grassp.tl.paga_dendrogram

paga_dendrogram(adata, groupby, *, neighbors_key=None, paga_model='v1.2', linkage_method='average', optimal_ordering=False, key_added=None, inplace=True)

Build a dendrogram from PAGA connectivity and store it in adata.uns.

Drop-in replacement for scanpy.tl.dendrogram() that uses adata.uns['paga']['connectivities'] as the similarity matrix instead of gene-expression correlations. The result dict has the same schema as scanpy.tl.dendrogram(), so it can be passed to any scanpy plotting function that accepts a dendrogram_key argument.

Parameters:

adata AnnData

AnnData object.

groupby str

Categorical adata.obs column to build the dendrogram for.

neighbors_key Optional[str] (default: None)

Passed to scanpy.tl.paga().

paga_model str (default: 'v1.2')

PAGA model version, passed to scanpy.tl.paga().

linkage_method str (default: 'average')

Linkage algorithm passed to scipy.cluster.hierarchy.linkage().

optimal_ordering bool (default: False)

Passed to scipy.cluster.hierarchy.linkage().

key_added Optional[str] (default: None)

Key written to adata.uns. Defaults to f"dendrogram_{groupby}".

inplace bool (default: True)

If True, write to adata.uns and return None. Otherwise return the dict.

Return type:

Optional[dict]

Returns:

dict or None Dendrogram dict when inplace=False, otherwise None.