Machine learning-assisted single-cell Raman fingerprinting for in situ and nondestructive classification of prokaryotes. Kanno et al.

In the natural environments, there are various phylogenetic groups of Bacteria and Archaea including unknown microorganisms. Non-destructive, single-cell level species discrimination will contribute to single-cell genomics, ecophysiological analysis at the single-cell level, and isolation. Using single-cell Raman spectra of six microorganisms (three bacteria and three archaea) and machine learning, we tested whether it is possible to identify microbial species at the single-cell level, and succeeded in classifying them with high accuracy.
The datasets are raw single-cell Raman spectra acquired from microbial cells suspended in phosphate buffer solution (PBS).

Contents:
1) Single-cell Raman spectra of 6 prokaryotic species
Escherichia coli JCM 20135
Bacillus subtilis JCM 1465T
Thermus thermophilus JCM 10941T
Thermococcus kodakarensisJCM 12380T
Sulfolobus acidocaldarius JCM 8929T
Nitrososphaera viennensis JCM 19564T
2) Raman spectra of PBS (10 times/file)
3) Neon standard
The numerical part of the file name is related to the numerical part of the folder name that contains the Raman spectra of the microorganisms.

# Column 1 in the file is only a reference for the Raman shift. Do not use it.
# Please follow the method described in the paper to remove cosmic rays, subtract PBS spectrum, use neon standards to integrate data from different days, make baseline corrections, and normalize the data.