scviva.tools.harreman.tools.compute_interaction_module_correlation#

scviva.tools.harreman.tools.compute_interaction_module_correlation(adata, cor_method='pearson', test=None, interaction_type='metabolite', only_sig_values=False, normalize_values=False, use_FDR=True, use_super_modules=False, ct_aware=None)[source]#

Compute correlations between interacting cell scores and module scores.

Parameters:
adata AnnData

Must contain: - uns['interacting_cell_results'] with parametric or non-parametric scores - obsm['module_scores'] or obsm['super_module_scores'] - uns['metabolites'] or uns['gene_pairs_sig_names']

cor_method Union[Literal['pearson'], Literal['spearman'], None]

Statistical method used to compute correlations.

test Union[Literal['parametric'], Literal['non-parametric'], None]

Which interacting-cell score set to use. - "parametric" uses uns['interacting_cell_results']['p'] - "non-parametric" uses uns['interacting_cell_results']['np']

interaction_type Union[Literal['metabolite'], Literal['gene_pair'], None]

Select whether to correlate: - metabolite scores, or - gene pair scores.

only_sig_values bool | None

If True, use only significant interacting cell score values (cs_sig_pval or cs_sig_FDR).

normalize_values bool | None

Apply min-max normalization to interacting cell score values per interaction.

use_FDR bool | None

If only_sig_values=True, determines whether to filter by FDR or raw p-values.

use_super_modules bool | None

Whether to use super-module scores instead of module scores.

ct_aware bool | None

Whether to use cell-type-aware interacting cell scores. If None, cell-type-aware scores are used only when standard scores are absent and matching cell-type-aware scores are present.

Return type:

None