forked from deweylab/RSEM
-
Notifications
You must be signed in to change notification settings - Fork 0
/
SingleQModel.h
546 lines (441 loc) · 15.1 KB
/
SingleQModel.h
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
#ifndef SINGLEQMODEL_H_
#define SINGLEQMODEL_H_
#include<cmath>
#include<cstdio>
#include<cassert>
#include<cstring>
#include<string>
#include<algorithm>
#include<sstream>
#include<iostream>
#include<vector>
#include "utils.h"
#include "my_assert.h"
#include "Orientation.h"
#include "LenDist.h"
#include "RSPD.h"
#include "QualDist.h"
#include "QProfile.h"
#include "NoiseQProfile.h"
#include "ModelParams.h"
#include "RefSeq.h"
#include "Refs.h"
#include "SingleReadQ.h"
#include "SingleHit.h"
#include "ReadReader.h"
#include "simul.h"
class SingleQModel {
public:
SingleQModel(Refs* refs = NULL) {
this->refs = refs;
M = (refs != NULL ? refs->getM() : 0);
memset(N, 0, sizeof(N));
estRSPD = false;
needCalcConPrb = true;
ori = new Orientation();
gld = new LenDist();
mld = NULL;
rspd = new RSPD(estRSPD);
qd = new QualDist();
qpro = new QProfile();
nqpro = new NoiseQProfile();
mean = -1.0; sd = 0.0;
mw = NULL;
seedLen = 0;
}
//If it is not a master node, only init & update can be used!
SingleQModel(ModelParams& params, bool isMaster = true) {
M = params.M;
memcpy(N, params.N, sizeof(params.N));
refs = params.refs;
estRSPD = params.estRSPD;
mean = params.mean; sd = params.sd;
seedLen = params.seedLen;
needCalcConPrb = true;
ori = NULL; gld = NULL; mld = NULL; rspd = NULL; qd = NULL; qpro = NULL; nqpro = NULL;
mw = NULL;
if (isMaster) {
gld = new LenDist(params.minL, params.maxL);
if (mean >= EPSILON) {
mld = new LenDist(params.mate_minL, params.mate_maxL);
}
if (!estRSPD) { rspd = new RSPD(estRSPD); }
qd = new QualDist();
}
ori = new Orientation(params.probF);
if (estRSPD) { rspd = new RSPD(estRSPD, params.B); }
qpro = new QProfile();
nqpro = new NoiseQProfile();
}
~SingleQModel() {
refs = NULL;
if (ori != NULL) delete ori;
if (gld != NULL) delete gld;
if (mld != NULL) delete mld;
if (rspd != NULL) delete rspd;
if (qd != NULL) delete qd;
if (qpro != NULL) delete qpro;
if (nqpro != NULL) delete nqpro;
if (mw != NULL) delete[] mw;
/*delete[] p1, p2;*/
}
//SingleQModel& operator=(const SingleQModel&);
void estimateFromReads(const char*);
//if prob is too small, just make it 0
double getConPrb(const SingleReadQ& read, const SingleHit& hit) const {
if (read.isLowQuality()) return 0.0;
double prob;
int sid = hit.getSid();
RefSeq &ref = refs->getRef(sid);
int fullLen = ref.getFullLen();
int totLen = ref.getTotLen();
int dir = hit.getDir();
int pos = hit.getPos();
int readLen = read.getReadLength();
int fpos = (dir == 0 ? pos : totLen - pos - readLen); // the aligned position reported in SAM file, should be a coordinate in forward strand
general_assert(fpos >= 0, "The alignment of read " + read.getName() + " to transcript " + itos(sid) + " starts at " + itos(fpos) + \
" from the forward direction, which should be a non-negative number! " + \
"It is possible that the aligner you use gave different read lengths for a same read in SAM file.");
general_assert(fpos + readLen <= totLen,"Read " + read.getName() + " is hung over the end of transcript " + itos(sid) + "! " \
+ "It is possible that the aligner you use gave different read lengths for a same read in SAM file.");
general_assert(readLen <= totLen, "Read " + read.getName() + " has length " + itos(readLen) + ", but it is aligned to transcript " \
+ itos(sid) + ", whose length (" + itos(totLen) + ") is shorter than the read's length!");
int seedPos = (dir == 0 ? pos : totLen - pos - seedLen); // the aligned position of the seed in forward strand coordinates
