1
|
|
|
"""Functions to calculate mean squared displacements from trajectory data |
2
|
|
|
|
3
|
|
|
This module includes functions to calculate mean squared displacements and |
4
|
|
|
additional measures from input trajectory datasets as calculated by the |
5
|
|
|
Trackmate ImageJ plugin. |
6
|
|
|
|
7
|
|
|
""" |
8
|
|
|
import warnings |
9
|
|
|
import random as rand |
10
|
|
|
|
11
|
|
|
import pandas as pd |
12
|
|
|
import numpy as np |
13
|
|
|
import numpy.ma as ma |
14
|
|
|
import scipy.stats as stats |
15
|
|
|
from scipy import interpolate |
16
|
|
|
import matplotlib.pyplot as plt |
17
|
|
|
from matplotlib.pyplot import cm |
18
|
|
|
import diff_classifier.aws as aws |
19
|
|
|
from scipy.ndimage.morphology import distance_transform_edt as eudist |
20
|
|
|
|
21
|
|
|
|
22
|
|
View Code Duplication |
def nth_diff(dataframe, n=1, axis=0): |
|
|
|
|
23
|
|
|
"""Calculates the nth difference between vector elements |
24
|
|
|
|
25
|
|
|
Returns a new vector of size N - n containing the nth difference between |
26
|
|
|
vector elements. |
27
|
|
|
|
28
|
|
|
Parameters |
29
|
|
|
---------- |
30
|
|
|
dataframe : pandas.core.series.Series of int or float |
31
|
|
|
Input data on which differences are to be calculated. |
32
|
|
|
n : int |
33
|
|
|
Function calculated xpos(i) - xpos(i - n) for all values in pandas |
34
|
|
|
series. |
35
|
|
|
axis : {0, 1} |
36
|
|
|
Axis along which differences are to be calculated. Default is 0. If 0, |
37
|
|
|
input must be a pandas series. If 1, input must be a numpy array. |
38
|
|
|
|
39
|
|
|
Returns |
40
|
|
|
------- |
41
|
|
|
diff : pandas.core.series.Series of int or float |
42
|
|
|
Pandas series of size N - n, where N is the original size of dataframe. |
43
|
|
|
|
44
|
|
|
Examples |
45
|
|
|
-------- |
46
|
|
|
>>> df = np.ones((5, 10)) |
47
|
|
|
>>> nth_diff(df) |
48
|
|
|
array([[0., 0., 0., 0., 0., 0., 0., 0., 0.], |
49
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0.], |
50
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0.], |
51
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0.], |
52
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0.]]) |
53
|
|
|
|
54
|
|
|
>>> df = np.ones((5, 10)) |
55
|
|
|
>>> nth_diff (df) |
56
|
|
|
array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], |
57
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], |
58
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], |
59
|
|
|
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]]) |
60
|
|
|
|
61
|
|
|
""" |
62
|
|
|
|
63
|
|
|
assert isinstance(n, int), "n must be an integer." |
64
|
|
|
|
65
|
|
|
if dataframe.ndim == 1: |
66
|
|
|
length = dataframe.shape[0] |
67
|
|
|
if n <= length: |
68
|
|
|
test1 = dataframe[:-n].reset_index(drop=True) |
69
|
|
|
test2 = dataframe[n:].reset_index(drop=True) |
70
|
|
|
diff = test2 - test1 |
71
|
|
|
else: |
72
|
|
|
diff = np.array([np.nan, np.nan]) |
73
|
|
|
else: |
74
|
|
|
length = dataframe.shape[0] |
75
|
|
|
if n <= length: |
76
|
|
|
if axis == 0: |
77
|
|
|
test1 = dataframe[:-n, :] |
78
|
|
|
test2 = dataframe[n:, :] |
79
|
|
|
else: |
80
|
|
|
test1 = dataframe[:, :-n] |
81
|
|
|
test2 = dataframe[:, n:] |
82
|
|
|
diff = test2 - test1 |
83
|
|
|
else: |
84
|
|
|
diff = np.array([np.nan, np.nan]) |
85
|
|
|
|
86
|
|
|
return diff |
87
|
|
|
|
88
|
|
|
|
89
|
|
View Code Duplication |
def msd_calc(track, length=10): |
|
|
|
|
90
|
|
|
"""Calculates mean squared displacement of input track. |
91
|
|
|
|
92
|
|
|
Returns numpy array containing MSD data calculated from an individual track. |
93
|
|
|
|
94
|
|
|
Parameters |
95
|
|
|
---------- |
96
|
|
|
track : pandas.core.frame.DataFrame |
97
|
|
|
Contains, at a minimum a 'Frame', 'X', and 'Y' column |
98
|
|
|
|
99
|
|
|
Returns |
100
|
|
|
------- |
101
|
|
|
new_track : pandas.core.frame.DataFrame |
102
|
|
|
Similar to input track. All missing frames of individual trajectories |
103
|
|
|
are filled in with NaNs, and two new columns, MSDs and Gauss are added: |
104
|
|
|
MSDs, calculated mean squared displacements using the formula |
105
|
|
|
MSD = <(xpos-x0)**2> |
106
|
|
|
Gauss, calculated Gaussianity |
107
|
|
|
|
108
|
|
|
Examples |
109
|
|
|
-------- |
110
|
|
|
>>> data1 = {'Frame': [1, 2, 3, 4, 5], |
111
|
|
|
... 'X': [5, 6, 7, 8, 9], |
112
|
|
|
... 'Y': [6, 7, 8, 9, 10]} |
113
|
|
|
>>> df = pd.DataFrame(data=data1) |
114
|
|
|
>>> new_track = msd.msd_calc(df, 5) |
115
|
|
|
|
116
|
|
|
>>> data1 = {'Frame': [1, 2, 3, 4, 5], |
117
|
|
|
... 'X': [5, 6, 7, 8, 9], |
118
|
|
|
... 'Y': [6, 7, 8, 9, 10]} |
119
|
|
|
>>> df = pd.DataFrame(data=data1) |
120
|
|
|
>>> new_track = msd.msd_calc(df) |
121
|
|
|
|
122
|
|
|
""" |
123
|
|
|
|
124
|
|
|
meansd = np.zeros(length) |
125
|
|
|
gauss = np.zeros(length) |
126
|
|
|
new_frame = np.linspace(1, length, length) |
127
|
|
|
old_frame = track['Frame'] |
128
|
|
|
oldxy = [track['X'], track['Y']] |
129
|
|
|
fxy = [interpolate.interp1d(old_frame, oldxy[0], bounds_error=False, |
