Discrimination of bacteria using pyrolysis-gas chromatography-differential mobility spectrometry (Py-GC-DMS) and chemometrics

Abstract

Discrimination of bacteria was investigated using pyrolysis-gas chromatography-differential mobility spectrometry (Py-GC-DMS). Three strains belonging to the genus Bacillus were investigated and these included two strains of Bacillus subtilis and a single Bacillus megaterium. These were chosen so as to evaluate the possibility of bacterial strain discrimination using Py-GC-DMS. The instrument was constructed in-house and the long-term reproducibility of the instrument was evaluated over a period of 60 days using a Scotch whisky quality control. To assess the reproducibility further each bacterium was cultured six times and each culture was analysed in replicate to give three analytical replicates. The DMS data were generated in both positive and negative modes, and the data in each mode were analysed independently of each other. The Py-GC-DMS data were pre-processed via correlation optimised warping (COW) and asymmetric least square (ALS) to align the DMS chromatograms and to remove any unavoidable baseline shifts, prior to normalisation. Processed chromatograms were analysed using principal component analysis (PCA) followed by supervised learning methodology using partial least squares for discriminant analysis (PLS-DA). It was found that the separations between B. subtilis and B. megaterium can be readily observed by PCA; however, strain discrimination within the two B. subtilis was only possible using supervised learning. As multiple biological replicates were analysed an exhaustive splitting of the training and test sets was undertaken and this allowed correct classification rates (CCRs) to be assessed for the 3375 test sets. It was found that with PLS-DA the negative ion mode DMS data were more discriminatory than the positive mode data.

Robust detection of P. aeruginosa and S. aureus acute lung infections by secondary electrospray ionization-mass spectrometry (SESI-MS) breathprinting: from initial infection to clearance

“…PLS is a supervised latent variable-based model that finds a linear model describing some predicted variables in terms of a set of observable variables. In this study, the observable variables (the SESI-MS peak intensities) were compressed into a few latent variables, called PLS factors, and then discriminant analyses were applied using these factors to predict the classification (the bacterial species that infected the mice) of each tested sample [24]. As can be seen in Fig.…”

Computational Methods for Metabolomic Data Analysis of Ion Mobility Spectrometry Data—Reviewing the State of the Art

“…In a related study, Westhoff et al embedded PCA into a comprehensive statistical analysis of IMS breath samples taken from 95 COPD patients and 35 healthy individuals including Mann-Whitney U test, correlation analysis and decision trees [66]. Cheung et al applied PCA to Py-GC-DMS data sets of two strains of B. subtilis and one strain of B. megaterium [67]. PCA proved to be sufficient to discriminate bacterial strains on species level, while separation of the two B. subtilis strains required chemometric methods using supervised classification.…”

Detection of infectious agents in the airways by ion mobility spectrometry of exhaled breath

“…More recently, the detection of different VOCs and their relations to medical questions was reported, including ion mobility spectrometry [22, [55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70][71]. Some of the VOCs were related to bacteria taken from headspace of cultures [58,72,73].…”