93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395 | def voronoi (adata, color_by=None, colors=None, x_coordinate='X_centroid', y_coordinate='Y_centroid',
imageid='imageid',subset=None, x_lim=None, y_lim=None, flip_y=True,
voronoi_edge_color='black', voronoi_line_width=0.1, voronoi_alpha=0.5, size_max=np.inf,
overlay_points=None, overlay_points_categories=None, overlay_drop_categories=None, overlay_points_colors=None,
overlay_point_size = 5, overlay_point_alpha= 1, overlay_point_shape=".", plot_legend=True, legend_size = 6, **kwargs):
"""
Parameters:
adata : Anndata object
color_by : string, optional
Color the voronoi diagram based on categorical variable (e.g. cell types or neighbourhoods).
Pass the name of the column which contains the categorical variable.
colors : string or Dict, optional
Custom coloring the voronoi diagram. The parameter accepts `sns color palettes` or a python dictionary
mapping the categorical variable with the required color.
x_coordinate : float, required
Column name containing the x-coordinates values.
y_coordinate : float, required
Column name containing the y-coordinates values.
flip_y : bool, optional
Flip the Y-axis if needed. Some algorithms output the XY with the Y-coordinates flipped.
If the image overlays do not align to the cells, try again by setting this to `False`.
imageid : string, optional
Column name of the column containing the image id.
subset : string, optional
imageid of a single image to be subsetted for plotting.
voronoi_edge_color : string, optional
A Matplotlib color for marking the edges of the voronoi.
If `facecolor` is passed, the edge color will always be the same as the face color.
voronoi_line_width : float, optional
The linewidth of the marker edges. Note: The default edgecolors is 'face'. You may want to change this as well.
voronoi_alpha : float, optional
The alpha blending value, between 0 (transparent) and 1 (opaque).
x_lim : list, optional
Pass the x-coordinates range [x1,x2].
y_lim : list, optional
Pass the y-coordinates range [y1,y2].
overlay_points : string, optional
It is possible to overlay a scatter plot on top of the voronoi diagram.
Pass the name of the column which contains categorical variable to be overlayed.
overlay_points_categories : list, optional
If the passed column in `overlay_points` contains multiple categories, however the user is only
interested in a subset of categories, those specific names can be passed as a list. By default all
categories will be overlayed on the voronoi diagram.
overlay_drop_categories : list, optional
Similar to `overlay_points_categories`. Here for ease of use, especially if large number of categories are present.
The user can drop a set of categories.
overlay_points_colors : string or dict, optional
Similar to `colors`.
User can pass in a
a) solid color (like `black`)
b) sns palettes name (like `Set1`)
c) python dictionary mapping the categories with custom colors
overlay_point_size : float, optional
Overlay scatter plot point size.
overlay_point_alpha : float, optional
The alpha blending value for the overlay, between 0 (transparent) and 1 (opaque).
overlay_point_shape : string, optional
The marker style. marker can be either an instance of the class or the text shorthand for a particular marker.
plot_legend : bool, optional
Define if the figure legend should be plotted.
Please note the figure legend may be out of view and you may need to resize the image to see it, especially
the legend for the scatter plot which will be on the left side of the plot.
legend_size : float, optional
Resize the legend if needed.
