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
|
#ifndef __SAMPLER__HPP__
#define __SAMPLER__HPP__
extern "C" {
#include "io_lib/scram.h"
#include "io_lib/os.h"
#undef max
#undef min
}
// for cpp-format
#include "spdlog/spdlog.h"
#include "spdlog/fmt/fmt.h"
#include <tbb/atomic.h>
#include <iostream>
#include <fstream>
#include <atomic>
#include <vector>
#include <random>
#include <memory>
#include <exception>
#include <thread>
#include <unordered_map>
#include <unordered_set>
#include <mutex>
#include <thread>
#include <condition_variable>
#include <tbb/concurrent_queue.h>
#include <boost/config.hpp>
#include <boost/timer/timer.hpp>
#include <boost/filesystem.hpp>
#include <boost/math/special_functions/digamma.hpp>
#include <boost/program_options.hpp>
#include <boost/pending/disjoint_sets.hpp>
#include <boost/math/distributions/normal.hpp>
#include "ClusterForest.hpp"
#include "AlignmentLibrary.hpp"
#include "MiniBatchInfo.hpp"
#include "BAMQueue.hpp"
#include "SalmonMath.hpp"
#include "FASTAParser.hpp"
#include "LibraryFormat.hpp"
#include "Transcript.hpp"
#include "ReadPair.hpp"
#include "ErrorModel.hpp"
#include "AlignmentModel.hpp"
#include "FragmentLengthDistribution.hpp"
#include "TranscriptCluster.hpp"
#include "SailfishUtils.hpp"
#include "SalmonUtils.hpp"
#include "SalmonConfig.hpp"
#include "SalmonOpts.hpp"
#include "OutputUnmappedFilter.hpp"
namespace salmon {
namespace sampler {
namespace bfs = boost::filesystem;
using salmon::math::LOG_0;
using salmon::math::LOG_1;
using salmon::math::logAdd;
using salmon::math::logSub;
template <typename FragT>
using AlignmentBatch = std::vector<FragT>;
template <typename FragT>
using MiniBatchQueue = tbb::concurrent_queue<MiniBatchInfo<FragT>*>;
template <typename FragT>
using OutputQueue = tbb::concurrent_bounded_queue<FragT*>;
template <typename FragT>
void sampleMiniBatch(AlignmentLibrary<FragT>& alnLib,
MiniBatchQueue<AlignmentGroup<FragT*>>& workQueue,
std::condition_variable& workAvailable,
std::mutex& cvmutex,
volatile bool& doneParsing,
std::atomic<size_t>& activeBatches,
const SalmonOpts& salmonOpts,
bool& burnedIn,
std::atomic<size_t>& processedReads,
OutputQueue<FragT>& outputQueue) {
// Seed with a real random value, if available
std::random_device rd;
auto log = spdlog::get("jointLog");
// Create a random uniform distribution
std::default_random_engine eng(rd());
std::uniform_real_distribution<> uni(0.0, 1.0 + std::numeric_limits<double>::min());
using salmon::math::LOG_0;
using salmon::math::logAdd;
using salmon::math::logSub;
bool useFSPD{salmonOpts.useFSPD};
auto& refs = alnLib.transcripts();
auto& clusterForest = alnLib.clusterForest();
auto& fragmentQueue = alnLib.fragmentQueue();
auto& alignmentGroupQueue = alnLib.alignmentGroupQueue();
auto& fragLengthDist = *(alnLib.fragmentLengthDistribution());
auto& alnMod = alnLib.alignmentModel();
std::vector<FragmentStartPositionDistribution>& fragStartDists =
alnLib.fragmentStartPositionDistributions();
const auto expectedLibraryFormat = alnLib.format();
std::chrono::microseconds sleepTime(1);
MiniBatchInfo<AlignmentGroup<FragT*>>* miniBatch = nullptr;
size_t numTranscripts = refs.size();
while (!doneParsing or !workQueue.empty()) {
bool foundWork = workQueue.try_pop(miniBatch);
// If work wasn't immediately available, then wait for it.
