forked from gpertea/stringtie
-
Notifications
You must be signed in to change notification settings - Fork 0
/
tablemaker.cpp
executable file
·403 lines (375 loc) · 13.6 KB
/
tablemaker.cpp
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
//#include "tablemaker.h"
#include "rlink.h"
#include <numeric>
extern GStr ballgown_dir;
int rc_cov_inc(int i) {
return ++i;
}
/*
void rc_update_tdata(BundleData& bundle, GffObj& scaff,
double cov, double fpkm) {
if (bundle.rc_data==NULL) return;
RC_BundleData& rc = *(bundle.rc_data);
if (rc.exons.size()==0) return;
RC_ScaffData& q = (RC_ScaffData*)scaff.uptr;
set<RC_ScaffData>::iterator tdata = rc.tdata.find(q);
if (tdata==rc.tdata.end()) {
fprintf(stderr, "Error: cannot locate bundle ref. transcript %s (%s:%d-%d)!\n",
scaff.getID(), scaff.getGSeqName(), scaff.start, scaff.end);
return;
}
(*tdata).cov=cov;
(*tdata).fpkm=fpkm;
}
*/
void BundleData::keepGuide(GffObj* t, GPVec<RC_TData>* rc_tdata,
GPVec<RC_Feature>* rc_edata, GPVec<RC_Feature>* rc_idata) {
if (rc_data==NULL) {
rc_init(t, rc_tdata, rc_edata, rc_idata);
}
keepguides.Add(t);
t->udata=(int)rc_data->addTranscript(*t);
}
struct COvlSorter {
bool operator() (pair<int, const RC_Feature*> i,
pair<int, const RC_Feature*> j) {
return (i.first>j.first); //sort in decreasing order of overlap length
}
} OvlSorter;
void rc_updateExonCounts(const RC_ExonOvl& exonovl, int nh) {
//this only gets read overlaps > 5bp and otherwise filtered in evalReadAln()
exonovl.feature->rcount++;
if (nh>1) {
exonovl.feature->mrcount += (1.0/nh);
exonovl.feature->movlcount += ((double)exonovl.ovlen/nh);
}
else { // nh<=1
exonovl.feature->mrcount++;
exonovl.feature->movlcount += exonovl.ovlen;
exonovl.feature->ucount++;
}
}
bool BundleData::evalReadAln(GReadAlnData& alndata, char& xstrand) {
//GBamRecord& brec, char& strand, int nh) {
if (rc_data==NULL) {
return false; //no ref transcripts available for this reads' region
}
GBamRecord& brec=*(alndata.brec);
int mate_pos=brec.mate_start();
int nh=alndata.nh;
if ((int)brec.end<rc_data->lmin || (int)brec.start>rc_data->rmax) {
return false; //hit outside coverage area
}
if (rc_data->g_exons.Count()==0 || rc_data->g_tdata.Count()==0)
return false; //nothing to do without transcripts
//check this read alignment against ref exons and introns
char strandbits=0;
bool result=false;
for (int i=0;i<brec.exons.Count();i++) {
if (ballgown)
rc_data->updateCov(xstrand, nh, brec.exons[i].start, brec.exons[i].len());
GArray<RC_ExonOvl> exonOverlaps(true, true); //overlaps sorted by decreasing length
if (rc_data->findOvlExons(exonOverlaps, brec.exons[i].start,
brec.exons[i].end, xstrand, mate_pos)) {
result=true;
int max_ovl=exonOverlaps[0].ovlen;
//alndata.g_exonovls.Add(new GVec<RC_ExonOvl>(exonOverlaps));
for (int k=0;k<exonOverlaps.Count();++k) {
//if (exonOverlaps[k].ovlen < 5) break; //ignore very short overlaps
