Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Resolving tissue complexity by multimodal spatial omics modeling with MISO

Nature Methods volume 22, pages 530–538 (2025)Cite this article

Subjects

An Author Correction to this article was published on 06 March 2025

An Author Correction to this article was published on 24 January 2025

This article has been updated

Abstract

Spatial molecular profiling has provided biomedical researchers valuable opportunities to better understand the relationship between cellular localization and tissue function. Effectively modeling multimodal spatial omics data is crucial for understanding tissue complexity and underlying biology. Furthermore, improvements in spatial resolution have led to the advent of technologies that can generate spatial molecular data with subcellular resolution, requiring the development of computationally efficient methods that can handle the resulting large-scale datasets. MISO (MultI-modal Spatial Omics) is a versatile algorithm for feature extraction and clustering, capable of integrating multiple modalities from diverse spatial omics experiments with high spatial resolution. Its effectiveness is demonstrated across various datasets, encompassing gene expression, protein expression, epigenetics, metabolomics and tissue histology modalities. MISO outperforms existing methods in identifying biologically relevant spatial domains, representing a substantial advancement in multimodal spatial omics analysis. Moreover, MISO’s computational efficiency ensures its scalability to handle large-scale datasets generated by subcellular resolution spatial omics technologies.

Access through your institution
Buy or subscribe

This is a preview of subscription content, access via your institution

Access options

Access through your institution

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: MISO workflow for analysis of spatial multi-omics dataset with paired histology image.
Fig. 2: Analysis of a 10x Visium bladder cancer spatial transcriptomics dataset.
Fig. 3: Analysis of large-scale spatial transcriptomics datasets.
Fig. 4: Analysis of a spatial ATAC–RNA-seq dataset from a mouse at E13.
Fig. 5: Analysis of spatial multi-omics datasets with three modalities.

Similar content being viewed by others

Data availability

We analyzed the following datasets: (1) 10x Visium human bladder cancer spatial transcriptomics data (GEO GSE246011, sample BLCA-B1); (2) 10x Xenium human gastric cancer spatial transcriptomics data. Data requests will be reviewed by Dr. Tae Hyun Hwang ([email protected]) and Dr. Jeong Hwan Park ([email protected]), and reasonable requests may be accommodated upon approval. IRB No. 30-2023-1; (3) 10x Visium HD human colorectal cancer data (https://www.10xgenomics.com/datasets/visium-hd-cytassist-gene-expression-libraries-of-human-crc); (4) Spatial ATAC-RNA-seq mouse embryonic day 13 (E13) data reported in Zhang et al.5 (https://cells.ucsc.edu/?ds=brain-spatial-omics); (5) Spatial transcriptomics and metabolomics mouse coronal brain data (https://upenn.box.com/s/3o8dq5j4x29ic6zo7iugdo83scnv4qis). (6) 10x Visium human tonsil gene and protein expression data (https://www.10xgenomics.com/resources/datasets/gene-protein-expression-library-of-human-tonsil-cytassist-ffpe-2-standard); (7) 10x Visium mouse anterior brain spatial transcriptomics data (https://www.10xgenomics.com/resources/datasets/mouse-brain-serial-section-1-sagittal-anterior-1-standard-1-1-0); (8) 10x Visium human breast cancer spatial transcriptomics data(https://www.10xgenomics.com/resources/datasets/human-breast-cancer-visium-fresh-frozen-whole-transcriptome-1-standard); (9) 10x Visium zebrafish melanoma spatial transcriptomics data reported in Hunter et al.41 (GSE159709); (10) 10x Visium mouse olfactory bulb spatial transcriptomics data (https://www.10xgenomics.com/resources/datasets/adult-mouse-olfactory-bulb-1-standard-1); (11) 10x Visium mouse coronal brain spatial transcriptomics data (https://www.10xgenomics.com/resources/datasets/mouse-brain-coronal-section-2-ffpe-2-standard); (12) 10x Visium human prostate cancer spatial transcriptomics data reported in Erickson et al.44 (https://data.mendeley.com/datasets/svw96g68dv/1); (13) Spatial CUT&Tag-RNA-seq (H3K27AC) mouse coronal brain data reported in Zhang et al.5 (https://cells.ucsc.edu/?ds=brain-spatial-omics); (14) Spatial CUT&Tag-RNA-seq (H3K27ME3) mouse coronal brain data reported in Zhang et al.5 (https://cells.ucsc.edu/?ds=brain-spatial-omics); (15) 10x Visium human breast cancer gene and protein expression data(https://www.10xgenomics.com/resources/datasets/gene-and-protein-expression-library-of-human-breast-cancer-cytassist-ffpe-2-standard); (16) Spatial CITE-seq mouse colon data reported in Liu et al.6 (GSE213264). Details of the datasets analyzed in this paper are described in Supplementary Table 1.