if (seedPos >= fullLen || ref.getMask(seedPos)) return 0.0;
int effL;
double value;
if (mld != NULL) {
int minL = std::max(readLen, gld->getMinL());
int maxL = std::min(totLen - pos, gld->getMaxL());
int pfpos; // possible fpos for fragment
value = 0.0;
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
pfpos = (dir == 0 ? pos : totLen - pos - fragLen);
effL = std::min(fullLen, totLen - fragLen + 1);
value += gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen) * mld->getAdjustedProb(readLen, fragLen);
}
}
else {
effL = std::min(fullLen, totLen - readLen + 1);
value = gld->getAdjustedProb(readLen, totLen) * rspd->getAdjustedProb(fpos, effL, fullLen);
}
prob = ori->getProb(dir) * value * qpro->getProb(read.getReadSeq(), read.getQScore(), ref, pos, dir);
if (prob < EPSILON) { prob = 0.0; }
prob = (mw[sid] < EPSILON ? 0.0 : prob / mw[sid]);
return prob;
}
double getNoiseConPrb(const SingleReadQ& read) {
if (read.isLowQuality()) return 0.0;
double prob = mld != NULL ? mld->getProb(read.getReadLength()) : gld->getProb(read.getReadLength());
prob *= nqpro->getProb(read.getReadSeq(), read.getQScore());
if (prob < EPSILON) { prob = 0.0; }
prob = (mw[0] < EPSILON ? 0.0 : prob / mw[0]);
return prob;
}
double getLogP() { return nqpro->getLogP(); }
void init();
void update(const SingleReadQ& read, const SingleHit& hit, double frac) {
if (read.isLowQuality() || frac < EPSILON) return;
const RefSeq& ref = refs->getRef(hit.getSid());
int dir = hit.getDir();
int pos = hit.getPos();
if (estRSPD) {
int fullLen = ref.getFullLen();
// Only use one strand to estimate RSPD
if (ori->getProb(0) >= ORIVALVE && dir == 0) {
rspd->update(pos, fullLen, frac);
}
if (ori->getProb(0) < ORIVALVE && dir == 1) {
int totLen = ref.getTotLen();
int readLen = read.getReadLength();
int pfpos, effL;
if (mld != NULL) {
int minL = std::max(readLen, gld->getMinL());
int maxL = std::min(totLen - pos, gld->getMaxL());
double sum = 0.0;
assert(maxL >= minL);
std::vector<double> frag_vec(maxL - minL + 1, 0.0);
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
pfpos = totLen - pos - fragLen;
effL = std::min(fullLen, totLen - fragLen + 1);
frag_vec[fragLen - minL] = gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen) * mld->getAdjustedProb(readLen, fragLen);
sum += frag_vec[fragLen - minL];
}
assert(sum >= EPSILON);
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
pfpos = totLen - pos - fragLen;
rspd->update(pfpos, fullLen, frac * (frag_vec[fragLen - minL] / sum));
}
}
else {
rspd->update(totLen - pos - readLen, fullLen, frac);
}
}
}
qpro->update(read.getReadSeq(), read.getQScore(), ref, pos, dir, frac);
}
void updateNoise(const SingleReadQ& read, double frac) {
if (read.isLowQuality() || frac < EPSILON) return;
nqpro->update(read.getReadSeq(), read.getQScore(), frac);
}
void finish();
void collect(const SingleQModel&);
//void copy(const SingleQModel&);
bool getNeedCalcConPrb() { return needCalcConPrb; }
void setNeedCalcConPrb(bool value) { needCalcConPrb = value; }
//void calcP1();
//void calcP2();
//double* getP1() { return p1; }
//double* getP2() { return p2; }
void read(const char*);
void write(const char*);
const LenDist& getGLD() { return *gld; }
void startSimulation(simul*, const std::vector<double>&);
bool simulate(READ_INT_TYPE, SingleReadQ&, int&);
void finishSimulation();
//Use it after function 'read' or 'estimateFromReads'
const double* getMW() {
assert(mw != NULL);
return mw;
}
int getModelType() const { return model_type; }
private:
static const int model_type = 1;
static const int read_type = 1;
int M;
READ_INT_TYPE N[3];
Refs *refs;
double mean, sd;