if (k && (exonOverlaps[k].mate_ovl < exonOverlaps[0].mate_ovl
|| exonOverlaps[k].ovlen+5<max_ovl) )
break; //ignore further overlaps after a mate matched or if they are shorter than max_overlap-5
if (exonOverlaps[k].feature->strand=='+') strandbits |= 0x01;
else if (exonOverlaps[k].feature->strand=='-') strandbits |= 0x02;
//TODO: perhaps we could use a better approach for non-overlapping ref exons
// spanned by this same read alignment
//counting this overlap for multiple exons if it's similarly large
//(in the shared region of overlapping exons)
rc_updateExonCounts(exonOverlaps[k], nh);
}
} //ref exon overlaps
if (i>0) { //intron processing
int j_l=brec.exons[i-1].end+1;
int j_r=brec.exons[i].start-1;
RC_Feature* ri=rc_data->findIntron(j_l, j_r, xstrand);
alndata.juncs.Add(new CJunction(j_l, j_r)); //don't set strand, etc. for now
if (ri) { //update guide intron counts
ri->rcount++;
ri->mrcount += (nh > 1) ? (1.0/nh) : 1;
if (nh==1) ri->ucount++;
alndata.juncs.Last()->guide_match=1;
}
} //intron processing
}
if (xstrand=='.' && strandbits && strandbits<3) {
xstrand = (strandbits==1) ? '+' : '-';
}
return result;
}
FILE* rc_fwopen(const char* fname) {
if (strcmp(fname,"-")==0) return stdout;
GStr fpath(ballgown_dir); //always ends with '/'
//fpath += '.';
fpath += fname;
//fpath += "/";
//fpath += fname;
fpath += ".ctab";
FILE* fh=fopen(fpath.chars(), "w");
if (fh==NULL) {
fprintf(stderr, "Error: cannot create file %s\n",
fpath.chars());
exit(1);
}
return fh;
}
FILE* rc_frenopen(const char* fname) {
//if (strcmp(fname,"-")==0) return stdout;
GStr fpath(ballgown_dir);
//fpath += '.';
fpath += fname;
//fpath += "/";
//fpath += fname;
fpath += ".ctab";
GStr fren(fpath);
fren += ".tmp";
//rename fpath to fren and open fren for reading
if (rename(fpath.chars(), fren.chars())!=0) {
GError("Error: cannot rename %s to %s!\n", fpath.chars(), fren.chars());
}
FILE* fh=fopen(fren.chars(), "r");
if (fh==NULL) {
GError("Error: cannot open file %s\n", fren.chars());
}
return fh;
}
void rc_frendel(const char* fname) {
GStr fpath(ballgown_dir);
//fpath += '.';
fpath += fname;
fpath += ".ctab";
fpath += ".tmp";
if (remove(fpath.chars())!=0) {
GMessage("Warning: could not remove file %s!\n",fpath.chars());
}
}
void rc_write_f2t(FILE* fh, map<uint, set<uint> >& f2t) {
for (map<uint, set<uint> >::iterator m=f2t.begin(); m!=f2t.end(); ++m) {
uint f_id=(*m).first;
set<uint>& tset = (*m).second;
for (set<uint>::iterator it=tset.begin();it!=tset.end();++it) {
uint t_id = *it;
fprintf(fh, "%u\t%u\n", f_id, t_id);
}
}
fflush(fh);
}
void rc_update_exons(RC_BundleData& rc) {
//update stdev etc. for all exons in bundle
for (int f=0;f<rc.g_exons.Count(); ++f) {
RC_Feature& exon = *(rc.g_exons[f]);
//assert( exon.l >= rc.lmin );
int L=exon.l-rc.lmin;
int xlen=exon.r-exon.l+1;
if (exon.l < rc.lmin) {
//shouldn't be here, bundle read-counting boundaries should be based on exons!
