Detection of Non-O157 Shiga Toxin-Producing Escherichia coli (STEC) Serogroups with Hyperspectral Microscope Imaging Technology

Abstract

Abstract. Non-O157 Shiga toxin-producing Escherichia coli (STEC) serogroups such as O26, O45, O103, O111, O121, and O145 are recognized as serious risks to human health due to their toxicity. The conventional microbiological detection method of cell counting on agar plates is laborious and time-consuming. Because optical methods are promising for real-time, in situ foodborne pathogen detection, a hyperspectral microscope imaging (HMI) method based on acousto-optic tunable filters (AOTF) was developed for detecting pathogenic bacteria with a capability to differentiate the spectral characteristics of each bacterial cell from microcolony samples. Using the AOTF-based HMI method, a total of 89 contiguous spectral images were acquired within approximately 45 s with 250 ms exposure time. In this study, we developed a protocol for successfully immobilizing live cells on glass slides to acquire quality spectral images of STEC bacterial cells using a modified drying method. Among the contiguous spectral imagery between 450 and 800 nm, the intensities at 458, 498, 522, 546, 570, 586, 670, and 690 nm were distinct for STEC bacteria under dark-field illumination. With two different classification algorithms, i.e., support vector machine (SVM) and sparse kernel-based ensemble learning (SKEL), STEC serogroup O45 could be classified with 92% detection accuracy. However, the mean accuracies in identifying the six STEC serogroups with SVM and SKEL were not high enough for use in classification models.