Dual fluorescence-enhanced AuAg-ZIF@Content structure cleavage-mediated ratiometric emission binding interpretable...

With the advancement of technology and the focus on health, biosensors have undergone a shift in focus from breakthroughs in a single area to an emphasis on improving integrated performance. Herein, aiming at two principal aspects of signal transduction and data processing in the development of fluorescent material-based biosensors, a generalized biosensing platform was developed that integrates dual-enhanced fluorescent materials, ratiometric detection, and interpretable machine learning algorithms. Initially, AuAg-ZIF was synthesized, whose fluorescence properties were doubly enhanced relative to AuNCs through the anti-galvanic reaction occurring between surface Ag(I) and core Au(0) and the confinement effect of ZIF-8. Furthermore, AuAg-ZIF@CD was prepared by encapsulating CD into AuAg-ZIF using the inner cavity of ZIF-8, which enables ratiometric fluorescence detection of different targets by the degradation responsiveness of ZIF-8 to acids, ATP, and phosphates. Subsequently, an interpretable machine learning prediction algorithm was constructed for automated data processing, which determines target concentrations within 5 s following image acquisition. Finally, to validate the platform for practical application, we achieved immediate, intelligent and accurate detection of glucose concentration-mediated ratiometric fluorescence changes in AuAg-ZIF@CD (R² = 0.95) based on AuAg-ZIF@CD degradation reactivity to acid. This was successfully confirmed in 48 clinical samples (AUC = 0.988).