Fluorescent Machine Learning Aided Classification of Pathogenic Bacteria Using the Excitation Emission Matrix

Characterization and classification of pathogenic bacteria using native fluorescence and spectral deconvolution

“…Recently, Mito et al optimized 280, 300, 380, and 480 nm as fluorescence excitation wavelengths and used three machine learning techniques for effective classification of bacteria[26]. Our group reported on the characterization and classification of clinically important eight bacterial species using EEM and PARAFAC analysis and achieved 100% accurate classification[45]. Although EEM techniques haveF I G U R E 6 Scatter plot showing the distribution of DF1 versus DF2 for eight bacterial strains using fluorescence emission spectral dataset at 280 nm excitation.…”

“…The classification results of stepwise multiple linear discriminant analysis using the original and cross validated dataset of fluorescence emission of eight bacteria classes. C. freundii E. faecalis E. cloacae K. pneumoniae P. mirabilis P. aeruginosa S. aureus Original (%) Classification results of bacterial species using fluorescence emission spectra Although the use of native fluorescence spectroscopic characterization of bacterial strains was reported as early as 1950[41,42], only from 1980 onwards, many attempted to characterize and classify different bacterial strains using different fluorescence spectroscopic techniques to optimize the excitation and/or emission wavelengths due to the emergence of various chemo-metric methods for accurate classification[26,[43][44][45]. However, still this area is under progress in the optimization of excitation wavelength(s), fluorescence spectroscopic techniques, and chemo-metric methods for better identification and classification.…”