Detection of bacterial signatures from culture supernatants using surface-enhanced Raman spectroscopy and machine learning

Bacterial growth inherently modifies the surrounding culture medium through the secretion and accumulation of metabolites and biomolecular byproducts, yet these extracellular components are typically treated as background in surface-enhanced Raman spectroscopy (SERS)–based bacterial analyses. Here, we demonstrate that cell-free bacterial culture supernatants themselves constitute robust and information-rich SERS fingerprints that enable reliable bacterial identification when coupled with machine-learning analysis. Using reproducible silver nanorod substrates fabricated by oblique-angle deposition, we systematically analyzed culture supernatants from multiple foodborne pathogens ( Escherichia coli O103, O111, O121, O157, and Salmonella STNR) and animal-associated Leptospira serovars (Pomona, Bratislava, Canicola, Grippotyphosa, and Icterohaemorrhagiae), alongside matched sterile growth-media controls (TSB and EMJH). Spectral analysis reveals reproducible bacterial-induced peak shifts, intensity redistribution, and the emergence of phosphate-, nucleic-acid-, and protein-associated vibrational features relative to sterile media, confirming that extracellular bacterial biomarkers are retained within the supernatants. Unsupervised principal component analysis reveals intrinsic clustering of bacterial supernatants, while supervised support vector machine classification achieves 99.29 ± 0.43% accuracy for foodborne pathogens, 96.19 ± 1.79 % accuracy for Leptospira serovars, and 97.83 ± 0.94 % accuracy for a combined multi-group dataset acquired across independent laboratories. By establishing bacterial culture supernatants as a reproducible and diagnostically informative SERS target rather than a confounding background, this work introduces a simplified cell-free strategy for bacterial identification that reduces direct handling of intact bacterial cells during the SERS measurement step and expands the practical scope of SERS-based biosensing.