risk-network 0.0.8b26__py3-none-any.whl → 0.0.9b26__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- risk/__init__.py +2 -2
- risk/annotations/__init__.py +2 -2
- risk/annotations/annotations.py +74 -47
- risk/annotations/io.py +47 -31
- risk/log/__init__.py +4 -2
- risk/log/{config.py → console.py} +5 -3
- risk/log/{params.py → parameters.py} +17 -42
- risk/neighborhoods/__init__.py +3 -5
- risk/neighborhoods/api.py +446 -0
- risk/neighborhoods/community.py +255 -77
- risk/neighborhoods/domains.py +62 -31
- risk/neighborhoods/neighborhoods.py +156 -160
- risk/network/__init__.py +1 -3
- risk/network/geometry.py +65 -57
- risk/network/graph/__init__.py +6 -0
- risk/network/graph/api.py +194 -0
- risk/network/{graph.py → graph/network.py} +87 -37
- risk/network/graph/summary.py +254 -0
- risk/network/io.py +56 -47
- risk/network/plotter/__init__.py +6 -0
- risk/network/plotter/api.py +54 -0
- risk/network/{plot → plotter}/canvas.py +7 -4
- risk/network/{plot → plotter}/contour.py +22 -19
- risk/network/{plot → plotter}/labels.py +69 -74
- risk/network/{plot → plotter}/network.py +170 -34
- risk/network/{plot/utils/color.py → plotter/utils/colors.py} +104 -112
- risk/network/{plot → plotter}/utils/layout.py +8 -5
- risk/risk.py +11 -500
- risk/stats/__init__.py +8 -4
- risk/stats/binom.py +51 -0
- risk/stats/chi2.py +69 -0
- risk/stats/hypergeom.py +27 -17
- risk/stats/permutation/__init__.py +1 -1
- risk/stats/permutation/permutation.py +44 -38
- risk/stats/permutation/test_functions.py +25 -17
- risk/stats/poisson.py +15 -9
- risk/stats/stats.py +15 -13
- risk/stats/zscore.py +68 -0
- {risk_network-0.0.8b26.dist-info → risk_network-0.0.9b26.dist-info}/METADATA +9 -5
- risk_network-0.0.9b26.dist-info/RECORD +44 -0
- {risk_network-0.0.8b26.dist-info → risk_network-0.0.9b26.dist-info}/WHEEL +1 -1
- risk/network/plot/__init__.py +0 -6
- risk/network/plot/plotter.py +0 -137
- risk_network-0.0.8b26.dist-info/RECORD +0 -37
- {risk_network-0.0.8b26.dist-info → risk_network-0.0.9b26.dist-info}/LICENSE +0 -0
- {risk_network-0.0.8b26.dist-info → risk_network-0.0.9b26.dist-info}/top_level.txt +0 -0
@@ -0,0 +1,446 @@
|
|
1
|
+
"""
|
2
|
+
risk/neighborhoods/api
|
3
|
+
~~~~~~~~~~~~~~~~~~~~~~
|
4
|
+
"""
|
5
|
+
|
6
|
+
import copy
|
7
|
+
from typing import Any, Dict, List, Tuple, Union
|
8
|
+
|
9
|
+
import networkx as nx
|
10
|
+
import numpy as np
|
11
|
+
from scipy.sparse import csr_matrix
|
12
|
+
|
13
|
+
from risk.log import logger, log_header, params
|
14
|
+
from risk.neighborhoods.neighborhoods import get_network_neighborhoods
|
15
|
+
from risk.stats import (
|
16
|
+
compute_binom_test,
|
17
|
+
compute_chi2_test,
|
18
|
+
compute_hypergeom_test,
|
19
|
+
compute_permutation_test,
|
20
|
+
compute_poisson_test,
|
21
|
+
compute_zscore_test,
|
22
|
+
)
|
23
|
+
|
24
|
+
|
25
|
+
class NeighborhoodsAPI:
|
26
|
+
"""Handles the loading of statistical results and annotation significance for neighborhoods.
|
27
|
+
|
28
|
+
The NeighborhoodsAPI class provides methods to load neighborhood results from statistical tests.
|
29
|
+
"""
|
30
|
+
|
31
|
+
def __init__() -> None:
|
32
|
+
pass
|
33
|
+
|
34
|
+
def load_neighborhoods_by_binom(
|
35
|
+
self,
|
36
|
+
network: nx.Graph,
|
37
|
+
annotations: Dict[str, Any],
|
38
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
39
|
+
louvain_resolution: float = 0.1,
|
40
|
+
leiden_resolution: float = 1.0,
|
41
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
42
|
+
null_distribution: str = "network",
|
43
|
+
random_seed: int = 888,
|
44
|
+
) -> Dict[str, Any]:
|
45
|
+
"""Load significant neighborhoods for the network using the binomial test.
