snap-stanford / UCE

UCE is a zero-shot foundation model for single-cell gene expression data
MIT License
120 stars 15 forks source link

Input data of processed data #29

Closed fengzhanying closed 3 months ago

fengzhanying commented 3 months ago

Hi, could you please provide the gene expression matrix that generated the IMA embedding file "IMA_sample.h5ad"? or a guideline of how to retrieve it from somewhere?

Yanay1 commented 3 months ago

Hello,

Unfortunately since it contains many different datasets and species we cannot come up with a common gene set for that anndata.

You can use https://cellxgene.cziscience.com/census-models to query UCE embeddings and gene corresponding gene expression from cell x gene which composed the majority of IMA (33 M / 36 M cells) for human and mouse.