omicverse.single.get_markers

omicverse.single.get_markers(adata, n_genes=10, key='rank_genes_groups', groups=None, return_type='dataframe', min_logfoldchange=None, min_score=None, min_pval_adj=None)[source]

Extract top marker genes from rank_genes_groups results.

Works with results produced by find_markers(), omicverse.single.cosg(), or sc.tl.rank_genes_groups.

Parameters:

  • adata (Annotated data matrix.)

  • n_genes (Maximum number of top marker genes to return per group.) – Default: 10.

  • key (Key in adata.uns holding the results.) – Default: 'rank_genes_groups'.

  • groups (One or more group names to extract. Accepts:) –

    • None — all groups (default).

    • '0' — single group as a string.

    • ['0', '1', '3'] — explicit list of groups.

  • return_type (Output format:) –

    • 'dataframe' — pandas.DataFrame with columns [group, rank, names, scores, logfoldchanges, pvals, pvals_adj] (columns present depend on which stats were computed).

    • 'dict' — {group: [gene1, gene2, ...]} mapping.

    Default: 'dataframe'.

  • min_logfoldchange (Keep only genes with) – |logfoldchange| ≥ min_logfoldchange. Default: None.

  • min_score (Keep only genes with score ≥ min_score. Default: None.)

  • min_pval_adj (Keep only genes with pvals_adj ≤ min_pval_adj.) – Default: None.

Return type:

Union[DataFrame, dict]

Returns:

  • pandas.DataFrame or dict depending on return_type.

  • Examples – >>> import omicverse as ov >>> ov.single.find_markers(adata, groupby=’leiden’) >>> # All clusters >>> df = ov.single.get_markers(adata, n_genes=10) >>> # Single cluster >>> df0 = ov.single.get_markers(adata, n_genes=5, groups=’0’) >>> # As dict for annotation workflows >>> d = ov.single.get_markers(adata, n_genes=5, return_type=’dict’)