Live Electrophoresis-Correlative AI Democratizes Single-cell Mass Spectrometry Proteomics

The goal of this project was to deepen proteome coverage on broadly accessible mass spectrometers. We developed electrophoresis-correlative (Eco) ion sorting mass spectrometry (MS) on a decade-old quadrupole orbitrap mass spectrometer (Q Exactive Plus, Thermo). Artificial intelligence (AI) was adapted (CHIMERYS, Thermo) to process the chimeric tandem mass spectra . The strategy was tested, configured, and validated on 1 ng to 250 pg of the HeLa proteome digest. Within a 15-min effective separation window, 1,758 proteins were identified cumulatively from single-cell equivalent proteome amounts (~250 pg). As demonstration, Eco-AI was deployed to profile the proteomic state of dorsal and ventral cell lineages in blastula stage (stage 8) Xenopus laevis embryos. 1,524 proteins were identified from n = 16 single cells. The quantitative protein profiles revealed emerging cell heterogeneity during differentiation.