-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSCOUR_Synthetic_random.m
401 lines (351 loc) · 16.1 KB
/
SCOUR_Synthetic_random.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
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
396
397
398
399
400
401
function SCOUR_Synthetic_random(nT,cov,num_IC,rep)
warning('off','all')
rng('shuffle');
fprintf('Starting machine learning framework...\n');
%% Initial Preparation of Synthetic Models
% Train on autogenerated data, test on BiggerModel and SmallerModel models
% BiggerModel and SmallerModel stoichiometric interactions
BiggerModel.S = sparse([ 1 -1 0 0 0 0 0 0 0 0 ;
0 1 -1 0 0 -1 0 0 0 0 ;
0 0 1 -1 -1 0 0 0 0 0 ;
0 0 0 1 0 0 0 0 -1 0 ;
0 0 0 0 1 0 0 0 0 0 ;
0 0 0 0 0 1 -1 -1 0 0 ;
0 0 0 0 0 0 1 0 0 0 ;
0 0 0 0 0 0 0 1 0 -1 ;
0 0 0 0 0 0 0 0 1 0 ;
0 0 0 0 0 0 0 0 0 1 ;]);
SmallerModel.S = sparse([ 1 -1 0 0 0 0 ;
0 1 -1 -1 -1 0 ;
0 0 1 0 0 0 ;
0 0 0 1 0 0 ;
0 0 0 0 1 -1 ;
0 0 0 0 0 1 ;]);
% Find BiggerModel mass action (MA) interactions
[row_BiggerModelMA col_BiggerModelMA] = find(BiggerModel.S < 0);
BiggerModel_MA = [row_BiggerModelMA col_BiggerModelMA];
unique_flux = unique(BiggerModel_MA(:,2));
count_1contMet = 1;
count_2contMet = 1;
for i = 1:length(unique_flux)
controller_met_idx = find(BiggerModel_MA(:,2) == unique_flux(i));
if length(BiggerModel_MA(controller_met_idx,1)) == 1
BiggerModel_1contMet_List(count_1contMet,:) = [BiggerModel_MA(controller_met_idx,1)' unique_flux(i)];
count_1contMet = count_1contMet + 1;
elseif length(BiggerModel_MA(controller_met_idx,1)) == 2
BiggerModel_2contMet_List(count_2contMet,:) = [BiggerModel_MA(controller_met_idx,1)' unique_flux(i)];
count_2contMet = count_2contMet + 1;
end
end
% Find SmallerModel mass action (MA) interactions
[row_SmallerModelMA col_SmallerModelMA] = find(SmallerModel.S < 0);
SmallerModel_MA = [row_SmallerModelMA col_SmallerModelMA];
unique_flux = unique(SmallerModel_MA(:,2));
count_1contMet = 1;
count_2contMet = 1;
for i = 1:length(unique_flux)
controller_met_idx = find(SmallerModel_MA(:,2) == unique_flux(i));
if length(SmallerModel_MA(controller_met_idx,1)) == 1
SmallerModel_1contMet_List(count_1contMet,:) = [SmallerModel_MA(controller_met_idx,1)' unique_flux(i)];
count_1contMet = count_1contMet + 1;
elseif length(SmallerModel_MA(controller_met_idx,1)) == 2
SmallerModel_2contMet_List(count_2contMet,:) = [SmallerModel_MA(controller_met_idx,1)' unique_flux(i)];
count_2contMet = count_2contMet + 1;
end
end
BiggerModel_prefix = 'BiggerModel';
SmallerModel_prefix = 'SmallerModel';
%% Identify 1 controller metabolite reactions
% List of BiggerModel mass action only interactions and find where they occur
% in list of all mass action interactions
true_1contMet_BiggerModel = [6,7;
4,9;
8,10];
count = 1;
for regIdx = 1:length(BiggerModel_1contMet_List)
for trueIdx = 1:size(true_1contMet_BiggerModel,1)
if isequal(BiggerModel_1contMet_List(regIdx,:),true_1contMet_BiggerModel(trueIdx,:))
BiggerModel_1contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% List of SmallerModel mass action only interactions and find where they
% occur in list of all mass action interactions
true_1contMet_SmallerModel = [2,5];
count = 1;
for regIdx = 1:length(SmallerModel_1contMet_List)
for trueIdx = 1:size(true_1contMet_SmallerModel,1)
if isequal(SmallerModel_1contMet_List(regIdx,:),true_1contMet_SmallerModel(trueIdx,:))
SmallerModel_1contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% BiggerModel
% Setup testing set and testing label
testingLabel_BiggerModel_1contMet = logical(zeros(size(BiggerModel_1contMet_List,1),1));
testingLabel_BiggerModel_1contMet(BiggerModel_1contMet_trueInRegIdx) = 1;
