cellxgene-census

Name: cellxgene-census
Author: LeonChaoX

41installs61forks

How to install cellxgene-census?

leonchaox/qinyan-academic-skills/cellxgene-census

$npx -y skills add LeonChaoX/qinyan-academic-skills --skill cellxgene-census

Installs into the current project.

Summary

Query the CELLxGENE Census (61M+ cells) programmatically. Use when you need expression data across tissues, diseases, or cell types from the largest curated single-cell atlas. Best for population-scale queries, reference atlas comparisons. For analyzing your own data use scanpy o

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 genomics data from CZ CELLxGENE Discover. This skill enables efficient querying and analysis of millions of cells across thousands of datasets.
6
7	The Census includes:
8	- 61+ million cells from human and mouse
9	- Standardized metadata (cell types, tissues, diseases, donors)
10	- Raw gene expression matrices
11	- Pre-calculated embeddings and statistics
12	- Integration with PyTorch, scanpy, and other analysis tools
13
14	## When to Use This Skill
15
16	This skill should be used when:
17	- Querying single-cell expression data by cell type, tissue, or disease
18	- Exploring available single-cell datasets and metadata
19	- Training machine learning models on single-cell data
20	- Performing large-scale cross-dataset analyses
21	- Integrating Census data with scanpy or other analysis frameworks
22	- Computing statistics across millions of cells
23	- Accessing pre-calculated embeddings or model predictions
24
25	## Installation and Setup
26
27	Install the Census API:
28	```bash
29	uv pip install cellxgene-census
30	```
31
32	For machine learning workflows, install additional dependencies:
33	```bash
34	uv pip install cellxgene-census[experimental]
35	```
36
37	## Core Workflow Patterns
38
39	### 1. Opening the Census
40
41	Always use the context manager to ensure proper resource cleanup:
42
43	```python
44	import cellxgene_census
45
46	# Open latest stable version
47	with cellxgene_census.open_soma() as census:
48	# Work with census data
49
50	# Open specific version for reproducibility
51	with cellxgene_census.open_soma(census_version="2023-07-25") as census:
52	# Work with census data
53	```
54
55	Key points:
56	- Use context manager (`with` statement) for automatic cleanup
57	- Specify `census_version` for reproducible analyses
58	- Default opens latest "stable" release
59
60	### 2. Exploring Census Information
61
62	Before querying expression data, explore available datasets and metadata.
63
64	Access summary information:
65	```python
66	# Get summary statistics
67	summary = census["census_info"]["summary"].read().concat().to_pandas()
68	print(f"Total cells: {summary['total_cell_count'][0]}")
69
70	# Get all datasets
71	datasets = census["census_info"]["datasets"].read().concat().to_pandas()
72
73	# Filter datasets by criteria
74	covid_datasets = datasets[datasets["disease"].str.contains("COVID", na=False)]
75	```
76
77	Query cell metadata to understand available data:
78	```python
79	# Get unique cell types in a tissue
80	cell_metadata = cellxgene_census.get_obs(
81	census,
82	"homo_sapiens",
83	value_filter="tissue_general == 'brain' and is_primary_data == True",
84	column_names=["cell_type"]
85	)
86	unique_cell_types = cell_metadata["cell_type"].unique()
87	print(f"Found {len(unique_cell_types)} cell types in brain")
88
89	# Count cells by tissue
90	tissue_counts = cell_metadata.groupby("tissue_general").size()
91	```
92
93	Important: Always filter for `is_primary_data == True` to avoid counting duplicate cells unless specifically analyzing duplicates.
94
95	### 3. Querying Expression Data (Small to Medium Scale)
96
97	For queries returning < 100k cells that fit in memory, use `get_anndata()`:
98
99	```python
100	# Basic query with cell type and tissue filters
101	adata = cellxgene_census.get_anndata(
102	census=census,
103	organism="Homo sapiens", # or "Mus musculus"
104	obs_value_filter="cell_type == 'B cell' and tissue_general == 'lung' and is_primary_data == True",
105	obs_column_names=["assay", "disease", "sex", "donor_id"],
106	)
107
108	# Query specific genes with multiple filters
109	adata = cellxgene_census.get_anndata(
110	census=census,
111	organism="Homo sapiens",
112	var_value_filter="feature_name in ['CD4', 'CD8A', 'CD19', 'FOXP3']",
113	obs_value_filter="cell_type == 'T cell' and disease == 'COVID-19' and is_primary_data == True",
114	obs_column_names=["cell_type", "tissue_general", "donor_id"],
115	)
116	```
117
118	Filter syntax:
119	- Use `obs_value_filter` for cell filtering
120	- Use `var_value_filter` for gene filtering
121	- Combine conditions with `and`, `or`
122	- Use `in` for multiple values: `tissue in ['lung', 'liver']`
123	- Select only needed columns with `obs_column_names`
124
125	Getting metadata separately:
126	```python
127	# Query cell metadata
128	cell_metadata = cellxgene_census.get_obs(
129	census, "homo_sapiens",
130	value_filter="disease == 'COVID-19' and is_primary_data == True",
131	column_names=["cell_type", "tissue_general", "donor_id"]
132	)
133
134	# Query gene metadata
135	gene_metadata = cellxgene_census.get_var(
136	census, "homo_sapiens",
137	value_filter="feature_name in ['CD4', 'CD8A']",
138	column_names=["feature_id", "feature_name", "feature_length"]
139	)
140	```
141
142	### 4. Large-Scale Q

