Abstract:
A spectroscopic method for spectroscopic detection and identification of bacteria in culture is disclosed. The method incorporates construction of at least one data set, which may be a spectrum, interference pattern, or scattering pattern, from a cultured sample suspected of containing said bacteria. The data set is corrected for the presence of water in the sample, spectral features are extracted using a principal components analysis, and the features are classified using a learning algorithm. In some embodiments of the invention, for example, to differentiate MRSA from MSSA, a multimodal analysis is performed in which identification of the bacteria is made based on a spectrum of the sample, an interference pattern used to determine cell wall thickness, and a scattering pattern used to determine cell wall roughness. An apparatus for performing the method is also disclosed, one embodiment of which incorporates a multiple sample analyzer.
Abstract:
A spectroscopic method for spectroscopic detection and identification of bacteria in culture is disclosed. The method incorporates construction of at least one data set, which may be a spectrum, interference pattern, or scattering pattern, from a cultured sample suspected of containing said bacteria. The data set is corrected for the presence of water in the sample, spectral features are extracted using a principal components analysis, and the features are classified using a learning algorithm. In some embodiments of the invention, for example, to differentiate MRSA from MSSA, a multimodal analysis is performed in which identification of the bacteria is made based on a spectrum of the sample, an interference pattern used to determine cell wall thickness, and a scattering pattern used to determine cell wall roughness. An apparatus for performing the method is also disclosed, one embodiment of which incorporates a multiple sample analyzer.
Abstract:
A spectroscopic method for spectroscopic detection and identification of bacteria in culture is disclosed. The method incorporates construction of at least one data set, which may be a spectrum, interference pattern, or scattering pattern, from a cultured sample suspected of containing said bacteria. The data set is corrected for the presence of water in the sample, spectral features are extracted using a principal components analysis, and the features are classified using a learning algorithm. In some embodiments of the invention, for example, to differentiate MRS A from MSSA, a multimodal analysis is performed in which identification of the bacteria is made based on a spectrum of the sample, an interference pattern used to determine cell wall thickness, and a scattering pattern used to determine cell wall roughness. An apparatus for performing the method is also disclosed, one embodiment of which incorporates a multiple sample analyzer.
Abstract:
A spectroscopic method for spectroscopic detection and identification of bacteria in culture is disclosed. The method incorporates construction of at least one data set, which may be a spectrum, interference pattern, or scattering pattern, from a cultured sample suspected of containing said bacteria. The data set is corrected for the presence of water in the sample, spectral features are extracted using a principal components analysis, and the features are classified using a learning algorithm. In some embodiments of the invention, for example, to differentiate MRS A from MSSA, a multimodal analysis is performed in which identification of the bacteria is made based on a spectrum of the sample, an interference pattern used to determine cell wall thickness, and a scattering pattern used to determine cell wall roughness. An apparatus for performing the method is also disclosed, one embodiment of which incorporates a multiple sample analyzer.
Abstract:
A spectroscopic method for spectroscopic detection and identification of bacteria in culture is disclosed. The method incorporates construction of at least one data set, which may be a spectrum, interference pattern, or scattering pattern, from a cultured sample suspected of containing said bacteria. The data set is corrected for the presence of water in the sample, spectral features are extracted using a principal components analysis, and the features are classified using a learning algorithm. In some embodiments of the invention, for example, to differentiate MRS A from MSSA, a multimodal analysis is performed in which identification of the bacteria is made based on a spectrum of the sample, an interference pattern used to determine cell wall thickness, and a scattering pattern used to determine cell wall roughness. An apparatus for performing the method is also disclosed, one embodiment of which incorporates a multiple sample analyzer.