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
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
|
// --------------------------------------------------------------------------
// OpenMS -- Open-Source Mass Spectrometry
// --------------------------------------------------------------------------
// Copyright The OpenMS Team -- Eberhard Karls University Tuebingen,
// ETH Zurich, and Freie Universitaet Berlin 2002-2013.
//
// This software is released under a three-clause BSD license:
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
// * Neither the name of any author or any participating institution
// may be used to endorse or promote products derived from this software
// without specific prior written permission.
// For a full list of authors, refer to the file AUTHORS.
// --------------------------------------------------------------------------
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
// ARE DISCLAIMED. IN NO EVENT SHALL ANY OF THE AUTHORS OR THE CONTRIBUTING
// INSTITUTIONS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
// OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
// WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
// OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
// ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// --------------------------------------------------------------------------
// $Maintainer: David Wojnar $
// $Authors: David Wojnar $
// --------------------------------------------------------------------------
//
#include <OpenMS/MATH/STATISTICS/PosteriorErrorProbabilityModel.h>
#include <OpenMS/FORMAT/TextFile.h>
#include <OpenMS/DATASTRUCTURES/String.h>
#include <algorithm>
#include <gsl/gsl_statistics.h>
#include <boost/math/special_functions/fpclassify.hpp>
using namespace std;
namespace OpenMS
{
namespace Math
{
PosteriorErrorProbabilityModel::PosteriorErrorProbabilityModel() :
DefaultParamHandler("PosteriorErrorProbabilityModel"), negative_prior_(0.5), max_incorrectly_(0), max_correctly_(0), smallest_score_(0)
{
defaults_.setValue("number_of_bins", 100, "Number of bins used for visualization. Only needed if each iteration step of the EM-Algorithm will be visualized", StringList::create("advanced"));
defaults_.setValue("output_plots", "false", "If true every step of the EM-algorithm will be written to a file as a gnuplot formula", StringList::create("advanced"));
defaults_.setValidStrings("output_plots", StringList::create("true,false"));
defaults_.setValue("output_name", "", "If output_plots is on, the output files will be saved in the following manner: <output_name>scores.txt for the scores and <output_name> which contains each step of the EM-algorithm e.g. output_name = /usr/home/OMSSA123 then /usr/home/OMSSA123_scores.txt, /usr/home/OMSSA123 will be written. If no directory is specified, e.g. instead of '/usr/home/OMSSA123' just OMSSA123, the files will be written into the working directory.", StringList::create("advanced,output file"));
defaults_.setValue("incorrectly_assigned", "Gumbel", "for 'Gumbel', the Gumbel distribution is used to plot incorrectly assigned sequences. For 'Gauss', the Gauss distribution is used.", StringList::create("advanced"));
defaults_.setValidStrings("incorrectly_assigned", StringList::create("Gumbel,Gauss"));
defaultsToParam_();
calc_incorrect_ = &PosteriorErrorProbabilityModel::getGumbel;
calc_correct_ = &PosteriorErrorProbabilityModel::getGauss;
getNegativeGnuplotFormula_ = &PosteriorErrorProbabilityModel::getGumbelGnuplotFormula;
getPositiveGnuplotFormula_ = &PosteriorErrorProbabilityModel::getGaussGnuplotFormula;
}
PosteriorErrorProbabilityModel::~PosteriorErrorProbabilityModel()
{
}
bool PosteriorErrorProbabilityModel::fit(std::vector<double> & search_engine_scores)
{
if (search_engine_scores.empty())
{
return false;
}
//-------------------------------------------------------------
// Initializing Parameters
//-------------------------------------------------------------
sort(search_engine_scores.begin(), search_engine_scores.end());
smallest_score_ = search_engine_scores[0];
vector<double> x_scores;
