Skip to content
DraconiticResearch
Search Sources API
186,125 studies31 sourcesUpdated 51 min ago
Peer-reviewedCC BYNature

The Virtual Tissues foundation model resolves spatial proteomics across scales

  • Johann Wenckstern1
  • Eeshaan Jain1
  • Benedikt von Querfurth1
  • Yexiang Cheng1
  • Kiril Vasilev2
  • Matteo Pariset2
  • Phil Fang Cheng3,4
  • Petros Liakopoulos3,4
  • Olivier Alain Michielin3,4
  • Andreas N. WICKI5,6
  • Gabriele Gut5,6
  • Charlotte Bunne7,8
8 affiliations
  1. 1School of Computer and Communication Sciences, EPFL, Lausanne, Switzerland
  2. 2Department of Computer Science, ETH, Zurich, Switzerland
  3. 3Department of Oncology, Geneva University Hospitals, Geneva, Switzerland
  4. 4Faculty of Medicine, University of Geneva, Geneva, Switzerland
  5. 5Department of Medical Oncology and Hematology, University Hospital Zurich, Zurich, Switzerland
  6. 6University of Zurich, Faculty of Medicine, Zurich, Switzerland
  7. 7School of Computer and Communication Sciences, EPFL, Lausanne, Switzerland. [email protected]
  8. 8Swiss Institute for Experimental Cancer Research, School of Life Sciences, EPFL, Lausanne, Switzerland. [email protected]
Open at sourceDOICite

Abstract

Spatial proteomics technologies have transformed our understanding of complex tissue architecture in cancer but present unique challenges for computational analysis 1 . Each study uses a different marker panel and protocol, and most methods are tailored to single cohorts, which limits knowledge transfer and robust biomarker discovery. Here we present Virtual Tissues (VirTues), a general-purpose foundation model for spatial proteomics that learns marker-aware, multi-scale representations of proteins, cells, niches and tissues directly from multiplex imaging data. From a single pretrained backbone, VirTues supports marker reconstruction, cell segmentation and typing, niche annotation, spatial biomarker discovery and patient stratification, including zero-shot annotation across heterogeneous panels and datasets. In triple-negative breast cancer, VirTues-derived biomarkers predict anti-PD-L1 chemo-immunotherapy response 2 and stratify disease-free survival in an independent cohort 3 , outperforming state-of-the-art biomarkers derived from the same datasets and current clinical stratification schemes.

Related

  1. Integrated Spatial Metabolomics and Proteomics from the Same Tissue Section Using a Conductive ITO-PET Slide
    h. wang, Y. Li
    bioRxivPreprintAll rights reservedFull textMolecular biology & gene regulationbioRxiv
  2. Characterization of spatial homogeneous regions in tissues with concordex
    Kayla C. Jackson, A. Sina Booeshaghi, Ángel Gálvez-Merchán +5
    BMC MethodsPeer-reviewedCC BYFull text8 citedMolecular biology & gene regulationOpenAlex
  3. Whole-transcriptome-scale isoform-resolved spatial imaging of single cells in tissues
    Limor Cohen, Aaron R. Halpern, Timothy R. Blosser +3
    CellPeer-reviewedCC BY15 citedMolecular biology & gene regulationOpenAlex
  4. Temporal and spatial atlas of eosinophil specialization across tissues
    Yanan Hu, Lianfu Wu, Sihao Qu +14
    Nature ImmunologyPeer-reviewedCC BY 4.0Full text15 citedProteomics & metabolomicsCell biologyHAL
  5. Spatial clustering of adhesion-deficient cells delaysclearance and prolongs geometric unjammingsignatures in epithelial tissues
    V. Manso, P. Guerrero, N. Brinas-Pascual
    bioRxivPreprintCC BY-NC-NDFull textBiophysicsbioRxiv

Cite

BibTeXRIS
BibTeX
@article{wenckstern2026virtual,
  title = {{The Virtual Tissues foundation model resolves spatial proteomics across scales}},
  author = {Johann Wenckstern and Eeshaan Jain and Benedikt von Querfurth and Yexiang Cheng and Kiril Vasilev and Matteo Pariset and Phil Fang Cheng and Petros Liakopoulos and Olivier Alain Michielin and Andreas N. WICKI and Gabriele Gut and Charlotte Bunne},
  journal = {Nature},
  year = {2026},
  month = aug,
  doi = {10.1038/s41586-026-10884-y},
  url = {https://doi.org/10.1038/s41586-026-10884-y},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - The Virtual Tissues foundation model resolves spatial proteomics across scales
AU  - Johann Wenckstern
AU  - Eeshaan Jain
AU  - Benedikt von Querfurth
AU  - Yexiang Cheng
AU  - Kiril Vasilev
AU  - Matteo Pariset
AU  - Phil Fang Cheng
AU  - Petros Liakopoulos
AU  - Olivier Alain Michielin
AU  - Andreas N. WICKI
AU  - Gabriele Gut
AU  - Charlotte Bunne
PY  - 2026
DA  - 2026/08/05
JO  - Nature
DO  - 10.1038/s41586-026-10884-y
UR  - https://doi.org/10.1038/s41586-026-10884-y
AB  - Spatial proteomics technologies have transformed our understanding of complex tissue architecture in cancer but present unique challenges for computational analysis 1 . Each study uses a different marker panel and protocol, and most methods are tailored to single cohorts, which limits knowledge transfer and robust biomarker discovery. Here we present Virtual Tissues (VirTues), a general-purpose foundation model for spatial proteomics that learns marker-aware, multi-scale representations of proteins, cells, niches and tissues directly from multiplex imaging data. From a single pretrained backbone, VirTues supports marker reconstruction, cell segmentation and typing, niche annotation, spatial biomarker discovery and patient stratification, including zero-shot annotation across heterogeneous panels and datasets. In triple-negative breast cancer, VirTues-derived biomarkers predict anti-PD-L1 chemo-immunotherapy response 2 and stratify disease-free survival in an independent cohort 3 , outperforming state-of-the-art biomarkers derived from the same datasets and current clinical stratification schemes.
SN  - 0028-0836
SN  - 1476-4687
ER  -