Vascular access of CD62L⁺ stem-like exhausted T cells overcomes resistance to PD-1 blockade by enabling intratumoral generation of cytolytic effectors

================================================================================PROCESSED SINGLE-CELL RNA-SEQ DATAOT-I CD8 T cells in B16-OVA and Pan02-OVA tumor models================================================================================ Provided for peer review of the manuscript: "Vascular access of CD62L+ stem-like exhausted T cells overcomes resistance to PD-1 blockade by enabling intratumoral generation of cytolytic effectors" Unpublished data - please do not redistribute. --------------------------------------------------------------------------------1. CONTENTS-------------------------------------------------------------------------------- 1.1 Cell Ranger filtered count matrices (Matrix Market format, one set per 10x GEM well) JH1_matrix.mtx.gz, JH1_barcodes.tsv.gz, JH1_features.tsv.gz B16-OVA, GEM well 1 JH2_matrix.mtx.gz, JH2_barcodes.tsv.gz, JH2_features.tsv.gz B16-OVA, GEM well 2 JH3_matrix.mtx.gz, JH3_barcodes.tsv.gz, JH3_features.tsv.gz Pan02-OVA Each matrix contains both Gene Expression and Antibody Capture (TotalSeq-B hashtag) features, as produced by Cell Ranger. 1.2 Annotated AnnData objects Analyzed_annotated_data.h5ad Final integrated, equalized and annotated dataset (22,706 cells). Preprocessed_pooled_data.h5ad Preprocessed, BBKNN-integrated object (35,002 cells). Input for notebook: Figure2_S2_5 B16-OVA_MNN_cleaned.h5ad Panc-OVA_MNN_cleaned.h5ad Preprocessed, MNN-integrated objects for each tumor experiment individually. Input for notebook: Figure5A 1.3 RNA velocity files B16OVA_Velocity.loom Panc-OVA_Velocity.loom B16OVA_cleaned_OrganID PancOVA_cleaned_OrganID velocyto output generated from the raw reads of the respective experiment. And underlying preprocessed and clustered gene expression UMAPs. The loom files of both B16-OVA runs (JH1, JH2) were concatenated. Input for notebooks: Figure5B/C 1.4 Per-cell metadata OTI_cell_metadata_balanced.tsv.gz One row per cell: GEM well, original 10x barcode, tissue (hashtag-derived), tumor model, Leiden cluster, UMAP coordinates. 1.5 Analysis notebooks Jupyter notebooks reproducing the figure panels from the objects above. Figure 4L is reproduced from published data by Yang et al. --------------------------------------------------------------------------------2. EXPERIMENTAL DESIGN-------------------------------------------------------------------------------- CD45.1+ PD-1-high Tim-3-low CD8 OT-I T cells were sorted from tumor,tumor-draining lymph node and spleen of B16-OVA or Pan02-OVA bearing mice.Cells from each tissue were labelled with distinct TotalSeq-B hashtagantibodies (BioLegend), pooled, and loaded onto a 10x Chromium chip (two wells for B16-OVA and one well for Panc-OVA). Tissue identity is recovered by hashtag demultiplexing; the "tissue" columnof the cell metadata table carries the resulting assignment. --------------------------------------------------------------------------------3. DATA PROCESSING-------------------------------------------------------------------------------- Library preparation Chromium Next GEM Single Cell 3' v3.1 with Feature Barcode technology for cell surface protein (10x Genomics). Sequencing JH1, JH2 Illumina NovaSeq X, paired-end 150 bp JH3 Illumina NovaSeq 6000, paired-end 150 bp Alignment and counting Cell Ranger 7.1.0, against the refdata-gex-mm10-2020-A mouse reference, with a feature reference describing the TotalSeq-B hashtag sequences. Downstream analysis (Scanpy) - Hashtag demultiplexing by Gaussian mixture model; HTO-low cells and doublets excluded. - QC thresholds set independently per experiment. - Ribosomal (Rps/Rpl), mitochondrial (mt-) and melanoma-derived ambient RNA genes excluded from HVG selection and differential expression testing. - Cell numbers equalized between experiments. - Batch-aware integration with BBKNN, followed by Leiden clustering and UMAP. - DEG scoring, visualizations and signature scorings - RNA velocity with velocyto and scVelo. -------------------------------------------------------------------------------- Analysis code is included in this Zenodo record. --------------------------------------------------------------------------------5. CONTACT-------------------------------------------------------------------------------- Anton MuehlbauerTechnical University of [e-mail] (c) 2026 Technical University of Munich.