Targeted SERS Nanoparticle Accumulation on Bacteria Enables On-Site Accurate Vibrio Typing in 10 min with AI-Powered SERS.

Abstract

Vibrio infections progress rapidly with high mortality, yet current diagnostics fail to meet the critical need for rapid, accurate, and field-deployable detection. To bridge this gap, we present a bioreceptor-free, AI-powered surface-enhanced Raman spectroscopy (SERS) platform for rapid Vibrio typing. Our approach leverages Na₂HPO₄-mediated synergistic anion/cation interactions to direct targeted accumulation of gold nanoparticles onto bacterial surfaces within 5 min, overcoming the uncontrollable aggregation and poor reproducibility of conventional methods. By integrating a gradient-weighted class activation mapping-assisted Vision Transformer model for intelligent spectral analysis with a custom-built portable Raman instrument for automated measurements, a remarkable typing accuracy of 97.62% is achieved across seven bacterial species (five Vibrio species and two interferents). The entire sample-to-result process takes only 10 min. This work establishes a new paradigm for SERS-based bacterial detection/typing and provides a promising solution for point-of-care diagnostics in public health, food safety, and environmental monitoring.