netmap.masking.internal.get_neighborhood_expression

netmap.masking.internal.get_neighborhood_expression(adata, knn_neighbours=10, required_neighbours=1, expression_threshold=0, layer='X')[source]

Checks if each gene is expressed in the k-nearest neighbors (kNN) of each cell.

Computes a kNN graph on adata, then for each cell counts how many neighbours express each gene above expression_threshold. Returns a binary matrix where 1 indicates the gene is expressed in at least required_neighbours neighbours.

Parameters:

  • adata (anndata.AnnData) – Expression AnnData with PCA in obsm['X_pca']. A kNN graph is computed internally via sc.pp.neighbors.

  • knn_neighbours (int) – Number of nearest neighbours for the kNN graph. Defaults to 10.

  • required_neighbours (int) – Minimum number of neighbours that must express a gene for the cell to be considered in the neighbourhood. Defaults to 1.

  • expression_threshold (float) – Minimum value to consider a gene expressed. Defaults to 0.

  • layer (str) – Layer to use for expression values — 'X' or a named layer. Defaults to 'X'.

Returns:

Binary matrix of shape

(n_cells, n_genes) where 1 means the gene is expressed in at least required_neighbours neighbours.

Return type:

scipy.sparse matrix or numpy.ndarray