Harmful algal blooms (HABs) are a natural phenomenon
caused by
outbreaks of algae, resulting in serious problems for aquatic ecosystems
and the coastal environment. Chaetoceros tenuissimus (C. tenuissimus) is one of the diatoms
responsible for HABs. The growth curve of C. tenuissimus can be observed from beginning to end of HABs: therefore, detailed
analysis is necessary to characterize each growth phase of C. tenuissimus. It is important to examine the phenotype
of each diatom cell individually, as they display heterogeneity even
in the same growth phase. Raman spectroscopy is a label-free technique
to elucidate biomolecular profiles and spatial information at the
cellular level. Multivariate data analysis (MVA) is an efficient method
for the analysis of complicated Raman spectra, to identify molecular
features. Here, we utilized Raman microspectroscopy to identify the
molecular information of each diatom cell, at the single-cell level.
The MVA, together with a support vector machine, which is a machine
learning technique, allowed the classification of proliferating and
nonproliferating cells. The classification includes polyunsaturated
fatty acids such as linoleic acid, eicosapentaenoic acid, and docosahexaenoic
acid. This study indicated that Raman spectroscopy is an appropriate
technique to examine C. tenuissimus at the single-cell level, providing relevant data to assess the
correlation between the molecular details obtained from the Raman
analysis, at each growth phase.