1. Institute of Physical Chemistry and Abbe Center of Photonics, Friedrich Schiller University of Jena, Jena, Germany;2. InfectoGnostics Forschungscampus Jena, Jena, Germany;3. Alere Technologies GmbH, Jena, Germany;4. National Reference Center for Mycobacteria, Research Center Borstel, Borstel, Germany;5. Present address: Laboratory Dr. Limbach, Heidelberg, Germany;6. Molecular and Experimental Mycobacteriology, Research Center Borstel, Borstel, Germany;7. +49 3641 948381+49 3641 948302;8. Leibniz‐Institute of Photonic Technology, Jena, Germany
Abstract:
In this study, Raman microspectroscopy has been utilized to identify mycobacteria to the species level. Because of the slow growth of mycobacteria, the per se cultivation‐independent Raman microspectroscopy emerges as a perfect tool for a rapid on‐the‐spot mycobacterial diagnostic test. Special focus was laid upon the identification of Mycobacterium tuberculosis complex (MTC) strains, as the main causative agent of pulmonary tuberculosis worldwide, and the differentiation between pathogenic and commensal nontuberculous mycobacteria (NTM). Overall the proposed model considers 26 different mycobacteria species as well as antibiotic susceptible and resistant strains. More than 8800 Raman spectra of single bacterial cells constituted a spectral library, which was the foundation for a two‐level classification system including three support vector machines. Our model allowed the discrimination of MTC samples in an independent validation dataset with an accuracy of 94% and could serve as a basis to further improve Raman microscopy as a first‐line diagnostic point‐of‐care tool for the confirmation of tuberculosis disease.