|
46
|
+
|
47
|
+
Args:
|
48
|
+
network (nx.Graph): The network graph.
|
49
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
50
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
51
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
52
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
53
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
54
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
55
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
56
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
57
|
+
Defaults to 0.5.
|
58
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
59
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
60
|
+
|
61
|
+
Returns:
|
62
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
63
|
+
"""
|
64
|
+
log_header("Running binomial test")
|
65
|
+
# Compute neighborhood significance using the binomial test
|
66
|
+
return self._load_neighborhoods_by_statistical_test(
|
67
|
+
network=network,
|
68
|
+
annotations=annotations,
|
69
|
+
distance_metric=distance_metric,
|
70
|
+
louvain_resolution=louvain_resolution,
|
71
|
+
leiden_resolution=leiden_resolution,
|
72
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
73
|
+
null_distribution=null_distribution,
|
74
|
+
random_seed=random_seed,
|
75
|
+
statistical_test_key="binom",
|
76
|
+
statistical_test_function=compute_binom_test,
|
77
|
+
)
|
78
|
+
|
79
|
+
def load_neighborhoods_by_chi2(
|
80
|
+
self,
|
81
|
+
network: nx.Graph,
|
82
|
+
annotations: Dict[str, Any],
|
83
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
84
|
+
louvain_resolution: float = 0.1,
|
85
|
+
leiden_resolution: float = 1.0,
|
86
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
87
|
+
null_distribution: str = "network",
|
88
|
+
random_seed: int = 888,
|
89
|
+
) -> Dict[str, Any]:
|
90
|
+
"""Load significant neighborhoods for the network using the chi-squared test.
|
91
|
+
|
92
|
+
Args:
|
93
|
+
network (nx.Graph): The network graph.
|
94
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
95
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
96
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
97
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
98
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
99
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
100
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
101
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
102
|
+
Defaults to 0.5.
|
103
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
104
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
105
|
+
|
106
|
+
Returns:
|
107
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
108
|
+
"""
|
109
|
+
log_header("Running chi-squared test")
|
110
|
+
# Compute neighborhood significance using the chi-squared test
|
111
|
+
return self._load_neighborhoods_by_statistical_test(
|
112
|
+
network=network,
|
113
|
+
annotations=annotations,
|
114
|
+
distance_metric=distance_metric,
|
115
|
+
louvain_resolution=louvain_resolution,
|
116
|
+
leiden_resolution=leiden_resolution,
|
117
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
118
|
+
null_distribution=null_distribution,
|
119
|
+
random_seed=random_seed,
|
120
|
+
statistical_test_key="chi2",
|
121
|
+
statistical_test_function=compute_chi2_test,
|
122
|
+
)
|
123
|
+
|
124
|
+
def load_neighborhoods_by_hypergeom(
|
125
|
+
self,
|
126
|
+
network: nx.Graph,
|
127
|
+
annotations: Dict[str, Any],
|
128
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
129
|
+
louvain_resolution: float = 0.1,
|
130
|
+
leiden_resolution: float = 1.0,
|
131
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
132
|
+
null_distribution: str = "network",
|
133
|
+
random_seed: int = 888,
|
134
|
+
) -> Dict[str, Any]:
|
135
|
+
"""Load significant neighborhoods for the network using the hypergeometric test.
|
136
|
+
|
137
|
+
Args:
|
138
|
+
network (nx.Graph): The network graph.
|
139
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
140
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
141
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
142
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
143
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
144
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
145
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
146
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
147
|
+
Defaults to 0.5.
|
148
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
149
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
150
|
+
|
151
|
+
Returns:
|
152
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
153
|
+
"""
|
154
|
+
log_header("Running hypergeometric test")
|
155
|
+
# Compute neighborhood significance using the hypergeometric test
|
156
|
+
return self._load_neighborhoods_by_statistical_test(
|
157
|
+
network=network,
|
158
|
+
annotations=annotations,
|
159
|
+
distance_metric=distance_metric,
|
160
|
+
louvain_resolution=louvain_resolution,
|
161
|
+
leiden_resolution=leiden_resolution,
|
162
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
163
|
+
null_distribution=null_distribution,
|
164
|
+
random_seed=random_seed,
|
165
|
+
statistical_test_key="hypergeom",
|
166
|
+
statistical_test_function=compute_hypergeom_test,
|
167
|
+
)
|
168
|
+
|
169
|
+
def load_neighborhoods_by_permutation(
|
170
|
+
self,
|
171
|
+
network: nx.Graph,
|
172
|
+
annotations: Dict[str, Any],
|
173
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
174
|
+
louvain_resolution: float = 0.1,
|
175
|
+
leiden_resolution: float = 1.0,
|
176
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
177
|
+
score_metric: str = "sum",
|
178
|
+
null_distribution: str = "network",
|
179
|
+
num_permutations: int = 1000,
|
180
|
+
random_seed: int = 888,
|
181
|
+
max_workers: int = 1,
|
182
|
+
) -> Dict[str, Any]:
|
183
|
+
"""Load significant neighborhoods for the network using the permutation test.
