pegasus.regress_out

pegasus.regress_out(data, attrs, rep='pca', n_comps=None)[source]

Regress out effects due to specific observational attributes.

Parameters

  • data (MultimodalData or UnimodalData object) – Annotated data matrix with rows for cells and columns for genes.

  • attrs (List[str]) – List of numeric cell attributes to be regressed out. They must exist in data.obs field.

  • rep (str, optional, default: pca) – This is to specify which embedding to be used for regressing out. The key 'X_'+rep must exist in data.obsm field. By default, use PCA embedding.

  • n_comps (int, optional, default: None) – Number of components of the embedding to be used. By default, use all components.

Return type

str

Returns

  • res_key (str) – The key to the resulting new embedding matrix in data.obsm. It’s 'X_'+rep+'_regressed'.

  • Update data.obsm –

    • data.obsm[pca_key]: The PCA matrix with effects of attributes specified regressed out.

Examples

>>> pg.regress_out(data, attrs=['G1/S', 'G2/M'])