130
|
|
|
fill_value=np.nan), |
131
|
|
|
interpolate.interp1d(old_frame, oldxy[1], bounds_error=False, |
132
|
|
|
fill_value=np.nan)] |
133
|
|
|
|
134
|
|
|
intxy = [ma.masked_equal(fxy[0](new_frame), np.nan), |
135
|
|
|
ma.masked_equal(fxy[1](new_frame), np.nan)] |
136
|
|
|
data1 = {'Frame': new_frame, |
137
|
|
|
'X': intxy[0], |
138
|
|
|
'Y': intxy[1] |
139
|
|
|
} |
140
|
|
|
new_track = pd.DataFrame(data=data1) |
141
|
|
|
|
142
|
|
|
for frame in range(0, length-1): |
|
|
|
|
143
|
|
|
xy = [np.square(nth_diff(new_track['X'], n=frame+1)), |
144
|
|
|
np.square(nth_diff(new_track['Y'], n=frame+1))] |
145
|
|
|
with warnings.catch_warnings(): |
146
|
|
|
warnings.simplefilter("ignore", category=RuntimeWarning) |
147
|
|
|
meansd[frame+1] = np.nanmean(xy[0] + xy[1]) |
148
|
|
|
gauss[frame+1] = np.nanmean(xy[0]**2 + xy[1]**2 |
149
|
|
|
)/(2*(meansd[frame+1]**2)) |
150
|
|
|
|
151
|
|
|
new_track['MSDs'] = pd.Series(meansd, index=new_track.index) |
152
|
|
|
new_track['Gauss'] = pd.Series(gauss, index=new_track.index) |
153
|
|
|
|
154
|
|
|
return new_track |
155
|
|
|
|
156
|
|
|
|
157
|
|
View Code Duplication |
def all_msds(data): |
|
|
|
|
158
|
|
|
"""Calculates mean squared displacements of a trajectory dataset |
159
|
|
|
|
160
|
|
|
Returns numpy array containing MSD data of all tracks in a trajectory |
161
|
|
|
pandas dataframe. |
162
|
|
|
|
163
|
|
|
Parameters |
164
|
|
|
---------- |
165
|
|
|
data : pandas.core.frame.DataFrame |
166
|
|
|
Contains, at a minimum a 'Frame', 'Track_ID', 'X', and |
167
|
|
|
'Y' column. Note: it is assumed that frames begins at 1, not 0 with this |
168
|
|
|
function. Adjust before feeding into function. |
169
|
|
|
|
170
|
|
|
Returns |
171
|
|
|
------- |
172
|
|
|
new_data : pandas.core.frame.DataFrame |
173
|
|
|
Similar to input data. All missing frames of individual trajectories |
174
|
|
|
are filled in with NaNs, and two new columns, MSDs and Gauss are added: |
175
|
|
|
MSDs, calculated mean squared displacements using the formula |
176
|
|
|
MSD = <(xpos-x0)**2> |
177
|
|
|
Gauss, calculated Gaussianity |
178
|
|
|
|
179
|
|
|
Examples |
180
|
|
|
-------- |
181
|
|
|
>>> data1 = {'Frame': [1, 2, 3, 4, 5, 1, 2, 3, 4, 5], |
182
|
|
|
... 'Track_ID': [1, 1, 1, 1, 1, 2, 2, 2, 2, 2], |
183
|
|
|
... 'X': [5, 6, 7, 8, 9, 1, 2, 3, 4, 5], |
184
|
|
|
... 'Y': [6, 7, 8, 9, 10, 2, 3, 4, 5, 6]} |
185
|
|
|
>>> df = pd.DataFrame(data=data1) |
186
|
|
|
>>> all_msds(df) |
187
|
|
|
|
188
|
|
|
""" |
189
|
|
|
|
190
|
|
|
trackids = data.Track_ID.unique() |
191
|
|
|
partcount = trackids.shape[0] |
192
|
|
|
length = int(max(data['Frame'])) |
193
|
|
|
new = {} |
194
|
|
|
new['length'] = partcount*length |
195
|
|
|
new['frame'] = np.zeros(new['length']) |
196
|
|
|
new['ID'] = np.zeros(new['length']) |
197
|
|
|
new['xy'] = [np.zeros(new['length']), |
198
|
|
|
np.zeros(new['length'])] |
199
|
|
|
meansd = np.zeros(new['length']) |
200
|
|
|
gauss = np.zeros(new['length']) |
201
|
|
|
|
202
|
|
|
for particle in range(0, partcount): |
|
|
|
|
203
|
|
|
single_track = data.loc[data['Track_ID'] == |
204
|
|
|
trackids[particle] |
205
|
|
|
].sort_values(['Track_ID', 'Frame'], |
206
|
|
|
ascending=[1, 1] |
207
|
|
|
).reset_index(drop=True) |
208
|
|
|
if particle == 0: |
209
|
|
|
index1 = 0 |
210
|
|
|
index2 = length |
211
|
|
|
else: |
212
|
|
|
index1 = index2 |
|
|
|
|
213
|
|
|
index2 = index2 + length |
214
|
|
|
new['single_track'] = msd_calc(single_track, length=length) |
215
|
|
|
new['frame'][index1:index2] = np.linspace(1, length, length) |
216
|
|
|
new['ID'][index1:index2] = particle+1 |
217
|
|
|
new['xy'][0][index1:index2] = new['single_track']['X'] |
218
|
|
|
new['xy'][1][index1:index2] = new['single_track']['Y'] |
219
|
|
|
meansd[index1:index2] = new['single_track']['MSDs'] |
220
|
|
|
gauss[index1:index2] = new['single_track']['Gauss'] |
221
|
|
|
|
222
|
|
|
data1 = {'Frame': new['frame'], |
223
|
|
|
'Track_ID': new['ID'], |
224
|
|
|
'X': new['xy'][0], |
225
|
|
|
'Y': new['xy'][1], |
226
|
|
|
'MSDs': meansd, |
227
|
|
|
'Gauss': gauss} |
228
|
|
|
new_data = pd.DataFrame(data=data1) |
229
|
|
|
|
230
|
|
|
return new_data |
231
|
|
|
|
232
|
|
|
|
233
|
|
View Code Duplication |
def make_xyarray(data, length=651): |
|
|
|
|
234
|
|
|
"""Rearranges xy position data into 2d arrays |
235
|
|
|
|
236
|
|
|
Rearranges xy data from input pandas dataframe into 2D numpy array. |
237
|
|
|
|
238
|
|
|
Parameters |
239
|
|
|
---------- |
240
|
|
|
data : pd.core.frame.DataFrame |
241
|
|
|
Contains, at a minimum a 'Frame', 'Track_ID', 'X', and |
242
|
|
|
'Y' column. |
243
|
|
|
length : int |
244
|
|
|
Desired length or number of frames to which to extend trajectories. |
245
|
|
|
Any trajectories shorter than the input length will have the extra space |
246
|
|
|
filled in with NaNs. |
247
|
|
|
|
248
|
|
|
Returns |
249
|
|
|
------- |
250
|
|
|
xyft : dict of np.ndarray |
251
|
|
|
Dictionary containing xy position data, frame data, and trajectory ID |
