Smartphone-Based Bimetallic Single-Atom Nanozyme Sensor Array Integrated with Deep Learning for Rapid Biothiol Detection

Abstract

Rapid and sensitive biothiol detection is critical for domains spanning food safety, environmental monitoring, and clinical diagnostics. Here, we report the development of a portable and highly sensitive point-of-care testing (POCT) platform that integrates bimetallic single-atom nanozymes (CuZn-N) with smartphone technology and deep learning algorithms. We synthesized CuZn-N bimetallic single-atom nanozymes exhibiting superior peroxidase-like activity and constructed a dual-channel colorimetric sensor array. The array detects biothiols via distinct colorimetric responses generated by analyte-specific inhibition of single-atom nanozymes (SAzyme) activity. By leveraging the high-resolution capabilities of smartphones and the advanced YOLO v5 deep learning algorithm, we developed a smartphone application, “ThiolSense”, which analyzes color changes in images to achieve sensitive and selective detection of cysteine (Cys), glutathione (GSH), and homocysteine (Hcy). The application employs image segmentation and feature extraction to process RGB color channel data, assisted with principal component analysis (PCA) and hierarchical clustering analysis (HCA) for enhanced accuracy. We demonstrated the platform’s potential in real-sample testing by quantifying biothiols in fetal bovine serum (FBS) and human serum (HS). The smartphone-equipped CuZn-N bimetallic single-atom nanozyme sensor array offers high sensitivity, portability, and user-friendly operation, facilitating rapid and accurate on-site detection. This innovative integration of nanomaterials, smartphone technology, and artificial intelligence presents a powerful tool for biosensing applications, with broad implications for critical fields requiring efficient biothiol detection.