Rapid Identification of Foodborne Pathogen Biofilms Using Fluorescence Sensor Array and Machine Learning Models

Abstract

Biofilms formed by common foodborne pathogens, including Salmonella, Staphylococcus aureus, andEscherichia coli, are persistent contamination sources on food-contact surfaces and pose major food safety concerns. In this study, a fluorescence-based nanosensor array composed of single-stranded DNA (ssDNA) probes and two-dimensional (2D) nanomaterials was developed for rapid biofilm characterization. Fluorescence recovery generated distinct response patterns following interactions with biofilm components collected from polystyrene and stainless-steel surfaces using a simple cotton swab method. Combined with machine learning analysis, the platform enabled biofilm identification and growth-stage classification within 30 min. Support vector machine models achieved 100% accuracy for bacterial species identification and 76% accuracy for biofilm growth-stage classification using best subset feature selection. These results demonstrate the potential of nanosensor arrays as a rapid and versatile tool for monitoring biofilms on food-contact surfaces.