cellxgene-census

You can query single-cell RNA-seq data by cell type and tissue using the CELLxGENE Census, which allows programmatic searching across 61 million cells and retrieving specific expression data as AnnData objects without downloading the full dataset.

Yes, you can stream large-scale expression matrices out-of-core using the cellxgene-census library, enabling direct integration with machine learning workflows like training PyTorch models on single-cell data.

The best way to retrieve human lung tissue expression data for B cells is by querying the CELLxGENE Census with specific metadata filters, returning the subsetted expression matrix as an AnnData object.

Yes, the cellxgene-census library supports searching for disease-specific gene markers by filtering the standardized single-cell observations across over 61 million cells using disease metadata.

You access population-scale single-cell datasets by querying the CELLxGENE Census programmatically, which standardizes over 61 million observations for bioinformatics research and ML integration.

Out-of-core streaming for single-cell RNA-seq queries requires the cellxgene-census library and is designed for large-scale data retrieval, though specific limitations depend on the memory constraints of your local machine.