/
HaplotypeCallerGenotypingEngine.java
812 lines (706 loc) · 49.8 KB
/
HaplotypeCallerGenotypingEngine.java
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
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
/*
* By downloading the PROGRAM you agree to the following terms of use:
*
* BROAD INSTITUTE
* SOFTWARE LICENSE AGREEMENT
* FOR ACADEMIC NON-COMMERCIAL RESEARCH PURPOSES ONLY
*
* This Agreement is made between the Broad Institute, Inc. with a principal address at 415 Main Street, Cambridge, MA 02142 (“BROAD”) and the LICENSEE and is effective at the date the downloading is completed (“EFFECTIVE DATE”).
*
* WHEREAS, LICENSEE desires to license the PROGRAM, as defined hereinafter, and BROAD wishes to have this PROGRAM utilized in the public interest, subject only to the royalty-free, nonexclusive, nontransferable license rights of the United States Government pursuant to 48 CFR 52.227-14; and
* WHEREAS, LICENSEE desires to license the PROGRAM and BROAD desires to grant a license on the following terms and conditions.
* NOW, THEREFORE, in consideration of the promises and covenants made herein, the parties hereto agree as follows:
*
* 1. DEFINITIONS
* 1.1 PROGRAM shall mean copyright in the object code and source code known as GATK3 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute.org/gatk on the EFFECTIVE DATE.
*
* 2. LICENSE
* 2.1 Grant. Subject to the terms of this Agreement, BROAD hereby grants to LICENSEE, solely for academic non-commercial research purposes, a non-exclusive, non-transferable license to: (a) download, execute and display the PROGRAM and (b) create bug fixes and modify the PROGRAM. LICENSEE hereby automatically grants to BROAD a non-exclusive, royalty-free, irrevocable license to any LICENSEE bug fixes or modifications to the PROGRAM with unlimited rights to sublicense and/or distribute. LICENSEE agrees to provide any such modifications and bug fixes to BROAD promptly upon their creation.
* The LICENSEE may apply the PROGRAM in a pipeline to data owned by users other than the LICENSEE and provide these users the results of the PROGRAM provided LICENSEE does so for academic non-commercial purposes only. For clarification purposes, academic sponsored research is not a commercial use under the terms of this Agreement.
* 2.2 No Sublicensing or Additional Rights. LICENSEE shall not sublicense or distribute the PROGRAM, in whole or in part, without prior written permission from BROAD. LICENSEE shall ensure that all of its users agree to the terms of this Agreement. LICENSEE further agrees that it shall not put the PROGRAM on a network, server, or other similar technology that may be accessed by anyone other than the LICENSEE and its employees and users who have agreed to the terms of this agreement.
* 2.3 License Limitations. Nothing in this Agreement shall be construed to confer any rights upon LICENSEE by implication, estoppel, or otherwise to any computer software, trademark, intellectual property, or patent rights of BROAD, or of any other entity, except as expressly granted herein. LICENSEE agrees that the PROGRAM, in whole or part, shall not be used for any commercial purpose, including without limitation, as the basis of a commercial software or hardware product or to provide services. LICENSEE further agrees that the PROGRAM shall not be copied or otherwise adapted in order to circumvent the need for obtaining a license for use of the PROGRAM.
*
* 3. PHONE-HOME FEATURE
* LICENSEE expressly acknowledges that the PROGRAM contains an embedded automatic reporting system (“PHONE-HOME”) which is enabled by default upon download. Unless LICENSEE requests disablement of PHONE-HOME, LICENSEE agrees that BROAD may collect limited information transmitted by PHONE-HOME regarding LICENSEE and its use of the PROGRAM. Such information shall include LICENSEE’S user identification, version number of the PROGRAM and tools being run, mode of analysis employed, and any error reports generated during run-time. Collection of such information is used by BROAD solely to monitor usage rates, fulfill reporting requirements to BROAD funding agencies, drive improvements to the PROGRAM, and facilitate adjustments to PROGRAM-related documentation.
