single_cell_data

`cell_metadata.csv.gz`

CSV file containing the metadata associated with each cell, including cell identifiers, cell area, stain intensities, and transcript QC metrics when calculated. Protein intensities are stored separately in cell_by_protein.csv.gz. This can be used to build single-cell analysis objects (Seurat/Scanpy) and perform downstream analyses. For more information, see data import.

name	type	description
`cell_id`	str	Unique cell identifier derived from the segmentation mask
`sample_id`	str	Sample identifier
`tissue_type`	str	Tissue type from sample metadata
`block`	str	Tissue block identifier from sample metadata
`seg_source`	str	Segmentation source used to generate the cell
`cell_x/y`	float	x/y coordinate of the cell centroid
`cell_area_um`	float	Area of the expanded cell segmentation in µm²
`nuclei_area_um`	float	Area of the nuclear segmentation in µm²
`total_counts`	int	Total transcript counts
`log1p_total_counts`	float	Log number of total transcripts
`n_genes_by_counts`	int	Number of unique genes detected
`log1p_n_genes_by_counts`	float	Log number of unique genes detected
`total_counts_ctrl`	int	Total control probe counts
`log1p_total_counts_ctrl`	float	Log number of control probe counts
`pct_counts_ctrl`	float	Percentage of counts assigned to control probes
`*stain_intensity_mean`	float	Mean intensity for nuclear and cytoplasmic stains
`<protein>_intensity_mean`	float	Mean intensity for a given protein

`cell_by_gene.csv.gz`

CSV file containing the cell x gene matrix. Each entry in the table is the count for a given gene in a given cell. Control probes are included as additional columns when present.

`clustering_umap.csv.gz`

CSV file containing pre-computed cluster annotations and UMAP coordinates for cells retained for analysis.

cell_id, UMAP1, and UMAP2 are required by the current Helpers validator. Cluster annotation columns depend on the analysis; current Helpers defaults are leiden_coarse and leiden_fine. They are not universal required column names.

`dgex.csv.gz`

CSV file containing differential gene expression (DGEx) results for the pre-computed clusters from the clustering_umap table.

All columns below are required by the current Helpers DGEx schema.

column	type	description
`gene_id`	str	Gene identifier
`score`	float	Z-score from Wilcoxon rank-sum test
`logfoldchange`	float	LogFoldChange for the given cluster compared to all other clusters combined
`pval`	float	P-value from Wilcoxon rank-sum test
`pval_adj`	float	Adjusted P-value
`pct_nz_group`	float	Fraction of cells with non-zero expression in the given cluster (0–1; e.g. 0.933 means 93.3%).
`pct_nz_reference`	float	Fraction of cells with non-zero expression outside the given cluster (0–1).
`cluster_id`	str	Leiden cluster identity
`leiden_res`	str	Name of the clustering annotation used for this result, such as `leiden_coarse` or `leiden_fine`; this is not a numeric resolution. Each result refers to a gene and cluster within that annotation.

`sc_processed.h5ad`

AnnData file containing filtered and processed RNA data with cell and gene metadata. The retained data is used to call clusters and calculate differential gene expression. Filtering is likely to remove cells and genes present in the raw CSV tables; please refer to data import to learn how to build an unprocessed/raw AnnData object.
When preprocessing completes, .X contains RNA values normalized by total counts per cell and transformed with log1p. .layers["counts"] contains raw counts for the same retained cells and genes. PCA, UMAP, and clustering results are included when the corresponding processing steps succeed. Outputs saved after an incomplete processing run may lack some layers or annotations. For multiomics data, .obsm["protein"] stores mean protein intensities for the retained cells, with column names in .uns["protein_names"]. These measurements are separate from the RNA matrix.

`cell_by_protein.csv.gz`

CSV file containing cell x protein intensity matrix. Each entry in the table is the average protein intensity for a given protein in a cell. These columns are separate from the stain intensity columns in cell_metadata.csv.gz which is provided for all runs.

`rna_protein_sc_correlation.csv`

CSV file containing a square correlation matrix between selected transcript counts and protein intensity means, computed per cell. Useful for quickly checking concordance between RNA and protein markers (e.g. CD4 counts vs CD4_intensity_mean).

`protein_sc_correlation.csv`

CSV file containing a protein-by-protein correlation matrix of per-cell intensity means, allowing assessment of protein marker co-expression.