omicverse.pp.regress

Contents

omicverse.pp.regress#

omicverse.pp.regress(adata, *, keys=None, layer=None, n_jobs=8, **kwargs)[source]#

Regress out technical covariates from each gene.

Parameters:

  • adata (anndata.AnnData) – AnnData object. The columns specified in keys must exist in adata.obs.

  • keys (list of str, optional) – Column names in adata.obs to regress out. Defaults to ['mito_perc', 'nUMIs'], which are computed by ov.pp.qc. Common additions include cell-cycle scores (S_score, G2M_score) and ribosomal gene percentage.

  • layer (default: None) – Expression layer to regress. None uses adata.X.

  • n_jobs (default: 8) – Parallel jobs for the CPU regressor. Default 8.

  • **kwargs – Extra options forwarded to scanpy.pp.regress_out on CPU (e.g. copy). Ignored when the GPU backend (rapids_singlecell) is active.

Returns:

Writes the residualised expression to adata.layers['regressed'].

Return type:

None

Contents