Pulse Brain · Growing Health Evidence Index
Tier 3 — Observational / field trialPeer-reviewed

Machine learning-assisted boronic acid colorimetric indicator displacement array for dual-functional tea fingerprinting and interference-resistant EGCG quantification

Y NI; Kexin Fan; Pengxia Song; Xiao Dong; Chaoyan Lou; Ming Guo; Qiao Deng; Xu Wang; Dan Wei

Food Chemistry X · 2026

Read source ↗ All evidence

Summary

This study presents a machine learning-enhanced colorimetric sensor array based on boronic acid-diol recognition and competitive indicator displacement for rapid, portable tea authentication and quality control. By screening dye indicators and phenylboronic acid receptors to generate 27-dimensional RGB fingerprints, the system achieved high accuracy in discriminating structurally similar catechins, classifying tea categories and varieties, and quantifying EGCG whilst minimising interference from other tea constituents. The smartphone-compatible platform offers a cost-effective alternative to laboratory instrumental analysis for on-site tea authentication and polyphenol assessment.

Regional applicability

The findings are directly applicable to United Kingdom tea supply chains and quality assurance, as tea is a significant commodity in UK food retail and cafes. The portable, cost-effective nature of the approach could facilitate rapid authentication and quality control at point-of-sale or in small retail settings, addressing consumer demand for tea authenticity and bioactive content verification.

Key measures

Discrimination accuracy (Linear Discriminant Analysis), classification accuracy for tea categories and varieties (Random Forest), EGCG quantification accuracy (XGBoost regression with R² and RMSE), spike recovery rates and relative standard deviations (RSDs)

Outcomes reported

The study developed a machine learning-assisted colorimetric sensor array using boronic acid-diol recognition to discriminate tea catechins and quantify EGCG (epigallocatechin gallate) with high accuracy. The system achieved 90.91% accuracy in discriminating 11 compounds, 88.9–100% accuracy in tea category and variety identification, and R² = 0.993 for EGCG quantification with spike recovery rates of 91.4–104.7%.

Theme
Measurement & metrics
Subject
Measurement methods & nutrient profiling
Study type
Research
Study design
Laboratory experiment
Source type
Peer-reviewed study
Status
Published
System type
Food supply chain
DOI
10.1016/j.fochx.2026.104778
Catalogue ID
NRmwavvkd5-004

Topic tags

Pulse AI · ask about this record

Dig deeper with Pulse AI.

Ask about this record, its theme or its relevance to UK farming and policy. Pulse AI uses selected catalogue evidence and cites the sources it draws on.