x_scores.resize(search_engine_scores.size());
std::vector<double>::iterator it = x_scores.begin();
for (std::vector<double>::iterator iti = search_engine_scores.begin(); iti < search_engine_scores.end(); ++it, ++iti)
{
*it = *iti + fabs(smallest_score_) + 0.001;
}
negative_prior_ = 0.7;
if (param_.getValue("incorrectly_assigned") == "Gumbel")
{
incorrectly_assigned_fit_param_.x0 = gsl_stats_mean(&x_scores[0], 1, ceil(0.5 * x_scores.size())) + x_scores[0];
incorrectly_assigned_fit_param_.sigma = gsl_stats_sd(&x_scores[0], 1, x_scores.size() - 1); //pow(gsl_stats_sd_with_fixed_mean(&probabilities[x_score_start], 1, probabilities.size() - x_score_start, gauss_fit_param_.x0),2);
incorrectly_assigned_fit_param_.A = 1 / sqrt(2 * 3.14159 * pow(incorrectly_assigned_fit_param_.sigma, 2));
//TODO: compute directly with gauss. Workaround:
calc_incorrect_ = &PosteriorErrorProbabilityModel::getGauss;
getNegativeGnuplotFormula_ = &PosteriorErrorProbabilityModel::getGumbelGnuplotFormula;
}
else
{
incorrectly_assigned_fit_param_.x0 = gsl_stats_mean(&x_scores[0], 1, ceil(0.5 * x_scores.size())) + x_scores[0];
incorrectly_assigned_fit_param_.sigma = gsl_stats_sd(&x_scores[0], 1, x_scores.size() - 1); //pow(gsl_stats_sd_with_fixed_mean(&probabilities[x_score_start], 1, probabilities.size() - x_score_start, gauss_fit_param_.x0),2);
incorrectly_assigned_fit_param_.A = 1 / sqrt(2 * 3.14159 * pow(incorrectly_assigned_fit_param_.sigma, 2));
calc_incorrect_ = &PosteriorErrorProbabilityModel::getGauss;
getNegativeGnuplotFormula_ = &PosteriorErrorProbabilityModel::getGaussGnuplotFormula;
}
getPositiveGnuplotFormula_ = &PosteriorErrorProbabilityModel::getGaussGnuplotFormula;
calc_correct_ = &PosteriorErrorProbabilityModel::getGauss;
Size x_score_start = std::min(x_scores.size() - 1, (Size) ceil(x_scores.size() * 0.7)); // if only one score is present, ceil(...) will yield 1, which is an invalid index
correctly_assigned_fit_param_.x0 = gsl_stats_mean(&x_scores[x_score_start], 1, x_scores.size() - x_score_start) + x_scores[x_score_start]; //(gauss_scores.begin()->getX() + (gauss_scores.end()-1)->getX())/2;
correctly_assigned_fit_param_.sigma = incorrectly_assigned_fit_param_.sigma;
correctly_assigned_fit_param_.A = 1.0 / sqrt(2 * 3.14159 * pow(correctly_assigned_fit_param_.sigma, 2));
DoubleReal maxlike(0);
vector<DoubleReal> incorrect_density;
vector<DoubleReal> correct_density;
fillDensities(x_scores, incorrect_density, correct_density);
maxlike = computeMaxLikelihood(incorrect_density, correct_density);
//-------------------------------------------------------------
// create files for output
//-------------------------------------------------------------
bool output_plots = param_.getValue("output_plots").toBool();
TextFile * file = NULL;
if (output_plots)
{
file = InitPlots(x_scores);
}
//-------------------------------------------------------------
// Estimate Parameters - EM algorithm
//-------------------------------------------------------------
bool stop_em_init = false;
do
{
//E-STEP
DoubleReal one_minus_sum_posterior = one_minus_sum_post(incorrect_density, correct_density);
DoubleReal sum_posterior = sum_post(incorrect_density, correct_density);
//new mean
DoubleReal sum_positive_x0 = sum_pos_x0(x_scores, incorrect_density, correct_density);
DoubleReal sum_negative_x0 = sum_neg_x0(x_scores, incorrect_density, correct_density);
DoubleReal positive_mean = sum_positive_x0 / one_minus_sum_posterior;
DoubleReal negative_mean = sum_negative_x0 / sum_posterior;
//new standard deviation
DoubleReal sum_positive_sigma = sum_pos_sigma(x_scores, incorrect_density, correct_density, positive_mean);
DoubleReal sum_negative_sigma = sum_neg_sigma(x_scores, incorrect_density, correct_density, negative_mean);
//update parameters
correctly_assigned_fit_param_.x0 = positive_mean;
if (sum_positive_sigma != 0 && one_minus_sum_posterior != 0)
{
correctly_assigned_fit_param_.sigma = sqrt(sum_positive_sigma / one_minus_sum_posterior);
correctly_assigned_fit_param_.A = 1 / sqrt(2 * 3.14159 * pow(correctly_assigned_fit_param_.sigma, 2));
}
incorrectly_assigned_fit_param_.x0 = negative_mean;
