Using Machine Learning to Design a FeMOF Bidirectional Regulator for Electrochemiluminescence Sensing of Tau Protein

The single-luminophore-based ratiometric electrochemiluminescence (ECL) sensor coupling bidirectional regulator has become a research hotspot in the detection field because of its simplicity and accuracy. However, the limited bidirectional regulator hinders its further development. In this study, by leveraging the robust predictive capabilities of machine learning, we prepared an Fe-based metal-organic framework (FeMOF) as a bidirectional regulator for modulating the dual-emission ECL signals of a single luminophore for the first time. The proof of concept was demonstrated by applying FeMOF to the classical luminophore Ru(bpy)₃²⁺, and the results showed its ability to enhance the cathode ECL signal (E_cathode) and inhibit the anode ECL signal (E_anode). As an example, a ratiometric ECL sensor for Tau protein (Tau) detection utilizing the FeMOF/Ru(bpy)₃²⁺ system was developed. The incorporation of a bidirectional regulator in the ECL system effectively mitigated erratic fluctuations or minor discrepancies between the two signals and showed a stronger correlation and stability of E_cathode/E_anode than before regulation. As a result, the ECL sensor showed good analytical performance with a detection limit as low as 3.38 fg mL^-1 (S/N = 3). Moreover, it was not only comparable in test results to the commercially available ELISA kit but also could well distinguish between normal and Alzheimer's disease (AD) patients (80% specificity and 90% sensitivity). Thus, the proposed strategy is promising to be extended to other ECL luminophores or MOFs, providing a new path for ratiometric ECL sensors.