Single-Cell Matrix Service

Build a custom expression matrix from a LungMAP single-cell atlas. Choose an atlas, narrow it to the cells you want by cell type, library or QC covariate, then take away either every gene for those cells as an .h5ad, or a named gene set as a ready-to-cluster matrix. Values are raw counts throughout; no normalized matrix is stored or returned. Controlled terms (disease, cell states), underlying studies and statistical methods are defined here.

Species

Narrow the cells

Select an atlas to load its covariates.

—
cells selected
Choose an atlas to begin.

Build custom data matrix (.h5ad)

The selected cells and every gene, raw counts in X, with every metadata column in obs. Opens directly in scanpy or anndata.

Metadata

select an atlas first

Find novel cell states

Builds a UMAP for entered and optionally correlated genes or AI search term using Leiden clustering for one selected cell type and datasets.

no genes entered