Code availability

The MISO algorithm was implemented in Python and is available on GitHub at https://github.com/kpcoleman/miso.

Change history

References

  1. Vandereyken, K., Sifrim, A., Thienpont, B. & Voet, T. Methods and applications for single-cell and spatial multi-omics. Nat. Rev. Genet. 24, 494–515 (2023).

    Article  CAS  PubMed  Google Scholar 

  2. Chen, K. H., Boettiger, A. N., Moffitt, J. R., Wang, S. & Zhuang, X. Spatially resolved, highly multiplexed RNA profiling in single cells. Science 348, aaa6090 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  3. Stickels, R. R. et al. Highly sensitive spatial transcriptomics at near-cellular resolution with Slide-seqV2. Nat. Biotechnol. 39, 313–319 (2021).

    Article  CAS  PubMed  Google Scholar 

  4. Eng, C.-H. L. et al. Transcriptome-scale super-resolved imaging in tissues by RNA seqFISH+. Nature 568, 235–239 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Zhang, D. et al. Spatial epigenome–transcriptome co-profiling of mammalian tissues. Nature 616, 113–122 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Liu, Y. et al. High-plex protein and whole transcriptome co-mapping at cellular resolution with spatial CITE-seq. Nat. Biotechnol. 41, 1405–1409 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Liao, S. et al. Integrated spatial transcriptomic and proteomic analysis of fresh frozen tissue based on Stereo-seq. Preprint at bioRxiv https://doi.org/10.1101/2023.04.28.538364 (2023).

  8. Ben-Chetrit, N. et al. Integration of whole transcriptome spatial profiling with protein markers. Nat. Biotechnol. 41, 788–793 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Vicari, M. et al. Spatial multimodal analysis of transcriptomes and metabolomes in tissues. Nat. Biotechnol. 42, 1046–1050 (2024).

    Article  CAS  PubMed  Google Scholar 

  10. Ghazanfar, S., Guibentif, C. & Marioni, J. C. Stabilized mosaic single-cell data integration using unshared features. Nat. Biotechnol. 42, 284–292 (2024).

    Article  CAS  PubMed  Google Scholar 

  11. Hao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol. 42, 293–304 (2024).

    Article  CAS  PubMed  Google Scholar 

  12. Bao, F. et al. Integrative spatial analysis of cell morphologies and transcriptional states with MUSE. Nat. Biotechnol. 40, 1200–1209 (2022).

    Article  CAS  PubMed  Google Scholar 

  13. Long, Y. et al. Deciphering spatial domains from spatial multi-omics with SpatialGlue. Nat. Methods 21, 1658–1667 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Jiang, J. et al. METI: deep profiling of tumor ecosystems by integrating cell morphology and spatial transcriptomics. Nat. Commun. 15, 7312 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Schumacher, T. N. & Thommen, D. S. Tertiary lymphoid structures in cancer. Science 375, eabf9419 (2022).

    Article  CAS  PubMed  Google Scholar 

  16. Sautès-Fridman, C. et al. Tertiary lymphoid structures in cancers: prognostic value, regulation, and manipulation for therapeutic intervention. Front. Immunol. 7, 407 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  17. Di Caro, G. et al. Occurrence of tertiary lymphoid tissue is associated with T-cell infiltration and predicts better prognosis in early-stage colorectal cancers. Clin. Cancer Res. 20, 2147–2158 (2014).