|
184
|
+
|
185
|
+
Args:
|
186
|
+
network (nx.Graph): The network graph.
|
187
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
188
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
189
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
190
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
191
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
192
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
193
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
194
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
195
|
+
Defaults to 0.5.
|
196
|
+
score_metric (str, optional): Scoring metric for neighborhood significance. Defaults to "sum".
|
197
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
198
|
+
num_permutations (int, optional): Number of permutations for significance testing. Defaults to 1000.
|
199
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
200
|
+
max_workers (int, optional): Maximum number of workers for parallel computation. Defaults to 1.
|
201
|
+
|
202
|
+
Returns:
|
203
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
204
|
+
"""
|
205
|
+
log_header("Running permutation test")
|
206
|
+
# Log and display permutation test settings, which is unique to this test
|
207
|
+
logger.debug(f"Neighborhood scoring metric: '{score_metric}'")
|
208
|
+
logger.debug(f"Number of permutations: {num_permutations}")
|
209
|
+
logger.debug(f"Maximum workers: {max_workers}")
|
210
|
+
# Compute neighborhood significance using the permutation test
|
211
|
+
return self._load_neighborhoods_by_statistical_test(
|
212
|
+
network=network,
|
213
|
+
annotations=annotations,
|
214
|
+
distance_metric=distance_metric,
|
215
|
+
louvain_resolution=louvain_resolution,
|
216
|
+
leiden_resolution=leiden_resolution,
|
217
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
218
|
+
null_distribution=null_distribution,
|
219
|
+
random_seed=random_seed,
|
220
|
+
statistical_test_key="permutation",
|
221
|
+
statistical_test_function=compute_permutation_test,
|
222
|
+
score_metric=score_metric,
|
223
|
+
num_permutations=num_permutations,
|
224
|
+
max_workers=max_workers,
|
225
|
+
)
|
226
|
+
|
227
|
+
def load_neighborhoods_by_poisson(
|
228
|
+
self,
|
229
|
+
network: nx.Graph,
|
230
|
+
annotations: Dict[str, Any],
|
231
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
232
|
+
louvain_resolution: float = 0.1,
|
233
|
+
leiden_resolution: float = 1.0,
|
234
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
235
|
+
null_distribution: str = "network",
|
236
|
+
random_seed: int = 888,
|
237
|
+
) -> Dict[str, Any]:
|
238
|
+
"""Load significant neighborhoods for the network using the Poisson test.
|
239
|
+
|
240
|
+
Args:
|
241
|
+
network (nx.Graph): The network graph.
|
242
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
243
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
244
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
245
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
246
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
247
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
248
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
249
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
250
|
+
Defaults to 0.5.
|
251
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
252
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
253
|
+
|
254
|
+
Returns:
|
255
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
256
|
+
"""
|
257
|
+
log_header("Running Poisson test")
|
258
|
+
# Compute neighborhood significance using the Poisson test
|
259
|
+
return self._load_neighborhoods_by_statistical_test(
|
260
|
+
network=network,
|
261
|
+
annotations=annotations,
|
262
|
+
distance_metric=distance_metric,
|
263
|
+
louvain_resolution=louvain_resolution,
|
264
|
+
leiden_resolution=leiden_resolution,
|
265
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
266
|
+
null_distribution=null_distribution,
|
267
|
+
random_seed=random_seed,
|
268
|
+
statistical_test_key="poisson",
|
269
|
+
statistical_test_function=compute_poisson_test,
|
270
|
+
)
|
271
|
+
|
272
|
+
def load_neighborhoods_by_zscore(
|
273
|
+
self,
|
274
|
+
network: nx.Graph,
|
275
|
+
annotations: Dict[str, Any],
|
276
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
277
|
+
louvain_resolution: float = 0.1,
|
278
|
+
leiden_resolution: float = 1.0,
|
279
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
280
|
+
null_distribution: str = "network",
|
281
|
+
random_seed: int = 888,
|
282
|
+
) -> Dict[str, Any]:
|
283
|
+
"""Load significant neighborhoods for the network using the Z-score test.
