Mechanistic Design of Graphdiyne-Based Multimodal Sensing Integrating Machine Learning and Photothermal Dynamics for...

Rapid and accurate detection of pathogenic bacteria remains essential for infection control and timely treatment. Here, we report a graphdiyne (GDY)-based self-powered biosensing platform that integrates CRISPR/Cas12a molecular recognition with GDY/Au nanoparticle-engineered bioelectrodes for multimodal detection and photothermal inactivation of Vibrio parahaemolyticus. The ultrathin GDY framework provides a high-surface-area scaffold for uniform Au nanoparticle dispersion, facilitating interfacial charge transfer and enhancing enzyme-mediated redox kinetics at the bioanode. Target-triggered CRISPR/Cas12a trans-cleavage regulates the release of glucose oxidase from a hairpin probe, enabling a self-powered electrochemical readout driven by glucose oxidation. In parallel, HRP-catalyzed oxidation of TMB generates oxTMB with strong near-infrared absorbance, providing complementary colorimetric and photothermal (808 nm) outputs and enabling in situ bacterial inactivation. The electrochemical, colorimetric, and thermal modes exhibit concentration-dependent responses with limits of detection of 0.34, 0.41, and 0.78 CFU mL^-1, respectively. Integration of multimodal signals via machine learning further enables infection grading with an overall accuracy of 97.71%. This multimodal diagnostic-therapeutic strategy demonstrates reliable performance in both in vitro and in vivo wound infection models, highlighting its potential for localized infection monitoring and point-of-care bacterial management. This study provides a proof-of-concept demonstration of an integrated self-powered multimodal biosensing platform for simultaneous bacterial detection and inactivation.