People who liked this also viewed

6 picks

Type

cellxgene-census

41installs61forks

View on GitHub

How to install cellxgene-census?

leonchaox/qinyan-academic-skills/cellxgene-census

$npx -y skills add LeonChaoX/qinyan-academic-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 genomics data from CZ CELLxGENE Discover. This skill enables efficient querying and analysis of millions of cells across thousands of datasets.
6
7	The Census includes:
8	- 61+ million cells from human and mouse
9	- Standardized metadata (cell types, tissues, diseases, donors)
10	- Raw gene expression matrices
11	- Pre-calculated embeddings and statistics
12	- Integration with PyTorch, scanpy, and other analysis tools
13
14	## When to Use This Skill
15
16	This skill should be used when:
17	- Querying single-cell expression data by cell type, tissue, or disease
18	- Exploring available single-cell datasets and metadata
19	- Training machine learning models on single-cell data
20	- Performing large-scale cross-dataset analyses
21	- Integrating Census data with scanpy or other analysis frameworks
22	- Computing statistics across millions of cells
23	- Accessing pre-calculated embeddings or model predictions
24
25	## Installation and Setup
26
27	Install the Census API:
28	```bash
29	uv pip install cellxgene-census
30	```
31
32	For machine learning workflows, install additional dependencies:
33	```bash
34	uv pip install cellxgene-census[experimental]
35	```
36
37	## Core Workflow Patterns
38
39	### 1. Opening the Census
40
41	Always use the context manager to ensure proper resource cleanup:
42
43	```python
44	import cellxgene_census
45
46	# Open latest stable version
47	with cellxgene_census.open_soma() as census:
48	# Work with census data
49
50	# Open specific version for reproducibility
51	with cellxgene_census.open_soma(census_version="2023-07-25") as census:
52	# Work with census data
53	```
54
55	Key points:
56	- Use context manager (`with` statement) for automatic cleanup
57	- Specify `census_version` for reproducible analyses
58	- Default opens latest "stable" release
59
60	### 2. Exploring Census Information
61
62	Before querying expression data, explore available datasets and metadata.
63
64	Access summary information:
65	```python
66	# Get summary statistics
67	summary = census["census_info"]["summary"].read().concat().to_pandas()
68	print(f"Total cells: {summary['total_cell_count'][0]}")
69
70	# Get all datasets
71	datasets = census["census_info"]["datasets"].read().concat().to_pandas()
72
73	# Filter datasets by criteria
74	covid_datasets = datasets[datasets["disease"].str.contains("COVID", na=False)]
75	```
76
77	Query cell metadata to understand available data:
78	```python
79	# Get unique cell types in a tissue
80	cell_metadata = cellxgene_census.get_obs(
81	census,
82	"homo_sapiens",
83	value_filter="tissue_general == 'brain' and is_primary_data == True",
84	column_names=["cell_type"]
85	)
86	unique_cell_types = cell_metadata["cell_type"].unique()
87	print(f"Found {len(unique_cell_types)} cell types in brain")
88
89	# Count cells by tissue
90	tissue_counts = cell_metadata.groupby("tissue_general").size()
91	```
92
93	Important: Always filter for `is_primary_data == True` to avoid counting duplicate cells unless specifically analyzing duplicates.
94
95	### 3. Querying Expression Data (Small to Medium Scale)
96
97	For queries returning < 100k cells that fit in memory, use `get_anndata()`:
98
99	```python
100	# Basic query with cell type and tissue filters
101	adata = cellxgene_census.get_anndata(
102	census=census,
103	organism="Homo sapiens", # or "Mus musculus"
104	obs_value_filter="cell_type == 'B cell' and tissue_general == 'lung' and is_primary_data == True",
105	obs_column_names=["assay", "disease", "sex", "donor_id"],
106	)
107
108	# Query specific genes with multiple filters
109	adata = cellxgene_census.get_anndata(
110	census=census,
111	organism="Homo sapiens",
112	var_value_filter="feature_name in ['CD4', 'CD8A', 'CD19', 'FOXP3']",
113	obs_value_filter="cell_type == 'T cell' and disease == 'COVID-19' and is_primary_data == True",
114	obs_column_names=["cell_type", "tissue_general", "donor_id"],
115	)
116	```
117
118	Filter syntax:
119	- Use `obs_value_filter` for cell filtering
120	- Use `var_value_filter` for gene filtering
121	- Combine conditions with `and`, `or`
122	- Use `in` for multiple values: `tissue in ['lung', 'liver']`
123	- Select only needed columns with `obs_column_names`
124
125	Getting metadata separately:
126	```python
127	# Query cell metadata
128	cell_metadata = cellxgene_census.get_obs(
129	census, "homo_sapiens",
130	value_filter="disease == 'COVID-19' and is_primary_data == True",
131	column_names=["cell_type", "tissue_general", "donor_id"]
132	)
133
134	# Query gene metadata
135	gene_metadata = cellxgene_census.get_var(
136	census, "homo_sapiens",
137	value_filter="feature_name in ['CD4', 'CD8A']",
138	column_names=["feature_id", "feature_name", "feature_length"]
139	)
140	```
141
142	### 4. Large-Scale Q

People who liked this also viewed

6 picks

Type