|
284
|
+
|
285
|
+
Args:
|
286
|
+
network (nx.Graph): The network graph.
|
287
|
+
annotations (Dict[str, Any]): The annotations associated with the network.
|
288
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
289
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
290
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
291
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
292
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
293
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
294
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
295
|
+
Defaults to 0.5.
|
296
|
+
null_distribution (str, optional): Type of null distribution ('network' or 'annotations'). Defaults to "network".
|
297
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
298
|
+
|
299
|
+
Returns:
|
300
|
+
Dict[str, Any]: Computed significance of neighborhoods.
|
301
|
+
"""
|
302
|
+
log_header("Running Z-score test")
|
303
|
+
# Compute neighborhood significance using the Z-score test
|
304
|
+
return self._load_neighborhoods_by_statistical_test(
|
305
|
+
network=network,
|
306
|
+
annotations=annotations,
|
307
|
+
distance_metric=distance_metric,
|
308
|
+
louvain_resolution=louvain_resolution,
|
309
|
+
leiden_resolution=leiden_resolution,
|
310
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
311
|
+
null_distribution=null_distribution,
|
312
|
+
random_seed=random_seed,
|
313
|
+
statistical_test_key="zscore",
|
314
|
+
statistical_test_function=compute_zscore_test,
|
315
|
+
)
|
316
|
+
|
317
|
+
def _load_neighborhoods_by_statistical_test(
|
318
|
+
self,
|
319
|
+
network: nx.Graph,
|
320
|
+
annotations: Dict[str, Any],
|
321
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
322
|
+
louvain_resolution: float = 0.1,
|
323
|
+
leiden_resolution: float = 1.0,
|
324
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
325
|
+
null_distribution: str = "network",
|
326
|
+
random_seed: int = 888,
|
327
|
+
statistical_test_key: str = "hypergeom",
|
328
|
+
statistical_test_function: Any = compute_hypergeom_test,
|
329
|
+
**kwargs,
|
330
|
+
):
|
331
|
+
"""Load and compute significant neighborhoods for the network using a specified statistical test.
|
332
|
+
|
333
|
+
Args:
|
334
|
+
network (nx.Graph): The input network graph.
|
335
|
+
annotations (Dict[str, Any]): Annotation data associated with the network, including a "matrix" key with annotation values.
|
336
|
+
distance_metric (Union[str, List, Tuple, np.ndarray], optional): The distance metric or clustering method to define neighborhoods.
|
337
|
+
Can be a string specifying one method (e.g., 'louvain', 'leiden') or a collection of methods.
|
338
|
+
Defaults to "louvain".
|
339
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
340
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
341
|
+
fraction_shortest_edges (Union[float, List, Tuple, np.ndarray], optional): Fraction of shortest edges to consider for creating subgraphs.
|
342
|
+
Can be a single value or a collection of thresholds for flexibility. Defaults to 0.5.
|
343
|
+
null_distribution (str, optional): The type of null distribution to use ('network' or 'annotations').
|
344
|
+
Defaults to "network".
|
345
|
+
random_seed (int, optional): Seed for random number generation to ensure reproducibility. Defaults to 888.
|
346
|
+
statistical_test_key (str, optional): Key or name of the statistical test to be applied (e.g., "hypergeom", "poisson").
|
347
|
+
Used for logging and debugging. Defaults to "hypergeom".
|
348
|
+
statistical_test_function (Any, optional): The function implementing the statistical test.
|
349
|
+
It should accept neighborhoods, annotations, null distribution, and additional kwargs.
|
350
|
+
Defaults to `compute_hypergeom_test`.
|
351
|
+
**kwargs: Additional parameters to be passed to the statistical test function.
|
352
|
+
|
353
|
+
Returns:
|
354
|
+
Dict[str, Any]: A dictionary containing the computed significance values for neighborhoods.