% Random prediction
predictedLabel_L2_BiggerModel_1contMet = logical(round(rand(length(testingLabel_BiggerModel_1contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_BiggerModel_1contMet = sum(predictedLabel_L2_BiggerModel_1contMet==testingLabel_BiggerModel_1contMet)/length(testingLabel_BiggerModel_1contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_BiggerModel_1contMet)
if testingLabel_BiggerModel_1contMet(k) == 1 && predictedLabel_L2_BiggerModel_1contMet(k) == 0
fn = fn + 1;
elseif testingLabel_BiggerModel_1contMet(k) == 0 && predictedLabel_L2_BiggerModel_1contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_BiggerModel_1contMet)
if testingLabel_BiggerModel_1contMet(k) == 1 && predictedLabel_L2_BiggerModel_1contMet(k) == 1
tp = tp + 1;
elseif testingLabel_BiggerModel_1contMet(k) == 0 && predictedLabel_L2_BiggerModel_1contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_BiggerModel_1contMet = tp / (fn+tp)
specificity_BiggerModel_1contMet = tn / (tn+fp)
ppv_BiggerModel_1contMet = tp / (tp+fp);
npv_BiggerModel_1contMet = tp / (tn+fn);
% SmallerModel
% Setup testing set and testing label
testingLabel_SmallerModel_1contMet = logical(zeros(size(SmallerModel_1contMet_List,1),1));
testingLabel_SmallerModel_1contMet(SmallerModel_1contMet_trueInRegIdx) = 1;
% Random prediction
predictedLabel_L2_SmallerModel_1contMet = logical(round(rand(length(testingLabel_SmallerModel_1contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_SmallerModel_1contMet = sum(predictedLabel_L2_SmallerModel_1contMet==testingLabel_SmallerModel_1contMet)/length(testingLabel_SmallerModel_1contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_SmallerModel_1contMet)
if testingLabel_SmallerModel_1contMet(k) == 1 && predictedLabel_L2_SmallerModel_1contMet(k) == 0
fn = fn + 1;
elseif testingLabel_SmallerModel_1contMet(k) == 0 && predictedLabel_L2_SmallerModel_1contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_SmallerModel_1contMet)
if testingLabel_SmallerModel_1contMet(k) == 1 && predictedLabel_L2_SmallerModel_1contMet(k) == 1
tp = tp + 1;
elseif testingLabel_SmallerModel_1contMet(k) == 0 && predictedLabel_L2_SmallerModel_1contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_SmallerModel_1contMet = tp / (fn+tp)
specificity_SmallerModel_1contMet = tn / (tn+fp)
ppv_SmallerModel_1contMet = tp / (tp+fp);
npv_SmallerModel_1contMet = tp / (tn+fn);
%% Remove 1 controller metabolite reactions
%BiggerModel_fluxes_to_remove = unique([1]);
BiggerModel_fluxes_to_remove = unique([1 7 9 10]);
if ~isequal(sort(BiggerModel_fluxes_to_remove),1:size(BiggerModel.S,2))
BiggerModel_regScheme_2contMet = createRegSchemeList(BiggerModel.S,BiggerModel_fluxes_to_remove);
end
%SmallerModel_fluxes_to_remove = unique([1]);
SmallerModel_fluxes_to_remove = unique([1 5]);
if ~isequal(sort(SmallerModel_fluxes_to_remove),1:size(SmallerModel.S,2))
SmallerModel_regScheme_2contMet = createRegSchemeList(SmallerModel.S,SmallerModel_fluxes_to_remove);
end
%% Identify 2 controller metabolite reactions
% List of BiggerModel interactions with two controller metabolites
true_2contMet_BiggerModel = [2,4,3;
3,8,4;
2,8,6];
count = 1;
for regIdx = 1:length(BiggerModel_regScheme_2contMet)
for trueIdx = 1:size(true_2contMet_BiggerModel,1)
if isequal(BiggerModel_regScheme_2contMet(regIdx,:),true_2contMet_BiggerModel(trueIdx,:))
BiggerModel_2contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% List of SmallerModel interactions with two controller metabolites
true_2contMet_SmallerModel = [2,5,4;
2,6,3];
count = 1;
for regIdx = 1:length(SmallerModel_regScheme_2contMet)
for trueIdx = 1:size(true_2contMet_SmallerModel,1)