if (exon.r<rc.lmin) continue;
xlen-=(rc.lmin-exon.l+1);
L=0;
}
if (rc.rmax<exon.r) {
if (exon.l>rc.rmax) continue; //should never happen
xlen-=(exon.r-rc.rmax+1);
}
int R=L+xlen;
vector<int>::iterator xcov_begin;
vector<int>::iterator xcov_end;
vector<float>::iterator xmcov_begin;
vector<float>::iterator xmcov_end;
if (exon.strand=='+' || exon.strand=='.') {
xcov_begin = rc.f_cov.begin()+L;
xcov_end = rc.f_cov.begin()+R;
xmcov_begin = rc.f_mcov.begin()+L;
xmcov_end = rc.f_mcov.begin()+R;
} else {
xcov_begin = rc.r_cov.begin()+L;
xcov_end = rc.r_cov.begin()+R;
xmcov_begin = rc.r_mcov.begin()+L;
xmcov_end = rc.r_mcov.begin()+R;
}
double avg = (double)accumulate(xcov_begin, xcov_end, 0) / xlen;
vector<double> diff(xlen);
transform(xcov_begin, xcov_end, diff.begin(),
bind2nd( minus<double>(), avg));
double sq_sum = inner_product(diff.begin(), diff.end(), diff.begin(), 0.0);
double stdev = sqrt(sq_sum / xlen);
double mavg = (double)accumulate(xmcov_begin, xmcov_end, 0) / xlen;
vector<double> mdiff(xlen);
transform(xmcov_begin, xmcov_end, mdiff.begin(),
bind2nd( minus<double>(), mavg));
sq_sum = inner_product(mdiff.begin(), mdiff.end(), mdiff.begin(), 0.0);
double mstdev = sqrt(sq_sum / xlen);
exon.avg=avg;
exon.stdev=stdev;
exon.mavg=mavg;
exon.mstdev=mstdev;
} //for each exon in bundle
}
void rc_write_RCfeature( GPVec<RC_TData>& rcdata, GPVec<RC_Feature>& features, FILE*& fdata, FILE*& f2t,
bool is_exon=false) {
for (int i=0;i<features.Count();++i) {
RC_Feature& f=*(features[i]);
const char* ref_name=rcdata[f.t_ids[0]-1]->ref_t->getGSeqName();
if (is_exon) {
fprintf(fdata, "%u\t%s\t%c\t%d\t%d\t%d\t%d\t%.2f\t%.4f\t%.4f\t%.4f\t%.4f\n",
f.id, ref_name, f.strand, f.l, f.r, f.rcount,
f.ucount, f.mrcount, f.avg, f.stdev, f.mavg, f.mstdev);
}
else { //introns
fprintf(fdata,"%u\t%s\t%c\t%d\t%d\t%d\t%d\t%.2f\n",f.id, ref_name,
f.strand, f.l, f.r, f.rcount, f.ucount, f.mrcount);
}
// f2t -------
for (int t=0;t<f.t_ids.Count();++t)
fprintf(f2t, "%u\t%u\n", f.id, f.t_ids[t]);
} //for each feature
fclose(fdata);
fclose(f2t);
}
/*void rc_write_counts(const char* refname, BundleData& bundle) {
RC_BundleData& rc = *bundle.rc_data;
//if (rc.exons.size()==0) return;
*
*/
void rc_writeRC(GPVec<RC_TData>& RC_data,
GPVec<RC_Feature>& RC_exons,
GPVec<RC_Feature>& RC_introns,
FILE* &f_tdata, FILE* &f_edata, FILE* &f_idata,
FILE* &f_e2t, FILE* &f_i2t) {
for (int t=0;t<RC_data.Count();++t) {
//File: t_data.ctab
//t_id tname chr strand start end num_exons gene_id gene_name cufflinks_cov cufflinks_fpkm
RC_TData& sd=*RC_data[t];
const char* refname = sd.ref_t->getGSeqName();
const char* genename= sd.ref_t->getGeneName();
const char* geneID= sd.ref_t->getGeneID();
if (genename==NULL) genename=".";
if (geneID==NULL) geneID=".";
fprintf(f_tdata, "%u\t%s\t%c\t%d\t%d\t%s\t%d\t%d\t%s\t%s\t%f\t%f\n",
sd.t_id, refname, sd.ref_t->strand, sd.l, sd.r, sd.ref_t->getID(),
sd.t_exons.Count(), sd.eff_len, geneID,
genename, sd.cov, sd.fpkm);
}//for each transcript
//fflush(f_tdata);
fclose(f_tdata);
//File: e_data.ctab
//e_id chr gstart gend rcount ucount mrcount
rc_write_RCfeature(RC_data, RC_exons, f_edata, f_e2t, true);
//File: i_data.ctab
//i_id chr gstart gend rcount ucount mrcount
rc_write_RCfeature(RC_data, RC_introns, f_idata, f_i2t);
}
