Foundation model for mass spectrometry proteomics

preprint · arXiv · 2025

preprint · arXiv · 2025. Justin Sanders et al. Mass spectrometry is the dominant technology in the field of proteomics, enabling high-throughput analysis of…
Date 2025-05-16
Type preprint
Venue arXiv
Publisher arXiv
Contribution algorithm
DOI 10.48550/arXiv.2505.10848
Citations (OpenAlex) 3

Abstract

Mass spectrometry is the dominant technology in the field of proteomics, enabling high-throughput analysis of the protein content of complex biological samples. Due to the complexity of the instrumentation and resulting data, sophisticated computational methods are required for the processing and interpretation of acquired mass spectra. Machine learning has shown great promise to improve the analysis of mass spectrometry data, with numerous purpose-built methods for improving specific steps in the data acquisition and analysis pipeline reaching widespread adoption. Here, we propose unifying various spectrum prediction tasks under a single foundation model for mass spectra. To this end, we pre-train a spectrum encoder using de novo sequencing as a pre-training task. We then show that using these pre-trained spectrum representations improves our performance on the four downstream tasks of spectrum quality prediction, chimericity prediction, phosphorylation prediction, and glycosylation status prediction. Finally, we perform multi-task fine-tuning and find that this approach improves the performance on each task individually. Overall, our work demonstrates that a foundation model for tandem mass spectrometry proteomics trained on de novo sequencing learns generalizable representations of spectra, improves performance on downstream tasks where training data is limited, and can ultimately enhance data acquisition and analysis in proteomics experiments.

Authors

  1. Justin Sanders · University of Washington
  2. Melih Yilmaz · University of Washington
  3. Jacob H. Russell · University of Washington
  4. Wout Bittremieux · Biomedical Informatics Research Center Antwerp, Indiana University, University of Antwerp, University of California San Diego, University of Washington
  5. William E. Fondrie · Talus Bioscience
  6. Nicholas M. Riley · University of Washington
  7. Sewoong Oh · University of Washington
  8. William Stafford Noble · University of Washington

Methods and tools

  • Casanovo Foundation: Foundation model for tandem-MS proteomics: pre-trained the Casanovo spectrum encoder on 30M labelled spectra from MassIVE-KB, then reused the encoder off-the-shelf for downstream tasks (de novo sequencing, spectrum quality, chimericity, phosphorylation, glycosylation prediction). Successor in spirit to Casanovo v1/v2/v5.

Data used

  • 2018CHPP (as deposited) · MSV000083978
  • A deep proteome and transcriptome abundance atlas of 29 healthy human tissues (MSV000083508) (as deposited) · MSV000083508
  • CHPP chr 1,8,20 proteome dataset using the HCC cell lines of Hep3B and MHCC97H, instrument is Q Exac (MSV000080254) (as deposited) · MSV000080254
  • CHPP chr 1,8,20 proteome dataset using the HCC cell lines of MHCC97H and HCCLM3, instrument is Q Exa (MSV000080255) (as deposited) · MSV000080255
  • Chr16-HPP. Shotgun Analysis improvement. JPR HPP Special issue 2013. CCD18 Cell line. (as deposited) · MSV000083961
  • Chr16-HPP. Shotgun Analysis improvement. JPR HPP Special issue 2013. Jurkat Cell line. (as deposited) · MSV000083966
  • Chr16-HPP. Shotgun Analysis improvement. JPR HPP Special issue 2013. Ramos Cell line. (as deposited) · MSV000083967
  • Chr18 Consortium MS-data for missing proteins mining (as deposited) · MSV000086439
  • Confetti: A Multi-protease Map of the HeLa Proteome for Comprehensive Proteomics (MSV000081607) (as deposited) · MSV000081607
  • Deep quantitative glycoproteomics reveals gut microbiome induced remodeling of the brain glycoproteome (as deposited) · PXD052447
  • HeLa proteome of 12,250 protein-coding genes (MSV000081563) (as deposited) · MSV000081563
  • Human bladder,colon,kidney,liver cancer LC MS/MS (as deposited) · MSV000081649
  • Human placental tissue LC-MS/MS (as deposited) · MSV000086385
  • Human testis off-line LC-MS/MS (as deposited) · MSV000086491
  • Human_Testis_QE-HF (as deposited) · MSV000083983
  • Quantitative PTM Maps of Human Pathologic Tau Identify Patient Heterogeneity and Define Critical Steps in Alzheimer’s Di (as deposited) · MSV000088405
  • Spatially and cell-type resolved quantitative proteomic atlas of healthy human skin (as deposited) · MSV000088236
  • Synaptosomal proteome from schizophrenia patients (as deposited) · MSV000086389
  • The exploration of Missing Proteins by a combination approach to enrich the low abundance membrane proteins from several (as deposited) · MSV000086369
  • The human phosphoproteome map based on PRIDE data (as deposited) · PXD012174
  • Update on virtual-experimental 2DE approach in chromosome-centric human proteome project (as deposited) · MSV000086448

Cites (7)

Cited by (1)

Seen in the charts

Back to the full map

Back to top