    Article  PubMed  Google Scholar 

  18. Asrir, A. et al. Tumor-associated high endothelial venules mediate lymphocyte entry into tumors and predict response to PD-1 plus CTLA-4 combination immunotherapy. Cancer Cell 40, 318–334 (2022).

    Article  CAS  PubMed  Google Scholar 

  19. Zhang, D. et al. Inferring super-resolution tissue architecture by integrating spatial transcriptomics with histology. Nat. Biotechnol. 42, 1372–1366 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. van Cutsem, E. et al. HER2 screening data from ToGA: targeting HER2 in gastric and gastroesophageal junction cancer. Gastric Cancer 18, 476–484 (2015).

    Article  CAS  PubMed  Google Scholar 

  21. Oliveira, M. F. et al. Characterization of immune cell populations in the tumor microenvironment of colorectal cancer using high definition spatial profiling. Preprint at bioRxiv https://doi.org/10.1101/2024.06.04.597233 (2024).

  22. Jantscheff, P. et al. Expression of CEACAM6 in resectable colorectal cancer: a factor of independent prognostic significance. Jo. Clin. Oncol. 21, 3638–3646 (2003).

    Article  CAS  Google Scholar 

  23. Burgos, M. et al. Prognostic value of the immune target CEACAM6 in cancer: a meta-analysis. Ther. Adv. Med. Oncol. 14, 17588359211072621 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Qi, J. et al. Single-cell and spatial analysis reveal interaction of FAP+ fibroblasts and SPP1+ macrophages in colorectal cancer. Nat. Commun. 13, 1742 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Ozato, Y. et al. Spatial and single-cell transcriptomics decipher the cellular environment containing HLA-G+ cancer cells and SPP1+ macrophages in colorectal cancer. Cell Rep. 42, 111929 (2023).

    Article  CAS  PubMed  Google Scholar 

  26. Bani-Yaghoub, M. et al. Role of Sox2 in the development of the mouse neocortex. Dev. Biol. 295, 52–66 (2006).

    Article  CAS  PubMed  Google Scholar 

  27. Koopman, A. D. et al. The association between GAD65 antibody levels and incident type 2 diabetes mellitus in an adult population: a meta-analysis. Metabolism 95, 1–7 (2019).

    Article  CAS  PubMed  Google Scholar 

  28. Petanjek, Z., Kostovic, I. & Esclapez, M. Primate-specific origins and migration of cortical GABAergic neurons. Front. Neuroanat. 3, 26 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

  29. Kronman, F. N. et al. Developmental mouse brain common coordinate framework. Nat. Commun. 15, 9072 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Hsueh, Y. -P., Wang, T. -F., Yang, F. -C. & Sheng, M. Nuclear translocation and transcription regulation by the membrane-associated guanylate kinase CASK/LIN-2. Nature 404, 298–302 (2000).

    Article  CAS  PubMed  Google Scholar 

  31. Bulfone, A. et al. T-brain-1: a homolog of Brachyury whose expression defines molecularly distinct domains within the cerebral cortex. Neuron 15, 63–78 (1995).

    Article  CAS  PubMed  Google Scholar 

  32. Puelles, L. et al. Pallial and subpallial derivatives in the embryonic chick and mouse telencephalon, traced by the expression of the genes Dlx-2, Emx-1, Nkx-2.1, Pax-6 and Tbr-1. J. Comp. Neurol. 424, 409–438 (2000).

    Article  CAS  PubMed  Google Scholar 

  33. Zhang, M. et al. Molecularly defined and spatially resolved cell atlas of the whole mouse brain. Nature 624, 343–354 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Hamilton, D., White, C., Rees, C., Wheeler, D. & Ascoli, G. Molecular fingerprinting of principal neurons in the rodent hippocampus: a neuroinformatics approach. J. Pharm. Biomed. Anal. 144, 269–278 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Lorente de Nó, R. Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system. Journal für Psychologie und Neurologie 46, 113–177 (1934).

  36. Li, X. G., Somogyi, P., Ylinen, A. & Buzsáki, G. The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neurol. 339, 181–208 (1994).

