omicverse.pp.highly_variable_genes

omicverse.pp.highly_variable_genes#

omicverse.pp.highly_variable_genes(adata, *, layer=None, n_top_genes=None, min_disp=0.5, max_disp=None, min_mean=0.0125, max_mean=3, span=0.3, n_bins=20, flavor='seurat', subset=False, inplace=True, batch_key=None, filter_unexpressed_genes=None, check_values=True, **kwargs)[source]#

Annotate highly variable genes (Satija 2015 / Zheng 2017 / Stuart 2019).

Parameters:

  • adata (anndata.AnnData) – Expression matrix. Expects log-normalised data except for flavor='seurat_v3' / 'seurat_v3_paper' which expect counts.

  • layer (str | None (default: None)) – Layer to use. None uses adata.X.

  • n_top_genes (int | None (default: None)) – Number of HVGs to keep. Required for flavor='seurat_v3'.

  • min_disp (float (default: 0.5)) – Dispersion / mean cutoffs for flavor='seurat' and flavor='cell_ranger'.

  • max_disp (default: None) – Dispersion / mean cutoffs for flavor='seurat' and flavor='cell_ranger'.

  • min_mean (float (default: 0.0125)) – Dispersion / mean cutoffs for flavor='seurat' and flavor='cell_ranger'.

  • max_mean (float (default: 3)) – Dispersion / mean cutoffs for flavor='seurat' and flavor='cell_ranger'.

  • span (float (default: 0.3)) – Loess smoothing span for flavor='seurat_v3'.

  • n_bins (int (default: 20)) – Dispersion normalisation bins.

  • flavor (str (default: 'seurat')) – 'seurat' (default), 'cell_ranger', 'seurat_v3', or 'seurat_v3_paper'.

  • subset (bool (default: False)) – If True, subset the AnnData to the HVGs in place.

  • inplace (bool (default: True)) – Write statistics into adata.var rather than returning a DataFrame.

  • batch_key (str | None (default: None)) – If set, HVG detection is done per batch.

  • filter_unexpressed_genes (bool | None (default: None)) – Drop genes with zero total expression before selection.

  • check_values (bool (default: True)) – Input validation (log-normalised vs counts).

  • **kwargs – Extra options forwarded to omicverse.pp._highly_variable_genes.highly_variable_genes.

Returns:

Mutates adata.var; also returns the stats DataFrame unless inplace=True.

Return type:

pandas.DataFrame | None

Contents