if (!foundWork) {
std::unique_lock<std::mutex> l(cvmutex);
workAvailable.wait(l, [&miniBatch, &workQueue, &doneParsing]() { return workQueue.try_pop(miniBatch) or doneParsing; });
}
if (miniBatch != nullptr) {
++activeBatches;
size_t batchReads{0};
std::vector<AlignmentGroup<FragT*>*>& alignmentGroups = *(miniBatch->alignments);
using TranscriptID = size_t;
using HitIDVector = std::vector<size_t>;
using HitProbVector = std::vector<double>;
std::unordered_map<TranscriptID, std::vector<FragT*>> hitList;
// Each alignment group corresponds to all of the potential
// mapping locations of a multi-mapping read
for (auto alnGroup : alignmentGroups) {
// Score all of these alignments and sample according to
// their probabilities
for (auto a : alnGroup->alignments()) {
auto transcriptID = a->transcriptID();
if (transcriptID < 0 or transcriptID >= refs.size()) {
log->warn("Invalid Transcript ID: {}\n", transcriptID);
}
hitList[transcriptID].emplace_back(a);
}
}
{
// Iterate over each group of alignments (a group consists of all alignments reported
// for a single read).
for (auto alnGroup : alignmentGroups) {
double sumOfAlignProbs{LOG_0};
// update the cluster-level properties
bool transcriptUnique{true};
auto firstTranscriptID = alnGroup->alignments().front()->transcriptID();
for (auto& aln : alnGroup->alignments()) {
auto transcriptID = aln->transcriptID();
auto& transcript = refs[transcriptID];
transcriptUnique = transcriptUnique and (transcriptID == firstTranscriptID);
double refLength = transcript.RefLength > 0 ? transcript.RefLength : 1.0;
double logFragProb = salmon::math::LOG_1;
if (!salmonOpts.noFragLengthDist) {
if(aln->fragLen() == 0) {
if (aln->isLeft() and transcript.RefLength - aln->left() < fragLengthDist.maxVal()) {
logFragProb = fragLengthDist.cmf(transcript.RefLength - aln->left());
} else if (aln->isRight() and aln->right() < fragLengthDist.maxVal()) {
logFragProb = fragLengthDist.cmf(aln->right());
}
} else {
logFragProb = fragLengthDist.pmf(static_cast<size_t>(aln->fragLen()));
}
}
// The alignment probability is the product of a
// transcript-level term (based on abundance and) an
// alignment-level term.
double logRefLength{salmon::math::LOG_0};
if (salmonOpts.noEffectiveLengthCorrection or !burnedIn) {
logRefLength = std::log(transcript.RefLength);
} else {
logRefLength = transcript.getCachedLogEffectiveLength();
}
double logAlignCompatProb =
(salmonOpts.useReadCompat) ?
(salmon::utils::logAlignFormatProb(
aln->libFormat(),
expectedLibraryFormat,
aln->pos(),
aln->fwd(), aln->mateStatus(), salmonOpts.incompatPrior)
) : LOG_1;
// Adjustment to the likelihood due to the
// error model
double errLike = salmon::math::LOG_1;
if (burnedIn and salmonOpts.useErrorModel) {
errLike = alnMod.logLikelihood(*aln, transcript);
}
// Allow for a non-uniform fragment start position distribution
double startPosProb = -logRefLength;
auto hitPos = aln->left();
if (useFSPD and burnedIn and hitPos < refLength) {
auto& fragStartDist =
fragStartDists[transcript.lengthClassIndex()];
startPosProb = fragStartDist(hitPos, refLength, logRefLength);
}
double auxProb = startPosProb + logFragProb +
aln->logQualProb() +
errLike + logAlignCompatProb;
double transcriptLogCount = transcript.mass(false);
if ( transcriptLogCount != LOG_0 and
auxProb != LOG_0 ) {
aln->logProb = transcriptLogCount + auxProb;
sumOfAlignProbs = logAdd(sumOfAlignProbs, aln->logProb);
} else {
aln->logProb = LOG_0;
}
}
// normalize the hits
if (sumOfAlignProbs == LOG_0) {
auto aln = alnGroup->alignments().front();
log->warn("0 probability fragment [{}] "
"encountered\n", aln->getName());
continue;
}
if (transcriptUnique) {
// avoid r-value ref until we figure out what's
// up with TBB 4.3
auto* alnPtr = alnGroup->alignments().front()->clone();
outputQueue.push(alnPtr);
} else { // read maps to multiple transcripts
double r = uni(eng);
double currentMass{0.0};
double massInc{0.0};
bool choseAlignment{false};
for (auto& aln : alnGroup->alignments()) {
if (aln->logProb == LOG_0) { continue; }
aln->logProb -= sumOfAlignProbs;