    Article  CAS  PubMed  Google Scholar 

  37. 10xGenomics. https://www.10xgenomics.com/resources/datasets/gene-protein-expression-library-of-human-tonsil-cytassist-ffpe-2-standard (2023).

  38. Lipponen, P. K. & Eskelinen, M. J. Cell proliferation of transitional cell bladder tumours determined by PCNA/cyclin immunostaining and its prognostic value. Br. J. Cancer 66, 171–176 (1992).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. 10xGenomics. https://www.10xgenomics.com/resources/datasets/mouse-brain-serial-section-1-sagittal-anterior-1-standard-1-1-0 (2020).

  40. 10xGenomics. https://www.10xgenomics.com/resources/datasets/human-breast-cancer-visium-fresh-frozen-whole-transcriptome-1-standard (2022).

  41. Hunter, M. V., Moncada, R., Weiss, J. M., Yanai, I. & White, R. M. Spatially resolved transcriptomics reveals the architecture of the tumor-microenvironment interface. Nat. Commun. 12, 6278 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. 10xGenomics. https://www.10xgenomics.com/resources/datasets/adult-mouse-olfactory-bulb-1-standard-1 (2022).

  43. 10xGenomics. https://www.10xgenomics.com/resources/datasets/mouse-brain-coronal-section-2-ffpe-2-standard (2022).

  44. Erickson, A. et al. Spatially resolved clonal copy number alterations in benign and malignant tissue. Nature 608, 360–367 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. 10xGenomics. https://www.10xgenomics.com/resources/datasets/gene-and-protein-expression-library-of-human-breast-cancer-cytassist-ffpe-2-standard (2023).

  46. Chen, R. J. et al. Scaling vision transformers to gigapixel images via hierarchical self-supervised learning. In Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition 16144–16155 (2022).

  47. Ng, A., Jordan, M. & Weiss, Y. On spectral clustering: analysis and an algorithm. Adv. Neural Inf. Process. Syst. 14, 849–856 (2001).

  48. Shaham, U. et al. Spectralnet: Spectral clustering using deep neural networks. Preprint at https://doi.org/10.48550/arXiv.1801.01587 (2018).

  49. Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In Proc. International Conference on Learning Representations (ICLR, 2015)

  50. Lowe, E. K., Cuomo, C., Voronov, D. & Arnone, M. I. Using ATAC-seq and RNA-seq to increase resolution in GRN connectivity. Methods Cell. Biol. 151, 115–126 (2019).

    Article  CAS  PubMed  Google Scholar 

  51. Høiem, T. S. et al. An optimized MALDI MSI protocol for spatial detection of tryptic peptides in fresh frozen prostate tissue. Proteomics 22, e2100223 (2022).

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

M. Li was partly supported by National Institutes of Health (NIH) grants R01HG013185, R01LM014592, R01EY030192, U19NS135582, R01HL171595 and U01CA294518. L.W. was partly supported by NIH grants R01CA266280, U01CA264583, U01CA294518 and U24CA274274; the start-up research fund provided by the University of Texas MD Anderson Cancer Center; The Break Through Cancer Foundation; and the Andrew Sabin Family Foundation. L.W. and J.J were also supported by NIH grant R01 CA254988. T.H.H. was partly supported by NIH grants R01CA276690 and U01CA294518, DOD grant CA190578, the Eric and Wendy Schmidt Foundation’s AI Innovation Award through the Mayo Clinic Foundation, and the Torrey Coast Foundation. X.Q. was partly supported by NIH grant K99NS135123. J.G. was partly supported by the Doris Duke Clinical Scientist Development Award (2018097), MD Anderson Faculty Scholar Award, the David H. Koch Center for Applied Research of Genitourinary Cancers, Wendy and Leslie Irvin Barnhart Fund, Joan and Herb Kelleher Charitable Foundation, KCA Advanced Discovery Award, the Williams TNT Fund, the V Foundation Translational Award, the DOD KCRP Translational Research Partnership Award, NIH/NCI R01 CA254988-01A1, NIH/National Cancer Institute (NCI) R01 CA269489-01A1 and NIH/NCI R01 CA282282-01; as well as in part by the Cancer Center Support Grant to MDACC (P30 CA016672) from the NCI, by MD Anderson’s Prometheus informatics system and by the Department of Genitourinary Medical Oncology’s Eckstein and Alexander Laboratories. J.D.R. was supported by Ludwig Cancer Research, the Penn Diabetes Research Center grant (P30-DK19525) and the Chan Zuckerberg Initiative DAF (2023-331955), an advised fund of Silicon Valley Community Foundation.

