liana.plotting.lric_lineplot

Contents

liana.plotting.lric_lineplot#

liana.plotting.lric_lineplot(radii, curves, *, overlay=False, ncols=4, colors=None, title=None, figure_size=None, ax=None, return_fig=True)#

Line plot for cross-PCF / LRIC g(r) curves with a baseline at 1.

Parameters:

  • radii – 1-D array of radius values shared by all curves. Individual curves may supply their own radii by passing (radii, g) tuples as values.

  • curves (dict) – {label: g} or {label: (radii, g)}. In small-multiples mode labels become subplot titles; in overlay mode they become legend entries.

  • overlay (bool (default: False)) – False (default) — one subplot per curve arranged in a grid. True — all curves drawn on a single axis.

  • ncols (int (default: 4)) – Number of columns for the small-multiples grid (ignored when overlay=True).

  • colors (default: None) – Colour specification. Accepts a single colour string (applied to all curves), a list (one per curve, in dict-insertion order), or a dict keyed by label. Defaults to the matplotlib tab10 cycle.

  • title (str | None (default: None)) – Figure-level super-title.

  • figure_size (tuple | None (default: None)) – Figure x,y size

  • ax (default: None) – Existing Axes to draw into. Only used when overlay=True; ignored otherwise.

  • return_fig (bool (default: True)) – bool whether to return the fig object.

Return type:

Figure | None

Returns:

Figure if return_fig is True, otherwise None.

Examples

radii and the curves come from liana.method.cross_pcf (cell-type co-localisation) or liana.method.lric (expression-weighted). The dotted line at g(r) = 1 marks spatial independence:

>>> import liana as li
>>> adata = li.testing.generate_toy_spatial()
>>> res = li.mt.cross_pcf(adata, groupby='bulk_labels',
...                       radius_step=50, annulus_width=50,
...                       inplace=False)
>>> curves = {f'{s} -> {r}': g
...           for (s, r), g in list(res['results'].items())[:2]}
>>> fig = li.pl.lric_lineplot(res['radii'], curves, overlay=True)

One subplot per curve without overlay.

Contents