massInc = std::exp(aln->logProb);
if (currentMass <= r and currentMass + massInc > r) {
// Write out this read
// avoid r-value ref until we figure out what's
// up with TBB 4.3
auto* alnPtr = aln->clone();
outputQueue.push(alnPtr);
currentMass += massInc;
choseAlignment = true;
break;
}
currentMass += massInc;
} // end alignment group
if (BOOST_UNLIKELY(!choseAlignment)) {
log->warn("[Sampler.hpp]: Failed to sample an alignment for this read; "
"this shouldn't happen\n"
"currentMass = {}, r = {}\n", currentMass, r);
}
} // non-unique read
++batchReads;
} // end read group
}// end timer
miniBatch->release(fragmentQueue, alignmentGroupQueue);
delete miniBatch;
--activeBatches;
processedReads += batchReads;
}
miniBatch = nullptr;
} // nothing left to process
}
/**
* Sample the alignments in the provided library given in current
* estimates of transcript abundance.
*/
template <typename FragT>
bool sampleLibrary(
AlignmentLibrary<FragT>& alnLib,
const SalmonOpts& salmonOpts,
bool burnedIn,
bfs::path& sampleFilePath,
bool sampleUnaligned) {
fmt::MemoryWriter msgStr;
auto log = spdlog::get("jointLog");
msgStr << "Sampling alignments; outputting results to "
<< sampleFilePath.string() << "\n";
log->info(msgStr.str());
auto& refs = alnLib.transcripts();
size_t numTranscripts = refs.size();
size_t miniBatchSize{1000};
size_t numObservedFragments{0};
MiniBatchQueue<AlignmentGroup<FragT*>> workQueue;
double logForgettingMass{std::log(1.0)};
double forgettingFactor{0.60};
size_t batchNum{0};
/**
* Output queue
*/
volatile bool consumedAllInput{false};
size_t defaultCapacity = 2000000;
OutputQueue<FragT> outQueue;
outQueue.set_capacity(defaultCapacity);
std::unique_ptr<OutputUnmappedFilter<FragT>> outFilt = nullptr;
if (sampleUnaligned) {
outFilt.reset(new OutputUnmappedFilter<FragT>(&outQueue));
}
// Reset our reader to the beginning
if (!alnLib.reset(false, outFilt.get())) {
fmt::print(stderr,
"\n\n======== WARNING ========\n"
"A provided alignment file "
"is not a regular file and therefore can't be read from "
"more than once.\n\n"
"Therefore, we cannot provide sample output alignments "
"from this file. No sampled BAM file will be generated. "
"Please consider re-running Salmon with these alignments "
"as a regular file!\n"
"==========================\n\n");
return false;
}
volatile bool doneParsing{false};
std::condition_variable workAvailable;
std::mutex cvmutex;
std::vector<std::thread> workers;
std::atomic<size_t> activeBatches{0};
std::atomic<size_t> processedReads{0};
size_t numProc{0};
for (uint32_t i = 0; i < salmonOpts.numQuantThreads; ++i) {
workers.emplace_back(sampleMiniBatch<FragT>,
std::ref(alnLib),
std::ref(workQueue),
std::ref(workAvailable), std::ref(cvmutex),
std::ref(doneParsing), std::ref(activeBatches),
std::ref(salmonOpts),
std::ref(burnedIn),
std::ref(processedReads),
std::ref(outQueue));
}
std::thread outputThread(
[&consumedAllInput, &alnLib, &outQueue, &log, sampleFilePath] () -> void {
scram_fd* bf = scram_open(sampleFilePath.c_str(), "wb");
scram_set_option(bf, CRAM_OPT_NTHREADS, 3);
scram_set_header(bf, alnLib.header());
scram_write_header(bf);
if (bf == nullptr) {
fmt::MemoryWriter errstr;
errstr << ioutils::SET_RED << "ERROR: "
<< ioutils::RESET_COLOR
<< "Couldn't open output bam file "
<< sampleFilePath.string() << ". Exiting\n";
log->warn(errstr.str());
std::exit(-1);
}
FragT* aln{nullptr};
while (!outQueue.empty() or !consumedAllInput) {
while (outQueue.try_pop(aln)) {
if (aln != nullptr) {
int ret = aln->writeToFile(bf);
if (ret != 0) {
std::cerr << "ret = " << ret << "\n";
fmt::MemoryWriter errstr;
errstr << ioutils::SET_RED << "ERROR:"
<< ioutils::RESET_COLOR << "Could not write "
<< "a sampled alignment to the output BAM "
<< "file. Please check that the file can "
<< "be created properly and that the disk "
<< "is not full. Exiting.\n";
log->warn(errstr.str());
std::exit(-1);
}
// Eventually, as we do in BAMQueue, we should
// have queue of bam1_t structures that can be
// re-used rather than continually calling
// new and delete.
