CZ CELLxGENE Discover

Overview

CZ CELLxGENE Discover is a free open-access data portal for curated, interoperable single-cell transcriptomics data, operated by the Chan Zuckerberg Initiative (CZI). Launched in 2021, it hosts community-contributed datasets standardised to a common data schema: all data must be deposited in AnnData (h5ad) format, with cell type annotations mapped to Cell Ontology, tissue annotations mapped to UBERON, and organism-level metadata standardised for interoperability. As of October 2024, the portal hosts over 1,550 datasets comprising 169.3 million cells (93.6 million unique after deduplication).

The platform has been adopted as a primary data sharing destination by major consortia including the NIH BICAN (Brain Initiative Cell Atlas Network), the Human Cell Atlas (HCA), and the Human Tumor Atlas Network (HTAN).

Key Features

The Explorer provides interactive single-dataset visualisation (UMAP, gene expression colouring, cluster annotation). Gene Expression provides cross-corpus views of gene expression by cell type and tissue across all datasets in the portal. Census is a programmatic API (Python and R) providing access to a concatenated, query-optimised snapshot of all non-spatial data in the portal, enabling large-scale computational analyses without downloading individual datasets. The Census July 2024 release contains over 44 million primary human cells and 16 million mouse cells.

Data Requirements

All submissions must conform to the CELLxGENE schema, which mandates Cell Ontology for cell type annotations, UBERON for tissue/organ terms, NCBI Taxonomy for organism, and AnnData as the container format. Raw counts must be included alongside normalised matrices. Spatial transcriptomics datasets are accepted with coordinate metadata. These requirements make CELLxGENE one of the strictest curation environments in single-cell data sharing and the largest source of Cell Ontology-annotated data.

Connections

• Data format: AnnData
• Cell type vocabulary: Cell Ontology
• Anatomy vocabulary: UBERON

Resources

• https://cellxgene.cziscience.com (portal)
• https://chanzuckerberg.github.io/cellxgene-census/ (Census API documentation)
• https://github.com/chanzuckerberg/cellxgene (source code)
• https://doi.org/10.1093/nar/gkae1142 (Abdulla et al. 2024, Nucleic Acids Research)
• https://cellxgene.cziscience.com/docs/04__Cite%20cellxgene%20discover/4_0__citation-policy (data citation policy)