Mass spectrometry-based proteomics faces a significant throughput bottleneck, with current methods analysing fewer than 200 samples per day despite growing demand for large-scale studies including single-cell proteomics. This limitation stems from overlapping mass spectral signals that occur when multiple samples are analysed in parallel using data-independent acquisition (DIA) methods. Existing software tools struggle with this complexity, particularly when using mass tags with small mass offsets.

To overcome this challenge, McDonnell et al. developed JMod, an open-source software platform designed to model and deconvolve overlapping mass spectral peaks, thereby enabling increased sample multiplexing in DIA proteomics. Their experimental approach used mixed human and yeast protein digests at single-cell input levels (200 pg human, 80 pg yeast) analysed on a Vanquish Neo coupled to an Orbitrap Exploris 480. IonOpticks Aurora Ultimate™ 25 cm×75 μm C18 UHPLC columns were used for peptides separation. The team from the Slavov lab tested PSMtag and diethylation labelling strategies with channels separated by both 4 Da and 2 Da mass offsets.

This quantitative study demonstrated that JMod enables analysis of 9-plex samples with 2 Da channel spacing, nearly doubling throughput compared to conventional 5-plex approaches with 4 Da spacing. The platform achieved similar quantitative accuracy for both spacing schemes whilst maintaining protein identification rates that scaled proportionally with plex size. JMod’s linear superposition modelling approach successfully deconvolved overlapping isotopic envelopes, producing approximately 30,000 protein data points from nine single-cell level samples in a single MS run.

This advancement addresses a critical scalability challenge in proteomics, potentially enabling population-scale studies and high-throughput single-cell analyses essential for biomedical research applications.

Publication
bioRxiv

Authors

Kevin McDonnell, Nathan Wamsley, Jason Derks, Sarah Sipe, Maddy Yeh, Harrison Specht, & Nikolai Slavov;

Title

JMod: Joint modeling of mass spectra for empowering multiplexed DIA proteomics