*
* 4. OWNERSHIP OF INTELLECTUAL PROPERTY
* LICENSEE acknowledges that title to the PROGRAM shall remain with BROAD. The PROGRAM is marked with the following BROAD copyright notice and notice of attribution to contributors. LICENSEE shall retain such notice on all copies. LICENSEE agrees to include appropriate attribution if any results obtained from use of the PROGRAM are included in any publication.
* Copyright 2012-2014 Broad Institute, Inc.
* Notice of attribution: The GATK3 program was made available through the generosity of Medical and Population Genetics program at the Broad Institute, Inc.
* LICENSEE shall not use any trademark or trade name of BROAD, or any variation, adaptation, or abbreviation, of such marks or trade names, or any names of officers, faculty, students, employees, or agents of BROAD except as states above for attribution purposes.
*
* 5. INDEMNIFICATION
* LICENSEE shall indemnify, defend, and hold harmless BROAD, and their respective officers, faculty, students, employees, associated investigators and agents, and their respective successors, heirs and assigns, (Indemnitees), against any liability, damage, loss, or expense (including reasonable attorneys fees and expenses) incurred by or imposed upon any of the Indemnitees in connection with any claims, suits, actions, demands or judgments arising out of any theory of liability (including, without limitation, actions in the form of tort, warranty, or strict liability and regardless of whether such action has any factual basis) pursuant to any right or license granted under this Agreement.
*
* 6. NO REPRESENTATIONS OR WARRANTIES
* THE PROGRAM IS DELIVERED AS IS. BROAD MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND CONCERNING THE PROGRAM OR THE COPYRIGHT, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF LATENT OR OTHER DEFECTS, WHETHER OR NOT DISCOVERABLE. BROAD EXTENDS NO WARRANTIES OF ANY KIND AS TO PROGRAM CONFORMITY WITH WHATEVER USER MANUALS OR OTHER LITERATURE MAY BE ISSUED FROM TIME TO TIME.
* IN NO EVENT SHALL BROAD OR ITS RESPECTIVE DIRECTORS, OFFICERS, EMPLOYEES, AFFILIATED INVESTIGATORS AND AFFILIATES BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND, INCLUDING, WITHOUT LIMITATION, ECONOMIC DAMAGES OR INJURY TO PROPERTY AND LOST PROFITS, REGARDLESS OF WHETHER BROAD SHALL BE ADVISED, SHALL HAVE OTHER REASON TO KNOW, OR IN FACT SHALL KNOW OF THE POSSIBILITY OF THE FOREGOING.
*
* 7. ASSIGNMENT
* This Agreement is personal to LICENSEE and any rights or obligations assigned by LICENSEE without the prior written consent of BROAD shall be null and void.
*
* 8. MISCELLANEOUS
* 8.1 Export Control. LICENSEE gives assurance that it will comply with all United States export control laws and regulations controlling the export of the PROGRAM, including, without limitation, all Export Administration Regulations of the United States Department of Commerce. Among other things, these laws and regulations prohibit, or require a license for, the export of certain types of software to specified countries.
* 8.2 Termination. LICENSEE shall have the right to terminate this Agreement for any reason upon prior written notice to BROAD. If LICENSEE breaches any provision hereunder, and fails to cure such breach within thirty (30) days, BROAD may terminate this Agreement immediately. Upon termination, LICENSEE shall provide BROAD with written assurance that the original and all copies of the PROGRAM have been destroyed, except that, upon prior written authorization from BROAD, LICENSEE may retain a copy for archive purposes.
* 8.3 Survival. The following provisions shall survive the expiration or termination of this Agreement: Articles 1, 3, 4, 5 and Sections 2.2, 2.3, 7.3, and 7.4.
