Skip to main content

Thank you for visiting 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.

Cell clustering for spatial transcriptomics data with graph neural networks


Spatial transcriptomics data can provide high-throughput gene expression profiling and the spatial structure of tissues simultaneously. Most studies have relied on only the gene expression information but cannot utilize the spatial information efficiently. Taking advantage of spatial transcriptomics and graph neural networks, we introduce cell clustering for spatial transcriptomics data with graph neural networks, an unsupervised cell clustering method based on graph convolutional networks to improve ab initio cell clustering and discovery of cell subtypes based on curated cell category annotation. On the basis of its application to five in vitro and in vivo spatial datasets, we show that cell clustering for spatial transcriptomics outperforms other spatial clustering approaches on spatial transcriptomics datasets and can clearly identify all four cell cycle phases from multiplexed error-robust fluorescence in situ hybridization data of cultured cells. From enhanced sequential fluorescence in situ hybridization data of brain, cell clustering for spatial transcriptomics finds functional cell subtypes with different micro-environments, which are all validated experimentally, inspiring biological hypotheses about the underlying interactions among the cell state, cell type and micro-environment.

Your institute does not have access to this article

Access options

Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: CCST workflow for cell subpopulation discovery.
Fig. 2: Spatial distribution of cells in different clustered groups in MERFISH dataset.
Fig. 3: Cell cycle phase identification.
Fig. 4: Performance of CCST on two annotated datasets.
Fig. 5: Identifying cell subgroups in interneuron cells of the seqFISH+ mouse OB dataset.
Fig. 6: Neighbour enrichment ratios and GO term analysis for each cell subtype of astrocytes, endothelial cells, and neural stem cells of the seqFISH+ mouse OB dataset.

Data availability

Source data for Figs. 26 are available with this manuscript. The datasets utilized in this study can be downloaded from: (1) MERFISH dataset5: or our Github link:; (2) SeqFISH+ dataset35:; (3) DLPFC dataset37:; (4) 10× Visium spatial transcriptomics data of human breast cancer: The annotation file can be found on the SEDR32 website:

Code availability

CCST is implemented in Python. The source code and the utilized MERFISH dataset can be downloaded from the supporting website: (ref. 50).


  1. Codeluppi, S. et al. Spatial organization of the somatosensory cortex revealed by osmFISH. Nat. Methods 15, 932–935 (2018).

    Article  Google Scholar 

  2. Moffitt, J. R. & Zhuang, X. RNA imaging with multiplexed error-robust fluorescence in situ hybridization (MERFISH). Methods Enzymol. 572, 1–49 (2016).

    Article  Google Scholar 

  3. Moffitt, J. R. et al. High-throughput single-cell gene-expression profiling with multiplexed error-robust fluorescence in situ hybridization. Proc. Natl. Acad. Sci. U.S.A. 113, 11046–11051 (2016).

    Article  Google Scholar 

  4. 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  Google Scholar 

  5. Xia, C., Fan, J., Emanuel, G., Hao, J. & Zhuang, X. Spatial transcriptome profiling by MERFISH reveals subcellular RNA compartmentalization and cell cycle-dependent gene expression. Proc. Natl. Acad. Sci. U.S.A. 116, 19490–19499 (2019).

    Article  Google Scholar 

  6. Eng, C.-H. L., Shah, S., Thomassie, J. & Cai, L. Profiling the transcriptome with RNA SPOTs. Nat. Methods 14, 1153–1155 (2017).

    Article  Google Scholar 

  7. Lubeck, E., Coskun, A. F., Zhiyentayev, T., Ahmad, M. & Cai, L. Single-cell in situ RNA profiling by sequential hybridization. Nat. Methods 11, 360–361 (2014).

    Article  Google Scholar 

  8. Wang, X. et al. Three-dimensional intact-tissue sequencing of single-cell transcriptional states. Science 361, aat5691 (2018).

    Article  Google Scholar 

  9. Ståhl, P. L. et al. Visualization and analysis of gene expression in tissue sections by spatial transcriptomics. Science 353, 78–82 (2016).

    Article  Google Scholar 

  10. Rodriques, S. G. et al. Slide-seq: a scalable technology for measuring genome-wide expression at high spatial resolution. Science 363, 1463–1467 (2019).

    Article  Google Scholar 

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

    Article  Google Scholar 

  12. Nichterwitz, S. et al. Laser capture microscopy coupled with Smart-seq2 for precise spatial transcriptomic profiling. Nat. Commun. 7, 12139 (2016).

  13. Zhao, E. et al. Spatial transcriptomics at subspot resolution with BayesSpace. Nat. Biotechnol. 39, 1375–1384 (2021).