Author information

Authors and Affiliations

  1. Department of Computational Biomedicine, Cedars-Sinai Medical Center, Los Angeles, CA, USA

    Kyle Coleman

  2. Statistical Center for Single-Cell and Spatial Genomics, Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Kyle Coleman, Amelia Schroeder, Melanie Loth, Daiwei Zhang, Hanying Yan & Mingyao Li

  3. Department of Pathology, Seoul Metropolitan Government-Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea

    Jeong Hwan Park

  4. Department of Pathology, College of Medicine, Kyung Hee University Hospital, Kyung Hee University, Seoul, Republic of Korea

    Ji-Youn Sung

  5. Department of Microbiology, Institute for Immunology, and Institute for Diabetes, Obesity and Metabolism, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Niklas Blank, Alexis J. Cowan & Christoph A. Thaiss

  6. Faculty of Biology, University of Freiburg, Freiburg, Germany

    Niklas Blank

  7. Division of Genetics and Genomics, Department of Pediatrics, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA

    Xuyu Qian

  8. Department of Genitourinary Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

    Jianfeng Chen & Jianjun Gao

  9. Department of Genomic Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

    Jiahui Jiang & Linghua Wang

  10. School of Public Health, The University of Texas Health Science Center at Houston, Houston, TX, USA

    Jiahui Jiang

  11. Department of Chemistry, Lewis Sigler Institute for Integrative Genomics, and Ludwig Institute for Cancer Research, Princeton University, Princeton, NJ, USA

    Laith Z. Samarah & Joshua D. Rabinowitz

  12. Department of Artificial Intelligence and Informatics, Mayo Clinic, Jacksonville, FL, USA

    Jean R. Clemenceau, Inyeop Jang, Minji Kim, Isabel Barnfather & Tae Hyun Hwang

  13. Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Yanxiang Deng, Edward B. Lee, Emma E. Furth & Mingyao Li

