cellxgene-census

Query large-scale single-cell datasets programmatically using the CELLxGENE Census to access cell metadata, gene metadata, and expression matrices across human and mouse tissues for cross-tissue analysis. It supports scalable queries and out-of-core processing for large-scale single-cell data retrieval.

Yes, you can integrate queried single-cell expression data with PyTorch workflows for machine learning experiments. The Census supports extracting expression matrices and metadata for direct use in ML pipelines, enabling cross-disease and cross-dataset analyses within PyTorch.

Retrieve gene-specific expression matrices by querying the Census X layers and var structures. This programmatic access allows you to extract targeted expression data across multiple datasets for downstream single-cell analysis.

Yes, the CELLxGENE Census works with Scanpy for downstream single-cell analysis. You can query cell metadata and expression matrices, then seamlessly integrate the retrieved data into Scanpy workflows for exploratory analysis and visualization.

Handle out-of-core processing for big single-cell queries by leveraging the Census TileDB-SOMA structures. This approach supports efficient memory management when querying large-scale single-cell datasets, allowing you to process data that exceeds available RAM.

List available datasets in the CELLxGENE Census by opening the Census with a context manager and running a simple metadata query. This exploratory metadata query reveals all accessible human and mouse single-cell datasets within the versioned collection.