void RC_TData::rc_addFeatures(uint& c_e_id, GList<RC_Feature>& exonSet, GPVec<RC_Feature>& exonTable,
uint& c_i_id, GList<RC_Feature>& intronSet, GPVec<RC_Feature>& intronTable) {
GASSERT(ref_t);
GffObj& m = *(ref_t);
int ecache_idx = exonSet.Count()-1;
int icache_idx = intronSet.Count()-1;
//int ecache_idx = e_idx_cache>=0 ? e_idx_cache : exonSet.Count()-1;
//int icache_idx = i_idx_cache>=0 ? i_idx_cache : intronSet.Count()-1;
for (int i = 0; i < m.exons.Count(); ++i) {
addFeature((int)m.exons[i]->start, (int)m.exons[i]->end, t_exons, c_e_id, exonSet, exonTable, ecache_idx);
//if (i==0) e_idx_cache=ecache_idx;
if (i>0) { //store intron
//if (i==1) i_idx_cache=icache_idx;
addFeature(m.exons[i-1]->end+1, m.exons[i]->start-1, t_introns, c_i_id,
intronSet, intronTable, icache_idx);
} //for each intron
} //for each exon
}
void RC_TData::addFeature(int fl, int fr, GPVec<RC_Feature>& fvec,
uint& f_id, GList<RC_Feature>& fset,
GPVec<RC_Feature>& fdata, int& cache_idx) {
//f_id is the largest f_id inserted so far in fset
bool add_new = true;
RC_Feature* newseg=new RC_Feature(fl, fr, ref_t->strand, 0, this->t_id);
//RC_Feature* newfeature=NULL;
int fit=cache_idx<0 ? fset.Count()-1 : cache_idx;
int fp_id=-1;
if (fset.Count()>0) {
if (*newseg < *(fset[fit])) {
bool eq=false;
while (*newseg < *(fset[fit]) || (eq = (*newseg==*(fset[fit])))) {
if (eq) {
add_new = false;
fp_id = fset[fit]->id; //fset[fit]->id;
break;
}
//newseg< fset[fit]
--fit;
if (fit<0) break; //newseg should be inserted at 0
} //while newseg<fset[fit]
if (add_new) ++fit;
// newseg < fset[fit+1]
//we'll insert newseg at position fit+1
}
else { //newseg >= *fset[fit]
bool eq=false;
while (*(fset[fit]) < *newseg || (eq = (*newseg==*(fset[fit])))) {
if (eq) {
add_new = false;
fp_id = fset[fit]->id;
break;
}
++fit;
if (fit==fset.Count()) {
//newseg should be appended to the list
break;
}
}
//if (fit<=fset.Count() && !add_new) {
//fset[fit-1] < newseg < fset[fit]
//we'll insert newseg at position fit
//}
}
} //check existing set
if (add_new) { //did not see this feature before
newseg->id = ++f_id;
if (fit<0) fit = fset.Add(newseg);
else fset.sortInsert(fit, newseg);
if (fit<0) {
GError("Error: feature %d-%d (%c) already in feature set!\n",
newseg->l, newseg->r, newseg->strand);
}
cache_idx=fit;
fp_id=fdata.Add(newseg)+1;
#ifdef DEBUG
if (fdata.Count()!=(int)f_id) {
GMessage("Error: fdata.Count=%d, f_id=%d\n", fdata.Count(), f_id);
}
#endif
GASSERT((uint)fdata.Count()==f_id);
}
else { //feature seen before, update its parent list
fdata[fp_id-1]->t_ids.Add(this->t_id);
delete newseg;
}
//fvec.push_back(newseg);
GASSERT(fdata[fp_id-1]->id==(uint)fp_id);
fvec.Add(fdata[fp_id-1]);
}
void Ballgown_setupFiles(FILE* &f_tdata, FILE* &f_edata, FILE* &f_idata,
FILE* &f_e2t, FILE* &f_i2t) {
if (f_tdata == NULL) {
//first call, create the files
f_tdata = rc_fwopen("t_data");
fprintf(f_tdata, "t_id\tchr\tstrand\tstart\tend\tt_name\tnum_exons\tlength\tgene_id\tgene_name\tcov\tFPKM\n");
f_edata = rc_fwopen("e_data");
fprintf(f_edata, "e_id\tchr\tstrand\tstart\tend\trcount\tucount\tmrcount\tcov\tcov_sd\tmcov\tmcov_sd\n");
f_idata = rc_fwopen("i_data");
fprintf(f_idata, "i_id\tchr\tstrand\tstart\tend\trcount\tucount\tmrcount\n");
f_e2t = rc_fwopen("e2t");
fprintf(f_e2t, "e_id\tt_id\n");
f_i2t = rc_fwopen("i2t");
fprintf(f_i2t, "i_id\tt_id\n");
}
}