Rapid meningitis pathogen typing via bioreceptor-free AI-SERS based on bacterial surface charge regulation

Rapid pathogen typing is essential for precise antimicrobial therapy; however, conventional methods are time-consuming. Although surface-enhanced Raman spectroscopy (SERS) provides quick detection, current methods rely on expensive recognition molecules or suffer from weak signals due to electrostatic repulsion-induced inefficient SERS-active nanoparticle adsorption with bacterial surfaces. Herein, we present a bioreceptor-free AI-powered SERS biosensing strategy for rapid and accurate pathogen typing. Through protonation under acidic conditions, the method regulates bacterial surface charge to overcome electrostatic repulsion between bacteria and SERS nanoparticles. It maintains bacterial viability and allows rapid in situ assembly of the SERS-active platform directly on bacterial surfaces, thereby preserving pathogen integrity, improving detection reliability, and significantly enhancing Raman signals. Integrated with machine learning, it delivers 99 % accuracy in typing three common bacterial meningitis pathogens, accompanied by a perfect AUC–ROC value of 1.0. The entire process, including sample pretreatment, surface charge modulation plus nanoparticle adsorption, SERS acquisition, and machine-learning typing, is completed in as little as 15 min. Cost-effective and user-friendly, this method eliminates the need for bioreceptors and complex processing, demonstrates robust performance in clinical samples, and shows broad applicability in public health, food safety, environmental monitoring, and emergency diagnostics.