if (sum_negative_sigma != 0 && sum_posterior != 0)
{
incorrectly_assigned_fit_param_.sigma = sqrt(sum_negative_sigma / sum_posterior);
incorrectly_assigned_fit_param_.A = 1 / sqrt(2 * 3.14159 * pow(incorrectly_assigned_fit_param_.sigma, 2));
}
//compute new prior probabilities negative peptides
fillDensities(x_scores, incorrect_density, correct_density);
sum_posterior = sum_post(incorrect_density, correct_density);
negative_prior_ = sum_posterior / x_scores.size();
DoubleReal new_maxlike(computeMaxLikelihood(incorrect_density, correct_density));
if (boost::math::isnan(new_maxlike - maxlike))
{
return false;
//throw Exception::UnableToFit(__FILE__,__LINE__,__PRETTY_FUNCTION__,"UnableToFit-PosteriorErrorProbability","Could not fit mixture model to data");
}
if (fabs(new_maxlike - maxlike) < 0.001)
{
stop_em_init = true;
sum_posterior = sum_post(incorrect_density, correct_density);
negative_prior_ = sum_posterior / x_scores.size();
}
if (output_plots)
{
String formula1, formula2, formula3;
formula1 = ((this)->*(getNegativeGnuplotFormula_))(incorrectly_assigned_fit_param_) + "* " + String(negative_prior_); //String(incorrectly_assigned_fit_param_.A) +" * exp(-(x - " + String(incorrectly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(incorrectly_assigned_fit_param_.sigma) + ") ** 2)"+ "*" + String(negative_prior_);
formula2 = ((this)->*(getPositiveGnuplotFormula_))(correctly_assigned_fit_param_) + "* (1 - " + String(negative_prior_) + ")"; //String(correctly_assigned_fit_param_.A) +" * exp(-(x - " + String(correctly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(correctly_assigned_fit_param_.sigma) + ") ** 2)"+ "* (1 - " + String(negative_prior_) + ")";
formula3 = getBothGnuplotFormula(incorrectly_assigned_fit_param_, correctly_assigned_fit_param_);
(*file) << ("plot \"" + (String)param_.getValue("output_name") + "_scores.txt\" with boxes, " + formula1 + " , " + formula2 + " , " + formula3);
}
//update maximum likelihood
maxlike = new_maxlike;
}
while (!stop_em_init);
//-------------------------------------------------------------
// Finished fitting
//-------------------------------------------------------------
//!!Workaround:
if (param_.getValue("incorrectly_assigned") == "Gumbel")
{
calc_incorrect_ = &PosteriorErrorProbabilityModel::getGumbel;
}
max_incorrectly_ = ((this)->*(calc_incorrect_))(incorrectly_assigned_fit_param_.x0, incorrectly_assigned_fit_param_);
max_correctly_ = ((this)->*(calc_correct_))(correctly_assigned_fit_param_.x0, correctly_assigned_fit_param_);
if (output_plots)
{
String formula1, formula2, formula3;
formula1 = ((this)->*(getNegativeGnuplotFormula_))(incorrectly_assigned_fit_param_) + "*" + String(negative_prior_); //String(incorrectly_assigned_fit_param_.A) +" * exp(-(x - " + String(incorrectly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(incorrectly_assigned_fit_param_.sigma) + ") ** 2)"+ "*" + String(negative_prior_);
formula2 = ((this)->*(getPositiveGnuplotFormula_))(correctly_assigned_fit_param_) + "* (1 - " + String(negative_prior_) + ")"; // String(correctly_assigned_fit_param_.A) +" * exp(-(x - " + String(correctly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(correctly_assigned_fit_param_.sigma) + ") ** 2)"+ "* (1 - " + String(negative_prior_) + ")";
formula3 = getBothGnuplotFormula(incorrectly_assigned_fit_param_, correctly_assigned_fit_param_);
(*file) << ("plot \"" + (String)param_.getValue("output_name") + "_scores.txt\" with boxes, " + formula1 + " , " + formula2 + " , " + formula3);
file->store((String)param_.getValue("output_name"));
delete file;
}
return true;
}
bool PosteriorErrorProbabilityModel::fit(std::vector<double> & search_engine_scores, vector<double> & probabilities)
{
bool return_value;
return_value = fit(search_engine_scores);
if (!return_value)
return false;
probabilities.resize(search_engine_scores.size());
vector<double>::iterator probs = probabilities.begin();
for (vector<double>::iterator scores = search_engine_scores.begin(); scores != search_engine_scores.end(); ++scores, ++probs)
{
*probs = computeProbability(*scores);
}
return true;
}