  14. Epigenetics Institute, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Yanxiang Deng

  15. Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

    Alexander Lazar

  16. The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences, Houston, TX, USA

    Linghua Wang

  17. Department of Human Genetics, School of Medicine, Emory University, Atlanta, GA, USA

    Jian Hu

Authors
  1. Kyle Coleman
    View author publications

    Search author on:PubMed Google Scholar

  2. Amelia Schroeder
    View author publications

    Search author on:PubMed Google Scholar

  3. Melanie Loth
    View author publications

    Search author on:PubMed Google Scholar

  4. Daiwei Zhang
    View author publications

    Search author on:PubMed Google Scholar

  5. Jeong Hwan Park
    View author publications

    Search author on:PubMed Google Scholar

  6. Ji-Youn Sung
    View author publications

    Search author on:PubMed Google Scholar

  7. Niklas Blank
    View author publications

    Search author on:PubMed Google Scholar

  8. Alexis J. Cowan
    View author publications

    Search author on:PubMed Google Scholar

  9. Xuyu Qian
    View author publications

    Search author on:PubMed Google Scholar

  10. Jianfeng Chen
    View author publications

    Search author on:PubMed Google Scholar

  11. Jiahui Jiang
    View author publications

    Search author on:PubMed Google Scholar

  12. Hanying Yan
    View author publications

    Search author on:PubMed Google Scholar

  13. Laith Z. Samarah
    View author publications

    Search author on:PubMed Google Scholar

  14. Jean R. Clemenceau
    View author publications

    Search author on:PubMed Google Scholar

  15. Inyeop Jang
    View author publications

    Search author on:PubMed Google Scholar

  16. Minji Kim
    View author publications

    Search author on:PubMed Google Scholar

  17. Isabel Barnfather
    View author publications

    Search author on:PubMed Google Scholar

  18. Joshua D. Rabinowitz
    View author publications

    Search author on:PubMed Google Scholar

  19. Yanxiang Deng
    View author publications

    Search author on:PubMed Google Scholar

  20. Edward B. Lee
    View author publications

    Search author on:PubMed Google Scholar

  21. Alexander Lazar
    View author publications

    Search author on:PubMed Google Scholar

  22. Jianjun Gao
    View author publications

    Search author on:PubMed Google Scholar

  23. Emma E. Furth
    View author publications

    Search author on:PubMed Google Scholar

  24. Tae Hyun Hwang
    View author publications

    Search author on:PubMed Google Scholar

  25. Linghua Wang
    View author publications

    Search author on:PubMed Google Scholar

  26. Christoph A. Thaiss
    View author publications

    Search author on:PubMed Google Scholar

  27. Jian Hu
    View author publications

    Search author on:PubMed Google Scholar

  28. Mingyao Li
    View author publications

    Search author on:PubMed Google Scholar

Contributions

This study was conceived of and led by M. Li and J.H. K.C. designed the model and algorithm with input from M. Li and J.H., implemented the MISO software and led data analyses. A.S., M. Loth and H.Y. performed data analyses. D.Z. proposed the histology image feature extraction approach. T.H.H., J.H.P., J.-Y.S., J.R.C., I.J., M.K. and I.B. generated and processed the Xenium gastric cancer data. J.H.P. annotated the Xenium gastric cancer data. J.-Y.S. annotated the Visium HD colon cancer data. L.W., J.G., J.C., A.L. and J.J. generated and processed the Visium bladder cancer data. A.L. annotated the Visium bladder cancer data. C.A.T., J.D.R., N.B., A.J.C. and L.Z.S. generated and processed the mouse brain spatial transcriptomics and metabolomics data. X.Q. annotated the hippocampus region and interpreted the results of the mouse brain spatial transcriptomics and metabolomics data. Y.D. provided input for the spatial CUT&Tag–RNA-seq data analysis. E.B.L. provided input for mouse brain data analysis. E.E.F. confirmed tissue annotation and provided input for interpretation of the Visium HD human colon cancer data. K.C. and M. Li wrote the paper with feedback from the other co-authors.

Corresponding authors

Correspondence to Kyle Coleman, Jian Hu or Mingyao Li.

Ethics declarations

Competing interests

M. Li receives research funding from Biogen unrelated to the current manuscript. M. Li and D.Z. are cofounders of OmicPath AI. T.H. is a cofounder of Kure.ai therapeutics and has received consulting fees from IQVIA; these affiliations and financial compensations are unrelated to the current paper. The other authors declare no competing interests.

Peer review

Peer review information

Nature Methods thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Rita Strack, in collaboration with the Nature Methods team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Intraclass correlation coefficient (ICC) results for all datasets provided in Figs. 2-5 that were evaluated using MISO, MUSE, and SpatialGlue.

The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided t-tests (n = 1250 ICC values for each group). For a vast majority of the modalities across all datasets, the MISO clustering results produced a higher ICC compared to the other methods. The only instance in which the MISO ICC was lower was for the RNA modality in the mouse hippocampus spatial transcriptomics and metabolomics dataset, where the ICC for SpatialGlue surpassed that of MISO. The likely cause of this is that, because the RNA data was of low quality, MISO did not use the RNA-specific terms in clustering, and only accounted for this modality through the RNAxImage and RNAxMetabolite interaction terms. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 2 Clustering results for a mouse anterior brain spatial transcriptomics dataset.

a, Allen Brain Atlas annotation of mouse anterior brain. b, Shown from left to right are clustering results from MISO, MUSE, and SpatialGlue, respectively. SpatialGlue is sensitive to weight specified for each modality. WG is the weight for gene expression and WH is the weight for histology. Adjusted Rand Index (ARI) is calculated between SpatialGlue clustering with different weights. c, RNA and image ICC distributions across all clusters and features for each method in the mouse anterior brain data (n = 1250 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. d, MISO and SpatialGlue RNA ICC distributions for all clusters corresponding to the cortical layers in the mouse anterior brain data (n = 50 ICC values for each group except SpatialGlue L2/3 and SpatialGlue L6, which contain n = 100 ICC values). Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. e, Illustration of image artifact. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 3 Clustering results for a human breast cancer spatial transcriptomics dataset.