if isequal(SmallerModel_regScheme_2contMet(regIdx,:),true_2contMet_SmallerModel(trueIdx,:))
SmallerModel_2contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% BiggerModel
% Setup testing set and testing label
testingLabel_BiggerModel_2contMet = logical(zeros(size(BiggerModel_regScheme_2contMet,1),1));
if exist('BiggerModel_2contMet_trueInRegIdx','var')
testingLabel_BiggerModel_2contMet(BiggerModel_2contMet_trueInRegIdx) = 1;
end
% Random prediction
predictedLabel_L2_BiggerModel_2contMet = logical(round(rand(length(testingLabel_BiggerModel_2contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_BiggerModel_2contMet = sum(predictedLabel_L2_BiggerModel_2contMet==testingLabel_BiggerModel_2contMet)/length(testingLabel_BiggerModel_2contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_BiggerModel_2contMet)
if testingLabel_BiggerModel_2contMet(k) == 1 && predictedLabel_L2_BiggerModel_2contMet(k) == 0
fn = fn + 1;
elseif testingLabel_BiggerModel_2contMet(k) == 0 && predictedLabel_L2_BiggerModel_2contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_BiggerModel_2contMet)
if testingLabel_BiggerModel_2contMet(k) == 1 && predictedLabel_L2_BiggerModel_2contMet(k) == 1
tp = tp + 1;
elseif testingLabel_BiggerModel_2contMet(k) == 0 && predictedLabel_L2_BiggerModel_2contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_BiggerModel_2contMet = tp / (fn+tp)
specificity_BiggerModel_2contMet = tn / (tn+fp)
ppv_BiggerModel_2contMet = tp / (tp+fp);
npv_BiggerModel_2contMet = tp / (tn+fn);
% SmallerModel
% Setup testing set and testing label
testingLabel_SmallerModel_2contMet = logical(zeros(size(SmallerModel_regScheme_2contMet,1),1));
if exist('SmallerModel_2contMet_trueInRegIdx','var')
testingLabel_SmallerModel_2contMet(SmallerModel_2contMet_trueInRegIdx) = 1;
end
% Random prediction
predictedLabel_L2_SmallerModel_2contMet = logical(round(rand(length(testingLabel_SmallerModel_2contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_SmallerModel_2contMet = sum(predictedLabel_L2_SmallerModel_2contMet==testingLabel_SmallerModel_2contMet)/length(testingLabel_SmallerModel_2contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_SmallerModel_2contMet)
if testingLabel_SmallerModel_2contMet(k) == 1 && predictedLabel_L2_SmallerModel_2contMet(k) == 0
fn = fn + 1;
elseif testingLabel_SmallerModel_2contMet(k) == 0 && predictedLabel_L2_SmallerModel_2contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_SmallerModel_2contMet)
if testingLabel_SmallerModel_2contMet(k) == 1 && predictedLabel_L2_SmallerModel_2contMet(k) == 1
tp = tp + 1;
elseif testingLabel_SmallerModel_2contMet(k) == 0 && predictedLabel_L2_SmallerModel_2contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_SmallerModel_2contMet = tp / (fn+tp)
specificity_SmallerModel_2contMet = tn / (tn+fp)
ppv_SmallerModel_2contMet = tp / (tp+fp);
npv_SmallerModel_2contMet = tp / (tn+fn);
%% Remove 2 controller metabolite reactions
if exist('BiggerModel_regScheme_2contMet','var')
%BiggerModel_fluxes_to_remove = unique([1]);
BiggerModel_fluxes_to_remove = unique([1 7 9 10 3 4 6]);
if ~isequal(sort(BiggerModel_fluxes_to_remove),1:size(BiggerModel.S,2))
[~,BiggerModel_regScheme_3contMet] = createRegSchemeList(BiggerModel.S,BiggerModel_fluxes_to_remove);
end
end
if exist('SmallerModel_regScheme_2contMet','var')
%SmallerModel_fluxes_to_remove = unique([1]);
SmallerModel_fluxes_to_remove = unique([1 5 4 3]);
if ~isequal(sort(SmallerModel_fluxes_to_remove),1:size(SmallerModel.S,2))
[~,SmallerModel_regScheme_3contMet] = createRegSchemeList(SmallerModel.S,SmallerModel_fluxes_to_remove);
end
end
%% Identify 3 controller metabolite reactions
