athena.plotting.visualization module¶

Summary¶

Functions:

`get_cmap`	Return the cmap and cmap labels for a given attribute if available, else a default
`infiltration`	Visualises a heatmap of the featuer intensity.
`interactions`	Visualise results from `interactions()` results.
`napari_viewer`	Starts interactive Napari viewer to visualise raw images and explore samples.
`ripleysK`	Visualise results from `ripleysK()` results.
`spatial`	Various functionalities to visualise samples.

Reference¶

spatial(so, spl, attr, *, mode='scatter', node_size=4, coordinate_keys=['x', 'y'], mask_key='cellmasks', graph_key='knn', edges=False, edge_width=0.5, edge_color='black', edge_zorder=2, background_color='white', ax=None, norm=None, set_title=True, cmap=None, cmap_labels=None, cbar=True, cbar_title=True, show=True, save=None, tight_layout=True)[source]¶

Various functionalities to visualise samples. Allows to visualise the samples and color observations according to features in either so.X or so.obs by setting the attr parameter accordingly. Furthermore, observations (cells) within a sample can be quickly visualised using scatter plots (requires extraction of centroids with extract_centroids()) or by their actual segmentatio mask by setting mode accordingly. Finally, the graph representation of the sample can be overlayed by setting edges=True and specifing the graph_key as in so.G[spl][graph_key].

For more examples on how to use this function have a look at the tutorial section.

Parameters

so – SpatialOmics instance
spl (str) – sample to visualise
attr (str) – feature to visualise
mode (str) – {scatter, mask}. In scatter mode, observations are represented by their centroid, in mask mode by their actual segmentation mask
node_size (float) – size of the node when plotting the graph representation
coordinate_keys (list) – column names in SpatialOmics.obs[spl] that indicates the x and y coordinates
mask_key (str) – key for the segmentation masks when in mask mode
graph_key (str) – which graph representation to use
edges (bool) – whether to plot the graph or not
edge_width (float) – width of edges
edge_color (str) – color of edges as string
edge_zorder (int) – z-order of edges
background_color (str) – background color of plot
ax (Optional[Axes]) – axes object in which to plot
norm – normalisation instance to normalise the values of attr
set_title (bool) – title of plot
cmap – colormap to use
cmap_labels (Optional[list]) – colormap labels to use
cbar (bool) – whether to plot a colorbar or not
cbar_title (bool) – whether to plot the attr name as title of the colorbar
show (bool) – whether to show the plot or not
save (Optional[str]) – path to the file in which the plot is saved
tight_layout (bool) – whether to apply tight_layout or not.

Examples

so = sh.dataset.imc()
sh.pl.spatial(so, 'slide_7_Cy2x2', 'meta_id', mode='mask')
sh.pl.spatial(so, 'slide_7_Cy2x2', 'meta_id', mode='scatter', edges=True)

napari_viewer(so, spl, attrs, censor=0.95, add_masks='cellmasks', attrs_key='target', index_key='fullstack_index')[source]¶

Starts interactive Napari viewer to visualise raw images and explore samples. attrs are measured features in the high dimensional images in so.images[spl]. All specified attrs should be in so.var[spl][attrs_key] along with the index in the high dimensional images. The column with the index in the high dimensional image where the measurement of an attribute is stored.

Parameters

so – SpatialOmics instance
spl (str) – sample to visualise
attrs (list) – list of attributes/features to add as channels to the viewer
censor (float) – percentil to use to censore pixle values in the raw images
add_masks – segmentation masks to add as channels to the viewer
attrs_key – key in so.var[spl] that defines the attrs names
index_key (str) – key in so.var[spl] that specifies the layer index in the high dimensional image in so.images[spl].