a, Pathologist manual annotation of tissue section. DCIS: Ductal carcinoma in situ. b, Shown from left to right are clustering results from MISO, MUSE, and SpatialGlue, respectively. Two patches were selected to highlight that MISO’s results agree better with histological patterns. c, RNA and image ICC distributions across all clusters and features for each method in the breast cancer data (n = 750 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. d, Spots plotted according to their RNA t-SNE coordinates and colored by the clustering results for each method. The MISO clustering results demonstrate coherence with respect to gene expression patterns and the annotated histological regions. e, RNA t-SNE plot for the breast cancer Visium dataset with spots colored according to total UMI count. MISO was able to localize a sub-cluster in the annotated invasive carcinoma region with much lower total UMI counts compared to other sub-clusters in this region. f, SpatialGlue clustering results when increasing the weight given to histology in the loss function. SpatialGlue was not able to detect the fat region of the tissue section when making the weight given to histology 10 or 50 times greater than that given to gene expression. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 4 Clustering results for a zebrafish melanoma spatial transcriptomics dataset.

a, Blurriness artifact in the H&E-stained histology image. b, Clustering results from MISO, MUSE, and SpatialGlue. MISO did not include the image-specific features in clustering because of the low quality of the image, but the image features were still accounted for in the RNAximage interaction terms. Clusters in the MUSE results are driven by the blurriness artifact. c, RNA and image ICC distributions across all clusters and features for each method (n = 400 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 5 Clustering results for a mouse olfactory bulb spatial transcriptomics dataset.

a, Clustering results from MISO, MUSE, and SpatialGlue. b, H&E-stained histology image of analyzed tissue section with layer annotation. The MISO results align well with the annotation, assigning clusters to each of the annotated layers. MUSE was not able to accurately localize clusters to the annotated layers, and instead separated spots in the upper region of the tissue section from those in the lower region. c, RNA and image ICC distributions across all clusters and features for each method (n = 400 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 6 Clustering results for a mouse coronal brain spatial transcriptomics dataset.

a, Clustering results from MISO, MUSE, and SpatialGlue. b, H&E-stained histology image of analyzed tissue section. c, RNA and image ICC distributions across all clusters and features for each method (n = 1000 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 7 Clustering results for a human prostate cancer spatial transcriptomics dataset.

a, Clustering results from MISO, MUSE, and SpatialGlue. b, H&E-stained histology image of analyzed tissue section. c, RNA and image ICC distributions across all clusters and features for each method (n = 750 ICC values for each group). The mean ICC for each method and each modality is printed on the corresponding box plot. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests. d, Clone annotation of cancer spots. e, ARI between the clone annotation and the clustering results across all cancer spots for MISO (0.51), MUSE (0.44), and SpatialGlue (0.50). f, Weighted F1 score for localization of clusters to the annotated clones for MISO (0.61), MUSE (0.52), and SpatialGlue (0.59). To calculate F1 score for a given method, a cluster was assigned to a clone if more than half of the spots from that cluster overlapped with the clone annotation. F1 score was weighted by the number of spots belonging to each clone in the annotation. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 8 Clustering results for a human breast cancer spatial gene and protein expression dataset.

Image features used as input for each method were extracted from an immunofluorescence image with 3 channels (DAPI, Vimentin, and PCNA) using a pre-trained InceptionV3 model. Patches corresponding to the omics spots were extracted from the immunofluorescence image, resized to 299×299 pixels, and normalized prior to extracting features for each patch using InceptionV3. a, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and image data as input. b, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and protein data as input. c, Clustering results from MISO, MUSE, and SpatialGlue when taking protein and image data as input. d, Clustering results from MISO when taking RNA, image, and protein data as input. e, Eosin-stained histology image of analyzed tissue section. f, RNA, image, and protein ICC distributions across all clusters and features for each method when taking each possible combination of modalities as input (n = 500 ICC values for each group). For each method, the mean ICC for each modality is printed on the corresponding box plot for its top-performing combination of modalities. Test statistics and p-values were obtained using one-sided (<,>) or two-sided (≈) t-tests when comparing each method’s top-performing results for a given modality. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 9 Clustering results for a mouse brain spatial transcriptomics (10x Visium) and metabolomics (MALDI-MSI) dataset, which was generated following the protocol described in Vicari et al. [9].

