Deep Learning-Enhanced DNAzyme-Driven Rolling-Circle Amplification Encoding for Multibacterial Detection.

Area of Science:

• Biotechnology
• Molecular Biology
• Food Safety

Background:

• Food-borne illnesses often stem from multiple pathogens, challenging conventional single-pathogen detection methods.
• Current diagnostic tools may lack the speed and multiplexing capability for real-time outbreak surveillance.

Purpose of the Study:

• To develop and validate a novel system for simultaneous detection and quantification of multiple live bacterial pathogens.
• To integrate DNAzyme amplification with smartphone-based imaging and AI for rapid pathogen screening.

Main Methods:

• A DNAzyme-driven rolling-circle amplification/molecular-beacon encoding system (DRM-ES) was designed.
• Bacteria-secreted proteins trigger DNAzyme cleavage, initiating rolling-circle amplification (RCA).
• Multiplexed molecular beacons generate distinct fluorescent signals (blue, green, red) captured by a smartphone and analyzed by a convolutional neural network (CNN).

Main Results:

• The DRM-ES achieved a sensitivity of 10^1–10^2 CFU/mL for detecting S. aureus, B. cocovenenans, and E. coli.
• The system demonstrated 100% positive and ≥95.2% negative agreement compared to traditional culture methods.
• 29 out of 32 real-world samples were correctly identified as naturally contaminated with the target bacteria using a 32-tube array.

Conclusions:

• The developed platform provides culture-comparable sensitivity and live-cell specificity for pathogen detection.
• DRM-ES offers a generalizable blueprint for large-scale, multiplex pathogen screening in various sample types.
• This approach enhances capabilities for rapid identification and quantification of foodborne pathogens.