The following cells were sorted for single-cell RNA-seq: gp33-tetramer+ live CD8+PD-1+CD44+ cells from spleens of day-28 LCMV Cl13-infected mice; SIINFEKL- tetramer+ live CD45+CD8+ cells from day-20 B16-OVA tumors and live PD-1+CD44+CD45+CD8+ cells from day-10 B16-OVA tumors. Cells were counted and loaded onto the Chromium Controller (10X Genomics) for a target recovery of 5,000 single cells. Samples were processed per the manufacturer’s protocol and sequenced on an Illumina NextSeq500 sequencer using a 75-bp kit with paired-end reads. Sample demultiplexing, barcode processing, alignment, filtering, UMI counting and aggregation of sequencing runs were performed using the Cell Ranger analysis pipeline (v.1.2). Downstream analyses were performed in R using the Seurat package44 (v.2.1.0).
For each cell, two quality control metrics were calculated: (1) the total number of genes detected, and (2) the proportion of UMIs contributed by mitochondrially encoded transcripts. Cells in which fewer than 200 genes were detected and in which mitochondrially encoded transcripts constituted greater than 10% of the total library were excluded from downstream analysis. Genes detected in fewer than three cells across the dataset were also excluded, yielding an expression matrix of 9,194 cells by 13,642 genes (LCMV), 11,212 cells by 14,496 genes (day 10 + 20 tumor), or 4,313 cells by 13,880 genes (day 20 tumor). Each gene expression measurement was normalized by total expression in the corresponding cell and multiplied by a scaling factor of 10,000. Mean and dispersion values were calculated for each gene across all cells; 1,653 genes (LCMV), 1,234 genes (day 10 + 20 tumor) or 914 genes (day 20 tumor) classified as highly variable. Highly variable genes were used for PCA. Principal components were determined to be significant (P < 0.01) using the jackstraw method, and tSNE was performed on these significant principle components using default parameters for 1,000 iterations for visualization in two dimensions. Unsupervised clustering was performed using a shared nearest neighbor modularity optimization-based algorithm45 . Differential expression analysis was performed between each cluster and all other cells using a Wilcoxon rank sum test. Single-cell signature scoring using FastProject46 (v.0.9.2) was performed with a curated version of the C7 Immune Signatures database from MSigDB. Significance of signature enrichment in single-cell datasets of one cluster relative to other clusters was determined using a Kolmogorov–Smirnov test. FastProject signature scoring of CD8+ T cells from tumors was performed in the scope of transcriptional signatures derived from the subpopulations we identified in LCMV or from signatures derived from the literature. Tumor cells were plotted on a two-dimensional axis on the basis of their scores for the terminally exhausted and progenitor exhausted signatures on the x axis and y axis, respectively.
For each cell, two quality control metrics were calculated: (1) the total number of genes detected, and (2) the proportion of UMIs contributed by mitochondrially encoded transcripts. Cells in which fewer than 200 genes were detected and in which mitochondrially encoded transcripts constituted greater than 10% of the total library were excluded from downstream analysis. Genes detected in fewer than three cells across the dataset were also excluded, yielding an expression matrix of 9,194 cells by 13,642 genes (LCMV), 11,212 cells by 14,496 genes (day 10 + 20 tumor), or 4,313 cells by 13,880 genes (day 20 tumor). Each gene expression measurement was normalized by total expression in the corresponding cell and multiplied by a scaling factor of 10,000. Mean and dispersion values were calculated for each gene across all cells; 1,653 genes (LCMV), 1,234 genes (day 10 + 20 tumor) or 914 genes (day 20 tumor) classified as highly variable. Highly variable genes were used for PCA. Principal components were determined to be significant (P < 0.01) using the jackstraw method, and tSNE was performed on these significant principle components using default parameters for 1,000 iterations for visualization in two dimensions. Unsupervised clustering was performed using a shared nearest neighbor modularity optimization-based algorithm45 . Differential expression analysis was performed between each cluster and all other cells using a Wilcoxon rank sum test. Single-cell signature scoring using FastProject46 (v.0.9.2) was performed with a curated version of the C7 Immune Signatures database from MSigDB. Significance of signature enrichment in single-cell datasets of one cluster relative to other clusters was determined using a Kolmogorov–Smirnov test. FastProject signature scoring of CD8+ T cells from tumors was performed in the scope of transcriptional signatures derived from the subpopulations we identified in LCMV or from signatures derived from the literature. Tumor cells were plotted on a two-dimensional axis on the basis of their scores for the terminally exhausted and progenitor exhausted signatures on the x axis and y axis, respectively.