252
|
|
|
data. Contains the following keys: |
253
|
|
|
farray, frames data (length x particles) |
254
|
|
|
tarray, trajectory ID data (length x particles) |
255
|
|
|
xarray, x position data (length x particles) |
256
|
|
|
yarray, y position data (length x particles) |
257
|
|
|
|
258
|
|
|
Examples |
259
|
|
|
-------- |
260
|
|
|
>>> data1 = {'Frame': [0, 1, 2, 3, 4, 2, 3, 4, 5, 6], |
261
|
|
|
... 'Track_ID': [1, 1, 1, 1, 1, 2, 2, 2, 2, 2], |
262
|
|
|
... 'X': [5, 6, 7, 8, 9, 1, 2, 3, 4, 5], |
263
|
|
|
... 'Y': [6, 7, 8, 9, 10, 2, 3, 4, 5, 6]} |
264
|
|
|
>>> df = pd.DataFrame(data=data1) |
265
|
|
|
>>> length = max(df['Frame']) + 1 |
266
|
|
|
>>> xyft = msd.make_xyarray(df, length=length) |
267
|
|
|
{'farray': array([[0., 0.], |
268
|
|
|
[1., 1.], |
269
|
|
|
[2., 2.], |
270
|
|
|
[3., 3.], |
271
|
|
|
[4., 4.], |
272
|
|
|
[5., 5.], |
273
|
|
|
[6., 6.]]), |
274
|
|
|
'tarray': array([[1., 2.], |
275
|
|
|
[1., 2.], |
276
|
|
|
[1., 2.], |
277
|
|
|
[1., 2.], |
278
|
|
|
[1., 2.], |
279
|
|
|
[1., 2.], |
280
|
|
|
[1., 2.]]), |
281
|
|
|
'xarray': array([[ 5., nan], |
282
|
|
|
[ 6., nan], |
283
|
|
|
[ 7., 1.], |
284
|
|
|
[ 8., 2.], |
285
|
|
|
[ 9., 3.], |
286
|
|
|
[nan, 4.], |
287
|
|
|
'yarray': [nan, 5.]]), |
288
|
|
|
array([[ 6., nan], |
289
|
|
|
[ 7., nan], |
290
|
|
|
[ 8., 2.], |
291
|
|
|
[ 9., 3.], |
292
|
|
|
[10., 4.], |
293
|
|
|
[nan, 5.], |
294
|
|
|
[nan, 6.]])} |
295
|
|
|
|
296
|
|
|
""" |
297
|
|
|
|
298
|
|
|
# Initial values |
299
|
|
|
first_p = int(min(data['Track_ID'])) |
300
|
|
|
particles = int(max(data['Track_ID'])) - first_p + 1 |
301
|
|
|
xyft = {} |
302
|
|
|
xyft['xarray'] = np.zeros((length, particles)) |
303
|
|
|
xyft['yarray'] = np.zeros((length, particles)) |
304
|
|
|
xyft['farray'] = np.zeros((length, particles)) |
305
|
|
|
xyft['tarray'] = np.zeros((length, particles)) |
306
|
|
|
xyft['qarray'] = np.zeros((length, particles)) |
307
|
|
|
xyft['snarray'] = np.zeros((length, particles)) |
308
|
|
|
xyft['iarray'] = np.zeros((length, particles)) |
309
|
|
|
|
310
|
|
|
track = data[data['Track_ID'] == first_p |
311
|
|
|
].sort_values(['Track_ID', 'Frame'], |
312
|
|
|
ascending=[1, 1]).reset_index(drop=True) |
313
|
|
|
new_frame = np.linspace(0, length-1, length) |
314
|
|
|
|
315
|
|
|
old_frame = track['Frame'].values.astype(float) |
316
|
|
|
oldxy = [track['X'].values, |
317
|
|
|
track['Y'].values, |
318
|
|
|
track['Quality'].values, |
319
|
|
|
track['SN_Ratio'].values, |
320
|
|
|
track['Mean_Intensity'].values] |
321
|
|
|
fxy = [interpolate.interp1d(old_frame, oldxy[0], bounds_error=False, |
322
|
|
|
fill_value=np.nan), |
323
|
|
|
interpolate.interp1d(old_frame, oldxy[1], bounds_error=False, |
324
|
|
|
fill_value=np.nan), |
325
|
|
|
interpolate.interp1d(old_frame, oldxy[2], bounds_error=False, |
326
|
|
|
fill_value=np.nan), |
327
|
|
|
interpolate.interp1d(old_frame, oldxy[3], bounds_error=False, |
328
|
|
|
fill_value=np.nan), |
329
|
|
|
interpolate.interp1d(old_frame, oldxy[4], bounds_error=False, |
330
|
|
|
fill_value=np.nan)] |
331
|
|
|
|
332
|
|
|
intxy = [fxy[0](new_frame), fxy[1](new_frame), fxy[2](new_frame), |
333
|
|
|
fxy[3](new_frame), fxy[4](new_frame)] |
334
|
|
|
|
335
|
|
|
# Fill in entire array |
336
|
|
|
xyft['xarray'][:, 0] = intxy[0] |
337
|
|
|
xyft['yarray'][:, 0] = intxy[1] |
338
|
|
|
xyft['farray'][:, 0] = new_frame |
339
|
|
|
xyft['tarray'][:, 0] = first_p |
340
|
|
|
xyft['qarray'][:, 0] = intxy[2] |
341
|
|
|
xyft['snarray'][:, 0] = intxy[3] |
342
|
|
|
xyft['iarray'][:, 0] = intxy[4] |
343
|
|
|
|
344
|
|
|
for part in range(first_p+1, first_p+particles): |
|
|
|
|
345
|
|
|
track = data[data['Track_ID'] == part |
346
|
|
|
].sort_values(['Track_ID', 'Frame'], |
347
|
|
|
ascending=[1, 1]).reset_index(drop=True) |
348
|
|
|
|
349
|
|
|
old_frame = track['Frame'] |
350
|
|
|
oldxy = [track['X'].values, |
351
|
|
|
track['Y'].values, |
352
|
|
|
track['Quality'].values, |
353
|
|
|
track['SN_Ratio'].values, |
354
|
|
|
track['Mean_Intensity'].values] |
355
|
|
|
|
356
|
|
|
fxy = [interpolate.interp1d(old_frame, oldxy[0], bounds_error=False, |
357
|
|
|
fill_value=np.nan), |
358
|
|
|
interpolate.interp1d(old_frame, oldxy[1], bounds_error=False, |
359
|
|
|
fill_value=np.nan), |
360
|
|
|
interpolate.interp1d(old_frame, oldxy[2], bounds_error=False, |
361
|
|
|
fill_value=np.nan), |
362
|
|
|
interpolate.interp1d(old_frame, oldxy[3], bounds_error=False, |
363
|
|
|
fill_value=np.nan), |
364
|
|
|
interpolate.interp1d(old_frame, oldxy[4], bounds_error=False, |
365
|
|
|
fill_value=np.nan)] |
366
|
|
|
|
367
|
|
|
intxy = [fxy[0](new_frame), fxy[1](new_frame), fxy[2](new_frame), |
368
|
|
|
fxy[3](new_frame), fxy[4](new_frame)] |
369
|
|
|
|
370
|
|
|
xyft['xarray'][:, part-first_p] = intxy[0] |
371
|
|
|
xyft['yarray'][:, part-first_p] = intxy[1] |
372
|
|
|
xyft['farray'][:, part-first_p] = new_frame |
373
|
|
|
xyft['tarray'][:, part-first_p] = part |
374
|
|
|
