Machine learning-assisted colorimetric sensor array using a microbially synthesized Prussian blue nanozyme with
Fang Chen1, Mengping Li1, Xinqi Xu1
1Laboratory of Micro & Nano Biosensing Technology in Food Safety, Hunan Provincial Key Laboratory of Food Science and Biotechnology, College of Food Science and Technology, Hunan Agricultural University, Changsha, 410128, China.
Biosensors & Bioelectronics
|July 20, 2026
Summary
A novel nanozyme-based sensor array using microbial synthesized Prussian blue nanozyme (Bio-PB) offers rapid, cost-effective detection of sulfur-containing metal salts (SCMs) in food. Machine learning decodes colorimetric fingerprints for accurate identification and quantification, enhancing food safety.
Area of Science:
- * Materials Science: Development of a bio-based nanozyme (Bio-PB) with enhanced peroxidase-like activity.
- * Analytical Chemistry: Creation of a multi-channel colorimetric sensor array for complex analyte detection.
- * Biotechnology: Microbial synthesis of Prussian blue nanozyme for enhanced catalytic properties.
Background:
- * Sulfur-containing metal salts (SCMs) are common food additives with potential health risks from excess residues.
- * Conventional analytical methods for SCMs are costly, time-consuming, and require specialized equipment.
- * There is a need for rapid, sensitive, and cost-effective methods for SCM detection in food matrices.
Purpose of the Study:
- * To develop a novel nanozyme-based colorimetric sensor array for SCM detection.
- * To utilize machine learning for accurate classification and identification of SCMs based on cross-reactive patterns.
- * To evaluate the sensor array's performance in complex food matrices for practical food safety applications.
Main Methods:
- * Construction of a sensor array integrating microbial synthesized Prussian blue nanozyme (Bio-PB) with three chromogenic substrates (TMB, OPD, ABTS).
- * Detection of SCMs based on competitive consumption of reactive oxygen species (ROS), generating unique colorimetric fingerprints.
- * Application of machine learning algorithms (LDA, HCA, PCA) for decoding response patterns and classifying SCMs.
Main Results:
- * Achieved 100% discrimination accuracy for five different SCMs and 98.78% accuracy for 82 unknown samples.
- * Enabled semi-quantification of SCMs within a concentration range of 10–180 μM and distinguished complex mixtures.
- * Demonstrated robust performance in real food matrices (milk, red wine, eggs) due to Bio-PB's specific binding and stability.
Conclusions:
- * The Bio-PB based sensor array offers a simple, rapid, and cost-effective platform for on-site SCM monitoring.
- * Machine learning-powered pattern recognition enhances the accuracy and reliability of SCM detection.
- * This approach presents a promising tool for improving food safety screening and preventing health risks associated with SCMs.
