High-Speed Diagnosis of Bacterial Pathogens at the Single Cell Level by Raman Microspectroscopy with Machine Learning Filters and Denoising Autoencoders

Abstract

Accurate and rapid identification of infectious bacteria is important in medicine. Raman microspectroscopy holds great promise in performing label-free identification at the single-cell level. However, due to the naturally weak Raman signal, it is a challenge to build extensive databases and achieve both accurate and fast identification. Here, we used signal-to-noise ratio (SNR) as a standard indicator for Raman data quality and performed bacterial identification using 11, 141 single-cell Raman spectra from nine bacterial strains. Subsequently, using two machine learning methods, a simple filter, and a neural networkbased denoising autoencoder (DAE), we demonstrated 92% (simple filter using 1 s/cell spectra) and 84% (DAE using 0.1 s/ cell spectra) identification accuracy. Our machine learning-aided Raman analysis paves the way for high-speed Raman microspectroscopic clinical diagnostics.

Technique Evolutions for Microorganism Detection in Complex Samples: A Review

“…Recent progress in artificial intelligence, and more specifically deep learning, opened the way to an automatic and robust analysis of images. This is also true for the analysis of microscopic images, and deep learning was applied to the detection and identification of microorganisms [103,104].…”

Artificial intelligence-aided rapid and accurate identification of clinical fungal infections by single-cell Raman spectroscopy

“…Therefore, the choice of an algorithm with logical design and minimal computational costs is essential to perform classification tasks using a large-scale dataset. Our recent study using 11,141 Raman spectra and a deep learning-based autoencoder has demonstrated ultrafast identification of bacterial pathogens with 97% accuracy and one-second acquisition time ( Xu et al, 2022 ). Compared to bacteria, the importance of artificial intelligence analysis of single-cell Raman spectra on fungi is still at an early stage, and only a few studies used a small number of standard strains or clinical isolates and a limited species coverage ( Witkowska et al, 2016 ; Žukovskaja et al, 2018 ; Pezzotti et al, 2021 ).…”

Optofluidic identification of single microorganisms using fiber‐optical‐tweezer‐based Raman spectroscopy with artificial neural network

“…Commonly used methods, such as drying and glutaraldehyde fixation, cannot maintain living activity of the microorganisms. [24][25][26][27][28] In recent years, optical trapping method has shown its effectiveness in boosting the biomolecular interaction in Raman spectroscopy and biosensing. [29,30] Therefore, laser tweezers Raman spectroscopy (LTRS) has been developed to achieve label-free and damage-free in situ detection.…”

“…However, such a combination faces the challenge of the strong microbial motility that makes the collected Raman signals extremely weak. Commonly used methods, such as drying and glutaraldehyde fixation, cannot maintain living activity of the microorganisms [24–28] . In recent years, optical trapping method has shown its effectiveness in boosting the biomolecular interaction in Raman spectroscopy and biosensing [29,30] .…”