xyft['qarray'][:, part-first_p] = intxy[2] |
375
|
|
|
xyft['snarray'][:, part-first_p] = intxy[3] |
376
|
|
|
xyft['iarray'][:, part-first_p] = intxy[4] |
377
|
|
|
|
378
|
|
|
return xyft |
379
|
|
|
|
380
|
|
|
|
381
|
|
View Code Duplication |
def all_msds2(data, frames=651): |
|
|
|
|
382
|
|
|
"""Calculates mean squared displacements of input trajectory dataset |
383
|
|
|
|
384
|
|
|
Returns numpy array containing MSD data of all tracks in a trajectory pandas |
385
|
|
|
dataframe. |
386
|
|
|
|
387
|
|
|
Parameters |
388
|
|
|
---------- |
389
|
|
|
data : pandas.core.frame.DataFrame |
390
|
|
|
Contains, at a minimum a 'Frame', 'Track_ID', 'X', and |
391
|
|
|
'Y' column. Note: it is assumed that frames begins at 0. |
392
|
|
|
|
393
|
|
|
Returns |
394
|
|
|
------- |
395
|
|
|
new_data : pandas.core.frame.DataFrame |
396
|
|
|
Similar to input data. All missing frames of individual trajectories |
397
|
|
|
are filled in with NaNs, and two new columns, MSDs and Gauss are added: |
398
|
|
|
MSDs, calculated mean squared displacements using the formula |
399
|
|
|
MSD = <(xpos-x0)**2> |
400
|
|
|
Gauss, calculated Gaussianity |
401
|
|
|
|
402
|
|
|
Examples |
403
|
|
|
-------- |
404
|
|
|
>>> data1 = {'Frame': [0, 1, 2, 3, 4, 0, 1, 2, 3, 4], |
405
|
|
|
... 'Track_ID': [1, 1, 1, 1, 1, 2, 2, 2, 2, 2], |
406
|
|
|
... 'X': [5, 6, 7, 8, 9, 1, 2, 3, 4, 5], |
407
|
|
|
... 'Y': [6, 7, 8, 9, 10, 2, 3, 4, 5, 6]} |
408
|
|
|
>>> df = pd.DataFrame(data=data1) |
409
|
|
|
>>> cols = ['Frame', 'Track_ID', 'X', 'Y', 'MSDs', 'Gauss'] |
410
|
|
|
>>> om flength = max(df['Frame']) + 1 |
411
|
|
|
>>> msd.all_msds2(df, frames=length)[cols] |
412
|
|
|
|
413
|
|
|
""" |
414
|
|
|
if data.shape[0] > 2: |
415
|
|
|
try: |
416
|
|
|
xyft = make_xyarray(data, length=frames) |
417
|
|
|
length = xyft['xarray'].shape[0] |
418
|
|
|
particles = xyft['xarray'].shape[1] |
419
|
|
|
|
420
|
|
|
meansd = np.zeros((length, particles)) |
421
|
|
|
gauss = np.zeros((length, particles)) |
422
|
|
|
|
423
|
|
|
for frame in range(0, length-1): |
|
|
|
|
424
|
|
|
xpos = np.square(nth_diff(xyft['xarray'], n=frame+1)) |
425
|
|
|
ypos = np.square(nth_diff(xyft['yarray'], n=frame+1)) |
426
|
|
|
|
427
|
|
|
with warnings.catch_warnings(): |
428
|
|
|
warnings.simplefilter("ignore", category=RuntimeWarning) |
429
|
|
|
meansd[frame+1, :] = np.nanmean(xpos + ypos, axis=0) |
430
|
|
|
gauss[frame+1, :] = np.nanmean(xpos**2 + ypos**2, axis=0 |
431
|
|
|
)/(2*(meansd[frame+1]**2)) |
432
|
|
|
|
433
|
|
|
data1 = {'Frame': xyft['farray'].flatten('F'), |
434
|
|
|
'Track_ID': xyft['tarray'].flatten('F'), |
435
|
|
|
'X': xyft['xarray'].flatten('F'), |
436
|
|
|
'Y': xyft['yarray'].flatten('F'), |
437
|
|
|
'MSDs': meansd.flatten('F'), |
438
|
|
|
'Gauss': gauss.flatten('F'), |
439
|
|
|
'Quality': xyft['qarray'].flatten('F'), |
440
|
|
|
'SN_Ratio': xyft['snarray'].flatten('F'), |
441
|
|
|
'Mean_Intensity': xyft['iarray'].flatten('F')} |
442
|
|
|
|
443
|
|
|
new_data = pd.DataFrame(data=data1) |
444
|
|
|
except ValueError: |
445
|
|
|
data1 = {'Frame': [], |
446
|
|
|
'Track_ID': [], |
447
|
|
|
'X': [], |
448
|
|
|
'Y': [], |
449
|
|
|
'MSDs': [], |
450
|
|
|
'Gauss': [], |
451
|
|
|
'Quality': [], |
452
|
|
|
'SN_Ratio': [], |
453
|
|
|
'Mean_Intensity': []} |
454
|
|
|
new_data = pd.DataFrame(data=data1) |
455
|
|
|
except IndexError: |
456
|
|
|
data1 = {'Frame': [], |
457
|
|
|
'Track_ID': [], |
458
|
|
|
'X': [], |
459
|
|
|
'Y': [], |
460
|
|
|
'MSDs': [], |
461
|
|
|
'Gauss': [], |
462
|
|
|
'Quality': [], |
463
|
|
|
'SN_Ratio': [], |
464
|
|
|
'Mean_Intensity': []} |
465
|
|
|
new_data = pd.DataFrame(data=data1) |
466
|
|
|
else: |
467
|
|
|
data1 = {'Frame': [], |
468
|
|
|
'Track_ID': [], |
469
|
|
|
'X': [], |
470
|
|
|
'Y': [], |
471
|
|
|
'MSDs': [], |
472
|
|
|
'Gauss': [], |
473
|
|
|
'Quality': [], |
474
|
|
|
'SN_Ratio': [], |
475
|
|
|
'Mean_Intensity': []} |
476
|
|
|
new_data = pd.DataFrame(data=data1) |
477
|
|
|
|
478
|
|
|
return new_data |
479
|
|
|
|
480
|
|
|
|
481
|
|
View Code Duplication |
def geomean_msdisp(prefix, umppx=0.16, fps=100.02, upload=True, |
|
|
|
|
482
|
|
|
remote_folder="01_18_Experiment", bucket='ccurtis.data', |
483
|
|
|
backup_frames=651): |
484
|
|
|
"""Comptes geometric averages of mean squared displacement datasets |
485
|
|
|
|
486
|
|
|
Calculates geometric averages and stadard errors for MSD datasets. Might |
487
|
|
|
error out if not formatted as output from all_msds2. |
488
|
|
|
|
489
|
|
|
Parameters |
490
|
|
|
---------- |
491
|
|
|
prefix : string |
492
|
|
|
Prefix of file name to be plotted e.g. features_P1.csv prefix is P1. |
493
|
|
|
umppx : float |
494
|
|
|
Microns per pixel of original images. |
495
|
|
|
fps : float |
496
|
|
|
Frames per second of video. |
497
|
|
|
upload : bool |
498
|
|
|
True if you want to upload to s3. |
499
|
|
|
remote_folder : string |
500
|
|
|
Folder in S3 bucket to upload to. |
501
|
|
|
bucket : string |
502
|
|
|
Name of S3 bucket to upload to. |
503