delete aln;
aln = nullptr;
}
}
}
scram_close(bf); // will delete the header itself
});
BAMQueue<FragT>& bq = alnLib.getAlignmentGroupQueue();
std::vector<AlignmentGroup<FragT*>*>* alignments = new std::vector<AlignmentGroup<FragT*>*>;
alignments->reserve(miniBatchSize);
AlignmentGroup<FragT*>* ag;
bool alignmentGroupsRemain = bq.getAlignmentGroup(ag);
while (alignmentGroupsRemain or alignments->size() > 0) {
if (alignmentGroupsRemain) { alignments->push_back(ag); }
// If this minibatch has reached the size limit, or we have nothing
// left to fill it up with
if (alignments->size() >= miniBatchSize or !alignmentGroupsRemain) {
// Don't need to update the batch number or log forgetting mass in this phase
MiniBatchInfo<AlignmentGroup<FragT*>>* mbi =
new MiniBatchInfo<AlignmentGroup<FragT*>>(batchNum, alignments, logForgettingMass);
workQueue.push(mbi);
{
std::unique_lock<std::mutex> l(cvmutex);
workAvailable.notify_one();
}
alignments = new std::vector<AlignmentGroup<FragT*>*>;
alignments->reserve(miniBatchSize);
}
if (numProc % 1000000 == 0) {
const char RESET_COLOR[] = "\x1b[0m";
char green[] = "\x1b[30m";
green[3] = '0' + static_cast<char>(fmt::GREEN);
char red[] = "\x1b[30m";
red[3] = '0' + static_cast<char>(fmt::RED);
fmt::print(stderr, "\r\r{}processed{} {} {}reads{}", green, red, numProc, green, RESET_COLOR);
}
++numProc;
alignmentGroupsRemain = bq.getAlignmentGroup(ag);
}
std::cerr << "\n";
// Free the alignments and the vector holding them
for (auto& aln : *alignments) {
aln->alignments().clear();
delete aln; aln = nullptr;
}
delete alignments;
doneParsing = true;
/**
* This could be a problem for small sets of alignments --- make sure the
* work queue is empty!!
* --- Thanks for finding a dataset that exposes this bug, Richard (Smith-Unna)!
*/
size_t t = 0;
while (!workQueue.empty()) {
std::unique_lock<std::mutex> l(cvmutex);
workAvailable.notify_one();
}
size_t tnum{0};
for (auto& t : workers) {
fmt::print(stderr, "killing thread {} . . . ", tnum++);
{
std::unique_lock<std::mutex> l(cvmutex);
workAvailable.notify_all();
}
t.join();
fmt::print(stderr, "done\r\r");
}
fmt::print(stderr, "\n");
consumedAllInput = true;
numObservedFragments += alnLib.numMappedFragments();
fmt::print(stderr, "# observed = {} mapped fragments.\033[F\033[F\033[F\033[F",
numObservedFragments);
fmt::print(stderr, "Waiting on output thread\n");
outputThread.join();
fmt::print(stderr, "done\n");
fmt::print(stderr, "\n\n\n\n");
return true;
}
} // namespace sampler
} // namespace salmon
#endif // __SAMPLER__HPP__
|
