HoloNet.tools.cluster_lr_based_on_ce

HoloNet.tools.cluster_lr_based_on_ce(ce_tensor, adata, lr_df, w_best, n_clusters=4, cluster_based='node_centrality_euclidean', centrality_measure='Eigenvector', cell_cci_centrality=None, **kwargs)

Cluster the LR pairs using the CE network.

Parameters

ce_tensor Tensor

A CE tensor (LR_pair_num * cell_num * cell_num)

adata AnnData

Annotated data matrix.

lr_df DataFrame

A preprocessed LR-gene dataframe. must contain three columns: ‘Ligand_gene_symbol’, ‘Receptor_gene_symbol’ and ‘LR_pair’.

w_best float

A distance parameter in edge weighting function controlling the covering region of ligands. ‘default_w_visium’ function provides a recommended value of w_best.

n_clusters int (default: 4)

Number of clusters

cluster_based str (default: 'node_centrality_euclidean')

Cluster methods, selected in ‘node_centrality_euclidean’, ‘node_centrality_physical’, and ‘edge_overlap’

centrality_measure str (default: 'Eigenvector')

Compute methods

cell_cci_centrality tensor | NoneOptional[tensor] (default: None)

Provided centrality tensor can save time. A tensor (LR_num * cell_num) for the centrality of each cell according to each LR pair.

kwargs

Parameters for compute centrality, see in ‘compute_ce_network_degree_centrality’ and

’compute_ce_network_eigenvector_centrality’ function

Return type

(DataFrame, AgglomerativeClustering)

Returns

: A LR-gene dataframe added the ‘cluster’ column. And a trained hierarchical clustering model