|
|
|
|
504
|
|
|
Returns |
505
|
|
|
------- |
506
|
|
|
geo_mean : numpy.ndarray |
507
|
|
|
Geometric mean of trajectory MSDs at all time points. |
508
|
|
|
geo_stder : numpy.ndarray |
509
|
|
|
Geometric standard errot of trajectory MSDs at all time points. |
510
|
|
|
|
511
|
|
|
""" |
512
|
|
|
|
513
|
|
|
merged = pd.read_csv('msd_{}.csv'.format(prefix)) |
514
|
|
|
try: |
515
|
|
|
particles = int(max(merged['Track_ID'])) |
516
|
|
|
frames = int(max(merged['Frame'])) |
517
|
|
|
ypos = np.zeros((particles+1, frames+1)) |
518
|
|
|
|
519
|
|
|
for i in range(0, particles+1): |
|
|
|
|
520
|
|
|
ypos[i, :] = merged.loc[merged.Track_ID == i, 'MSDs']*umppx*umppx |
521
|
|
|
xpos = merged.loc[merged.Track_ID == i, 'Frame']/fps |
522
|
|
|
|
523
|
|
|
geo_mean = np.nanmean(ma.log(ypos), axis=0) |
524
|
|
|
geo_stder = ma.masked_equal(stats.sem(ma.log(ypos), axis=0, |
525
|
|
|
nan_policy='omit'), 0.0) |
526
|
|
|
|
527
|
|
|
except ValueError: |
528
|
|
|
geo_mean = np.nan*np.ones(backup_frames) |
529
|
|
|
geo_stder = np.nan*np.ones(backup_frames) |
530
|
|
|
|
531
|
|
|
np.savetxt('geomean_{}.csv'.format(prefix), geo_mean, delimiter=",") |
532
|
|
|
np.savetxt('geoSEM_{}.csv'.format(prefix), geo_stder, delimiter=",") |
533
|
|
|
|
534
|
|
|
if upload: |
535
|
|
|
aws.upload_s3('geomean_{}.csv'.format(prefix), |
536
|
|
|
remote_folder+'/'+'geomean_{}.csv'.format(prefix), |
537
|
|
|
bucket_name=bucket) |
538
|
|
|
aws.upload_s3('geoSEM_{}.csv'.format(prefix), |
539
|
|
|
remote_folder+'/'+'geoSEM_{}.csv'.format(prefix), |
540
|
|
|
bucket_name=bucket) |
541
|
|
|
|
542
|
|
|
return geo_mean, geo_stder |
543
|
|
|
|
544
|
|
|
|
545
|
|
View Code Duplication |
def binning(experiments, wells=4, prefix='test'): |
|
|
|
|
546
|
|
|
"""Split set of input experiments into groups. |
547
|
|
|
|
548
|
|
|
Parameters |
549
|
|
|
---------- |
550
|
|
|
experiments : list of str |
551
|
|
|
List of experiment names. |
552
|
|
|
wells : int |
553
|
|
|
Number of groups to divide experiments into. |
554
|
|
|
|
555
|
|
|
Returns |
556
|
|
|
------- |
557
|
|
|
slices : int |
558
|
|
|
Number of experiments per group. |
559
|
|
|
bins : dict of list of str |
560
|
|
|
Dictionary, keys corresponding to group names, and elements containing |
561
|
|
|
lists of experiments in each group. |
562
|
|
|
bin_names : list of str |
563
|
|
|
List of group names |
564
|
|
|
|
565
|
|
|
""" |
566
|
|
|
|
567
|
|
|
total_videos = len(experiments) |
568
|
|
|
bins = {} |
569
|
|
|
slices = int(total_videos/wells) |
570
|
|
|
bin_names = [] |
571
|
|
|
|
572
|
|
|
for num in range(0, wells): |
|
|
|
|
573
|
|
|
slice1 = num*slices |
574
|
|
|
slice2 = (num+1)*(slices) |
575
|
|
|
pref = '{}_W{}'.format(prefix, num) |
576
|
|
|
bins[pref] = experiments[slice1:slice2] |
577
|
|
|
bin_names.append(pref) |
578
|
|
|
return slices, bins, bin_names |
579
|
|
|
|
580
|
|
|
|
581
|
|
View Code Duplication |
def precision_weight(group, geo_stder): |
|
|
|
|
582
|
|
|
"""Calculates precision-based weights from input standard error data |
583
|
|
|
|
584
|
|
|
Calculates precision weights to be used in precision-averaged MSD |
585
|
|
|
calculations. |
586
|
|
|
|
587
|
|
|
Parameters |
588
|
|
|
---------- |
589
|
|
|
group : list of str |
590
|
|
|
List of experiment names to average. Each element corresponds to a key |
591
|
|
|
in geo_stder and geomean. |
592
|
|
|
geo_stder : dict of numpy.ndarray |
593
|
|
|
Each entry in dictionary corresponds to the standard errors of an MSD |
594
|
|
|
profile, the key corresponding to an experiment name. |
595
|
|
|
|
596
|
|
|
Returns |
597
|
|
|
------- |
598
|
|
|
weights: numpy.ndarray |
599
|
|
|
Precision weights to be used in precision averaging. |
600
|
|
|
w_holder : numpy.ndarray |
601
|
|
|
Precision values of each video at each time point. |
602
|
|
|
|
603
|
|
|
""" |
604
|
|
|
|
605
|
|
|
frames = np.shape(geo_stder[group[0]])[0] |
606
|
|
|
slices = len(group) |
607
|
|
|
video_counter = 0 |
608
|
|
|
w_holder = np.zeros((slices, frames)) |
609
|
|
|
for sample in group: |
610
|
|
|
w_holder[video_counter, :] = 1/(geo_stder[sample]*geo_stder[sample]) |
611
|
|
|
video_counter = video_counter + 1 |
612
|
|
|
|
613
|
|
|
w_holder = ma.masked_equal(w_holder, 0.0) |
614
|
|
|
w_holder = ma.masked_equal(w_holder, 1.0) |
615
|
|
|
|
616
|
|
|
weights = ma.sum(w_holder, axis=0) |
617
|
|
|
|
618
|
|
|
return weights, w_holder |
619
|
|
|
|
620
|
|
|
|
621
|
|
View Code Duplication |
def precision_averaging(group, geomean, geo_stder, weights, save=True, |
|
|
|
|
622
|
|
|
bucket='ccurtis.data', folder='test', |
623
|
|
|
experiment='test'): |
624
|
|
|
"""Calculates precision-weighted averages of MSD datasets. |
625
|
|
|
|
626
|
|
|
Parameters |
627
|
|
|
---------- |
628
|
|
|
group : list of str |
629
|
|
|
List of experiment names to average. Each element corresponds to a key |
630
|
|
|
in geo_stder and geomean. |
631
|
|
|
geomean : dict of numpy.ndarray |
632
|
|
|
