Nondestructive microbial discrimination using single-cell Raman spectra and random forest machine learning algorithm

Nanako Kanno; Shingo Kato; Moriya Ohkuma; Motomu Matsui; Wataru Iwasaki; Shinsuke Shigeto

doi:10.1016/j.xpro.2022.101812

Nondestructive microbial discrimination using single-cell Raman spectra and random forest machine learning algorithm

STAR Protoc. 2022 Nov 3;3(4):101812. doi: 10.1016/j.xpro.2022.101812. eCollection 2022 Dec 16.

Authors

Nanako Kanno¹, Shingo Kato², Moriya Ohkuma², Motomu Matsui³, Wataru Iwasaki³, Shinsuke Shigeto⁴

Affiliations

¹ Department of Chemistry, School of Science, Kwansei Gakuin University, Sanda, Hyogo 669-1330, Japan. Electronic address: nkanno@kwansei.ac.jp.
² Japan Collection of Microorganisms, RIKEN BioResource Research Center, Tsukuba, Ibaraki 305-0074, Japan.
³ Department of Biological Sciences, Graduate School of Science, The University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan; Department of Integrated Biosciences, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Chiba 277-0882, Japan.
⁴ Department of Chemistry, School of Science, Kwansei Gakuin University, Sanda, Hyogo 669-1330, Japan. Electronic address: shigeto@kwansei.ac.jp.

Abstract

Raman microspectroscopy is a powerful tool for obtaining biomolecular information from single microbial cells in a nondestructive manner. Here, we detail steps to discriminate prokaryotic species using single-cell Raman spectra acquisitions followed by data preprocessing and random forest model tuning. In addition, we describe the steps required to evaluate the model. This protocol requires minimal preprocessing of Raman spectral data, making it accessible to non-spectroscopists, yet allows intuitive visualization of feature importance. For complete details on the use and execution of this protocol, please refer to Kanno et al. (2021).

Keywords: Bioinformatics; Biophysics; Chemistry; Microbiology; Single cell.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Algorithms
Machine Learning*
Serogroup
Spectrum Analysis, Raman* / methods