Neural Networks: A Comprehensive Foundation
Neural Networks: A Comprehensive Foundation
Data reduction in headspace analysis of blood and urine samples for robust bacterial identification
Computer Methods and Programs in Biomedicine
| Hi-index | 0.00 |
An investigation has been made into the identification of different strains of bacteria through the analysis of their volatile chemical headspace. Cultured samples were prepared and compared with clinical blood samples with known infection. Two different analytical instruments are combined to sample the volatile headspaces of samples, namely a mass spectrometer and a conducting polymer based electronic nose. Data were analysed using both linear parametric (principal components analysis) and non-linear non-parametric (radial basic function) techniques. Our results show that it is possible to discriminate between different strains of bacteria and in some cases between methicillin-resistant S. aureus (MRSA) and methicillin-sensitive S. aureus (MSSA). However, tests on inoculates prepared from closely related clinical isolates of S. aureus characterised by multilocus sequence typing (MLST), suggest that the acquisition of methicillin resistance per se cannot be detected by its volatile headspace alone. Thus the ability to discriminate between these bacterial strains appears to be more related to their overall genetic evolutionary distance or underlying differences in metabolism. This observation shows that the discriminating power of an electronic nose is limited to a subset of the subgroups of strains of MRSA and MSSA.