Machine Learning-Assisted Ratiometric Fluorescence Electrospun Nanofiber Films for Portable and Intelligent Monitoring of Multiple Alkylresorcinol Homologues in Whole Wheat Foods

Ruiqing Sun ¹, Yujuan Xie ², Xinpeng Zhou ³

⁰Key Laboratory of Geriatric Nutrition and Health, Ministry of Education, Beijing Technology and Business University, Beijing 100048, China.

Acs Applied Materials & Interfaces +

|

December 2, 2025

View abstract on PubMed

Summary

This study introduces a novel sensing platform using machine learning and fluorescence materials for detecting alkylresorcinol (AR) homologues. The system enables rapid, on-site authentication of whole wheat products, enhancing food safety.

Area Of Science

Analytical Chemistry
Materials Science
Food Science

Background

Authenticating whole wheat products is challenging due to difficulties in monitoring multiple alkylresorcinol (AR) homologues.
Existing methods struggle with complex food matrices, necessitating advanced analytical techniques.

Purpose Of The Study

To develop a novel sensing platform for intelligent, rapid, and on-site detection of AR homologues.
To enable accurate whole wheat product authentication using machine learning and fluorescence technology.

Main Methods

Developed a ratiometric fluorescence (FL) sensing platform using dual-emission g-C3N4/Ru materials.
Integrated machine learning (ML) algorithms, specifically a random forest-back-propagation neural network (RF-BPNN), for quantitative analysis.
Utilized RGB feature extraction and stratified sampling for model training and validation.

Main Results

Achieved highly sensitive detection of five AR homologues with ultralow limits of detection (2.1-9.9 ng·mL⁻¹).
Demonstrated excellent quantitative detection range (0-60 μg·mL⁻¹) and high prediction accuracy (R² = 0.9822) using the RF-BPNN algorithm.
Successfully applied the system to commercial wheat samples for AR detection and whole wheat authentication.

Conclusions

The integrated ML-FL sensing platform overcomes traditional analytical limitations for AR detection.
This approach offers a promising tool for intelligent, rapid, and on-site verification of whole wheat product authenticity.
The system has significant potential for food safety monitoring and quality control applications.

Keywords:

alkylresorcinols electrospun nanofiber film machine learning ratiometric fluorescence sensor whole wheat foods