int seedLen;
//double *p1, *p2; P_i' & P_i'';
bool estRSPD; // true if estimate RSPD
bool needCalcConPrb; //true need, false does not need
Orientation *ori;
LenDist *gld, *mld;
RSPD *rspd;
QualDist *qd;
QProfile *qpro;
NoiseQProfile *nqpro;
simul *sampler; // for simulation
double *theta_cdf; // for simulation
double *mw; // for masking
void calcMW();
};
void SingleQModel::estimateFromReads(const char* readFN) {
int s;
char readFs[2][STRLEN];
SingleReadQ read;
int n_warns = 0;
mld != NULL ? mld->init() : gld->init();
for (int i = 0; i < 3; i++)
if (N[i] > 0) {
genReadFileNames(readFN, i, read_type, s, readFs);
ReadReader<SingleReadQ> reader(s, readFs, refs->hasPolyA(), seedLen); // allow calculation of calc_lq() function
READ_INT_TYPE cnt = 0;
while (reader.next(read)) {
if (!read.isLowQuality()) {
mld != NULL ? mld->update(read.getReadLength(), 1.0) : gld->update(read.getReadLength(), 1.0);
qd->update(read.getQScore());
if (i == 0) { nqpro->updateC(read.getReadSeq(), read.getQScore()); }
}
else if (read.getReadLength() < seedLen)
if (++n_warns <= MAX_WARNS)
fprintf(stderr, "Warning: Read %s is ignored due to read length (= %d) < seed length (= %d)!\n", read.getName().c_str(), read.getReadLength(), seedLen);
++cnt;
if (verbose && cnt % 1000000 == 0) { std::cout<< cnt<< " READS PROCESSED"<< std::endl; }
}
if (verbose) { std::cout<< "estimateFromReads, N"<< i<< " finished."<< std::endl; }
}
if (n_warns > 0) fprintf(stderr, "Warning: There are %d reads ignored in total.\n", n_warns);
mld != NULL ? mld->finish() : gld->finish();
if (mean >= EPSILON) { //mean should be > 0
assert(mld->getMaxL() <= gld->getMaxL());
gld->setAsNormal(mean, sd, std::max(mld->getMinL(), gld->getMinL()), gld->getMaxL());
}
qd->finish();
nqpro->calcInitParams();
mw = new double[M + 1];
calcMW();
}
void SingleQModel::init() {
if (estRSPD) rspd->init();
qpro->init();
nqpro->init();
}
void SingleQModel::finish() {
if (estRSPD) rspd->finish();
qpro->finish();
nqpro->finish();
needCalcConPrb = true;
if (estRSPD) calcMW();
}
void SingleQModel::collect(const SingleQModel& o) {
if (estRSPD) rspd->collect(*(o.rspd));
qpro->collect(*(o.qpro));
nqpro->collect(*(o.nqpro));
}
//Only master node can call
void SingleQModel::read(const char* inpF) {
int val;
FILE *fi = fopen(inpF, "r");
general_assert(fi != NULL, "Cannot open " + cstrtos(inpF) + "! It may not exist.");
assert(fscanf(fi, "%d", &val) == 1);
assert(val == model_type);
ori->read(fi);
gld->read(fi);
assert(fscanf(fi, "%d", &val) == 1);
if (val > 0) {
if (mld == NULL) mld = new LenDist();
mld->read(fi);
}
rspd->read(fi);
qd->read(fi);
qpro->read(fi);
nqpro->read(fi);
if (fscanf(fi, "%d", &val) == 1) {
if (M == 0) M = val;
if (M == val) {
mw = new double[M + 1];
for (int i = 0; i <= M; i++) assert(fscanf(fi, "%lf", &mw[i]) == 1);
}
}
fclose(fi);
}
//Only master node can call. Only be called at EM.cpp
void SingleQModel::write(const char* outF) {
FILE *fo = fopen(outF, "w");
fprintf(fo, "%d\n", model_type);
fprintf(fo, "\n");
ori->write(fo); fprintf(fo, "\n");
gld->write(fo); fprintf(fo, "\n");
if (mld != NULL) {
fprintf(fo, "1\n");
mld->write(fo);
}
else { fprintf(fo, "0\n"); }
fprintf(fo, "\n");
rspd->write(fo); fprintf(fo, "\n");
qd->write(fo); fprintf(fo, "\n");
qpro->write(fo); fprintf(fo, "\n");
nqpro->write(fo);
if (mw != NULL) {
fprintf(fo, "\n%d\n", M);
for (int i = 0; i < M; i++) {
fprintf(fo, "%.15g ", mw[i]);
}
fprintf(fo, "%.15g\n", mw[M]);
}
fclose(fo);
}
void SingleQModel::startSimulation(simul* sampler, const std::vector<double>& theta) {
this->sampler = sampler;
theta_cdf = new double[M + 1];
for (int i = 0; i <= M; i++) {
