Split AnnData object by levels in a specified meta data trait

splitAnnData splits a AnnData object by a given metadata column and write out the splitted data subsets into separate .h5ad files.

Usage

splitAnnData(anndata_file, split.by, levels, conda_env = "scicsr")

Arguments

anndata_file

filename pointing to the AnnData file.

split.by

column name in the metadata (i.e. in SeuratObj@meta.data considering the Seurat object, or adata.obs considering the AnnData object) indicating distinct levels to split the AnnData.

levels

vector of values which can be found in the column given by split.by. Subsets of the Anndata by each value of levels will be made and written out into .h5ad files.

conda_env

character, if not NULL this named conda environment is used to perform the split. (Default: 'scicsr'). If NULL, no conda environment will be used, the program assumes the python packages scanpy and scvelo are installed in the local python)

Value

a vector of output .h5ad filenames with the indicated subset of the data (given in the filenames, see examples.)

Details

splitAnnData uses the R reticulate package to run python commands which subset the data by the metadata column ( i.e. column in adata.obs) and write out each susbet as separate AnnData objects in .h5ad files.