π-SPECFormer: A Mass Spectrometry-based Foundation Model for Decoding the Dark Proteome

Abstract

In mass-spectrometry-based proteomics, a substantial proportion of acquired spectra remain unidentified, concealing a vast reservoir of biological variations commonly termed the "dark proteome", potentially obscuring sequence mutations, non-canonical peptides and post-translational modifications. Existing deep-learning methods largely rely on supervised training with peptide-spectrum matches (PSMs), limiting their ability to learn from the much larger pool of unannotated spectra. Here we introduce {pi}-SPECFormer, a mass-spectrometry foundation model pre-trained on 291 million spectra through a two-stage framework that combines self-supervised and PSMs-guided representation learning. {pi}-SPECFormer transfers effectively to de novo peptide sequencing, spectrum clustering and modification detection, consistently outperforming task-specific baselines. When applied to proteomic profiles from 103 individuals with lung adenocarcinoma, the model identified nine high-confidence mutated peptides, including six that were absent from the reference databases. {pi}-SPECFormer thus enables unbiased decoding of molecular complexity beyond the boundaries of the canonical proteome.

Proteome-wide crosslinking mass spectrometry reveals novel components of essential complexes in Toxoplasma
S. Butterworth, A. L. Gin, S. Shikha +10
bioRxiv1 Sep 2026PreprintCC BY-NCFull textMicrobiologybioRxiv
Coupling of oceanographic state to the dark proteome: a foundation for genome-informed marine productivity modeling
D. R. Nelson, M. Plouviez, S. Daakour +4
bioRxiv23 Sep 2026PreprintCC BYFull textBioinformatics & computational biologybioRxiv
Limitations of Genomic Foundation Models for Decoding Regulatory Mechanisms in ALS
A. Talukder, A. Kaplan
bioRxiv1 Oct 2026PreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
Detection of Carbapenemase-Producing Bacteria by Using an Ultra-Performance Liquid Chromatography-Tandem Mass Spectrometry Method
Anne Carricajo, Paul Verhoeven, Salim Guezzou +2
Antimicrobial Agents and Chemotherapy5 Oct 2026Peer-reviewedLicence not statedFull text18 citedCell biology MicrobiologyHAL
Brain Landscape In Situ Crosslinking Mass Spectrometry (BLIS-XL-MS) Enables Global Analysis of Protein Structural Remodeling in the Brain
Y.-Z. Wang, J. Xu, A. Contractor +1
bioRxiv16 Sep 2026PreprintAll rights reservedFull textNeurosciencebioRxiv

Cite

BibTeX RIS

BibTeX

@misc{ling2026πspecformer,
  title = {{π-SPECFormer: A Mass Spectrometry-based Foundation Model for Decoding the Dark Proteome}},
  author = {T. Ling and T. Liu and F. Xu and Y. Zhang and J. Xia and X. Huang and S. Chen and L. Xie and R. Xu and T. Yang and K. Li and G. Wang and Z. Liang and J. Zhu and S. Z. Li and F. He and Y. Wang and C. Chang},
  howpublished = {bioRxiv},
  year = {2026},
  month = nov,
  doi = {10.64898/2026.11.06.764812},
  url = {https://www.biorxiv.org/content/10.64898/2026.11.06.764812v1},
  note = {CC BY-NC-ND},
}

RIS

TY  - UNPB
TI  - π-SPECFormer: A Mass Spectrometry-based Foundation Model for Decoding the Dark Proteome
AU  - T. Ling
AU  - T. Liu
AU  - F. Xu
AU  - Y. Zhang
AU  - J. Xia
AU  - X. Huang
AU  - S. Chen
AU  - L. Xie
AU  - R. Xu
AU  - T. Yang
AU  - K. Li
AU  - G. Wang
AU  - Z. Liang
AU  - J. Zhu
AU  - S. Z. Li
AU  - F. He
AU  - Y. Wang
AU  - C. Chang
PY  - 2026
DA  - 2026/11/06
PB  - bioRxiv
DO  - 10.64898/2026.11.06.764812
UR  - https://www.biorxiv.org/content/10.64898/2026.11.06.764812v1
AB  - In mass-spectrometry-based proteomics, a substantial proportion of acquired spectra remain unidentified, concealing a vast reservoir of biological variations commonly termed the "dark proteome", potentially obscuring sequence mutations, non-canonical peptides and post-translational modifications. Existing deep-learning methods largely rely on supervised training with peptide-spectrum matches (PSMs), limiting their ability to learn from the much larger pool of unannotated spectra. Here we introduce {pi}-SPECFormer, a mass-spectrometry foundation model pre-trained on 291 million spectra through a two-stage framework that combines self-supervised and PSMs-guided representation learning. {pi}-SPECFormer transfers effectively to de novo peptide sequencing, spectrum clustering and modification detection, consistently outperforming task-specific baselines. When applied to proteomic profiles from 103 individuals with lung adenocarcinoma, the model identified nine high-confidence mutated peptides, including six that were absent from the reference databases. {pi}-SPECFormer thus enables unbiased decoding of molecular complexity beyond the boundaries of the canonical proteome.
ER  -

Abstract

Related

Cite