Example:
```python
sm.pl.voronoi(adata, color_by='phenotype', colors=None,
x_coordinate='X_position', y_coordinate='Y_position',
imageid='ImageId',subset=None,
voronoi_edge_color = 'black',voronoi_line_width = 0.2,
voronoi_alpha = 0.5, size_max=np.inf,
overlay_points='phenotype', overlay_points_categories=None,
overlay_drop_categories=None,
overlay_point_size = 5, overlay_point_alpha= 1,
overlay_point_shape=".", plot_legend=False, legend_size=6)
```
"""
# create the data frame needed
data = adata.obs
# Subset the image of interest
if subset is not None:
data = data[data[imageid] == subset]
# subset coordinates if needed
if x_lim is not None:
x1 = x_lim [0]
if len(x_lim) < 2:
x2 = max(data[x_coordinate])
else:
x2 = x_lim [1]
if y_lim is not None:
y1 = y_lim [0]
if len(y_lim) < 2:
y2 = min(data[y_coordinate])
else:
y2 = y_lim [1]
# do the actuall subsetting
if x_lim is not None:
data = data[data[x_coordinate] >= x1]
data = data[data[x_coordinate] <= x2]
if y_lim is not None:
data = data[data[y_coordinate] <= y1]
data = data[data[y_coordinate] >= y2]
# create an extra column with index information
data['index_info'] = np.arange(data.shape[0])
# generate the x and y coordinates
points = data[[x_coordinate,y_coordinate]].values
# invert the Y-axis
if flip_y is True:
points[:,1] = max(points[:,1])-points[:,1]
# Generate colors
if color_by is None:
colors = np.repeat('#e5e5e5', len(data))
# elif color_by is None and colors is not None:
# if isinstance(colors,str):
# colors = np.repeat(colors, len(data))
elif color_by is not None and colors is None:
# auto color the samples
if len(np.unique(data[color_by])) <= 9:
c = sns.color_palette('Set1')[0:len(np.unique(data[color_by]))]
if len(np.unique(data[color_by])) > 9 and len(np.unique(data[color_by])) <= 20:
c = sns.color_palette('tab20')[0:len(np.unique(data[color_by]))]
if len(np.unique(data[color_by])) > 20:
# For large categories generate random colors
np.random.seed(0) ; c = np.random.rand(len(np.unique(data[color_by])),3).tolist()
# merge colors with phenotypes/ categories of interest
p = np.unique(data[color_by])
c_p = dict(zip(p, c))
# map to colors
colors = list(map(c_p.get, list(data[color_by].values)))
elif color_by is not None and colors is not None:
# check if colors is a dictionary or a sns color scale
if isinstance(colors,str):
if len(sns.color_palette(colors)) < len(np.unique(data[color_by])):
raise ValueError(str(colors) + ' includes a maximun of ' + str(len(sns.color_palette(colors))) + ' colors, while your data need ' + str(len(np.unique(data[color_by]))) + ' colors')
else:
c = sns.color_palette(colors)[0:len(np.unique(data[color_by]))]
# merge colors with phenotypes/ categories of interest
p = np.unique(data[color_by])
c_p = dict(zip(p, c))
if isinstance(colors,dict):
if len(colors) < len(np.unique(data[color_by])):
raise ValueError('Color mapping is not provided for all categories. Please check')
else:
c_p = colors
# map to colors
colors = list(map(c_p.get, list(data[color_by].values)))
# create the voronoi object
vor = Voronoi(points)
# trim the object
regions, vertices = voronoi_finite_polygons_2d(vor)
# plotting
pts = MultiPoint([Point(i) for i in points])
mask = pts.convex_hull
new_vertices = []
if type(voronoi_alpha)!=list:
voronoi_alpha = [voronoi_alpha]*len(points)
areas = []
for i,(region,alph) in enumerate(zip(regions,voronoi_alpha)):
polygon = vertices[region]
shape = list(polygon.shape)
shape[0] += 1
p = Polygon(np.append(polygon, polygon[0]).reshape(*shape)).intersection(mask)
areas+=[p.area]
if p.area <size_max:
poly = np.array(list(zip(p.boundary.coords.xy[0][:-1], p.boundary.coords.xy[1][:-1])))
new_vertices.append(poly)
if voronoi_edge_color == 'facecolor':
plt.fill(*zip(*poly), alpha=alph, edgecolor=colors[i], linewidth = voronoi_line_width , facecolor = colors[i])
plt.xticks([]) ; plt.yticks([]);
else:
plt.fill(*zip(*poly), alpha=alph, edgecolor=voronoi_edge_color, linewidth = voronoi_line_width, facecolor = colors[i])
plt.xticks([]) ; plt.yticks([]);
#plt.xlim([1097.5,1414.5])
#plt.ylim([167.3,464.1])
# Add scatter on top of the voronoi if user requests
if overlay_points is not None:
