spatialleiden.spatialleiden

spatialleiden.spatialleiden(adata, *, resolution=1, latent_neighbors=None, spatial_neighbors=None, key_added='spatialleiden', directed=True, use_weights=True, n_iterations=-1, partition_type=<class 'leidenalg.VertexPartition.RBConfigurationVertexPartition'>, layer_ratio=1, latent_neighbors_key='connectivities', spatial_neighbors_key='spatial_connectivities', latent_partition_kwargs=None, spatial_partition_kwargs=None, random_state=42)

Perform SpatialLeiden clustering.

This is a wrapper around spatialleiden.multiplex_leiden() that uses anndata.AnnData as input and works with two layers; one latent space and one spatial layer.

Parameters:

  • adata (AnnData)

  • resolution (float | tuple[float, float], optional) – Resolution for the latent space and spatial layer, respectively. A single float applies to both layers.

  • latent_neighbors (sparray | spmatrix | ndarray) – Matrix of row-wise neighbor definitions in the latent space layer i.e. cij is the connectivity of i \(\to\) j.

  • spatial_neighbors (sparray | spmatrix | ndarray) – Matrix of row-wise neighbor definitions in the spatial layer i.e. cij is the connectivity of i \(\to\) j.

  • key_added (str, optional) – Key to store the clustering results in anndata.AnnData.obs

  • directed (bool | tuple[bool, bool], optional) – Whether to use a directed graph for latent space and spatial neighbors, respectively. A single bool applies to both layers.

  • use_weights (bool | tuple[bool, bool], optional) – Whether to use weights for the edges for latent space and spatial neighbors, respectively. A single bool applies to both layers.

  • n_iterations (int, optional) – Number of iterations to run the Leiden algorithm. If the number is negative it runs until convergence.

  • partition_type (Type[leidenalg.VertexPartition.MutableVertexPartition], optional) – A leidenalg.VertexPartition.MutableVertexPartition to be used.

  • layer_ratio (float, optional) – The ratio of the weighting of the layers; latent space vs spatial. A higher ratio will increase relevance of the spatial neighbors and lead to more spatially homogeneous clusters.

  • latent_neighbors_key (str, optional) – Key to use for the latent neighbor connectivities in anndata.AnnData.obsp. Only used if latent_neighbors is None.

  • spatial_neighbors_key (str, optional) – Key to use for the spatial neighbor connectivities in anndata.AnnData.obsp. Only used if spatial_neighbors is None.

  • latent_partition_kwargs (dict | None, optional) – Keyword arguments for the latent space partition.

  • spatial_partition_kwargs (dict | None, optional) – Keyword arguments for the spatial partition.

  • random_state (int, optional) – Random seed.