void PosteriorErrorProbabilityModel::fillDensities(vector<double> & x_scores, vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
if (incorrect_density.size() != x_scores.size())
{
incorrect_density.resize(x_scores.size());
correct_density.resize(x_scores.size());
}
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
vector<DoubleReal>::iterator correct = correct_density.begin();
for (vector<double>::iterator scores = x_scores.begin(); scores != x_scores.end(); ++scores, ++incorrect, ++correct)
{
*incorrect = ((this)->*(calc_incorrect_))(*scores, incorrectly_assigned_fit_param_);
*correct = ((this)->*(calc_correct_))(*scores, correctly_assigned_fit_param_);
}
}
DoubleReal PosteriorErrorProbabilityModel::computeMaxLikelihood(vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
DoubleReal maxlike(0);
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect)
{
maxlike += log10(negative_prior_ * (*incorrect) + (1 - negative_prior_) * (*correct));
}
return maxlike;
}
DoubleReal PosteriorErrorProbabilityModel::one_minus_sum_post(vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
DoubleReal one_min(0);
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect)
{
one_min += 1 - ((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)));
}
return one_min;
}
DoubleReal PosteriorErrorProbabilityModel::sum_post(vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
DoubleReal post(0);
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect)
{
post += ((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)));
}
return post;
}
DoubleReal PosteriorErrorProbabilityModel::sum_pos_x0(vector<double> & x_scores, vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
DoubleReal pos_x0(0);
vector<double>::iterator the_x = x_scores.begin();
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect, ++the_x)
{
pos_x0 += ((1 - ((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)))) * (*the_x));
}
return pos_x0;
}
DoubleReal PosteriorErrorProbabilityModel::sum_neg_x0(vector<double> & x_scores, vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density)
{
DoubleReal neg_x0(0);
vector<double>::iterator the_x = x_scores.begin();
vector<DoubleReal>::iterator correct = correct_density.begin();
for (vector<DoubleReal>::iterator incorrect = incorrect_density.begin(); incorrect < incorrect_density.end(); ++correct, ++incorrect, ++the_x)
{
neg_x0 += ((((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)))) * (*the_x));
}
return neg_x0;
}
DoubleReal PosteriorErrorProbabilityModel::sum_pos_sigma(vector<double> & x_scores, vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density, DoubleReal positive_mean)
{
DoubleReal pos_sigma(0);
vector<double>::iterator the_x = x_scores.begin();
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect, ++the_x)
{
pos_sigma += ((1 - ((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)))) * pow((*the_x) - positive_mean, 2));
}
return pos_sigma;
}
DoubleReal PosteriorErrorProbabilityModel::sum_neg_sigma(vector<double> & x_scores, vector<DoubleReal> & incorrect_density, vector<DoubleReal> & correct_density, DoubleReal positive_mean)
{
DoubleReal neg_sigma(0);
vector<double>::iterator the_x = x_scores.begin();
vector<DoubleReal>::iterator incorrect = incorrect_density.begin();
for (vector<DoubleReal>::iterator correct = correct_density.begin(); correct < correct_density.end(); ++correct, ++incorrect, ++the_x)
{
neg_sigma += ((((negative_prior_ * (*incorrect)) / ((negative_prior_ * (*incorrect)) + (1 - negative_prior_) * (*correct)))) * pow((*the_x) - positive_mean, 2));
}
return neg_sigma;
}
DoubleReal PosteriorErrorProbabilityModel::computeProbability(DoubleReal score)
{
score = score + fabs(smallest_score_) + 0.001;
DoubleReal x_neg;
DoubleReal x_pos;
// the score is smaller than the peak of incorrectly assigned sequences. To ensure that the probabilities wont rise again use the incorrectly assigned peak for computation
if (score < incorrectly_assigned_fit_param_.x0)
{
x_neg = max_incorrectly_;
x_pos = ((this)->*(calc_correct_))(score, correctly_assigned_fit_param_);
}
// same as above. However, this time to ensure that probabilities wont drop again.
