Detection and Identification of Bacteria in an Isolated System with Near-Infrared Spectroscopy and Multivariate Analysis

Abstract

Near-infrared (NIR) transflectance spectra of Listeria innocua FH, Lactococcus lactis, Pseudomonas fluorescens, Pseudomonas mendocina, and Pseudomonas putida suspensions were collected and investigated for their potential use in the identification and classification of bacteria. Unmodified spectral data were transformed (first and second derivative) using the Savitzsky-Golay algorithm. Principal component analysis (PCA), partial least-squares discriminant analysis (PLS2-DA), and soft independent modeling of class analogy (SIMCA) were used in the analysis. Using either full cross-validation or separate calibration and prediction data sets, PLS2 regression classified the five bacterial suspensions with 100% accuracy at species level. At Pseudomonas genus level, PLS2 regression classified the three Pseudomonas species with 100% accuracy. In the case of SIMCA, prediction of an unknown sample set produced correct classification rates of 100% except for L. innocua FH (77%). At genus level, SIMCA produced correct classification rates of 96.7, 100, and 100% for P. fluorescens, P. mendocina, and P. putida, respectively. This successful investigation suggests that NIR spectroscopy can become a useful, rapid, and noninvasive tool for bacterial identification.