Each entry in dictionary corresponds to an MSD profiles, they key |
633
|
|
|
corresponding to an experiment name. |
634
|
|
|
geo_stder : dict of numpy.ndarray |
635
|
|
|
Each entry in dictionary corresponds to the standard errors of an MSD |
636
|
|
|
profile, the key corresponding to an experiment name. |
637
|
|
|
weights : numpy.ndarray |
638
|
|
|
Precision weights to be used in precision averaging. |
639
|
|
|
|
640
|
|
|
Returns |
641
|
|
|
------- |
642
|
|
|
geo : numpy.ndarray |
643
|
|
|
Precision-weighted averaged MSDs from experiments specified in group |
644
|
|
|
geo_stder : numpy.ndarray |
645
|
|
|
Precision-weighted averaged SEMs from experiments specified in group |
646
|
|
|
|
647
|
|
|
""" |
648
|
|
|
|
649
|
|
|
frames = np.shape(geo_stder[group[0]])[0] |
650
|
|
|
slices = len(group) |
651
|
|
|
|
652
|
|
|
video_counter = 0 |
653
|
|
|
geo_holder = np.zeros((slices, frames)) |
654
|
|
|
gstder_holder = np.zeros((slices, frames)) |
655
|
|
|
w_holder = np.zeros((slices, frames)) |
656
|
|
|
for sample in group: |
657
|
|
|
w_holder[video_counter, :] = (1/(geo_stder[sample]*geo_stder[sample]) |
658
|
|
|
)/weights |
659
|
|
|
geo_holder[video_counter, :] = w_holder[video_counter, : |
660
|
|
|
] * geomean[sample] |
661
|
|
|
gstder_holder[video_counter, :] = 1/(geo_stder[sample]*geo_stder[sample] |
662
|
|
|
) |
663
|
|
|
video_counter = video_counter + 1 |
664
|
|
|
|
665
|
|
|
w_holder = ma.masked_equal(w_holder, 0.0) |
666
|
|
|
w_holder = ma.masked_equal(w_holder, 1.0) |
667
|
|
|
geo_holder = ma.masked_equal(geo_holder, 0.0) |
668
|
|
|
geo_holder = ma.masked_equal(geo_holder, 1.0) |
669
|
|
|
gstder_holder = ma.masked_equal(gstder_holder, 0.0) |
670
|
|
|
gstder_holder = ma.masked_equal(gstder_holder, 1.0) |
671
|
|
|
|
672
|
|
|
geo = ma.sum(geo_holder, axis=0) |
673
|
|
|
geo_stder = ma.sqrt((1/ma.sum(gstder_holder, axis=0))) |
674
|
|
|
|
675
|
|
|
if save: |
676
|
|
|
geo_f = 'geomean_{}.csv'.format(experiment) |
677
|
|
|
gstder_f = 'geoSEM_{}.csv'.format(experiment) |
678
|
|
|
np.savetxt(geo_f, geo, delimiter=',') |
679
|
|
|
np.savetxt(gstder_f, geo_stder, delimiter=',') |
680
|
|
|
aws.upload_s3(geo_f, '{}/{}'.format(folder, geo_f), bucket_name=bucket) |
681
|
|
|
aws.upload_s3(gstder_f, '{}/{}'.format(folder, gstder_f), |
682
|
|
|
bucket_name=bucket) |
683
|
|
|
|
684
|
|
|
geodata = Bunch(geomean=geo, geostd=geo_stder, weighthold=w_holder, |
685
|
|
|
geostdhold=gstder_holder) |
686
|
|
|
|
687
|
|
|
return geodata |
688
|
|
|
|
689
|
|
|
|
690
|
|
View Code Duplication |
def plot_all_experiments(experiments, bucket='ccurtis.data', folder='test', |
|
|
|
|
691
|
|
|
yrange=(10**-1, 10**1), fps=100.02, |
692
|
|
|
xrange=(10**-2, 10**0), upload=True, |
693
|
|
|
outfile='test.png', exponential=True, |
694
|
|
|
labels=None, log=True): |
695
|
|
|
"""Plots precision-weighted averages of MSD datasets. |
696
|
|
|
|
697
|
|
|
Plots pre-calculated precision-weighted averages of MSD datasets calculated |
698
|
|
|
from precision_averaging and stored in an AWS S3 bucket. |
699
|
|
|
|
700
|
|
|
Parameters |
701
|
|
|
---------- |
702
|
|
|
group : list of str |
703
|
|
|
List of experiment names to plot. Each experiment must have an MSD and |
704
|
|
|
SEM file associated with it in s3. |
705
|
|
|
bucket : str |
706
|
|
|
S3 bucket from which to download data. |
707
|
|
|
folder : str |
708
|
|
|
Folder in s3 bucket from which to download data. |
709
|
|
|
yrange : list of float |
710
|
|
|
Y range of plot |
711
|
|
|
xrange: list of float |
712
|
|
|
X range of plot |
713
|
|
|
upload : bool |
714
|
|
|
True to upload to S3 |
715
|
|
|
outfile : str |
716
|
|
|
Filename of output image |
717
|
|
|
|
718
|
|
|
""" |
719
|
|
|
|
720
|
|
|
n = len(experiments) |
721
|
|
|
|
722
|
|
|
if labels==None: |
723
|
|
|
labels = experiments |
724
|
|
|
|
725
|
|
|
color = iter(cm.viridis(np.linspace(0, 0.9, n))) |
726
|
|
|
|
727
|
|
|
fig = plt.figure(figsize=(8.5, 8.5)) |
728
|
|
|
ax = fig.add_subplot(111) |
729
|
|
|
plt.xlim(xrange[0], xrange[1]) |
730
|
|
|
plt.ylim(yrange[0], yrange[1]) |
731
|
|
|
plt.xlabel('Tau (s)', fontsize=25) |
732
|
|
|
plt.ylabel(r'Mean Squared Displacement ($\mu$m$^2$)', fontsize=25) |
733
|
|
|
|
734
|
|
|
geo = {} |
735
|
|
|
gstder = {} |
736
|
|
|
counter = 0 |
737
|
|
|
for experiment in experiments: |
738
|
|
|
aws.download_s3('{}/geomean_{}.csv'.format(folder, experiment), |
739
|
|
|
'geomean_{}.csv'.format(experiment), bucket_name=bucket) |
740
|
|
|
aws.download_s3('{}/geoSEM_{}.csv'.format(folder, experiment), |
741
|
|
|
'geoSEM_{}.csv'.format(experiment), bucket_name=bucket) |
742
|
|
|
|
743
|
|
|
geo[counter] = np.genfromtxt('geomean_{}.csv'.format(experiment)) |
744
|
|
|
gstder[counter] = np.genfromtxt('geoSEM_{}.csv'.format(experiment)) |
745
|
|
|
geo[counter] = ma.masked_equal(geo[counter], 0.0) |
746
|
|
|
gstder[counter] = ma.masked_equal(gstder[counter], 0.0) |
747
|
|
|
|
748
|
|
|
frames = np.shape(gstder[counter])[0] |
749
|
|
|