Examples:

so = sh.dataset.imc()
spl = so.spl.index[0]

# add all measured features to the napari viewer
sh.pl.napari_viewer(so, spl, attrs=so.var[spl]['target'], add_masks=so.masks[spl].keys())

# specify the column containing the ``attrs`` names
sh.pl.napari_viewer(so, spl, attrs=so.var[spl]['target'], attrs_key='target')

# specify the column containing the ``attrs`` names and the columns that specifies the layer index
sh.pl.napari_viewer(so, spl, attrs=so.var[spl]['target'], attrs_key='target', index_key='fullstack_index')

interactions(so, spl, attr, mode='proportion', prediction_type='diff', graph_key='knn', linewidths=0.5, cmap=None, norm=None, ax=None, show=True, cbar=True)[source]¶

Visualise results from interactions() results.

Parameters

so – SpatialOmics instance
spl – Spl for which to compute the metric
attr – Categorical feature in SpatialOmics.obs to use for the grouping
mode – One of {classic, histoCAT, proportion}, see notes
prediction_type – prediction_type: One of {observation, pvalue, diff}
graph_key – Specifies the graph representation to use in so.G[spl]
linewidths – Space between tiles
cmap – colormap to use
norm – normalisation to use
ax – axes object to use
show – whether to show the plot
cbar – wheter to show the colorbar

Examples

# compute Ripley's K
so = sh.dataset.imc()
spl = so.spl.index[0]

# build graph
sh.graph.build_graph(so, spl, builder_type='knn', mask_key='cellmasks')

# compute & plot interactions
sh.neigh.interactions(so, spl, 'meta_id', mode='proportion', prediction_type='observation')
sh.pl.interactions(so, spl, 'meta_id', mode='proportion', prediction_type='observation')

get_cmap(so, attr, data)[source]¶

Return the cmap and cmap labels for a given attribute if available, else a default

Parameters

so (IMCData) – so object form which to fetch the data
spl (str) – spl for which to get data
attr (str) – attribute for which to get the cmap and cmap labels if available

Returns

Return type

cmap and cmap labels for attribute

ripleysK(so, spl, attr, ids, *, mode='K', correction='ripley', key=None, ax=None, legend='auto')[source]¶

Visualise results from ripleysK() results.

Parameters

so – SpatialOmics instance
spl (str) – Spl for which to compute the metric
attr (str) – Categorical feature in SpatialOmics.obs to use for the grouping
ids – The category in the categorical feature attr, for which Ripley’s K should be plotted
mode – {K, csr-deviation}. If K, Ripley’s K is estimated, with csr-deviation the deviation from a poission process is computed.
correction – Correction method to use to correct for boarder effects, see [1].
key – key to use in so.uns[‘ripleysK’] for the plot, if None it is constructed from spl,attr,ids,mode and correction
ax – axes to use for the plot

Examples

# compute Ripley's K
so = sh.dataset.imc()
sh.neigh.ripleysK(so, so.spl.index[0], 'meta_id', 1, mode='csr-deviation', radii=radii)
sh.pl.ripleysK(so, so.spl.index[0], 'meta_id', [1], mode='csr-deviation')

infiltration(so, spl, attr='infiltration', step_size=10, interpolation='gaussian', cmap='plasma', collision_strategy='mean', ax=None, show=True)[source]¶

Visualises a heatmap of the featuer intensity.

Approximates the sample with a grid representation and colors each grid square according to the value of the attribute. If multiple observations map to the same grid square a the aggregation specified in collision_strategy is employed (any value accepted by pandas aggregate function, i.e. ‘mean’, ‘max’, …)

Parameters

so – SpatialOmics instance
spl (str) – Spl for which to compute the metric
attr (str) – feature in SpatialOmics.obs to plot
step_size (int) – grid step size
interpolation (str) – interpolation method to use between grid values, see [1]
cmap (str) – colormap to use
collision_strategy – aggragation strategy to use if multiple obseravtion values map to the same grid value
ax – axes to use for the plot
show – whether to show the plot or not. Will be set to False if axes is provided.

Examples

so = sh.dataset.imc()
spl = so.spl.index[0]

# build graph
sh.graph.build_graph(so, spl, builder_type='knn', mask_key='cellmasks')

sh.neigh.infiltration(so, spl, 'meta_id', graph_key='knn')
sh.pl.infiltration(so, spl, step_size=10)

Notes

1: https://matplotlib.org/stable/gallery/images_contours_and_fields/interpolation_methods.html