omicverse.pp.qc

omicverse.pp.qc(adata, **kwargs)[source]

Perform quality control on a dictionary of AnnData objects.

Parameters:

  • adata – AnnData object

  • mode – The filtering method to use. Valid options are ‘seurat’

  • 'seurat'. (and 'mads'. Default is)

  • min_cells – The minimum number of cells for a sample to pass QC. Default is 3.

  • min_genes – The minimum number of genes for a cell to pass QC. Default is 200.

  • max_cells_ratio – The maximum number of cells ratio for a sample to pass QC. Default is 1.

  • max_genes_ratio – The maximum number of genes ratio for a cell to pass QC. Default is 1.

  • nmads – The number of MADs to use for MADs filtering. Default is 5.

  • doublets – Whether to perform doublet detection. Default is True.

  • doublets_method – The doublet detection method to use. Options are ‘scrublet’ or ‘sccomposite’. Default is ‘scrublet’.

  • filter_doublets – Whether to filter out doublets (True) or just flag them (False). Default is True.

  • path_viz – The path to save the QC plots. Default is None.

  • tresh – A dictionary of QC thresholds. The keys should be ‘mito_perc’,

  • 'nUMIs' – Only used if mode is ‘seurat’. Default is None.

  • 'detected_genes'. (and) – Only used if mode is ‘seurat’. Default is None.

  • mt_startswith – The prefix of mitochondrial genes. Default is ‘MT-‘.

  • mt_genes – The list of mitochondrial genes. Default is None.

  • None (if mt_genes is not)

  • ignored. (mt_startswith will be)

Returns:

An AnnData object containing cells that passed QC filters.

Return type:

adata

Examples

>>> import omicverse as ov
>>> adata = ov.pp.qc(adata, tresh={'mito_perc': 0.2, 'nUMIs': 500, 'detected_genes': 250})
>>> adata = ov.pp.qc(adata, mode='mads', nmads=5, doublets=True)