Utils functions

vega.utils.setup_anndata(adata, batch_key=None, categorical_covariate_keys=None, copy=False)[source]

Creates VEGA fields in input Anndata object for mask. Also creates SCVI field which will be used for batch and covariates.

Parameters

  • adata (AnnData) – Scanpy single-cell object

  • copy (bool) – Whether to return a copy or change in place

  • batch_key (Optional[str]) – Observation to be used as batch

  • categorical_covariate_keys (Union[str, list, None]) – Observation to use as covariate keys

Returns

updated object if copy is True

Return type

adata

vega.utils.create_mask(adata, gmt_paths=None, add_nodes=1, min_genes=0, max_genes=1000, copy=False)[source]

Initialize mask M for GMV from one or multiple .gmt files.

Parameters

  • adata (AnnData) – Scanpy single-cell object.

  • gmt_paths (Union[str, list, None]) – One or several paths to .gmt files.

  • add_nodes (int) – Additional latent nodes for capturing additional variance.

  • min_genes (int) – Minimum number of genes per GMV.

  • max_genes (int) – Maximum number of genes per GMV.

  • copy (bool) – Whether to return a copy of the updated Anndata object.

Returns

Scanpy single-cell object.

Return type

adata

vega.utils.preprocess_anndata(adata, n_top_genes=5000, copy=False)[source]

Simple (default) Scanpy preprocessing function before autoencoders.

Parameters

  • adata (AnnData) – Scanpy single-cell object

  • n_top_genes (int) – Number of highly variable genes to retain

  • copy (bool) – Return a copy or in place

Returns

Preprocessed Anndata object

Return type

adata