To make the super-resolution spatial molecular data inferred by iStar more suitable for input to all methods, we merged superpixels obtained from iStar to create 4,687 pseudo-spots of size 128×128 pixels, containing paired gene expression and metabolite information. Because the RNA data is of low quality (e), the RNA-specific features extracted by MISO were not used to produce any of the results provided, but RNA was still accounted for in its interactions with metabolomics and image features. For all applicable results, metabolomics data were normalized by total intensity and log transformed. a, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and histology image data as input. b, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and metabolomics data as input. c, Clustering results from MISO, MUSE, and SpatialGlue when taking metabolomics and histology image data as input. d, Clustering results from MISO when taking RNA, histology image, and metabolomics data as input. e, Total UMI counts across all spots in the dataset. The UMI counts are low because the tissue section was analyzed using MALDI-MSI prior to Visium. f, Total metabolite intensities across all spots in the dataset. g, H&E-stained histology image of analyzed tissue section. h, RNA, image, and metabolomics ICC distributions across all clusters and features for each method when taking each possible combination of modalities as input (n = 1500 ICC values for each group). For each method, the mean ICC for each modality is printed on the corresponding box plot for its top-performing combination of modalities. Test statistics and p-values were obtained using one-sided t-tests when comparing each method’s top-performing results for a given modality. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Extended Data Fig. 10 Clustering results for large-scale mouse brain spatial transcriptomics (10x Visium) and metabolomics (MALDI-MSI) dataset described in Extended Data Fig. 9.

Results were obtained using RNA, metabolomics, and image features as input, but RNA was only accounted for in the interaction terms due to its low quality. For all applicable results, other than the MUSE results obtained when taking metabolomics and image data as input, metabolomics data were normalized by total intensity and log transformed. MUSE was not able to evaluate the combination of metabolomics and image data when the metabolomics data were log normalized, so this step was not utilized to obtain the corresponding results. The dataset used to generate the results in (a-e) contains 74,851 pseudo-spots of size 32×32 pixels. The dataset used to generate the results in (f) contains 299,350 pseudo-spots of size 16×16 pixels. Due to memory requirements, MISO was the only method that could evaluate the dataset with pseudo-spots of size 16×16 pixels. a, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and histology image data as input. b, Clustering results from MISO, MUSE, and SpatialGlue when taking RNA and metabomics data as input. c, Clustering results from MISO, MUSE, and SpatialGlue when taking metabolomics and histology image data as input. d, Clustering results from MISO when taking RNA, histology image, and metabolomics data as input. e, RNA, image, and metabolomics ICC distributions across all clusters and features for each method when taking each possible combination of modalities as input (n = 1500 ICC values for each group). For each method, the mean ICC for each modality is printed on the corresponding box plot for its top-performing combination of modalities. Test statistics and p-values were obtained using one-sided t-tests when comparing each method’s top-performing results for a given modality. f, Clustering results from MISO when taking RNA, histology image, and metabolomics data from the dataset with pseudo-spots of size 16×16 pixels as input. Test statistics and p-values were obtained using one-sided t-tests when comparing each method’s top-performing results for a given modality. Box plots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5x interquartile range; points, outliers.

Supplementary information

Supplementary Information

Supplementary Tables 1 and 2 and Figs. 1–6.

Reporting Summary

Peer Review File

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Coleman, K., Schroeder, A., Loth, M. et al. Resolving tissue complexity by multimodal spatial omics modeling with MISO. Nat Methods 22, 530–538 (2025). https://doi.org/10.1038/s41592-024-02574-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41592-024-02574-2

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing