Bacterial proteomes are predictable from cellular Raman spectra*

Whole-cell Raman spectra reflect its molecular composition and allow us to distinguish cellular states in a label-free and non-destructive manner. Previously, we showed that in S. pombe and E. coli, cellular Raman spectra and transcriptomes were linked quantitatively and that the transcriptomes could be inferred from Raman spectra (Kobayashi-Kirschvink et al., Cell Systems, 2018). We now ask whether Raman spectra could also be connected to other types of omics information. Using quantitative proteome data of E. coli cultured under various conditions (Schmidt et al., Nature Biotechnology, 2016), we show that proteomes can be linked to and reconstructed from cellular Raman spectra. These results suggest that cellular Raman spectra have the potential to unravel and integrate multi-omic states of cells non-destructively, and therefore show great promise in establishing live-cell omics.