if overlay_points_categories is None:
d = data
if overlay_points_categories is not None:
# convert to list if needed (cells to keep)
if isinstance(overlay_points_categories,str):
overlay_points_categories = [overlay_points_categories]
# subset cells needed
d = data[data[overlay_points].isin(overlay_points_categories)]
if overlay_drop_categories is not None:
# conver to list if needed (cells to drop)
if isinstance(overlay_drop_categories,str):
overlay_drop_categories = [overlay_drop_categories]
# subset cells needed
d = d[-d[overlay_points].isin(overlay_drop_categories)]
# Find the x and y coordinates for the overlay category
#points_scatter = d[[x_coordinate,y_coordinate]].values
points_scatter = points[d.index_info.values]
# invert the Y-axis
#points_scatter[:,1] = max(points_scatter[:,1])-points_scatter[:,1]
# Generate colors for the scatter plot
if overlay_points_colors is None and color_by == overlay_points:
# Borrow color from vornoi
wanted_keys = np.unique(d[overlay_points]) # The keys to extract
c_p_scatter = dict((k, c_p[k]) for k in wanted_keys if k in c_p)
elif overlay_points_colors is None and color_by != overlay_points:
# Randomly generate colors for all the categories in scatter plot
# auto color the samples
if len(np.unique(d[overlay_points])) <= 9:
c_scatter = sns.color_palette('Set1')[0:len(np.unique(d[overlay_points]))]
if len(np.unique(d[overlay_points])) > 9 and len(np.unique(d[overlay_points])) <= 20:
c_scatter = sns.color_palette('tab20')[0:len(np.unique(d[overlay_points]))]
if len(np.unique(d[overlay_points])) > 20:
# For large categories generate random colors
np.random.seed(1) ; c_scatter = np.random.rand(len(np.unique(d[overlay_points])),3).tolist()
# merge colors with phenotypes/ categories of interest
p_scatter = np.unique(d[overlay_points])
c_p_scatter = dict(zip(p_scatter, c_scatter))
elif overlay_points_colors is not None:
# check if the overlay_points_colors is a pallete
if isinstance(overlay_points_colors,str):
try:
c_scatter = sns.color_palette(overlay_points_colors)[0:len(np.unique(d[overlay_points]))]
if len(sns.color_palette(overlay_points_colors)) < len(np.unique(d[overlay_points])):
raise ValueError(str(overlay_points_colors) + ' pallete includes a maximun of ' + str(len(sns.color_palette(overlay_points_colors))) + ' colors, while your data (overlay_points_colors) need ' + str(len(np.unique(d[overlay_points]))) + ' colors')
except:
c_scatter = np.repeat(overlay_points_colors,len(np.unique(d[overlay_points]))) #[overlay_points_colors]
# create a dict
p_scatter = np.unique(d[overlay_points])
c_p_scatter = dict(zip(p_scatter, c_scatter))
if isinstance(overlay_points_colors,dict):
if len(overlay_points_colors) < len(np.unique(d[overlay_points])):
raise ValueError('Color mapping is not provided for all categories. Please check overlay_points_colors')
else:
c_p_scatter = overlay_points_colors
# map to colors
colors_scatter = list(map(c_p_scatter.get, list(d[overlay_points].values)))
#plt.scatter(x = points_scatter[:,0], y = points_scatter[:,1], s= overlay_point_size, alpha= overlay_point_alpha, c= colors_scatter, marker=overlay_point_shape)
plt.scatter(x = points_scatter[:,0], y = points_scatter[:,1], s= overlay_point_size, alpha= overlay_point_alpha, c= colors_scatter, marker=overlay_point_shape,**kwargs)
plt.xticks([]) ; plt.yticks([]);
if plot_legend is True:
# Add legend to voronoi
patchList = []
for key in c_p:
data_key = mpatches.Patch(color=c_p[key], label=key)
patchList.append(data_key)
first_legend = plt.legend(handles=patchList, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0., prop={'size': legend_size})
plt.tight_layout()
# Add the legend manually to the current Axes.
ax = plt.gca().add_artist(first_legend)
if overlay_points is not None:
# Add legend to scatter
patchList_scatter = []
for key in c_p_scatter:
data_key_scatter = mpatches.Patch(color=c_p_scatter[key], label=key)
patchList_scatter.append(data_key_scatter)
plt.legend(handles=patchList_scatter, bbox_to_anchor=(-0.05, 1), loc=1, borderaxespad=0., prop={'size': legend_size})
|