* 8.4 Notice. Any notices under this Agreement shall be in writing, shall specifically refer to this Agreement, and shall be sent by hand, recognized national overnight courier, confirmed facsimile transmission, confirmed electronic mail, or registered or certified mail, postage prepaid, return receipt requested. All notices under this Agreement shall be deemed effective upon receipt.
* 8.5 Amendment and Waiver; Entire Agreement. This Agreement may be amended, supplemented, or otherwise modified only by means of a written instrument signed by all parties. Any waiver of any rights or failure to act in a specific instance shall relate only to such instance and shall not be construed as an agreement to waive any rights or fail to act in any other instance, whether or not similar. This Agreement constitutes the entire agreement among the parties with respect to its subject matter and supersedes prior agreements or understandings between the parties relating to its subject matter.
* 8.6 Binding Effect; Headings. This Agreement shall be binding upon and inure to the benefit of the parties and their respective permitted successors and assigns. All headings are for convenience only and shall not affect the meaning of any provision of this Agreement.
* 8.7 Governing Law. This Agreement shall be construed, governed, interpreted and applied in accordance with the internal laws of the Commonwealth of Massachusetts, U.S.A., without regard to conflict of laws principles.
*/
package org.broadinstitute.gatk.tools.walkers.haplotypecaller;
import com.google.java.contract.Ensures;
import com.google.java.contract.Requires;
import htsjdk.variant.variantcontext.*;
import org.broadinstitute.gatk.utils.contexts.ReferenceContext;
import org.broadinstitute.gatk.utils.genotyper.AlleleList;
import org.broadinstitute.gatk.utils.genotyper.IndexedAlleleList;
import org.broadinstitute.gatk.utils.genotyper.SampleList;
import org.broadinstitute.gatk.utils.refdata.RefMetaDataTracker;
import org.broadinstitute.gatk.tools.walkers.genotyper.*;
import org.broadinstitute.gatk.tools.walkers.genotyper.afcalc.AFCalculatorProvider;
import org.broadinstitute.gatk.utils.GenomeLoc;
import org.broadinstitute.gatk.utils.GenomeLocParser;
import org.broadinstitute.gatk.utils.Utils;
import org.broadinstitute.gatk.utils.collections.Pair;
import org.broadinstitute.gatk.utils.genotyper.ReadLikelihoods;
import org.broadinstitute.gatk.utils.haplotype.EventMap;
import org.broadinstitute.gatk.utils.haplotype.Haplotype;
import org.broadinstitute.gatk.utils.sam.GATKSAMRecord;
import org.broadinstitute.gatk.utils.variant.GATKVCFConstants;
import org.broadinstitute.gatk.utils.variant.GATKVariantContextUtils;
import java.util.*;
/**
* {@link HaplotypeCaller}'s genotyping strategy implementation.
*/
public class HaplotypeCallerGenotypingEngine extends GenotypingEngine<AssemblyBasedCallerArgumentCollection> {
protected static final int ALLELE_EXTENSION = 2;
private static final String phase01 = "0|1";
private static final String phase10 = "1|0";
private MergeVariantsAcrossHaplotypes crossHaplotypeEventMerger;
protected final boolean doPhysicalPhasing;
private final GenotypingModel genotypingModel;
private final PloidyModel ploidyModel;
/**
* {@inheritDoc}
* @param configuration {@inheritDoc}
* @param samples {@inheritDoc}
* @param genomeLocParser {@inheritDoc}
* @param doPhysicalPhasing whether to try physical phasing.
*/
public HaplotypeCallerGenotypingEngine(final AssemblyBasedCallerArgumentCollection configuration, final SampleList samples, final GenomeLocParser genomeLocParser, final AFCalculatorProvider afCalculatorProvider, final boolean doPhysicalPhasing) {
super(configuration,samples,genomeLocParser,afCalculatorProvider);
if (genomeLocParser == null)
throw new IllegalArgumentException("the genome location parser provided cannot be null");
this.doPhysicalPhasing= doPhysicalPhasing;
ploidyModel = new HomogeneousPloidyModel(samples,configuration.genotypeArgs.samplePloidy);
genotypingModel = new InfiniteRandomMatingPopulationModel();
}
/**
* Change the merge variant across haplotypes for this engine.
