silhouette_computation

besca.tl.sig.silhouette_computation(adata: AnnData, cluster: str = 'dblabel', emb: str = 'X_umap', verbose: bool = False) → silhouette_in

Compute the average and per cell (ie samples) silhouette score for the cluster label (should be present in dataobs) (level 3 annotation), computed level 2 annotation and a random cell assignbation. Return a silhouette_in object

Parameters:

• adata (anndata.AnnData) –

• cluster ('str') – clustering to evaluate (should be a column in adata.obs)

• emb (str) – embedding to use for computing the euclidian distance. should be a key of obsm

Returns:

• silhouette_in dataclass object – Example

• ——-

• >>> import besca as bc

• >>> adata = bc.datasets.simulated_pbmc3k_processed()

• >>> sils = bc.tl.sig.silhouette_computation (adata)

• >>> figure = sils.show_samples.get_figure()