SERS nanowire chip and machine learning enabled instant identification and classification of clinically relevant wild-type and antibiotic resistant bacteria at species and strain level
The world health organization considers antimicrobial resistance (AMR) to be a critical global public health problem. Conventional culture-based methods that are used to detect and identify bacterial infection are slow. Thus, there is a growing need for the development of robust, cost-effective, and fast diagnostic solutions for the identification of pathogens. Surface-enhanced Raman spectroscopy (SERS) can be used to identify target analytes with sensitivity down to the single-molecule level. Here, we developed a SERS chip by optimizing the entire fabrication pipeline of the metal-assisted chemical etching (MACE) method. The MACE approach offers a large-scale, densely packed silver (Ag) nanostructure on top of silicon nanowires (Si-NWs) with a large aspect ratio that significantly enhances the Raman signal due to localised surface plasmonic enhancement. The optimised SERS chips exhibited sensitivity down to 10-12 M concentration of R6G molecule and detected reproducible Raman spectra of bacteria down to a concentration of 100 colony forming units (CFU)/ml, which is a thousand times lower than the clinical threshold of bacterial infections like UTI (105 CFU/ml). A Siamese neural network model was used to classify SERS Raman spectra from bacteria specimens. The trained model identified 12 different bacterial species, including those which are causative agents for tuberculosis and urinary tract infection (UTI). Next, the SERS chips and another Siamese neural network model were used to differentiate antibiotic-resistant strains from susceptible strains of E. coli. The enhancement offered by SERS chip enabled acquisitions of Raman spectra of bacteria directly in the synthetic urine by spiking the sample with only 103 CFU/ml E. coli. Thus, the present study lays the ground for the identification and quantification of bacteria on SERS chips, thereby offering a potential future use for rapid, reproducible, label-free, and low limit detection of clinical pathogens.