cellxgene-census

Name: cellxgene-census
Author: K-Dense-AI

808installs3.1kforks

How to install cellxgene-census?

k-dense-ai/scientific-agent-skills/cellxgene-census

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill cellxgene-census

Installs into the current project.

Summary

Query the CZ CELLxGENE Census programmatically for versioned public single-cell and spatial transcriptomics data. Use when you need population-scale cell metadata, gene expression slices, Census summary counts, source H5AD URIs/downloads, embeddings, spatial Census data, or refer

Files · 1

View on GitHub

SKILL.md

1	# CZ CELLxGENE Census
2
3	## Overview
4
5	The CZ CELLxGENE Census provides programmatic access to a comprehensive, versioned collection of standardized single-cell and spatial transcriptomics data from CZ CELLxGENE Discover. This skill enables efficient querying and analysis of public Census releases without downloading whole datasets first.
6
7	The Census includes:
8	- 217+ million total cells and 125+ million unique cells in the 2025-11-08 stable LTS release
9	- 1,845 datasets in the 2025-11-08 stable LTS release
10	- Human, mouse, marmoset, rhesus macaque, and chimpanzee data in the current schema
11	- Standardized metadata (cell types, tissues, diseases, donors)
12	- Raw gene expression matrices and source H5AD lookup/download helpers
13	- Pre-calculated summary counts, embeddings, and spatial data
14	- Integration with AnnData, Scanpy, TileDB-SOMA, TileDB-SOMA-ML, and other analysis tools
15
16	## When to Use This Skill
17
18	This skill should be used when:
19	- Querying single-cell expression data by cell type, tissue, or disease
20	- Exploring available single-cell datasets and metadata
21	- Training machine learning models on single-cell data
22	- Performing large-scale cross-dataset analyses
23	- Integrating Census data with scanpy or other analysis frameworks
24	- Computing statistics across millions of cells
25	- Accessing pre-calculated embeddings or model predictions
26
27	## Installation and Setup
28
29	Install the Census API:
30	```bash
31	uv pip install "cellxgene-census==1.17.*"
32	```
33
34	For spatial workflows:
35	```bash
36	uv pip install "cellxgene-census[spatial]==1.17.*" "spatialdata[extra]>=0.2.5"
37	```
38
39	For PyTorch model training, use TileDB-SOMA-ML. The old `cellxgene_census.experimental.ml` loaders are deprecated:
40
41	```bash
42	uv pip install "cellxgene-census==1.17.*" tiledbsoma-ml
43	```
44
45	## Core Workflow Patterns
46
47	### 1. Opening the Census
48
49	Always use the context manager to ensure proper resource cleanup:
50
51	```python
52	import cellxgene_census
53
54	# Open latest stable version
55	with cellxgene_census.open_soma() as census:
56	# Work with census data
57
58	# Open the current LTS version for reproducibility
59	with cellxgene_census.open_soma(census_version="2025-11-08") as census:
60	# Work with census data
61	```
62
63	Key points:
64	- Use context manager (`with` statement) for automatic cleanup
65	- Specify `census_version` for reproducible analyses
66	- `stable` opens the current LTS Census release; `latest` opens the newest weekly release retained for a shorter period
67
68	### 2. Exploring Census Information
69
70	Before querying expression data, explore available datasets and metadata.
71
72	Access summary information:
73	```python
74	# Get summary statistics as label/value rows
75	summary = census["census_info"]["summary"].read().concat().to_pandas()
76	summary_values = summary.set_index("label")["value"]
77	print(f"Total cells: {int(summary_values['total_cell_count']):,}")
78	print(f"Unique cells: {int(summary_values['unique_cell_count']):,}")
79
80	# Get all datasets
81	datasets = census["census_info"]["datasets"].read().concat().to_pandas()
82
83	# Get precomputed counts by organism, cell type, tissue, disease, and assay
84	summary_counts = census["census_info"]["summary_cell_counts"].read().concat().to_pandas()
85	tissue_counts = summary_counts[summary_counts["category"].eq("tissue_general")]
86	```
87
88	Query cell metadata to understand available data:
89	```python
90	# Get unique cell types in a tissue
91	cell_metadata = cellxgene_census.get_obs(
92	census,
93	"homo_sapiens",
94	value_filter="tissue_general == 'brain' and is_primary_data == True",
95	column_names=["cell_type"]
96	)
97	unique_cell_types = cell_metadata["cell_type"].unique()
98	print(f"Found {len(unique_cell_types)} cell types in brain")
99
100	# Count cells by tissue
101	tissue_metadata = cellxgene_census.get_obs(
102	census,
103	"homo_sapiens",
104	value_filter="is_primary_data == True",
105	column_names=["tissue_general"],
106	)
107	tissue_counts = tissue_metadata["tissue_general"].value_counts()
108	```
109
110	Important: Always filter for `is_primary_data == True` to avoid counting duplicate cells unless specifically analyzing duplicates.
111
112	### 3. Querying Expression Data (Small to Medium Scale)
113
114	For queries returning < 100k cells that fit in memory,

