E Dublin, Belfield, Dublin four, Ireland; [email protected] (J.-L.X.); [email protected] (A.H.-L.); [email protected] (V.C.) School of Agriculture and Meals Science, University College Dublin, Belfield, Dublin 4, Ireland; [email protected] (S.L.); [email protected] (M.F.); [email protected] (A.G.M.S.) Institute of Food and Well being, University College Dublin, Belfield, Dublin 4, Ireland Correspondence: [email protected]: Xu, J.-L.; Herrero-Langreo, A.; Lamba, S.; Ferone, M.; Scannell, A.G.M.; Caponigro, V.; Gowen, A.A. Characterisation and Classification of Foodborne Bacteria Employing Reflectance FTIR Microscopic Imaging. Molecules 2021, 26, 6318. 10.3390/molecules26206318 Academic Editor: Barry K. Lavine Received: 9 September 2021 Accepted: 9 October 2021 Published: 19 OctoberAbstract: This perform investigates the application of reflectance Fourier transform infrared (FTIR) microscopic imaging for rapid, and non-invasive detection and classification amongst Bacillus subtilis and Escherichia coli cell suspensions dried onto metallic substrates (stainless steel (STS) and Viral Proteins Biological Activity aluminium (Al) slides) in the optical density (OD) concentration selection of 0.001 to ten. Benefits showed that reflectance FTIR of samples with OD reduced than 0.1 did not present an acceptable spectral signal to enable classification. Two modelling methods were devised to evaluate model trans-Ned 19 Technical Information overall performance, transferability and consistency amongst concentration levels. Modelling tactic 1 includes instruction the model with half of the sample set, consisting of all concentrations, and applying it for the remaining half. Using this approach, for the STS substrate, the very best model was achieved employing support vector machine (SVM) classification, delivering an accuracy of 96 and Matthews correlation coefficient (MCC) of 0.93 for the independent test set. For the Al substrate, the most effective SVM model created an accuracy and MCC of 91 and 0.82, respectively. Furthermore, the aforementioned greatest model built from one substrate was transferred to predict the bacterial samples deposited on the other substrate. Final results revealed an acceptable predictive capability when transferring the STS model to samples on Al (accuracy = 82). Nevertheless, the Al model could not be adapted to bacterial samples deposited on STS (accuracy = 57). For modelling approach 2, models have been created utilizing one particular concentration level and tested on the other concentrations for every single substrate. Results proved that models constructed from samples with moderate (1 OD) concentration may be adapted to other concentrations with great model generalization. Prediction maps revealed the heterogeneous distribution of biomolecules due to the coffee ring effect. This function demonstrated the feasibility of applying FTIR to characterise spectroscopic fingerprints of dry bacterial cells on substrates of relevance for meals processing. Key phrases: FTIR; foodborne bacteria; classification; machine finding out; stainless steelPublisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.1. Introduction Foodborne bacteria are of important concern and pose a critical threat to public overall health and food safety. In line with the Planet Overall health Organization, an estimated 2.2 million men and women die of foodborne ailments every year [1]. The increasing awareness of the overall health dangers of foodborne illnesses has accounted for the greater efforts to create fast and sensitive approaches for pathogen detection an.
