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

peer-reviewed · Nature Methods · 2026

peer-reviewed · Nature Methods · 2026. Jun A et al. Here we present MassNet, inspired by large-scale data corpora of text and images; MassNet is an artificial…
Date 2026-10-14
Type peer-reviewed
Venue Nature Methods
Contribution algorithm
DOI 10.1038/s41592-026-03247-y
Citations (OpenAlex) 0
Venue 2-year citedness 22.29

Preprint version: MassNet: billion-scale AI-friendly mass spectral corpus enables robust de novo peptide sequencing (2025-06-20, bioRxiv)

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.

Authors

  1. Jun A · Westlake University
  2. Xiang Zhang (Shanghai AI Lab) · Fudan University, Shanghai Artificial Intelligence Laboratory, University of British Columbia
  3. Xiaofan Zhang · Westlake University
  4. Jiaqi Wei · Shanghai Artificial Intelligence Laboratory, Zhejiang University
  5. Te Zhang · Westlake University
  6. Yamin Deng · Westlake University
  7. Pu Liu · Westlake Omics (Hangzhou) Biotechnology Co., Ltd.
  8. Zongxiang Nie · Westlake University
  9. Yiwen Yu · Soochow University, Westlake University
  10. Yi Chen · Westlake University
  11. Nanqing Dong · Shanghai Artificial Intelligence Laboratory, Shanghai Innovation Institute
  12. Zhiqiang Gao · Shanghai Artificial Intelligence Laboratory
  13. Siqi Sun · Fudan University, Shanghai Artificial Intelligence Laboratory
  14. Tiannan Guo · Westlake Institute for Advanced Study, Westlake University

Methods and tools

Data deposited

Data used

  • MassIVE-KB (version not stated)
  • Nine-species benchmark (original (DeepNovo, 2017)) · MSV000081382, PXD003868 (provenance), PXD004325 (provenance), PXD004424 (provenance), PXD004467 (provenance), PXD004536 (provenance), PXD004565 (provenance), PXD004947 (provenance), PXD004948 (provenance), PXD005025 (provenance)
  • ProteomeTools (version not stated)
  • Proteomics identifies new therapeutic targets of early-stage hepatocellular carcinoma (as deposited) · IPX0000937000
  • Ultra-fast label-free quantification and comprehensive proteome coverage with narrow-window data-independent acquisition (as deposited) · PXD046453

Cites (27)

Seen in the charts

Back to the full map

Back to top