People who liked this also viewed

6 picks

Type

cellxgene-census

808installs3.1kforks

View on GitHub

How to install cellxgene-census?

k-dense-ai/scientific-agent-skills/cellxgene-census

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill cellxgene-census

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# CZ CELLxGENE Census
2
3	## Overview
4
5	The CZ CELLxGENE Census provides programmatic access to a comprehensive, versioned collection of standardized single-cell and spatial transcriptomics data from CZ CELLxGENE Discover. This skill enables efficient querying and analysis of public Census releases without downloading whole datasets first.
6
7	The Census includes:
8	- 217+ million total cells and 125+ million unique cells in the 2025-11-08 stable LTS release
9	- 1,845 datasets in the 2025-11-08 stable LTS release
10	- Human, mouse, marmoset, rhesus macaque, and chimpanzee data in the current schema
11	- Standardized metadata (cell types, tissues, diseases, donors)
12	- Raw gene expression matrices and source H5AD lookup/download helpers
13	- Pre-calculated summary counts, embeddings, and spatial data
14	- Integration with AnnData, Scanpy, TileDB-SOMA, TileDB-SOMA-ML, and other analysis tools
15
16	## When to Use This Skill
17
18	This skill should be used when:
19	- Querying single-cell expression data by cell type, tissue, or disease
20	- Exploring available single-cell datasets and metadata
21	- Training machine learning models on single-cell data
22	- Performing large-scale cross-dataset analyses
23	- Integrating Census data with scanpy or other analysis frameworks
24	- Computing statistics across millions of cells
25	- Accessing pre-calculated embeddings or model predictions
26
27	## Installation and Setup
28
29	Install the Census API:
30	```bash
31	uv pip install "cellxgene-census==1.17.*"
32	```
33
34	For spatial workflows:
35	```bash
36	uv pip install "cellxgene-census[spatial]==1.17.*" "spatialdata[extra]>=0.2.5"
37	```
38
39	For PyTorch model training, use TileDB-SOMA-ML. The old `cellxgene_census.experimental.ml` loaders are deprecated:
40
41	```bash
42	uv pip install "cellxgene-census==1.17.*" tiledbsoma-ml
43	```
44
45	## Core Workflow Patterns
46
47	### 1. Opening the Census
48
49	Always use the context manager to ensure proper resource cleanup:
50
51	```python
52	import cellxgene_census
53
54	# Open latest stable version
55	with cellxgene_census.open_soma() as census:
56	# Work with census data
57
58	# Open the current LTS version for reproducibility
59	with cellxgene_census.open_soma(census_version="2025-11-08") as census:
60	# Work with census data
61	```
62
63	Key points:
64	- Use context manager (`with` statement) for automatic cleanup
65	- Specify `census_version` for reproducible analyses
66	- `stable` opens the current LTS Census release; `latest` opens the newest weekly release retained for a shorter period
67
68	### 2. Exploring Census Information
69
70	Before querying expression data, explore available datasets and metadata.
71
72	Access summary information:
73	```python
74	# Get summary statistics as label/value rows
75	summary = census["census_info"]["summary"].read().concat().to_pandas()
76	summary_values = summary.set_index("label")["value"]
77	print(f"Total cells: {int(summary_values['total_cell_count']):,}")
78	print(f"Unique cells: {int(summary_values['unique_cell_count']):,}")
79
80	# Get all datasets
81	datasets = census["census_info"]["datasets"].read().concat().to_pandas()
82
83	# Get precomputed counts by organism, cell type, tissue, disease, and assay
84	summary_counts = census["census_info"]["summary_cell_counts"].read().concat().to_pandas()
85	tissue_counts = summary_counts[summary_counts["category"].eq("tissue_general")]
86	```
87
88	Query cell metadata to understand available data:
89	```python
90	# Get unique cell types in a tissue
91	cell_metadata = cellxgene_census.get_obs(
92	census,
93	"homo_sapiens",
94	value_filter="tissue_general == 'brain' and is_primary_data == True",
95	column_names=["cell_type"]
96	)
97	unique_cell_types = cell_metadata["cell_type"].unique()
98	print(f"Found {len(unique_cell_types)} cell types in brain")
99
100	# Count cells by tissue
101	tissue_metadata = cellxgene_census.get_obs(
102	census,
103	"homo_sapiens",
104	value_filter="is_primary_data == True",
105	column_names=["tissue_general"],
106	)
107	tissue_counts = tissue_metadata["tissue_general"].value_counts()
108	```
109
110	Important: Always filter for `is_primary_data == True` to avoid counting duplicate cells unless specifically analyzing duplicates.
111
112	### 3. Querying Expression Data (Small to Medium Scale)
113
114	For queries returning < 100k cells that fit in memory,

People who liked this also viewed

6 picks

Type