else if (score > correctly_assigned_fit_param_.x0)
{
x_neg = ((this)->*(calc_incorrect_))(score, incorrectly_assigned_fit_param_);
x_pos = max_correctly_;
}
// if its in between use the normal formula
else
{
x_neg = ((this)->*(calc_incorrect_))(score, incorrectly_assigned_fit_param_);
x_pos = ((this)->*(calc_correct_))(score, correctly_assigned_fit_param_);
}
return (negative_prior_ * x_neg) / ((negative_prior_ * x_neg) + (1 - negative_prior_) * x_pos);
}
TextFile * PosteriorErrorProbabilityModel::InitPlots(vector<double> & x_scores)
{
TextFile * file = new TextFile;
String output;
std::vector<DPosition<2> > points;
Int number_of_bins = param_.getValue("number_of_bins");
points.resize(number_of_bins);
DPosition<2> temp;
double dividing_score = (x_scores.back() - x_scores[0]) / number_of_bins;
temp.setX(dividing_score / 2);
temp.setY(0);
Int bin = 0;
points[bin] = temp;
double temp_divider = dividing_score;
for (std::vector<double>::iterator it = x_scores.begin(); it < x_scores.end(); ++it)
{
if (temp_divider - *it >= 0 && bin < number_of_bins - 1)
{
points[bin].setY(points[bin].getY() + 1);
}
else if (bin == number_of_bins - 1)
{
points[bin].setY(points[bin].getY() + 1);
}
else
{
temp.setX((temp_divider + temp_divider + dividing_score) / 2);
temp.setY(1);
++bin;
points[bin] = temp;
temp_divider += dividing_score;
}
}
for (vector<DPosition<2> >::iterator it = points.begin(); it < points.end(); ++it)
{
it->setY(it->getY() / (x_scores.size() * dividing_score));
}
TextFile data_points;
for (vector<DPosition<2> >::iterator it = points.begin(); it < points.end(); ++it)
{
String temp = it->getX();
temp += "\t";
temp += it->getY();
data_points << temp;
}
data_points.store((String)param_.getValue("output_name") + "_scores.txt");
output = "set output \"" + (String)param_.getValue("output_name") + ".ps\"";
(*file) << "set terminal postscript color solid linewidth 2.0 rounded";
//(*file)<<"set style empty solid 0.5 border -1";
//(*file)<<"set style function lines";
(*file) << "set xlabel \"discriminant score\"";
(*file) << "set ylabel \"density\"";
//TODO: (*file)<<"set title ";
(*file) << "set key off";
(*file) << (output);
String formula1, formula2;
formula1 = ((this)->*(getNegativeGnuplotFormula_))(incorrectly_assigned_fit_param_) + "* " + String(negative_prior_); //String(incorrectly_assigned_fit_param_.A) +" * exp(-(x - " + String(incorrectly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(incorrectly_assigned_fit_param_.sigma) + ") ** 2)"+ "*" + String(negative_prior_);
formula2 = ((this)->*(getPositiveGnuplotFormula_))(correctly_assigned_fit_param_) + "* (1 - " + String(negative_prior_) + ")"; //String(correctly_assigned_fit_param_.A) +" * exp(-(x - " + String(correctly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(correctly_assigned_fit_param_.sigma) + ") ** 2)"+ "* (1 - " + String(negative_prior_) + ")";
(*file) << ("plot \"" + (String)param_.getValue("output_name") + "_scores.txt\" with boxes, " + formula1 + " , " + formula2);
return file;
}
const String PosteriorErrorProbabilityModel::getGumbelGnuplotFormula(const GaussFitter::GaussFitResult & params) const
{
// build a formula with the fitted parameters for gnuplot
stringstream formula;
formula << "(1/" << params.sigma << ") * " << "exp(( " << params.x0 << "- x)/" << params.sigma << ") * exp(-exp((" << params.x0 << " - x)/" << params.sigma << "))";
return formula.str();
}
const String PosteriorErrorProbabilityModel::getGaussGnuplotFormula(const GaussFitter::GaussFitResult & params) const
{
stringstream formula;
formula << params.A << " * exp(-(x - " << params.x0 << ") ** 2 / 2 / (" << params.sigma << ") ** 2)";
return formula.str();
}
const String PosteriorErrorProbabilityModel::getBothGnuplotFormula(const GaussFitter::GaussFitResult & incorrect, const GaussFitter::GaussFitResult & correct) const
{
stringstream formula;
