A Single-Cell Tumor Immune Atlas for Precision Oncology

<strong>Preprint version of the Single-Cell Tumor Immune Atlas</strong> This upload contains: <strong>TICAtlas.rds:</strong> an rds file containing a Seurat object with the whole Atlas (317111 cells, RNA and integrated assays, PCA and UMAP reductions) <strong>TICAtlas.h5ad:</strong> an h5ad file with the whole Atlas (317111 cells, RNA assay, PCA and UMAP) <strong>TICAtlas_RNA.rds:</strong> an rds file containing a Seurat object of the whole Atlas but only the RNA assay (317111 cells, UMAP embedding) <strong>TICAtlas_downsampled_1000.rds:</strong> an rds file containing a downsampled version of the Seurat object of the whole Atlas (24834 cells, RNA and integrated assay, PCA and UMAP reductions) <strong>TICAtlas_downsampled_1000.h5ad:</strong> an rds file containing a downsampled version of the Seurat object of the whole Atlas (24834 cells, RNA assay, PCA and UMAP reductions) <strong>TICAtlas_metadata.csv: </strong>a comma-separated text file with the metadata for each of the cells For the <strong>h5ad</strong> files, the .X slot contains the normalized data, while the .X.raw slot contains the raw counts as they were in the original datasets. All the files contain the following patient/sample metadata variables: patient: assigned patient identifiers gender: the patient's gender (male/female/unknown) source: dataset of origin subtype: cancer type (abbreviations as indicated in the preprint) cluster_kmeans_k6: patients clusters, NA if filtered out cell_type: annotated cell type for each of the cells If you have any issues with the metadata you can use the <strong>TICAtlas_metadata.csv </strong>file. For more information, read our preprint and check our GitHub. h5ad files can be read with Python using Scanpy, rds files can be read in R using Seurat. For format conversion between AnnData and Seurat we recommend SeuratDisk. For other single-cell data formats you can use sceasy.