Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 53 min ago
Peer-reviewedCC BY-NC-ND 4.0Nature Methods

Accurate and well-powered case–control analysis of spatial molecular data

  • Yakir A. Reshef1,2,3,4
  • Lakshay Sood1,2,3,4
  • Michelle Curtis1,2,3,4
  • Laurie Rumker1,2,3,4,5
  • Daniel J. Stein1,2,3,4,5
  • Mukta G. Palshikar1,2,3,4
  • Saba Nayar6,7
  • Andrew Filer6
  • Anna Helena Jonsson8
  • Ilya Korsunsky1,2,3,4
  • Soumya Raychaudhuri1,2,3,4,5
8 affiliations
  1. 1Center for Data Sciences, Brigham and Women’s Hospital, Boston, MA USA
  2. 2Division of Rheumatology, Inflammation, and Immunity, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA USA
  3. 3Broad Institute of MIT and Harvard, Cambridge, MA USA
  4. 4Division of Genetics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA USA
  5. 5Department of Biomedical Informatics, Harvard Medical School, Boston, MA USA
  6. 6NIHR Birmingham Biomedical Research Centre, University Hospitals Birmingham NHS Foundation Trust and Department of Inflammation and Ageing, College of Medicine & Health, University of Birmingham, Birmingham, UK
  7. 7Birmingham Tissue Analytics, College of Medicine and Health, University of Birmingham, Birmingham, UK
  8. 8University of Colorado Anschutz Medical Campus, Division of Rheumatology, Aurora, CO USA
Read full textPDFDOICite

Abstract

As spatial molecular data grow in scope, there is a pressing need to identify disease-associated spatial structures. Current approaches typically make restrictive assumptions such as representing tissue regions by abundances of discrete cell types and samples by abundances of discrete niches; this risks overlooking important signals. Here we introduce variational inference-based microniche analysis (VIMA), a method combining deep learning with principled statistics to discover disease-associated spatial features with greater flexibility and precision. VIMA trains an ensemble of variational autoencoders to summarize the contents of every small tissue patch in a dataset via numeric ‘fingerprints’. It uses these to define many data-dependent, overlapping ‘microniches’ and meta-analyzes them to identify microniches whose abundance correlates significantly with case−control status. We confirm VIMA’s calibration, power and spatial accuracy in simulations. We then apply VIMA to spatial datasets spanning three diseases and spatial modalities, recapitulating known biology and identifying new spatial features of disease.

Related

  1. SpaMOAL is a deep learning method that enables accurate spatial domain identification from multi-omics data
    Jinxia Wang, Yuying Huo, Rui Zhao +4
    PLOS BiologyPeer-reviewedCC BY 4.0Full textMolecular biology & gene regulationGenetics & genomicsPLOS
  2. AutoHMD: scalable and accurate validation of antibody-antigen complexes by heated molecular dynamics.
    Julia S Souza, Aline O Albuquerque, Diego S Almeida +3
    Bioinformatics (Oxford, England)Peer-reviewedCC BYFull textMolecular biology & gene regulationEurope PMCCrossref
  3. Lipid-organized tetramerization of PLPP1 defines the molecular basis of spatial lipid phosphate signaling
    Yunhai Li, Qingning Yuan, Zhenyu Cao +11
    Nature CommunicationsPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
  4. Molecular phenotypes and spatial archetypes: A new framework for cancer-associated fibroblasts
    Yunhe Liu, Xiongfeng Chen, Yibo Dai +14
    Cancer CellPeer-reviewedCC BY-NC-ND19 citedMolecular biology & gene regulationOpenAlex
  5. Spatial transcriptomics unveils the in situ cellular and molecular hallmarks of the lung in fatal COVID-19
    Carlos A. Garcia-Prieto, Eva Musulen, Daniela Grases +9
    Nature CommunicationsPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossrefEurope PMC

Cite

BibTeXRIS
BibTeX
@article{reshef2026accurate,
  title = {{Accurate and well-powered case–control analysis of spatial molecular data}},
  author = {Yakir A. Reshef and Lakshay Sood and Michelle Curtis and Laurie Rumker and Daniel J. Stein and Mukta G. Palshikar and Saba Nayar and Andrew Filer and Anna Helena Jonsson and Ilya Korsunsky and Soumya Raychaudhuri},
  journal = {Nature Methods},
  year = {2026},
  month = oct,
  doi = {10.1038/s41592-026-03236-1},
  url = {https://doi.org/10.1038/s41592-026-03236-1},
  note = {CC BY-NC-ND 4.0},
}
RIS
TY  - JOUR
TI  - Accurate and well-powered case–control analysis of spatial molecular data
AU  - Yakir A. Reshef
AU  - Lakshay Sood
AU  - Michelle Curtis
AU  - Laurie Rumker
AU  - Daniel J. Stein
AU  - Mukta G. Palshikar
AU  - Saba Nayar
AU  - Andrew Filer
AU  - Anna Helena Jonsson
AU  - Ilya Korsunsky
AU  - Soumya Raychaudhuri
PY  - 2026
DA  - 2026/10/06
JO  - Nature Methods
DO  - 10.1038/s41592-026-03236-1
UR  - https://doi.org/10.1038/s41592-026-03236-1
AB  - As spatial molecular data grow in scope, there is a pressing need to identify disease-associated spatial structures. Current approaches typically make restrictive assumptions such as representing tissue regions by abundances of discrete cell types and samples by abundances of discrete niches; this risks overlooking important signals. Here we introduce variational inference-based microniche analysis (VIMA), a method combining deep learning with principled statistics to discover disease-associated spatial features with greater flexibility and precision. VIMA trains an ensemble of variational autoencoders to summarize the contents of every small tissue patch in a dataset via numeric ‘fingerprints’. It uses these to define many data-dependent, overlapping ‘microniches’ and meta-analyzes them to identify microniches whose abundance correlates significantly with case−control status. We confirm VIMA’s calibration, power and spatial accuracy in simulations. We then apply VIMA to spatial datasets spanning three diseases and spatial modalities, recapitulating known biology and identifying new spatial features of disease.
SN  - 1548-7091
SN  - 1548-7105
ER  -