symphonypy.tl.per_cell_confidence

symphonypy.tl.per_cell_confidence(adata_query: AnnData, adata_ref: AnnData, ref_basis_adjusted: str = 'X_pca_harmony', query_basis_adjusted: str = 'X_pca_harmony', transferred_primary_basis: str = 'X_pca_reference', obs: str = 'symphony_per_cell_dist')

Calculates the weighted Mahalanobis distance for query cells to reference clusters. Higher distance metric indicates less confidence. Saves the metric to adata_query.obs[obs]

Parameters:

• adata_query (AnnData) – query adata object mapped to adata_ref with Symphony

• adata_ref (AnnData) – reference adata object (with Harmony object in adata_ref.uns)

• ref_basis_adjusted (str, optional) – adata_ref.obsm[ref_basis_adjusted] should contain resulting (harmony integrated if batch was present) reference representation, defaults to “X_pca_harmony”

• query_basis_adjusted (str, optional) – adata_query.obsm[query_basis_adjusted] should contain symphony adjusted query representation, defaults to “X_pca_harmony”

• transferred_primary_basis (str, optional) – adata_query.obsm[transferred_primary_basis] should contain pre-Symphony reference PC query representation, defaults to “X_pca_reference”

• obs (str, optional) – at adata_query.obs[obs] confidence metric will be saved, defaults to “symphony_per_cell_dist”