xpos = np.linspace(0, frames-1, frames)/fps |
750
|
|
|
c = next(color) |
751
|
|
|
|
752
|
|
|
if exponential: |
753
|
|
|
ax.plot(xpos, np.exp(geo[counter]), c=c, linewidth=6, |
754
|
|
|
label=labels[counter]) |
755
|
|
|
ax.fill_between(xpos, np.exp(geo[counter] - 1.96*gstder[counter]), |
756
|
|
|
np.exp(geo[counter] + 1.96*gstder[counter]), |
757
|
|
|
color=c, alpha=0.4) |
758
|
|
|
|
759
|
|
|
else: |
760
|
|
|
ax.plot(xpos, geo[counter], c=c, linewidth=6, |
761
|
|
|
label=labels[counter]) |
762
|
|
|
ax.fill_between(xpos, geo[counter] - 1.96*gstder[counter], |
763
|
|
|
geo[counter] + 1.96*gstder[counter], color=c, |
764
|
|
|
alpha=0.4) |
765
|
|
|
|
766
|
|
|
counter = counter + 1 |
767
|
|
|
|
768
|
|
|
if log: |
769
|
|
|
ax.set_xscale("log") |
770
|
|
|
ax.set_yscale("log") |
771
|
|
|
|
772
|
|
|
plt.legend(frameon=False, loc=2, prop={'size': 16}) |
773
|
|
|
|
774
|
|
|
if upload: |
775
|
|
|
fig.savefig(outfile, bbox_inches='tight') |
776
|
|
|
aws.upload_s3(outfile, folder+'/'+outfile, bucket_name=bucket) |
777
|
|
|
|
778
|
|
|
|
779
|
|
View Code Duplication |
def checkerboard_mask(dims=(512, 512), squares=50, width=25): |
|
|
|
|
780
|
|
|
"""Creates a 2D Boolean checkerboard mask |
781
|
|
|
|
782
|
|
|
Creates a Boolean array of evenly spaced squares. |
783
|
|
|
Whitespace is set to True. |
784
|
|
|
|
785
|
|
|
Parameters |
786
|
|
|
---------- |
787
|
|
|
|
788
|
|
|
dims : tuple of int |
789
|
|
|
Dimensions of desired Boolean array |
790
|
|
|
squares : int |
791
|
|
|
Dimensions of in individual square in array |
792
|
|
|
width : int |
793
|
|
|
Dimension of spacing between squares |
794
|
|
|
|
795
|
|
|
Returns |
796
|
|
|
---------- |
797
|
|
|
|
798
|
|
|
zeros : numpy.ndarray of bool |
799
|
|
|
2D Boolean array of evenly spaced squares |
800
|
|
|
|
801
|
|
|
""" |
802
|
|
|
zeros = np.zeros(dims) == 0 |
803
|
|
|
square_d = squares |
804
|
|
|
|
805
|
|
|
loy = width |
806
|
|
|
hiy = loy + square_d |
807
|
|
|
|
808
|
|
|
for j in range(50): |
|
|
|
|
809
|
|
|
|
810
|
|
|
lox = width |
811
|
|
|
hix = lox + square_d |
812
|
|
|
for i in range(50): |
813
|
|
|
|
814
|
|
|
if hix < 512 and hiy < 512: |
815
|
|
|
zeros[loy:hiy, lox:hix] = False |
816
|
|
|
elif hix < 512: |
817
|
|
|
zeros[loy:512-1, lox:hix] = False |
818
|
|
|
elif hiy < 512: |
819
|
|
|
zeros[loy:hiy, lox:512-1] = False |
820
|
|
|
else: |
821
|
|
|
zeros[loy:512-1, lox:512-1] = False |
822
|
|
|
break |
823
|
|
|
|
824
|
|
|
lox = hix + width |
825
|
|
|
hix = lox + square_d |
826
|
|
|
|
827
|
|
|
loy = hiy + width |
828
|
|
|
hiy = loy + square_d |
829
|
|
|
|
830
|
|
|
return zeros |
831
|
|
|
|
832
|
|
|
|
833
|
|
View Code Duplication |
def random_walk(nsteps=100, seed=None, start=(0, 0), step=1, mask=None, |
|
|
|
|
834
|
|
|
stuckprob=0.5): |
835
|
|
|
"""Creates 2d random walk trajectory. |
836
|
|
|
|
837
|
|
|
Parameters |
838
|
|
|
---------- |
839
|
|
|
nsteps : int |
840
|
|
|
Number of steps for trajectory to move. |
841
|
|
|
seed : int |
842
|
|
|
Seed for pseudo-random number generator for reproducability. |
843
|
|
|
start : tuple of int or float |
844
|
|
|
Starting xy coordinates at which the random walk begins. |
845
|
|
|
step : int or float |
846
|
|
|
Magnitude of single step |
847
|
|
|
mask : numpy.ndarray of bool |
848
|
|
|
Mask of barriers contraining diffusion |
849
|
|
|
stuckprop : float |
850
|
|
|
Probability of "particle" adhering to barrier when it makes contact |
851
|
|
|
|
852
|
|
|
Returns |
853
|
|
|
------- |
854
|
|
|
x : numpy.ndarray |
855
|
|
|
Array of x coordinates of random walk. |
856
|
|
|
y : numpy.ndarray |
857
|
|
|
Array of y coordinates of random walk. |
858
|
|
|
|
859
|
|
|
""" |
860
|
|
|
|
861
|
|
|
if type(mask) is np.ndarray: |
862
|
|
|
while not mask[start[0], start[1]]: |
863
|
|
|
start = (start[0], start[1]+1) |
864
|
|
|
eumask = eudist(~mask) |
865
|
|
|
|
866
|
|
|
np.random.seed(seed=seed) |
867
|
|
|
|
868
|
|
|
x = np.zeros(nsteps) |
869
|
|
|
y = np.zeros(nsteps) |
870
|
|
|
x[0] = start[0] |
871
|
|
|
y[0] = start[1] |
872
|
|
|
|
873
|
|
|
# Checks to see if a mask is being used first |
874
|
|
|
if not type(mask) is np.ndarray: |
875
|
|
|
for i in range(1, nsteps): |
|
|
|
|
876
|
|
|
val = rand.randint(1, 4) |
877
|
|
|
if val == 1: |
878
|
|
|
x[i] = x[i - 1] + step |
879
|
|
|
y[i] = y[i - 1] |
880
|
|
|
elif val == 2: |
881
|
|
|
x[i] = x[i - 1] - step |
882
|
|
|
y[i] = y[i - 1] |
883
|
|
|
elif val == 3: |
884
|
|
|
x[i] = x[i - 1] |
885
|
|
|
y[i] = y[i - 1] + step |
886
|
|
|
else: |
887
|
|
|
x[i] = x[i - 1] |
888
|
|
|
y[i] = y[i - 1] - step |
889
|
|
|
else: |
890
|
|
|
# print("Applied mask") |
891
|
|
|
for i in range(1, nsteps): |
892
|
|
|
val = rand.randint(1, 4) |
893
|
|
|
# If mask is being used, checks if entry is in mask or not |
894
|
|
|
if mask[int(x[i-1]), int(y[i-1])]: |
895
|
|
|
if val == 1: |
896
|
|
|
x[i] = x[i - 1] + step |
897
|
|
|
y[i] = y[i - 1] |
898
|
|
|
elif val == 2: |
899
|
|
|
x[i] = x[i - 1] - step |
