-
Notifications
You must be signed in to change notification settings - Fork 66
/
SingleAligner.cpp
411 lines (350 loc) · 13.5 KB
/
SingleAligner.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
404
405
406
407
408
409
410
411
/*++
Module Name:
SingleAligner.cpp
Abstract:
Functions for running the single end aligner sub-program.
Authors:
Matei Zaharia, February, 2012
Environment:
User mode service.
Revision History:
Adapted from cSNAP, which was in turn adapted from the scala prototype
--*/
#include "stdafx.h"
#include "options.h"
#include "BaseAligner.h"
#include "Compat.h"
#include "RangeSplitter.h"
#include "GenomeIndex.h"
#include "SAM.h"
#include "Tables.h"
#include "AlignerContext.h"
#include "AlignerOptions.h"
#include "FASTQ.h"
#include "Util.h"
#include "SingleAligner.h"
#include "MultiInputReadSupplier.h"
using namespace std;
using util::stringEndsWith;
SingleAlignerContext::SingleAlignerContext(AlignerExtension* i_extension)
: AlignerContext(0, NULL, NULL, i_extension)
{
}
AlignerStats*
SingleAlignerContext::newStats()
{
return new AlignerStats();
}
void
SingleAlignerContext::runTask()
{
ParallelTask<SingleAlignerContext> task(this);
task.run();
}
void SingleAlignerContext::runIterationThread()
{
Read * read = NULL;
#ifdef _MSC_VER
__try {
#endif // _MSC_VER
runIterationThreadImpl(read);
#ifdef _MSC_VER
} __except (EXCEPTION_EXECUTE_HANDLER) {
if (read == NULL) {
fprintf(stderr, "SNAP crashed before processing a read\n");
} else {
fprintf(stderr, "SNAP crashed while processing single-end read with ID %.*s\n", read->getIdLength(), read->getId());
fprintf(stderr, "@%.*s\n%.*s\n+\n%.*s\n", read->getIdLength(), read->getId(), read->getDataLength(), read->getData(),
read->getDataLength(), read->getQuality());
}
fflush(stderr);
soft_exit(1);
}
#endif // _MSC_VER
}
void
SingleAlignerContext::runIterationThreadImpl(Read *& read)
{
PreventMachineHibernationWhileThisThreadIsAlive();
ReadSupplier *supplier = readSupplierGenerator->generateNewReadSupplier();
if (NULL == supplier) {
//
// No work for this thread to do.
//
return;
}
if (extension->runIterationThread(supplier, this)) {
delete supplier;
return;
}
if (index == NULL) {
// no alignment, just input/output
while (NULL != (read = supplier->getNextRead())) {
stats->totalReads++;
SingleAlignmentResult result;
result.status = NotFound;
result.direction = FORWARD;
result.mapq = 0;
result.score = 0;
result.location = InvalidGenomeLocation;
result.usedAffineGapScoring = false;
result.basesClippedBefore = 0;
result.basesClippedAfter = 0;
if (options->passFilter(read, NotFound, read->getDataLength() < minReadLength || read->countOfNs() > maxDist, false)) {
stats->notFound++;
if (NULL != readWriter) {
readWriter->writeReads(readerContext, read, &result, 1, true);
}
} else {
stats->filtered++;
}
extension->writeRead(read, &result);
}
delete supplier;
return;
}
int maxReadSize = MAX_READ_LENGTH;
SingleAlignmentResult *alignmentResults = NULL;
bool alignmentResultsReallocated = false;
_int64 alignmentResultBufferCount;
if (maxSecondaryAlignmentAdditionalEditDistance < 0) {
alignmentResultBufferCount = 1;
} else {
alignmentResultBufferCount = 32; // Just a nice number that's not too small. We reallocate on demand.