formula << negative_prior_ << "*" << ((this)->*(getNegativeGnuplotFormula_))(incorrect) << " + (1-" << negative_prior_ << ")*" << ((this)->*(getPositiveGnuplotFormula_))(correct);
return formula.str();
}
void PosteriorErrorProbabilityModel::plotTargetDecoyEstimation(vector<double> & target, vector<double> & decoy)
{
TextFile file;
String output;
std::vector<DPosition<3> > points;
Int number_of_bins = param_.getValue("number_of_bins");
points.resize(number_of_bins);
DPosition<3> temp;
sort(target.begin(), target.end());
sort(decoy.begin(), decoy.end());
double dividing_score = (max(target.back(), decoy.back()) /*scores.back()*/ - min(target[0], decoy[0]) /*scores[0]*/) / number_of_bins;
temp[0] = (dividing_score / 2);
temp[1] = 0;
temp[2] = 0;
Int bin = 0;
points[bin] = temp;
double temp_divider = dividing_score;
for (std::vector<double>::iterator it = target.begin(); it < target.end(); ++it)
{
*it = *it + fabs(smallest_score_) + 0.001;
if (temp_divider - *it >= 0 && bin < number_of_bins - 1)
{
points[bin][1] = (points[bin][1] + 1);
}
else if (bin == number_of_bins - 1)
{
points[bin][1] = (points[bin][1] + 1);
}
else
{
temp[0] = ((temp_divider + temp_divider + dividing_score) / 2);
temp[1] = 1;
++bin;
points[bin] = temp;
temp_divider += dividing_score;
}
}
bin = 0;
temp_divider = dividing_score;
for (std::vector<double>::iterator it = decoy.begin(); it < decoy.end(); ++it)
{
*it = *it + fabs(smallest_score_) + 0.001;
if (temp_divider - *it >= 0 && bin < number_of_bins - 1)
{
points[bin][2] = (points[bin][2] + 1);
}
else if (bin == number_of_bins - 1)
{
points[bin][2] = (points[bin][2] + 1);
}
else
{
// temp[0] = ((temp_divider + temp_divider + dividing_score)/2);
// temp[2] = 1;
++bin;
points[bin][2] = 1;
temp_divider += dividing_score;
}
}
for (vector<DPosition<3> >::iterator it = points.begin(); it < points.end(); ++it)
{
// if((*it)[1] > (*it)[2])
// {(*it)[1] = (*it)[1] + (*it)[2];}
/* else{/(*it)[2] = (*it)[1] + (*it)[2];//}*/
(*it)[1] = ((*it)[1] / ((decoy.size() + target.size()) * dividing_score));
(*it)[2] = ((*it)[2] / ((decoy.size() + target.size()) * dividing_score));
}
TextFile data_points;
for (vector<DPosition<3> >::iterator it = points.begin(); it < points.end(); ++it)
{
String temp = (*it)[0];
temp += "\t";
temp += (*it)[1];
temp += "\t";
temp += (*it)[2];
data_points << temp;
}
data_points.store((String)param_.getValue("output_name") + "_target_decoy_scores.txt");
output = String("set output \"") + (String)param_.getValue("output_name") + "_target_decoy.ps\"";
(file) << "set terminal postscript color solid linewidth 2.0 rounded";
//(*file)<<"set style empty solid 0.5 border -1";
//(*file)<<"set style function lines";
(file) << "set xlabel \"discriminant score\"";
(file) << "set ylabel \"density\"";
//TODO: (*file)<<"set title ";
(file) << "set key off";
(file) << (output);
String formula1, formula2;
formula1 = getGumbelGnuplotFormula(getIncorrectlyAssignedFitResult()) + "* " + String(getNegativePrior()); //String(incorrectly_assigned_fit_param_.A) +" * exp(-(x - " + String(incorrectly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(incorrectly_assigned_fit_param_.sigma) + ") ** 2)"+ "*" + String(negative_prior_);
formula2 = getGaussGnuplotFormula(getCorrectlyAssignedFitResult()) + "* (1 - " + String(getNegativePrior()) + ")"; //String(correctly_assigned_fit_param_.A) +" * exp(-(x - " + String(correctly_assigned_fit_param_.x0) + ") ** 2 / 2 / (" + String(correctly_assigned_fit_param_.sigma) + ") ** 2)"+ "* (1 - " + String(negative_prior_) + ")";
(file) << ("plot \"" + (String)param_.getValue("output_name") + "_target_decoy_scores.txt\" using 1:3 with boxes fill solid 0.8 noborder, \"" + (String)param_.getValue("output_name") + "_target_decoy_scores.txt\" using 1:2 with boxes, " + formula1 + " , " + formula2);
file.store((String)param_.getValue("output_name") + "_target_decoy");
}
} //namespace Math
} // namespace OpenMS
|