    Article  Google Scholar 

  14. Andersson, A. et al. Single-cell and spatial transcriptomics enables probabilistic inference of cell type topography. Commun. Biol. 3, 565 (2020).

  15. Pal, B. et al. Construction of developmental lineage relationships in the mouse mammary gland by single-cell RNA profiling. Nat. Commun. 8, 1627 (2017).

  16. Arnol, D., Schapiro, D., Bodenmiller, B., Saez-Rodriguez, J. & Stegle, O. Modeling cell–cell interactions from spatial molecular data with spatial variance component analysis. Cell Rep. 29, 202–211 (2019).

    Article  Google Scholar 

  17. Yuan, Y. & Bar-Joseph, Z. GCNG: graph convolutional networks for inferring gene interaction from spatial transcriptomics data. Genome Biol. 21, 300 (2020).

  18. Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902 (2019).

    Article  Google Scholar 

  19. Abdelaal, T. et al. A comparison of automatic cell identification methods for single-cell RNA sequencing data. Genome Biol. 20, 194 (2019).

  20. Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15 (2018).

    Article  Google Scholar 

  21. Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587 (2021).

    Article  Google Scholar 

  22. Blondel, V. D., Guillaume, J.-L., Lambiotte, R. & Lefebvre, E. Fast unfolding of communities in large networks. J. Stat. Mech. Theory Exp. 2008, P10008 (2008).

    Article  Google Scholar 

  23. Traag, V. A., Waltman, L. & Van Eck, N. J. From Louvain to Leiden: guaranteeing well-connected communities. Sci. Rep. 9, 5233 (2019).

  24. Shekhar, K. et al. Comprehensive classification of retinal bipolar neurons by single-cell transcriptomics. Cell 166, 1308–1323 (2016).

    Article  Google Scholar 

  25. Pandey, S., Shekhar, K., Regev, A. & Schier, A. F. Comprehensive identification and spatial mapping of habenular neuronal types using single-cell RNA-seq. Curr. Biol. 28, 1052–1065 (2018).

    Article  Google Scholar 

  26. Zhu, Q., Shah, S., Dries, R., Cai, L. & Yuan, G.-C. Identification of spatially associated subpopulations by combining scRNAseq and sequential fluorescence in situ hybridization data. Nat. Biotechnol. 36, 1183–1190 (2018).

    Article  Google Scholar 

  27. Stoltzfus, C. R. et al. CytoMAP: a spatial analysis toolbox reveals features of myeloid cell organization in lymphoid tissues. Cell Rep. 31, 107523 (2020).

    Article  Google Scholar 

  28. Dries, R. et al. Giotto: a toolbox for integrative analysis and visualization of spatial expression data. Genome Biol. 22, 78 (2021).

  29. Pham D., et al. stLearn: integrating spatial location, tissue morphology and gene expression to find cell types, cell–cell interactions and spatial trajectories within undissociated tissues. Prerint at bioRxiv (2020).

  30. Teng, H., Yuan, Y. & Bar-Joseph, Z. Clustering spatial transcriptomics data. Bioinformatics 38, 997–1004 (2022).

    Article  Google Scholar 

  31. Hu, J. et al. SpaGCN: integrating gene expression, spatial location and histology to identify spatial domains and spatially variable genes by graph convolutional network. Nat. Methods 18, 1342–1351 (2021).

    Article  Google Scholar 

  32. Fu, H., et al. Unsupervised spatial embedded deep representation of spatial transcriptomics. Preprint at bioRxiv (2021).

  33. Chen Y., Zhou S., Li M., Zhao F., & Qi J. STEEL enables high-resolution delineation of spatiotemporal transcriptomic data. Preprint at research square (2022).

  34. Kipf T. N. & Welling M. Semi-supervised classification with graph convolutional networks. In Proc. International Conference on Learning Representations (2017).

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

    Article  Google Scholar 

  36. Veličković P., et al. Deep graph infomax. In Proc. International Conference on Learning Representations (2019).

  37. Maynard, K. R. et al. Transcriptome-scale spatial gene expression in the human dorsolateral prefrontal cortex. Nat. Neurosci. 24, 425–436 (2021).

    Article  Google Scholar 

  38. Donjerkovic, D. & Scott, D. W. Regulation of the G1 phase of the mammalian cell cycle. Cell Res. 284, C349–364 (2000).

    Google Scholar 

  39. Tripathi, V. et al. Long noncoding RNA MALAT1 controls cell cycle progression by regulating the expression of oncogenic transcription factor B-MYB. PLoS Genet. 9, e1003368 (2013).

    Article  Google Scholar 

  40. Wang, J. et al. MALAT1 promotes cell proliferation in gastric cancer by recruiting SF2/ASF. Biomed. Pharmacother. 68, 557–564 (2014).

