Nanosensor-Based Pattern-Generating Probe Accelerates Sepsis Diagnosis.

Abstract

Biomimetic optical sensor arrays hold promising potential in differentiating nuances within intricate mixtures and biosamples. Nonetheless, developing a standalone pattern-generating sensor for multianalyte identification in clinical biofluids, without the necessity of complicated synthesis, presents significant challenges. Here, we introduce a nanosensor-inspired sensor array (NanoSA), composed of a nanoassembly incorporating three sensors, that exhibits multiple intermolecular fluorescence resonance energy transfer (FRET) effects. This pattern-generating nanoassembly, featuring six channels, parallel sensing, and rapid response, identifies 24 septic bacteria within 30 s (96.9% accuracy). Meanwhile, the nanoassembly shows excellent results in discriminating between multiplex bacterial infections, varying concentrations of bacteria, and those in diverse biofluids (serum and urine). Optimized by nine machine learning algorithms, the multilayer perceptron (MLP) model achieved a 97.6% test accuracy in differentiating clinical sepsis samples infected by five distinct bacteria within 30 s, highlighting the potential of the NanoSA as a crucial tool for the rapid clinical diagnosis.