|
355
|
+
"""
|
356
|
+
# Log null distribution type
|
357
|
+
logger.debug(f"Null distribution: '{null_distribution}'")
|
358
|
+
# Log neighborhood analysis parameters
|
359
|
+
params.log_neighborhoods(
|
360
|
+
distance_metric=distance_metric,
|
361
|
+
louvain_resolution=louvain_resolution,
|
362
|
+
leiden_resolution=leiden_resolution,
|
363
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
364
|
+
statistical_test_function=statistical_test_key,
|
365
|
+
null_distribution=null_distribution,
|
366
|
+
random_seed=random_seed,
|
367
|
+
**kwargs,
|
368
|
+
)
|
369
|
+
|
370
|
+
# Make a copy of the network to avoid modifying the original
|
371
|
+
network = copy.deepcopy(network)
|
372
|
+
# Load neighborhoods based on the network and distance metric
|
373
|
+
neighborhoods = self._load_neighborhoods(
|
374
|
+
network,
|
375
|
+
distance_metric,
|
376
|
+
louvain_resolution=louvain_resolution,
|
377
|
+
leiden_resolution=leiden_resolution,
|
378
|
+
fraction_shortest_edges=fraction_shortest_edges,
|
379
|
+
random_seed=random_seed,
|
380
|
+
)
|
381
|
+
# Apply statistical test function to compute neighborhood significance
|
382
|
+
neighborhood_significance = statistical_test_function(
|
383
|
+
neighborhoods=neighborhoods,
|
384
|
+
annotations=annotations["matrix"],
|
385
|
+
null_distribution=null_distribution,
|
386
|
+
**kwargs,
|
387
|
+
)
|
388
|
+
|
389
|
+
# Return the computed neighborhood significance
|
390
|
+
return neighborhood_significance
|
391
|
+
|
392
|
+
def _load_neighborhoods(
|
393
|
+
self,
|
394
|
+
network: nx.Graph,
|
395
|
+
distance_metric: Union[str, List, Tuple, np.ndarray] = "louvain",
|
396
|
+
louvain_resolution: float = 0.1,
|
397
|
+
leiden_resolution: float = 1.0,
|
398
|
+
fraction_shortest_edges: Union[float, List, Tuple, np.ndarray] = 0.5,
|
399
|
+
random_seed: int = 888,
|
400
|
+
) -> csr_matrix:
|
401
|
+
"""Load significant neighborhoods for the network.
|
402
|
+
|
403
|
+
Args:
|
404
|
+
network (nx.Graph): The network graph.
|
405
|
+
distance_metric (str, List, Tuple, or np.ndarray, optional): The distance metric(s) to use. Can be a string for one
|
406
|
+
metric or a list/tuple/ndarray of metrics ('greedy_modularity', 'louvain', 'leiden', 'label_propagation',
|
407
|
+
'markov_clustering', 'walktrap', 'spinglass'). Defaults to 'louvain'.
|
408
|
+
louvain_resolution (float, optional): Resolution parameter for Louvain clustering. Defaults to 0.1.
|
409
|
+
leiden_resolution (float, optional): Resolution parameter for Leiden clustering. Defaults to 1.0.
|
410
|
+
fraction_shortest_edges (float, List, Tuple, or np.ndarray, optional): Shortest edge rank fraction threshold(s) for creating subgraphs.
|
411
|
+
Can be a single float for one threshold or a list/tuple of floats corresponding to multiple thresholds.
|
412
|
+
Defaults to 0.5.
|
413
|
+
random_seed (int, optional): Seed for random number generation. Defaults to 888.
|
414
|
+
|
415
|
+
Returns:
|
416
|
+
csr_matrix: Sparse neighborhood matrix calculated based on the selected distance metric.
|
417
|
+
"""
|
418
|
+
# Display the chosen distance metric
|
419
|
+
if distance_metric == "louvain":
|
420
|
+
for_print_distance_metric = f"louvain (resolution={louvain_resolution})"
|
421
|
+
elif distance_metric == "leiden":
|
422
|
+
for_print_distance_metric = f"leiden (resolution={leiden_resolution})"
|
423
|
+
else:
|
424
|
+
for_print_distance_metric = distance_metric
|
425
|
+
|
426
|
+
# Log and display neighborhood settings
|
427
|
+
logger.debug(f"Distance metric: '{for_print_distance_metric}'")
|
428
|
+
logger.debug(f"Edge length threshold: {fraction_shortest_edges}")
|
429
|
+
logger.debug(f"Random seed: {random_seed}")
|
430
|
+
|
431
|
+
# Compute neighborhoods
|
432
|
+
neighborhoods = get_network_neighborhoods(
|
433
|
+
network,
|
434
|
+
distance_metric,
|
435
|
+
fraction_shortest_edges,
|
436
|
+
louvain_resolution=louvain_resolution,
|
437
|
+
leiden_resolution=leiden_resolution,
|
438
|
+
random_seed=random_seed,
|
439
|
+
)
|
440
|
+
|
441
|
+
# Ensure the neighborhood matrix is in sparse format
|
442
|
+
if not isinstance(neighborhoods, csr_matrix):
|
443
|
+
neighborhoods = csr_matrix(neighborhoods)
|
444
|
+
|
445
|
+
# Return the sparse neighborhood matrix
|
446
|
+
return neighborhoods
|