Culture-Independent Raman Spectroscopic Identification of Bacterial Pathogens from Clinical Samples Using Deep Transfer Learning

• Published in: Analytical chemistry (Washington) 2022-10, Vol.94 (42), p.14745-14754
• Main Authors: Singh, Saumya, Kumbhar, Dipak, Reghu, Dhanya, Venugopal, Shwetha J., Rekha, P. T., Mohandas, Silpa, Rao, Shruti, Rangaiah, Ambica, Chunchanur, Sneha K., Saini, Deepak Kumar, Umapathy, Siva
• Format: Article
• Language: English
• Subjects: Bacteria; Cell culture; Chemistry; Clinical isolates; Clinical microbiology; Datasets; Deep learning; Feature extraction; Heterogeneity; Identification methods; Machine learning; Microbiology; Nucleic acids; Pathogens; Raman spectroscopy; Robustness; Species classification; Spectra; Spectroscopy; Spectrum analysis; Sputum; Transfer learning
• ISSN: 0003-2700; 1520-6882
• DOI: 10.1021/acs.analchem.2c03391

The rapid identification of bacterial pathogens in clinical samples like blood, urine, pus, and sputum is the need of the hour. Conventional bacterial identification methods like culturing and nucleic acid-based amplification have limitations like poor sensitivity, high cost, slow turnaround time, etc. Raman spectroscopy, a label-free and noninvasive technique, has overcome these drawbacks by providing rapid biochemical signatures from a single bacterium. Raman spectroscopy combined with chemometric methods has been used effectively to identify pathogens. However, a robust approach is needed to utilize Raman features for accurate classification while dealing with complex data sets such as spectra obtained from clinical isolates, showing high sample-to-sample heterogeneity. In this study, we have used Raman spectroscopy-based identification of pathogens from clinical isolates using a deep transfer learning approach at the single-cell level resolution. We have used the data-augmentation method to increase the volume of spectra needed for deep-learning analysis. Our ResNet model could specifically extract the spectral features of eight different pathogenic bacterial species with a 99.99% classification accuracy. The robustness of our model was validated on a set of blinded data sets, a mix of cultured and noncultured bacterial isolates of various origins and types. Our proposed ResNet model efficiently identified the pathogens from the blinded data set with high accuracy, providing a robust and rapid bacterial identification platform for clinical microbiology.