We're excited to share the publication of our latest work in the Journal of Proteome Research:
"SWAPS: A Modular Deep-Learning Empowered Peptide Identity Propagation Framework Beyond Match-Between-Run"
Read the full article here

This study is part of the ORIGIN grant. In this study, we present SWAPS, a novel MS1-centric peptide identity propagation (PIP) framework that moves beyond the limitations of traditional match-between-runs (MBR) strategies. Unlike conventional methods, which are often confined to highly similar experimental conditions, SWAPS leverages recent advances in peptide property prediction, deep learning, and large-scale proteomics libraries to enable robust identification across diverse LC gradient lengths and experimental setups.

Applied to LC gradients of 30, 15, and 7.5 minutes, SWAPS boosts precursor-level identifications by 46.3%, 86.2%, and 112.1%, respectively, compared to standard MS2-based approaches. Importantly, it achieves this while maintaining high quantitative accuracy and offering new insights into the performance of current predictive models.

With its modular design, SWAPS is built for future development—ready to incorporate improved predictors and enable even deeper sequence-level proteome analysis.

This work pushes the boundaries of MS1-based quantification and opens the door for more flexible, accurate, and scalable proteomic workflows.

Check out the git repo here!

This is the beginning of our MS1-centric proteomics workflow, protein-level inferences and optimization from acquisition method will follow.