symphonypy.tl.map_embedding

symphonypy.tl.map_embedding(adata_query: AnnData, adata_ref: AnnData, key: list[str] | str | None = None, lamb: float | np.array | None = None, sigma: float | np.array = 0.1, use_genes_column: str | None = 'highly_variable', transferred_adjusted_basis: str = 'X_pca_harmony', transferred_primary_basis: str = 'X_pca_reference', ref_basis_loadings: str = 'PCs', K: int | None = None, reference_primary_basis: str = 'X_pca') → None

Actually runs Symphony algorithm for mapping adata_query to adata_ref. Will use Harmony object from adata_ref.uns if present, otherwise will firstly run k-means clusterization Harmony step without batch correction.

Parameters:

  • adata_query (AnnData) – query AnnData object

  • adata_ref (AnnData) – reference AnnData object (to account for batch effect in reference first run pp.harmony_integrate).

  • key (list[str] | str | None, optional) – which of the columns from adata_query.obs to consider as batch keys, defaults to None

  • lamb (float | np.array | None, optional) – Entropy regularization parameter for soft k-means, defaults to None

  • sigma (float | np.array, optional) – Soft k-means kernel bandwidth. Only used when the reference Harmony object does not carry its own sigma (i.e. the no-batch soft-kmeans path). When adata_ref was produced by pp.harmony_integrate, the reference’s sigma is reused, as Symphony requires. Default 0.1.

  • use_genes_column (str | None, optional) – adata_ref.var[use_genes_column] genes will be used to map query embeddings to reference, defaults to “highly_variable”

  • transferred_adjusted_basis (str, optional) – in adata_query.obsm[transferred_adjusted_basis] symphony-adjusted coords will be saved, defaults to “X_pca_harmony”

  • transferred_primary_basis (str, optional) – in adata_query.obsm[transferred_primary_basis] query mapping to reference coordinates will be saved, defaults to “X_pca_reference”

  • ref_basis_loadings (str, optional) – adata_ref.varm[ref_basis_loadings] will be used as gene loadings to map adata_query to adata_ref coordinates, defaults to “PCs”

  • K (int | None, optional) – Number of clusters to use for k-means clustering. Only used if harmony integration was not performed on adata_ref, defaults to None

  • reference_primary_basis (str, optional) – which reference embedding to use for k-means clustering. Only used if harmony integration was not performed on adata_ref, defaults to “X_pca”