation products in Chinese black tea. 27 compounds were confirmed at MSI Level 1 using an in-house library, while 16 oxidation products were annotated at Level 2 by integrating public libraries with SIRIUS-based formula prediction, structure ranking, and class annotation. For compounds absent from libraries, a pathway-informed in-silico pipeline constrained structural space and enabled tentative annotation at MSI Levels 3–4. The approach was further exemplified using oxidation products of theasinensin A and theaflavic acid, reducing the number of chemically plausible candidates for these unknown features.