}
size_t alignmentResultBufferSize = sizeof(*alignmentResults) * alignmentResultBufferCount;
BigAllocator *allocator = new BigAllocator(BaseAligner::getBigAllocatorReservation(index, true, maxHits, maxReadSize, index->getSeedLength(), numSeedsFromCommandLine, seedCoverage, maxSecondaryAlignmentsPerContig, extraSearchDepth)
+ alignmentResultBufferSize, 16); // FIXME: Used larger allocation granularity for __m128i that needs to be aligned at 16 byte boundaries
BaseAligner *aligner = new (allocator) BaseAligner(
index,
maxHits,
maxDist,
maxReadSize,
numSeedsFromCommandLine,
seedCoverage,
minWeightToCheck,
extraSearchDepth,
noUkkonen,
noOrderedEvaluation,
noTruncation,
useAffineGap,
ignoreAlignmentAdjustmentForOm,
altAwareness,
emitALTAlignments,
maxScoreGapToPreferNonALTAlignment,
maxSecondaryAlignmentsPerContig,
NULL, // LV (no need to cache in the single aligner)
NULL, // reverse LV
matchReward,
subPenalty,
gapOpenPenalty,
gapExtendPenalty,
fivePrimeEndBonus,
threePrimeEndBonus,
stats,
allocator);
alignmentResults = (SingleAlignmentResult *)allocator->allocate(alignmentResultBufferSize);
allocator->checkCanaries();
aligner->setExplorePopularSeeds(options->explorePopularSeeds);
aligner->setStopOnFirstHit(options->stopOnFirstHit);
#ifdef _MSC_VER
if (options->useTimingBarrier) {
if (0 == InterlockedDecrementAndReturnNewValue(nThreadsAllocatingMemory)) {
AllowEventWaitersToProceed(memoryAllocationCompleteBarrier);
} else {
WaitForEvent(memoryAllocationCompleteBarrier);
}
}
#endif // _MSC_VER
// Align the reads.
_uint64 lastReportTime = timeInMillis();
_uint64 readsWhenLastReported = 0;
_int64 startTime = timeInMillis();
while (NULL != (read = supplier->getNextRead())) {
_int64 readFinishedTime;
if (options->profile) {
readFinishedTime = timeInMillis();
stats->millisReading += (readFinishedTime - startTime);
}
stats->totalReads++;
if (AlignerOptions::useHadoopErrorMessages && stats->totalReads % 10000 == 0 && timeInMillis() - lastReportTime > 10000) {
fprintf(stderr,"reporter:counter:SNAP,readsAligned,%llu\n",stats->totalReads - readsWhenLastReported);
readsWhenLastReported = stats->totalReads;
lastReportTime = timeInMillis();
}
// Skip the read if it has too many Ns or trailing 2 quality scores.
if (read->getDataLength() < minReadLength || read->countOfNs() > maxDist) {
if (!options->passFilter(read, NotFound, true, false)) {
stats->filtered++;
} else {
if (NULL != readWriter) {
SingleAlignmentResult result;
result.status = NotFound;
result.location = InvalidGenomeLocation;
result.mapq = 0;
result.direction = FORWARD;
result.clippingForReadAdjustment = 0;
result.usedAffineGapScoring = false;
result.basesClippedBefore = 0;
result.basesClippedAfter = 0;
result.supplementary = false;
readWriter->writeReads(readerContext, read, &result, 1, true, useAffineGap);
}
stats->uselessReads++;
}
continue;
}
#if TIME_HISTOGRAM
_int64 startTime = timeInNanos();
#endif // TIME_HISTOGRAM
_int64 nSecondaryResults = 0;
#ifdef LONG_READS
int oldMaxK = aligner->getMaxK();
if (options->maxDistFraction > 0.0) {
aligner->setMaxK(min(MAX_K, (int)(read->getDataLength() * options->maxDistFraction)));
}
#endif
SingleAlignmentResult firstALTResult;
while (!aligner->AlignRead(read, alignmentResults, &firstALTResult, maxSecondaryAlignmentAdditionalEditDistance, alignmentResultBufferCount - 1, &nSecondaryResults, maxSecondaryAlignments, alignmentResults + 1, 0, NULL, NULL, useSoftClipping)) {
//
// Out of secondary alignment buffer. Reallocate.
//
if (alignmentResultsReallocated) {
BigDealloc(alignmentResults);
alignmentResults = NULL;
}
alignmentResultBufferCount *= 2;
alignmentResultBufferSize = alignmentResultBufferCount * sizeof(SingleAlignmentResult);
alignmentResults = (SingleAlignmentResult *)BigAlloc(alignmentResultBufferSize);
alignmentResultsReallocated = true;
}
#ifdef LONG_READS
aligner->setMaxK(oldMaxK);
#endif
_int64 alignFinishedTime;
if (options->profile) {
alignFinishedTime = timeInMillis();
stats->millisAligning += (alignFinishedTime - readFinishedTime);
}
#if TIME_HISTOGRAM
_int64 runTime = timeInNanos() - startTime;
int timeBucket = min(30, cheezyLogBase2(runTime));
stats->countByTimeBucket[timeBucket]++;
stats->nanosByTimeBucket[timeBucket] += runTime;
#endif // TIME_HISTOGRAM
allocator->checkCanaries();
bool containsPrimary = true;
if (NULL != readWriter) {
//
// Remove any reads that don't pass the filter, then send the remainder down to the writer.
