No abstract

Genomic deoxyribounucleic acid (DNA) extracted from Brucella and Bacillus genera including Bacillus anthracis was investigated for the first time using Raman spectroscopy coupled with deep learning technique. Since DNA sequence is unique and independent of growth phases of bacteria, Raman spectroscopy can be a potential molecular diagnostic tool to identify different pathogens. Additionally, pure cellular components such as DNA provide pure Raman spectra and are not corrupted by spectral features from other cell components which is usually the case in whole organism detection. In this work, 15 DNA samples (two from Brucella genus and 13 from Bacillus genus) were studied. Raman signatures revealed unique features for Brucella and Bacillus genus bacteria. We propose an artificial intelligence (AI) based method, convolutional neural network (CNN) to discriminate all 15 DNA samples. The results reveal that Bacillus anthracis has distinct Raman DNA signatures compared to Bacillus cereus and Bacillus thuringiensis and could be discriminated from the latter two using principal component analysis (PCA), hierarchical cluster analysis (HCA), principal component-linear discriminant analysis (PC-LDA). In addition to these multivariate analysis techniques, we show that using convolutional neural network (CNN) architecture all 15 DNA samples could be discriminated with 100% accuracy.

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.