liana.pl.elbow

liana.pl.elbow(adata=None, errors=None, rank=None, figure_size=(5, 4), return_fig=True)

Plot the NMF reconstruction error per rank, with the estimated elbow.

Parameters:

adata AnnData | None (default: None)

AnnData on which liana.ms.nmf() was run with n_components=None; reads adata.uns['nmf_errors'] and adata.uns['nmf_rank'].

errors DataFrame | None (default: None)

Error curve as returned by liana.ms.estimate_elbow(); takes precedence over adata.uns['nmf_errors'].

rank int | None (default: None)

Estimated rank to mark with a dashed line; takes precedence over adata.uns['nmf_rank'].

figure_size tuple[float, float] (default: (5, 4))

Figure x,y size

return_fig bool (default: True)

bool whether to return the fig object.

Return type:

plotnine.ggplot | None

Returns:

The resulting elbow plot.

Raises:

ValueError – If no error curve is available, e.g. nmf was run with a fixed n_components.

Examples

>>> import liana as li
>>> adata = li.ds.generate_toy_spatial()
>>> lrdata = li.mt.bivariate(adata, resource_name="consensus", local_name="cosine", global_name=None, n_perms=None)
>>> li.ms.nmf(lrdata, n_components=None, k_range=range(1, 6), random_state=0, max_iter=200)
>>> p = li.pl.elbow(lrdata)