Single Cell CPTAC RCC

These data include a subset of single-cell samples from the CPTAC Renal Cell Carcinoma data processed using the following steps:

Loom files were read into R and converted into SingleCellExperiment objects.

Ensembl gene ID's were matched to HGNC symbols, chromosome name, starting position, and ending position via Biomart using the scater R package.

Size factors were computed using the scran R package.

UMAP dimensions were computed using the scater R package.

Probes without matching HGNC symbols were removed.

Where duplicate HGNC symbols were present, the gene with the maximum normalized range was retained.

Cell types were inferred using the scMRMA R package.

Cells with less than or equal to 1,000 features were removed.

Cells with mitochondrial reads greater than or equal to 50% were removed.

Expression data for podocytes and macrophages were saved separately for each sample.

For podocytes and macrophages for each sample, expression data were projected onto the first 100 principal components using the irlba R package. The results are available from CPTAC_RCC_PCA.zip.

For macrophages for each sample, a differentiation trajectory was estimated using the monocle3 R package, and plots were colored by combined expression of the M0 markers CSF1R, CD14, CD68, and CD11B, the M1 markers CD86, MARC0, CXCL9, CXCL10, CXCL11, NOS2, SOCS1, and CD64, and the M2 markers TGM2, CD23, ARG1, CCL22, CD163, and CD206 (from PMC8268869). Pseudotime starting points were annotated in monocle3 using visual inspection of plots. Only samples in which a visible trajectory from M0 -> M1 -> M2 was evident using these markers were retained.

For podocytes for each sample, a differentiation trajectory was estimated using the monocle3 R package, and plots were colored by combined expression of the dedifferentiation markers DACH1 (from PMC5908116) and PTPRO (from PMID9639039. Pseudotime starting points were annotated in monocle3 using visual inspection of plots. Only samples in which a visible trajectory of dedifferentiation was evident using these markers were retained. These pseudotime assignments and the macrophage assignments are available from CPTAC_RCC_pseudotime_all_cells.zip.

For podocytes and macrophages for each sample, 10-fold cross-validation matrices for expression, PCA, and pseudotime were generated across 5 random splits, for a total of 50 files per cell type, per sample. These files are available from CPTAC_RCC_expression_all_cells.zip.

Expression data were subset to include 63 randomly-selected podocytes and 63 randomly-selected macrophages to ensure balanced data. These expression data are available from CPTAC_RCC_expression.zip. Pseudotimes were also subset and are available in CPTAC_RCC_pseudotime.zip.

Expression data were projected onto the first 100 principal components. These data are available from CPTAC_RCC_expression_dimReduced.zip.