    Article  Google Scholar 

  41. Merlot, S., Gosti, F., Guerrier, D., Vavasseur, A. & Giraudat, J. The ABI1 and ABI2 protein phosphatases 2C act in a negative feedback regulatory loop of the abscisic acid signalling pathway. Plant J. 25, 295–303 (2001).

    Article  Google Scholar 

  42. Mahdessian, D. et al. Spatiotemporal dissection of the cell cycle with single-cell proteogenomics. Nature 590, 649–654 (2021).

    Article  Google Scholar 

  43. Sakaue-Sawano, A. et al. Visualizing spatiotemporal dynamics of multicellular cell-cycle progression. Cell 132, 487–498 (2008).

    Article  Google Scholar 

  44. Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods 16, 1289–1296 (2019).

    Article  Google Scholar 

  45. Cheng, C. et al. Cloning, expression and characterization of a novel human VMP gene. Mol. Biol. Rep. 29, 281–286 (2002).

    Article  Google Scholar 

  46. Li, S. et al. Endothelial cell-derived GABA signaling modulates neuronal migration and postnatal behavior. Cell Res. 28, 221–248 (2018).

    Article  Google Scholar 

  47. Russ, A. P. et al. Eomesodermin is required for mouse trophoblast development and mesoderm formation. Nature 404, 95–99 (2000).

    Article  Google Scholar 

  48. Taberner, L., Bañón, A. & Alsina, B. Sensory neuroblast quiescence depends on vascular cytoneme contacts and sensory neuronal differentiation requires initiation of blood flow. Cell Rep. 32, 107903 (2020).

    Article  Google Scholar 

  49. Hie, B., Bryson, B. & Berger, B. Efficient integration of heterogeneous single-cell transcriptomes using Scanorama. Nat. Biotechnol. 37, 685–691 (2019).

    Article  Google Scholar 

  50. Li J., Chen S., Pan X., Yuan Y., & Shen H.-B. Cell clustering for spatial transcriptomics data with graph neural networks. Zenodo (2022).

Download references


This work was supported by grants from the National Natural Science Foundation of China (no. 61725302 to H.S.), 62073219 (to H.S.), 62103262 (to Y.Y.) and 61903248 (to X.P.) and the Shanghai Pujiang Programme (no. 21PJ1407700 to Y.Y.).

Author information

Authors and Affiliations



H.S. and Y.Y. conceived and supervised the study. Y.Y. designed experiments. J.L. developed the computational model and conducted data analysis. Y.Y., H.S. and X.P. provided advice on data analysis. Y.Y. and S.C. proposed the proper computational model. J.L. drafted the manuscript. Y.Y. and H.S. revised the manuscript.

Corresponding authors

Correspondence to Ye Yuan or Hong-Bin Shen.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Computational Science thanks Xin Zhou and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Handling editor: Kaitlin McCardle, in collaboration with the Nature Computational Science 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 Comparison on sample 151676 of DLPFC.

Annotation and cluster labels obtained by CCST and prior methods on sample 151676 of DLPFC. Metrics including ARI, FMI and NMI are annotated on the bottom of each figure. Numbers in the legend refer to cluster labels.

Source data

Extended Data Fig. 2 Comparison on 10x Visium spatial transcriptomics data of human breast cancer.

Annotation and cluster labels obtained by CCST and prior methods on 10x Visium spatial transcriptomics data of human breast cancer. Metrics including ARI, FMI and NMI are annotated on the bottom of each figure. Numbers in the legend refer to cluster labels.

Source data

Supplementary information

Supplementary information Comparison on 10x Visium spatial transcriptomics data of human breast cancer.

Supplementary Figs. 1–30, Sections 1–15 and Tables 1–4.

Reporting summary

Supplementary Data 1. Comparison on 10x Visium spatial transcriptomics data of human breast cancer.

The top 200 significantly DE genes of each cell group obtained by CCST on the MERFISH dataset.

Source data

Source Data Fig. 2.

Statistical source data of each subfigure is listed in each sheet.

Source Data Fig. 3.

Statistical source data of each subfigure is listed in each sheet.

Source Data Fig. 4.

Raw data of boxplots is listed in each sheet.

Source Data Fig. 5.

Statistical source data of each subfigure is listed in each sheet.

Source Data Fig. 6.

Statistical source data of each subfigure is listed in each sheet.

Source Data Extended Data Fig. 1.

Cell cluster labels.

Source Data Extended Data Fig. 2.

Cell cluster labels.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Li, J., Chen, S., Pan, X. et al. Cell clustering for spatial transcriptomics data with graph neural networks. Nat Comput Sci 2, 399–408 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

Further reading


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