PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
December 8, 2025Analytical Chemistry4 citations

Integrating AI-assisted SERS Biosensing and Photoactivated Antibacterial Therapy in Au@Cu 2– x Se for Combating Multidrug-Resistant Bacteria

View Full Paper
RYRuiling YuanHZHonghong ZhanYLYang Liu

Key Points

  • This research aims to develop a multifunctional platform for identifying and combating multidrug-resistant bacteria.
  • Developed a Au@Cu2-xSe core-shell nanoplatform for biosensing and antibacterial therapy.
  • Employed surface-enhanced Raman spectroscopy (SERS) for detecting bacterial metabolites.
  • Applied a convolutional neural network for classification and identification of pathogens.
  • Achieved >95% accuracy in identifying six pathogenic species.
  • Obtained >99% accuracy in Gram-type differentiation.
  • Enabled rapid detection and eradication of Gram-positive and Gram-negative bacteria.

Abstract

The escalating global crisis of multidrug-resistant (MDR) bacteria demands innovative strategies that bypass conventional antibiotic limitations. This study introduced a multifunctional Au@Cu2-xSe core-shell nanoplatform integrating artificial intelligence (AI)-assisted surface-enhanced Raman spectroscopy (SERS) biosensing with photoactivated antibacterial therapy for combating MDR pathogens. The Au@Cu2-xSe nanoparticles exhibit exceptional photothermal conversion efficiency and robust photodynamic activity, generating cytotoxic reactive oxygen species (ROS) under light excitation. This dual photothermal-photodynamic mechanism enables rapid, broad-spectrum eradication of both Gram-positive and Gram-negative bacteria by disrupting cellular integrity and inducing oxidative damage, overcoming intrinsic resistance barriers. Complementing this therapeutic action, we engineered large-area, uniform nanobowl-array surface-enhanced Raman spectroscopy (SERS) substrates functionalized with Au@Cu2-xSe hotspots. These substrates amplified bacterial Raman signals, enabling label-free detection of trace metabolites and biomarkers at single-cell resolution. To address spectral complexity, an AI driven workflow was developed: Raw spectra underwent rigorous preprocessing, followed by dimensionality reduction and classification via a tailored 1D convolutional neural network. The model achieved >95% accuracy in identifying six pathogenic species and >99% accuracy in Gram-type differentiation. This integrated platform bridges rapid, precise pathogen identification with targeted photomechanical sterilization, offering a promising "diagnose-and-treat" paradigm for point-of-care management of MDR infections.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Yuan et al. (2025) studied this question.

synapsesocial.com/papers/69362f6c4fa91c937236e037https://doi.org/10.1021/acs.analchem.5c04671
Ask AI
Helpful
Bookmark
Share
View Full Paper