# Chemical Attributes of UK-Grown Tea and Identifying Catechin and Metabolite Dynamics in Green and Black Tea Using Metabolomics and Machine Learning

## TL;DR

This study explores how processing methods affect the chemical makeup of UK-grown green and black tea, using advanced techniques to identify key compounds and pathways.

## Contribution

The study introduces a novel approach combining metabolomics and machine learning to analyze tea processing effects on chemical profiles.

## Key findings

- Green tea processing was associated with 272 m/z features identified via random forest regression.
- Black tea processing yielded 209 and 605 discriminatory m/z features for different oxidation times.
- K-means clustering grouped features into 15 clusters linked to metabolic pathways like flavonoid biosynthesis.

## Abstract

The Dartmoor Estate Tea plantation in Devon, UK, benefits from a unique microclimate and diverse soil conditions, which, together with its different processing methods, contribute to the distinctive flavours and chemical profiles of its teas. Objectives: The chemical diversity of Dartmoor tea was assessed via samples collected during processing of green and black tea. Methods: Leaf samples were collected during the processing of green and black tea and analysed using Flow Infusion Electrospray Ionisation Mass Spectrometry (FIE-MS). Results: For green tea processing, random forest regression identified features associated with the processing steps, resulting in a total of 272 m/z explanatory features. The analysis of black tea processing (4 h and overnight oxidation prior to roasting) yielded 209 discriminatory m/z features (4 h) and the model for the overnight oxidation and roasting treatments yielded 605 discriminatory m/z features. K-means clustering was performed on the percentage of relative abundance of the discriminatory m/z features. This grouped the discriminatory m/z features into 15 clusters of features showing similar trends across the processing stages. Functional and structural enrichment analysis was performed on each of the clusters and significant metabolic pathways included metabolism and biosynthesis of flavonoids, amino acids and lipids, the Pentose phosphate pathway, and the TCA cycle. Many discriminatory features were putatively classified as catechin-derived flavan-3-ols and flavonol glycosides. Conclusions: This research highlights the complex role that processing plays in shaping tea quality. It provides valuable insights into the metabolic pathways that influence tea production and emphasises how these factors determine the final chemical profile and sensory characteristics of tea.