API for disease cell atlas approximations

Disease cell atlases are single-cell omics datasets that provide insights into diseased tissues across multiple organs and conditions. A disease cell atlas approximation is a lightweight, lossy compression of such a cell atlas, retaining key disease-specific information while reducing data size and complexity. This API, initially sourced from the CellxGene Census, enables researchers to explore cell type and gene expression data between disease states, sexes, and other conditions. It helps answer biological questions such as:

What cell types are present in blood samples from adult patients with influenza across all datasets?
How does the proportion of alveolar type 2 cells in the lung change between healthy individuals and those with COVID-19
What are the top 20 most differentially expressed genes in kidney-related disease?

Version

The most recent version of the API is v1.

Interfaces

There are multiple ways to access atlas approximations programmatically:

Authors

This project is developed and maintained by: Fabio Zanini, Ying Xu

Affiliated with [Fabilab](https://fabilab.org).

Citation

Xu et al. (2024). Lightweight and scalable approximations democratise access to single cell atlases. bioRxiv. doi:10.1101/2024.01.03.573994.

Data sources

CZ CELLxGENE: Discover (Census). CZ CELLxGENE Discover: A single-cell data platform for scalable exploration, analysis and modeling of aggregated data CZI Single-Cell Biology, et al. bioRxiv 2023.10.30; doi: https://doi.org/10.1101/2023.10.30.563174