//
for (int i = 0; i <= nSecondaryResults; i++) {
if (!options->passFilter(read, alignmentResults[i].status, false, i != 0 || !containsPrimary)) {
if (i == 0) {
containsPrimary = false;
}
//
// Copy the last result here.
//
alignmentResults[i] = alignmentResults[nSecondaryResults];
nSecondaryResults--;
//
// And back up so it gets checked.
//
i--;
}
} // For each result
stats->extraAlignments += nSecondaryResults + (containsPrimary ? 0 : 1); // If it doesn't contain the primary, then it's a secondary.
readWriter->writeReads(readerContext, read, alignmentResults, nSecondaryResults + 1, containsPrimary, useAffineGap);
if (altAwareness && firstALTResult.status != NotFound && options->passFilter(read, firstALTResult.status, false, false)) {
readWriter->writeReads(readerContext, read, &firstALTResult, 1, false, useAffineGap);
}
} // If we're writing reads at all
if (options->profile) {
startTime = timeInMillis();
stats->millisWriting = (startTime - alignFinishedTime);
}
if (containsPrimary) {
updateStats(stats, read, alignmentResults[0].status, alignmentResults[0].score, alignmentResults[0].mapq);
} else {
stats->filtered++;
}
}
aligner->~BaseAligner(); // This calls the destructor without calling operator delete, allocator owns the memory.
if (supplier != NULL) {
delete supplier;
}
if (alignmentResultsReallocated) {
BigDealloc(alignmentResults);
}
delete allocator; // This is what actually frees the memory.
}
void
SingleAlignerContext::updateStats(
AlignerStats* stats,
Read* read,
AlignmentResult result,
int score,
int mapq)
{
if (isOneLocation(result)) {
stats->singleHits++;
} else if (result == MultipleHits) {
stats->multiHits++;
} else {
_ASSERT(result == NotFound);
stats->notFound++;
}
if (result != NotFound) {
_ASSERT(mapq >= 0 && mapq <= AlignerStats::maxMapq);
stats->mapqHistogram[mapq]++;
}
}
void
SingleAlignerContext::typeSpecificBeginIteration()
{
if (1 == options->nInputs) {
//
// We've only got one input, so just connect it directly to the consumer.
//
readSupplierGenerator = options->inputs[0].createReadSupplierGenerator(options->numThreads, readerContext);
} else {
//
// We've got multiple inputs, so use a MultiInputReadSupplier to combine the individual inputs.
//
ReadSupplierGenerator **generators = new ReadSupplierGenerator *[options->nInputs];
// use separate context for each supplier, initialized from common
for (int i = 0; i < options->nInputs; i++) {
ReaderContext context(readerContext);
generators[i] = options->inputs[i].createReadSupplierGenerator(options->numThreads, context);
}
readSupplierGenerator = new MultiInputReadSupplierGenerator(options->nInputs,generators);
}
ReaderContext* context = readSupplierGenerator->getContext();
readerContext.header = context->header;
readerContext.headerBytes = context->headerBytes;
readerContext.headerLength = context->headerLength;
readerContext.headerMatchesIndex = context->headerMatchesIndex;
readerContext.numRGLines = context->numRGLines;
readerContext.rgLines = context->rgLines;
readerContext.rgLineOffsets = context->rgLineOffsets;
}
void
SingleAlignerContext::typeSpecificNextIteration()
{
if (readerContext.header != NULL) {
delete [] readerContext.header;
readerContext.header = NULL;
readerContext.headerLength = readerContext.headerBytes = 0;
readerContext.headerMatchesIndex = false;
}
if (readerContext.rgLines != NULL) {
delete[] readerContext.rgLines;
delete[] readerContext.rgLineOffsets;
readerContext.numRGLines = 0;
readerContext.rgLines = NULL;
readerContext.rgLineOffsets = NULL;
}
delete readSupplierGenerator;
readSupplierGenerator = NULL;
}