% List of BiggerModel interactions with three controller metabolites
true_3contMet_BiggerModel = [1,2,7,2;
3,6,10,5;
5,6,7,8];
count = 1;
for regIdx = 1:length(BiggerModel_regScheme_3contMet)
for trueIdx = 1:size(true_3contMet_BiggerModel,1)
if isequal(BiggerModel_regScheme_3contMet(regIdx,:),true_3contMet_BiggerModel(trueIdx,:))
BiggerModel_3contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% List of SmallerModel interactions with three controller metabolites
true_3contMet_SmallerModel = [1,4,5,2;
3,5,6,6];
count = 1;
for regIdx = 1:length(SmallerModel_regScheme_3contMet)
for trueIdx = 1:size(true_3contMet_SmallerModel,1)
if isequal(SmallerModel_regScheme_3contMet(regIdx,:),true_3contMet_SmallerModel(trueIdx,:))
SmallerModel_3contMet_trueInRegIdx(count,1) = regIdx;
count = count + 1;
end
end
end
% BiggerModel
% Setup testing set and testing label
testingLabel_BiggerModel_3contMet = logical(zeros(size(BiggerModel_regScheme_3contMet,1),1));
if exist('BiggerModel_3contMet_trueInRegIdx','var')
testingLabel_BiggerModel_3contMet(BiggerModel_3contMet_trueInRegIdx) = 1;
end
% Random prediction
predictedLabel_L2_BiggerModel_3contMet = logical(round(rand(length(testingLabel_BiggerModel_3contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_BiggerModel_3contMet = sum(predictedLabel_L2_BiggerModel_3contMet==testingLabel_BiggerModel_3contMet)/length(testingLabel_BiggerModel_3contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_BiggerModel_3contMet)
if testingLabel_BiggerModel_3contMet(k) == 1 && predictedLabel_L2_BiggerModel_3contMet(k) == 0
fn = fn + 1;
elseif testingLabel_BiggerModel_3contMet(k) == 0 && predictedLabel_L2_BiggerModel_3contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_BiggerModel_3contMet)
if testingLabel_BiggerModel_3contMet(k) == 1 && predictedLabel_L2_BiggerModel_3contMet(k) == 1
tp = tp + 1;
elseif testingLabel_BiggerModel_3contMet(k) == 0 && predictedLabel_L2_BiggerModel_3contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_BiggerModel_3contMet = tp / (fn+tp)
specificity_BiggerModel_3contMet = tn / (tn+fp)
ppv_BiggerModel_3contMet = tp / (tp+fp);
npv_BiggerModel_3contMet = tp / (tn+fn);
% SmallerModel
% Setup testing set and testing label
testingLabel_SmallerModel_3contMet = logical(zeros(size(SmallerModel_regScheme_3contMet,1),1));
if exist('SmallerModel_3contMet_trueInRegIdx','var')
testingLabel_SmallerModel_3contMet(SmallerModel_3contMet_trueInRegIdx) = 1;
end
% Random prediction
predictedLabel_L2_SmallerModel_3contMet = logical(round(rand(length(testingLabel_SmallerModel_3contMet),1)))
% Accuracy, sensitivity, and specificity calculations
predictionAccuracy_SmallerModel_3contMet = sum(predictedLabel_L2_SmallerModel_3contMet==testingLabel_SmallerModel_3contMet)/length(testingLabel_SmallerModel_3contMet)
fp = 0;
fn = 0;
for k = 1:length(testingLabel_SmallerModel_3contMet)
if testingLabel_SmallerModel_3contMet(k) == 1 && predictedLabel_L2_SmallerModel_3contMet(k) == 0
fn = fn + 1;
elseif testingLabel_SmallerModel_3contMet(k) == 0 && predictedLabel_L2_SmallerModel_3contMet(k) == 1
fp = fp + 1;
end
end
tp = 0;
tn = 0;
for k = 1:length(testingLabel_SmallerModel_3contMet)
if testingLabel_SmallerModel_3contMet(k) == 1 && predictedLabel_L2_SmallerModel_3contMet(k) == 1
tp = tp + 1;
elseif testingLabel_SmallerModel_3contMet(k) == 0 && predictedLabel_L2_SmallerModel_3contMet(k) == 0
tn = tn + 1;
end
end
sensitivity_SmallerModel_3contMet = tp / (fn+tp)
specificity_SmallerModel_3contMet = tn / (tn+fp)
ppv_SmallerModel_3contMet = tp / (tp+fp);
npv_SmallerModel_3contMet = tp / (tn+fn);
save(sprintf('synthetic_random_results_IC-%02d_nT-%03d_cov-%02d_rep-%02d.mat',num_IC,nT,cov,rep));