dotools_py.pl.heatmap

dotools_py.pl.heatmap(adata, x_axis, features, y_axis=None, xticks_order=None, yticks_order=None, layer=None, logcounts=True, figsize=(5, 6), ax=None, swap_axes=True, title='', title_fontproperties=None, palette='Reds', xticks_properties=None, xticks_rotation=45, yticks_properties=None, yticks_rotation=0, cluster_x_axis=False, cluster_y_axis=False, legend_title='LogMean(nUMI)\\nin group', path=None, filename='Heatmap.svg', show=True, add_stats=None, test='wilcoxon', correction_method='benjamini-hochberg', df_pvals=None, stats_x_size=None, square_x_size=None, pval_cutoff=0.05, log2fc_cutoff=0.0, z_score=None, clustering_method='complete', clustering_metric='euclidean', linewidth=0.1, vmin=None, vcenter=None, vmax=None, annot_fontsize=12, annot_ratio=0.35, **kwargs)

Makes a 2d or 3d heatmap.

There are two type of visualization:

• 2d heatmap: X_axis shows x_axis categories and Y_axis the features. The color represents the logarithmize mean

  nUMI.

• 3d dotplot: X_axis shows x_axis categories and Y_axis shows y_axis categories. For each feature the

  x_axis categories will be duplicated, to show how is the expressing across 2 categorical columns in .obs. The color represents the logarithmize mean nUMI.

Differential gene expression analysis between the different groups can be performed.

Parameters:

adata AnnData

Annotated data matrix.

x_axis str

Name of a categorical column in adata.obs to groupby.

features str | list

A valid feature in adara.var_names or column in adata.obs with continuous values. The order of the features is maintained.

y_axis str | None (default: None)

A valid feature in adara.var_names or column in adata.obs with continuous values.

xticks_order list | None (default: None)

Order for the categories in adata.obs[x_axis]

yticks_order list | None (default: None)

Order for the categories in adata.obs[y_axis]

layer str | None (default: None)

Name of the AnnData object layer that wants to be plotted. By default, adata.X is plotted.

logcounts bool (default: True)

Set to True if the input data is in logspace.

figsize tuple (default: (5, 6))

Figure size, the format is (width, height).

ax Axes | None (default: None)

Matplotlib axes to use for plotting. If not set, a new figure will be generated.

swap_axes bool (default: True)

Whether to swap the x_axis categories and features. Only used if y_axis is set to None.

title str (default: '')

Title for the figure.

title_fontproperties Optional[Dict[Literal['size', 'weight'], str | int]] (default: None)

Dictionary which should contain ‘size’ and ‘weight’ to define the fontsize and fontweight of the title of the figure.

palette str (default: 'Reds')

String denoting matplotlib colormap.

xticks_properties dict | None (default: None)

Dictionary which should contain ‘size’ and ‘weight’ to define the fontsize and fontweight of the font of the x-axis.

xticks_rotation int (default: 45)

Rotation of the x-ticks.

yticks_properties dict | None (default: None)

Dictionary which should contain ‘size’ and ‘weight’ to define the fontsize and fontweight of the font of the y-axis.

yticks_rotation int (default: 0)

Rotations of the y-ticks.

cluster_x_axis bool (default: False)

Hierarchically clustering the x-axis.

cluster_y_axis bool (default: False)

Hierarchically clustering the y-axis.

legend_title str (default: 'LogMean(nUMI)\\nin group')

Title for the colorbar.

path str | PathLike[str] | Path | None (default: None)

Path to the folder to save the figure.

filename str (default: 'Heatmap.svg')

Name of file to use when saving the figure.

show bool (default: True)

If set to False, returns a dictionary with the matplotlib axes.

add_stats Optional[Literal['x_axis', 'y_axis']] (default: None)

Add statistical annotation. Will add a square with an ‘*’ in the center if the expression is significantly different in a group with respect to the others.

test Literal['wilcoxon', 't-test'] (default: 'wilcoxon')

Name of the method to test for significance.

correction_method Literal['benjamini-hochberg', 'bonferroni'] (default: 'benjamini-hochberg')

Correction method for multiple testing.

df_pvals DataFrame | None (default: None)

Dataframe with the pvals.

stats_x_size float | None (default: None)

Scaling factor to control the size of the asterisk.

square_x_size dict | None (default: None)

Size and thickness of the square.

pval_cutoff float (default: 0.05)

Cutoff for the p-value.

log2fc_cutoff float (default: 0.0)

Minimum cutoff for the log2FC.

z_score Optional[Literal['x_axis', 'y_axis']] (default: None)

Apply z-score transformation.

clustering_method str (default: 'complete')

Linkage method to use for calculating clusters. See scipy.cluster.hierarchy.linkage.

clustering_metric str (default: 'euclidean')

Distance metric to use for the data. See scipy.spatial.distance.pdist.

linewidth float (default: 0.1)

Linewidth for the border of cells.

vmin float | None (default: None)

The value representing the lower limit of the color scale.

vcenter float | None (default: None)

The value representing the center of the color scale.

vmax float | None (default: None)

The value representing the upper limit of the color scale.

annot_fontsize float (default: 12)

Fontsize of the features text in 3d heatmaps.

annot_ratio float (default: 0.35)

Fraction of the figure reserved for the feature text in 3d heatmaps.

kwargs

Additional arguments pass to sns.heatmap.

Returns:

Depending on show, returns the plot if set to True or a dictionary with the axes.

Example

Create a 2d heatmap and add statistical information

import dotools_py as do
adata = do.dt.example_10x_processed()
do.pl.heatmap(adata, 'annotation', ['CD4', 'CD79A'], add_stats="x_axis")

Create a 3d heatmap grouping also by condition

do.pl.heatmap(adata, 'condition',  ['CD4', 'CD79A'], 'annotation', figsize=(6, 4))

# Add Statistical significance for groups with pvals < 0.05 and log2fc > 0.0
# Note, the object is quite small, some groups cannot be tested for having one condition only.
do.pl.heatmap(adata, 'condition', ['CD4', 'CD79A'], 'annotation', figsize=(6, 4), add_stats='y_axis')