crested.pl.corr.heatmap

crested.pl.corr.heatmap(adata, model_names=None, split='test', log_transform=False, vmin=None, vmax=None, reorder=False, cmap='coolwarm', cbar=True, cbar_kws=None, plot_kws=None, ax=None, **kwargs)

Plot correlation heatmaps of predictions vs ground truth for all cell types.

Parameters:

• adata (AnnData) – AnnData object containing the data in X and predictions in layers.

• model_names (str | list[str] | None (default: None)) – Model name or list of model names (adata.layers) to plot for predictions heatmap. Default is to plot all models in adata.layers.

• split (str | None (default: 'test')) – ‘train’, ‘val’, ‘test’ subset or None. If None, will use all targets. If not None, expects a “split” column in adata.var.

• log_transform (bool (default: False)) – Whether to log-transform the data before plotting.

• vmin (float | None (default: None)) – Minimum value for heatmap color scale.

• vmax (float | None (default: None)) – Maximum value for heatmap color scale.

• reorder (bool (default: False)) – Whether or not to order the clases by similarity (boolean).

• cmap (str | Colormap (default: 'coolwarm')) – Colormap to use.

• cbar (bool (default: True)) – whether to draw a colorbar.

• cbar_kws (dict | None (default: None)) – Extra keyword arguments passed to the colorbar. Default is {'label': "Pearson correlations (of log1p-transformed values)"}

• plot_kws (dict | None (default: None)) – Extra keyword arguments passed to heatmap(). Adjusted defaults compared to the base function are square=True and fmt='.2f'.

• ax (Axes | None (default: None)) – Axis to plot values on. If not supplied, creates a figure from scratch. Can only be supplied if plotting a single model.

• width – Width of the newly created figure if ax=None. Default is 10 per model to plot, or 8 if cbar=False.

• height – Height of the newly created figure if ax=None. Default is 8 per model to plot.

• kwargs – Additional arguments passed to render_plot() to control the final plot output. Please see render_plot() for details. Custom defaults for correlations_predictions: xtick_rotation=90, layout='compressed', title=list(adata.obs_names).

Return type:

tuple[Figure, Axes] | tuple[Figure, list[Axes]] | None

See also

crested.pl.render_plot, crested.pl.corr.heatmap_self, crested.pl.corr.violin

Examples

>>> crested.pl.corr.heatmap(
...     adata,
...     model_names=None,
...     split="test",
...     log_transform=True,
...     vmin=0.4,
...     vmax=0.85,
...     suptitle="Correlations: predictions vs ground truth",
... )