theta_cdf[i] = theta[i];
if (i > 0) theta_cdf[i] += theta_cdf[i - 1];
}
rspd->startSimulation(M, refs);
qd->startSimulation();
qpro->startSimulation();
nqpro->startSimulation();
}
bool SingleQModel::simulate(READ_INT_TYPE rid, SingleReadQ& read, int& sid) {
int dir, pos, readLen, fragLen;
std::string name;
std::string qual, readseq;
std::ostringstream strout;
sid = sampler->sample(theta_cdf, M + 1);
if (sid == 0) {
dir = pos = 0;
readLen = (mld != NULL ? mld->simulate(sampler, -1) : gld->simulate(sampler, -1));
qual = qd->simulate(sampler, readLen);
readseq = nqpro->simulate(sampler, readLen, qual);
}
else {
RefSeq &ref = refs->getRef(sid);
dir = ori->simulate(sampler);
fragLen = gld->simulate(sampler, ref.getTotLen());
if (fragLen < 0) return false;
int effL = std::min(ref.getFullLen(), ref.getTotLen() - fragLen + 1);
pos = rspd->simulate(sampler, sid, effL);
if (pos < 0) return false;
if (dir > 0) pos = ref.getTotLen() - pos - fragLen;
if (mld != NULL) {
readLen = mld->simulate(sampler, fragLen);
if (readLen < 0) return false;
qual = qd->simulate(sampler, readLen);
readseq = qpro->simulate(sampler, readLen, pos, dir, qual, ref);
}
else {
qual = qd->simulate(sampler, fragLen);
readseq = qpro->simulate(sampler, fragLen, pos, dir, qual, ref);
}
}
strout<<rid<<"_"<<dir<<"_"<<sid<<"_"<<pos;
name = strout.str();
read = SingleReadQ(name, readseq, qual);
return true;
}
void SingleQModel::finishSimulation() {
delete[] theta_cdf;
rspd->finishSimulation();
qd->finishSimulation();
qpro->finishSimulation();
nqpro->finishSimulation();
}
void SingleQModel::calcMW() {
double probF, probR;
assert((mld == NULL ? gld->getMinL() : mld->getMinL()) >= seedLen);
memset(mw, 0, sizeof(double) * (M + 1));
mw[0] = 1.0;
probF = ori->getProb(0);
probR = ori->getProb(1);
for (int i = 1; i <= M; i++) {
RefSeq& ref = refs->getRef(i);
int totLen = ref.getTotLen();
int fullLen = ref.getFullLen();
double value = 0.0;
int minL, maxL;
int effL, pfpos;
int end = std::min(fullLen, totLen - seedLen + 1);
double factor;
for (int seedPos = 0; seedPos < end; seedPos++)
if (ref.getMask(seedPos)) {
//forward
minL = gld->getMinL();
maxL = std::min(gld->getMaxL(), totLen - seedPos);
pfpos = seedPos;
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
effL = std::min(fullLen, totLen - fragLen + 1);
factor = (mld == NULL ? 1.0 : mld->getAdjustedCumulativeProb(std::min(mld->getMaxL(), fragLen), fragLen));
value += probF * gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen) * factor;
}
//reverse
minL = gld->getMinL();
maxL = std::min(gld->getMaxL(), seedPos + seedLen);
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
pfpos = seedPos - (fragLen - seedLen);
effL = std::min(fullLen, totLen - fragLen + 1);
factor = (mld == NULL ? 1.0 : mld->getAdjustedCumulativeProb(std::min(mld->getMaxL(), fragLen), fragLen));
value += probR * gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen) * factor;
}
}
//for reverse strand masking
for (int seedPos = end; seedPos <= totLen - seedLen; seedPos++) {
minL = std::max(gld->getMinL(), seedPos + seedLen - fullLen + 1);
maxL = std::min(gld->getMaxL(), seedPos + seedLen);
for (int fragLen = minL; fragLen <= maxL; fragLen++) {
pfpos = seedPos - (fragLen - seedLen);
effL = std::min(fullLen, totLen - fragLen + 1);
factor = (mld == NULL ? 1.0 : mld->getAdjustedCumulativeProb(std::min(mld->getMaxL(), fragLen), fragLen));
value += probR * gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen) * factor;
}
}
mw[i] = 1.0 - value;
if (mw[i] < 1e-8) {
// fprintf(stderr, "Warning: %dth reference sequence is masked for almost all positions!\n", i);
mw[i] = 0.0;
}
}
}
#endif /* SINGLEQMODEL_H_ */