2026-10-14T05:32:39Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33676305 2026-09-12T18:07:31Z category_1 category_4 category_146 category_16 category_873 category_39 category_46 category_734 portal_63 item_type_3 month_year_09_2026

Rapid Eco-AI CE–MS Enables Sub‑7 Min Low-Input Proteome Profiling with Quantitative Validation Fei Zhou (86034) Bowen Shen (2831456) Isabelle S. Luz (6020573) Laura G. Rodriguez (23826906) Steven J. Prior (13569983) Wagner Fontes (573686) Peter Nemes (1289418) Biophysics Biochemistry Space Science Physiology Chemical Sciences not elsewhere classified Ecology Immunology Biological Sciences not elsewhere classified supported robust identification sensitivity remains challenging scarce primary cells measured mean b maintaining substantial peptide legacy orbitrap platform effective separation window conservative background filtering balancing analytical speed 7 min separation digest per run derived digest analysis 11 min injection ∼ 2 pg input proteome profiling sustained high peptide identified 135 proteins ce – ms 2 </ sup quantitative validation low quantitative validation proteome coverage hela digest rapid ce injection runtimes cumulatively identified quantitative precision 835 proteins 75 pg 347 proteins 300 pg z </ time tracking time correlations principle transfer principle application optimized ms optimized electrophoresis median coefficients match conversion mass spectrometry free quantification expected values chimeric ms assisted interpretation artificial intelligence advanced rapidly Low-input and single-cell proteomics by mass spectrometry (MS) have advanced rapidly, but balancing analytical speed and sensitivity remains challenging, especially for small, protein-scarce primary cells such as neutrophils. Here, we report Rapid Eco–AI, an optimized electrophoresis-correlative (Eco) capillary electrophoresis–mass spectrometry (CE–MS) workflow with artificial intelligence (AI)-assisted data processing for fast, high-efficiency, low-input proteome profiling. Relative to our 15 min Eco–AI method, Rapid Eco–AI compresses the effective separation window to 5–7 min while maintaining substantial peptide and proteome coverage, efficient precursor ion sampling, and quantitative precision. Real-time tracking of charge-dependent m/z and migration-time correlations, together with an optimized MS2 strategy, sustained high peptide-spectrum-match conversion under rapid CE, while AI-assisted interpretation of chimeric MS2 spectra (CHIMERYS) supported robust identification and label-free quantification. Implemented on a legacy Orbitrap platform (Q Exactive Plus), Rapid Eco–AI cumulatively identified 1,634 and 1,347 proteins from 300 pg of single-cell-equivalent HeLa digest across technical triplicates using 7 and 5 min effective separations, corresponding to <15 and <11 min injection-to-injection runtimes, respectively. Using the 7 min separation, 835 proteins were cumulatively identified from 75 pg of HeLa digest. Hybrid HeLa–yeast proteome mixtures validated quantitative performance, with median coefficients of variation <5% and measured mean B/A ratios of 0.57 and 1.60 for expected values of 0.60 and 1.80, respectively. As a proof-of-principle application, we analyzed fractional digest loads from 13 freshly isolated single human neutrophils, corresponding to ∼2 pg of digest per run on a manual CE–MS platform, and identified 135 proteins after conservative background filtering. Collectively, these results establish Rapid Eco–AI as a fast CE–MS workflow for low-input proteome profiling with quantitative validation and proof-of-principle transfer to neutrophil-derived digest analysis. 2026-09-12T00:00:00Z Dataset Dataset 10.1021/acs.analchem.6c03046.s001 https://figshare.com/articles/dataset/Rapid_Eco-AI_CE_MS_Enables_Sub_7_Min_Low-Input_Proteome_Profiling_with_Quantitative_Validation/33676305 CC BY-NC 4.0