900
|
|
|
y[i] = y[i - 1] |
901
|
|
|
elif val == 3: |
902
|
|
|
x[i] = x[i - 1] |
903
|
|
|
y[i] = y[i - 1] + step |
904
|
|
|
else: |
905
|
|
|
x[i] = x[i - 1] |
906
|
|
|
y[i] = y[i - 1] - step |
907
|
|
|
# If it does cross into a False area, probability to be stuck |
908
|
|
|
elif np.random.rand() > stuckprob: |
909
|
|
|
x[i] = x[i - 1] |
910
|
|
|
y[i] = y[i - 1] |
911
|
|
|
|
912
|
|
|
while eumask[int(x[i]), int(y[i])] > 0: |
|
|
|
|
913
|
|
|
vals = np.zeros(4) |
914
|
|
|
vals[0] = eumask[int(x[i] + step), int(y[i])] |
915
|
|
|
vals[1] = eumask[int(x[i] - step), int(y[i])] |
916
|
|
|
vals[2] = eumask[int(x[i]), int(y[i] + step)] |
917
|
|
|
vals[3] = eumask[int(x[i]), int(y[i] - step)] |
918
|
|
|
vali = np.argmin(vals) |
919
|
|
|
|
920
|
|
|
if vali == 0: |
921
|
|
|
x[i] = x[i] + step |
922
|
|
|
y[i] = y[i] |
923
|
|
|
elif vali == 1: |
924
|
|
|
x[i] = x[i] - step |
925
|
|
|
y[i] = y[i] |
926
|
|
|
elif vali == 2: |
927
|
|
|
x[i] = x[i] |
928
|
|
|
y[i] = y[i] + step |
929
|
|
|
else: |
930
|
|
|
x[i] = x[i] |
931
|
|
|
y[i] = y[i] - step |
932
|
|
|
# Otherwise, particle is stuck on "cell" |
933
|
|
|
else: |
934
|
|
|
x[i] = x[i - 1] |
935
|
|
|
y[i] = y[i - 1] |
936
|
|
|
|
937
|
|
|
return x, y |
938
|
|
|
|
939
|
|
|
|
940
|
|
|
# def random_walk(nsteps=100, seed=1, start=(0, 0)): |
941
|
|
|
# """Creates 2d random walk trajectory. |
942
|
|
|
# |
943
|
|
|
# Parameters |
944
|
|
|
# ---------- |
945
|
|
|
# nsteps : int |
946
|
|
|
# Number of steps for trajectory to move. |
947
|
|
|
# seed : int |
948
|
|
|
# Seed for pseudo-random number generator for reproducability. |
949
|
|
|
# start : tuple of int or float |
950
|
|
|
# Starting xy coordinates at which the random walk begins. |
951
|
|
|
# |
952
|
|
|
# Returns |
953
|
|
|
# ------- |
954
|
|
|
# x : numpy.ndarray |
955
|
|
|
# Array of x coordinates of random walk. |
956
|
|
|
# y : numpy.ndarray |
957
|
|
|
# Array of y coordinates of random walk. |
958
|
|
|
# |
959
|
|
|
# """ |
960
|
|
|
# |
961
|
|
|
# rand.seed(a=seed) |
962
|
|
|
# |
963
|
|
|
# x = np.zeros(nsteps) |
964
|
|
|
# y = np.zeros(nsteps) |
965
|
|
|
# x[0] = start[0] |
966
|
|
|
# y[0] = start[1] |
967
|
|
|
# |
968
|
|
|
# for i in range(1, nsteps): |
969
|
|
|
# val = rand.randint(1, 4) |
970
|
|
|
# if val == 1: |
971
|
|
|
# x[i] = x[i - 1] + 1 |
972
|
|
|
# y[i] = y[i - 1] |
973
|
|
|
# elif val == 2: |
974
|
|
|
# x[i] = x[i - 1] - 1 |
975
|
|
|
# y[i] = y[i - 1] |
976
|
|
|
# elif val == 3: |
977
|
|
|
# x[i] = x[i - 1] |
978
|
|
|
# y[i] = y[i - 1] + 1 |
979
|
|
|
# else: |
980
|
|
|
# x[i] = x[i - 1] |
981
|
|
|
# y[i] = y[i - 1] - 1 |
982
|
|
|
# |
983
|
|
|
# return x, y |
984
|
|
|
|
985
|
|
|
|
986
|
|
View Code Duplication |
def random_traj_dataset(nframes=100, nparts=30, seed=1, fsize=(0, 512), |
|
|
|
|
987
|
|
|
ndist=(1, 2)): |
988
|
|
|
"""Creates a random population of random walks. |
989
|
|
|
|
990
|
|
|
Parameters |
991
|
|
|
---------- |
992
|
|
|
nframes : int |
993
|
|
|
Number of frames for each random trajectory. |
994
|
|
|
nparts : int |
995
|
|
|
Number of particles in trajectory dataset. |
996
|
|
|
seed : int |
997
|
|
|
Seed for pseudo-random number generator for reproducability. |
998
|
|
|
fsize : tuple of int or float |
999
|
|
|
Scope of points over which particles may start at. |
1000
|
|
|
ndist : tuple of int or float |
1001
|
|
|
Parameters to generate normal distribution, mu and sigma. |
1002
|
|
|
|
1003
|
|
|
Returns |
1004
|
|
|
------- |
1005
|
|
|
dataf : pandas.core.frame.DataFrame |
1006
|
|
|
Trajectory data containing a 'Frame', 'Track_ID', 'X', and |
1007
|
|
|
'Y' column. |
1008
|
|
|
|
1009
|
|
|
""" |
1010
|
|
|
|
1011
|
|
|
frames = [] |
1012
|
|
|
trackid = [] |
1013
|
|
|
x = [] |
1014
|
|
|
y = [] |
1015
|
|
|
start = [0, 0] |
1016
|
|
|
pseed = seed |
1017
|
|
|
|
1018
|
|
|
for i in range(nparts): |
|
|
|
|
1019
|
|
|
rand.seed(a=i+pseed) |
1020
|
|
|
start[0] = rand.randint(fsize[0], fsize[1]) |
1021
|
|
|
rand.seed(a=i+3+pseed) |
1022
|
|
|
start[1] = rand.randint(fsize[0], fsize[1]) |
1023
|
|
|
rand.seed(a=i+5+pseed) |
1024
|
|
|
weight = rand.normalvariate(mu=ndist[0], sigma=ndist[1]) |
1025
|
|
|
|
1026
|
|
|
trackid = np.append(trackid, np.array([i]*nframes)) |
1027
|
|
|
xi, yi = random_walk(nsteps=nframes, seed=i) |
1028
|
|
|
x = np.append(x, weight*xi+start[0]) |
1029
|
|
|
y = np.append(y, weight*yi+start[1]) |
1030
|
|
|
frames = np.append(frames, np.linspace(0, nframes-1, nframes)) |
1031
|
|
|
|
1032
|
|
|
datai = {'Frame': frames, |
1033
|
|
|
'Track_ID': trackid, |
1034
|
|
|
'X': x, |
1035
|
|
|
'Y': y, |
1036
|
|
|
'Quality': nframes*nparts*[10], |
1037
|
|
|
'SN_Ratio': nframes*nparts*[0.1], |
1038
|
|
|
'Mean_Intensity': nframes*nparts*[120]} |
1039
|
|
|
dataf = pd.DataFrame(data=datai) |
1040
|
|
|
|
1041
|
|
|
return dataf |
1042
|
|
|
|
1043
|
|
|
|
1044
|
|
|
class Bunch: |
1045
|
|
|
def __init__(self, **kwds): |
1046
|
|
|
self.__dict__.update(kwds) |
1047
|
|
|
|