Skip to content
DraconiticResearch
Search Sources API
203,952 studies31 sourcesUpdated 51 min ago
Peer-reviewedCC BY-NC-ND 4.0Nature Methods

MassNet: billion-scale AI-ready mass spectral corpus empowering scalable deep learning in proteomics

  • Jun A1,2,3
  • Xiang Zhang4,5
  • Xiaofan Zhang1,2,3
  • Jiaqi Wei6,7
  • Te Zhang1,2,3
  • Yamin Deng1,2,3
  • Pu Liu8
  • Zongxiang Nie2
  • Yiwen Yu1,9
  • Yi Chen2
  • Nanqing Dong6,10
  • Zhiqiang Gao6
2 more authors
  • Siqi Sun4,6
  • Tiannan Guo1,2,3
10 affiliations
  1. 1Affiliated Hangzhou First People’s Hospital, State Key Laboratory of Medical Proteomics, School of Medicine, Westlake University, Hangzhou, China
  2. 2Westlake Center for Intelligent Proteomics, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, China
  3. 3Research Center for Industries of the Future, School of Life Sciences, Westlake University, Hangzhou, China
  4. 4Research Institute of Intelligent Complex Systems, Fudan University, Shanghai, China
  5. 5University of British Columbia, Vancouver, British Columbia Canada
  6. 6Shanghai Artificial Intelligence Laboratory, Shanghai, China
  7. 7College of Computer Science and Technology, Zhejiang University, Hangzhou, China
  8. 8Westlake Omics Biotechnology Co. Ltd, Hangzhou, China
  9. 9Suzhou Medical College, Soochow University, Suzhou, China
  10. 10Shanghai Innovation Institute, Shanghai, China
Read full textPDFDOICite

Abstract

Here we present MassNet, inspired by large-scale data corpora of text and images; MassNet is an artificial intelligence (AI)-ready data corpus for mass spectrometry (MS)-based proteomics. The corpus contains 1.5 billion tandem MS spectra from high-quality data-dependent acquisition (DDA) MS data across 35 species. Approximately 1.7 million peptide precursors are from human samples. We further present an AI-ready DDA data format called Mass Spectrometry DDA Tensor (MSDT) for compact and standardized storage, curation and reuse of data from multiple sources, which can be optionally converted to a Lightning Memory-Mapped Database for accelerated input/output in deep learning. On the basis of MassNet, we develop XuanjiNovo, a nonautoregressive transformer model that leverages a curriculum learning strategy for de novo peptide sequencing. XuanjiNovo achieves 38.35–143.57% improvement over state-of-the-art methods on human Astral data. MassNet also supports additional downstream tasks, including retention time prediction and PSM rescoring. MassNet represents a foundational MS-based proteomics data resource, marking a critical milestone for AI proteomics research.

Related

  1. Making the Impossible Possible: Uniting TMT and DIA for Scalable Low-Input-Proteomics
    T. Thierer, M. J. Birklbauer, V. Dorfer +2
    bioRxivPreprintCC BYFull textPhysical chemistrybioRxiv
  2. Tesorai Search: Cloud-based Database Search Engine Boosts Identifications for Mass Spectrometry Proteomics With a Pretrained Peptide-spectrum Deep-learning Model
    Maximilien Burq, Dejan Štepec, Juan Restrepo +8
    Journal of Molecular BiologyPeer-reviewedLicence not stated8 citedPhysical chemistryOpenAlex
  3. A Lightweight Universal Machine-Learning Interatomic Potential via Knowledge Distillation for Scalable Atomistic Simulations
    Sangmin Oh, J.H. You, Jaesun Kim +4
    Journal of Chemical Information and ModelingPeer-reviewedLicence not statedFull text9 citedMaterials & polymersOpenAlex
  4. AI and Machine Learning for Proteomics-Driven Drug Discovery: Methods, Tools, and Best Practices
    Suman Basak
    Current Issues in Molecular BiologyPeer-reviewedCC BY11 citedPhysical chemistryOpenAlex
  5. Integrated proteomics and machine learning for identifying candidate serum biomarkers in acute myocardial infarction-complicated cardiogenic shock: a prospective exploratory study
    Xi Wang, Qian-feng Xiao, Fang-yang Huang +6
    Clinical ProteomicsPeer-reviewedCC BY-NC-ND 4.0Full textPhysical chemistryPMC

Cite

BibTeXRIS
BibTeX
@article{jun2026massnet,
  title = {{MassNet: billion-scale AI-ready mass spectral corpus empowering scalable deep learning in proteomics}},
  author = {Jun A and Xiang Zhang and Xiaofan Zhang and Jiaqi Wei and Te Zhang and Yamin Deng and Pu Liu and Zongxiang Nie and Yiwen Yu and Yi Chen and Nanqing Dong and Zhiqiang Gao and Siqi Sun and Tiannan Guo},
  journal = {Nature Methods},
  year = {2026},
  month = oct,
  doi = {10.1038/s41592-026-03247-y},
  url = {https://doi.org/10.1038/s41592-026-03247-y},
  note = {CC BY-NC-ND 4.0},
}
RIS
TY  - JOUR
TI  - MassNet: billion-scale AI-ready mass spectral corpus empowering scalable deep learning in proteomics
AU  - Jun A
AU  - Xiang Zhang
AU  - Xiaofan Zhang
AU  - Jiaqi Wei
AU  - Te Zhang
AU  - Yamin Deng
AU  - Pu Liu
AU  - Zongxiang Nie
AU  - Yiwen Yu
AU  - Yi Chen
AU  - Nanqing Dong
AU  - Zhiqiang Gao
AU  - Siqi Sun
AU  - Tiannan Guo
PY  - 2026
DA  - 2026/10/14
JO  - Nature Methods
DO  - 10.1038/s41592-026-03247-y
UR  - https://doi.org/10.1038/s41592-026-03247-y
AB  - Here we present MassNet, inspired by large-scale data corpora of text and images; MassNet is an artificial intelligence (AI)-ready data corpus for mass spectrometry (MS)-based proteomics. The corpus contains 1.5 billion tandem MS spectra from high-quality data-dependent acquisition (DDA) MS data across 35 species. Approximately 1.7 million peptide precursors are from human samples. We further present an AI-ready DDA data format called Mass Spectrometry DDA Tensor (MSDT) for compact and standardized storage, curation and reuse of data from multiple sources, which can be optionally converted to a Lightning Memory-Mapped Database for accelerated input/output in deep learning. On the basis of MassNet, we develop XuanjiNovo, a nonautoregressive transformer model that leverages a curriculum learning strategy for de novo peptide sequencing. XuanjiNovo achieves 38.35–143.57% improvement over state-of-the-art methods on human Astral data. MassNet also supports additional downstream tasks, including retention time prediction and PSM rescoring. MassNet represents a foundational MS-based proteomics data resource, marking a critical milestone for AI proteomics research.
SN  - 1548-7091
SN  - 1548-7105
ER  -