Visium spatial transcriptomics data for individual oral squamous cell carcinoma (OSCC) patients

How to CiteIf you use this dataset in your research, please cite our primary publication in PLOS Genetics:Furudate K, Kasai S, Yoshizawa T, Sasaki Y, Fujikura K, et al. (2025) Spatial colocalization and molecular crosstalk of myofroblastic CAFs and tumor cells shape lymph node metastasis in oral squamous cell carcinoma. PLOS Genetics 21(9): e1011791. https://doi.org/10.1371/journal.pgen.1011791Dataset Citation (Optional but Recommended): For data traceability, you may also cite this Figshare dataset directly using the citation generated by the "Cite" button on this page. However, please ensure the primary publication above is also cited.Dataset DescriptionThis dataset contains the 10x Genomics Space Ranger output files for four individual human oral squamous cell carcinoma (OSCC) samples, supporting the findings of our study: "Spatial colocalization and molecular crosstalk of myofibroblastic CAFs and tumor cells shape lymph node metastasis in oral squamous cell carcinoma" (https://doi.org/10.1371/journal.pgen.1011791).For each of the four samples (Sample A [HUH001-P], Sample B [HUH001-P2], Sample C [HUH001-met], and Sample D [HUH002-P]), the following files and folders are provided:The filtered feature-barcode matrix in HDF5 format (filtered_feature_bc_matrix.h5).A spatial subfolder containing:High-resolution tissue image (tissue_hires_image.png)Low-resolution tissue image (tissue_lowres_image.png)Fiducial alignment image (aligned_fiducials.jpg)Tissue detection image (detected_tissue_image.jpg)Scale factors JSON file (scalefactors_json.json)Tissue spot coordinates CSV file (tissue_positions_list.csv)A patho_annot subfolder containing annotation files in CSV format for each respective sample. These files link annotations to spot barcodes and provide detailed information for each spot, including columns for:Barcode (the unique spot identifier)patho (primary pathological annotation ID)graph-based cluster (computationally assigned cluster ID based on gene expression)patho_diag (detailed pathological diagnostic information or feature)category (a broader classification category for the spot, e.g., tumor, peritumor, non_tumor)description (additional descriptive text or notes for the spot)These files allow for the direct import, reprocessing, and reanalysis of individual samples using standard spatial transcriptomics software packages such as Seurat (e.g., the Load10X_Spatial function) or Scanpy/Squidpy (e.g., the read_visium function).Interactive Data Exploration with Loupe BrowserFor users who prefer interactive visualization without programming, this dataset also includes Loupe Browser (.cloupe) files for each of the four samples. Loupe Browser is a free desktop application from 10x Genomics that allows for easy exploration of the spatial gene expression data.To get started:Download and install Loupe Browser from the 10x Genomics website:https://www.10xgenomics.com/support/software/loupe-browser/downloadsOpen the .cloupe file for the sample you wish to explore (e.g., A.cloupe).Import pathologist annotations: To view our custom pathology annotations on the tissue image, navigate to the "Categories" tab, click the three-dot menu (...) next to a clustering (e.g., "Graph-Based"), and select "Import Categories". Then, select the corresponding annotation file from the patho_annot folder provided in this deposit.You can now explore gene expression spatially and overlay it with our pathological annotations. Please note: The gene expression data visualized in Loupe Browser is based on raw, un-normalized UMI counts.Related Data RepositoriesThe raw sequencing reads for these samples are available at DDBJ under BioProject accession PRJDB13905 (https://ddbj.nig.ac.jp/search/entry/bioproject/PRJDB13905).A processed and integrated AnnData object (.h5ad) combining these four samples is available at GEA under accession E-GEAD-511 (https://ddbj.nig.ac.jp/public/ddbj_database/gea/experiment/E-GEAD-000/E-GEAD-511/).Supplementary spatial metadata components (pickled Python objects) for use with the integrated AnnData object are available at Figshare (https://doi.org/10.6084/m9.figshare.20408067).The trained scVI model used for data integration is available at Figshare (https://doi.org/10.6084/m9.figshare.20279025.v1).All patient-derived images included in this dataset were de-identified prior to deposition.