*
* @param crossHaplotypeEventMerger new merger, can be {@code null}.
*/
public void setCrossHaplotypeEventMerger(final MergeVariantsAcrossHaplotypes crossHaplotypeEventMerger) {
this.crossHaplotypeEventMerger = crossHaplotypeEventMerger;
}
@Override
protected String callSourceString() {
return "HC_call";
}
@Override
protected boolean forceKeepAllele(final Allele allele) {
return allele == GATKVCFConstants.NON_REF_SYMBOLIC_ALLELE ||
configuration.genotypingOutputMode == GenotypingOutputMode.GENOTYPE_GIVEN_ALLELES ||
configuration.emitReferenceConfidence != ReferenceConfidenceMode.NONE;
}
@Override
protected boolean forceSiteEmission() {
return configuration.outputMode == OutputMode.EMIT_ALL_SITES || configuration.genotypingOutputMode == GenotypingOutputMode.GENOTYPE_GIVEN_ALLELES;
}
/**
* Carries the result of a call to #assignGenotypeLikelihoods
*/
public static class CalledHaplotypes {
private final List<VariantContext> calls;
private final Set<Haplotype> calledHaplotypes;
public CalledHaplotypes(final List<VariantContext> calls, final Set<Haplotype> calledHaplotypes) {
if ( calls == null ) throw new IllegalArgumentException("calls cannot be null");
if ( calledHaplotypes == null ) throw new IllegalArgumentException("calledHaplotypes cannot be null");
if ( Utils.xor(calls.isEmpty(), calledHaplotypes.isEmpty()) )
throw new IllegalArgumentException("Calls and calledHaplotypes should both be empty or both not but got calls=" + calls + " calledHaplotypes=" + calledHaplotypes);
this.calls = calls;
this.calledHaplotypes = calledHaplotypes;
}
/**
* Get the list of calls made at this location
* @return a non-null (but potentially empty) list of calls
*/
public List<VariantContext> getCalls() {
return calls;
}
/**
* Get the set of haplotypes that we actually called (i.e., underlying one of the VCs in getCalls().
* @return a non-null set of haplotypes
*/
public Set<Haplotype> getCalledHaplotypes() {
return calledHaplotypes;
}
}
/**
* Main entry point of class - given a particular set of haplotypes, samples and reference context, compute
* genotype likelihoods and assemble into a list of variant contexts and genomic events ready for calling
*
* The list of samples we're working with is obtained from the readLikelihoods
*
* @param haplotypes Haplotypes to assign likelihoods to
* @param readLikelihoods Map from reads->(haplotypes,likelihoods)
* @param perSampleFilteredReadList Map from sample to reads that were filtered after assembly and before calculating per-read likelihoods.
* @param ref Reference bytes at active region
* @param refLoc Corresponding active region genome location
* @param activeRegionWindow Active window
* @param genomeLocParser GenomeLocParser
* @param activeAllelesToGenotype Alleles to genotype
* @param emitReferenceConfidence whether we should add a <NON_REF> alternative allele to the result variation contexts.
*
* @return A CalledHaplotypes object containing a list of VC's with genotyped events and called haplotypes
*
*/
@Requires({"refLoc.containsP(activeRegionWindow)", "haplotypes.size() > 0"})
@Ensures("result != null")
// TODO - can this be refactored? this is hard to follow!
public CalledHaplotypes assignGenotypeLikelihoods( final List<Haplotype> haplotypes,
final ReadLikelihoods<Haplotype> readLikelihoods,
final Map<String, List<GATKSAMRecord>> perSampleFilteredReadList,
final byte[] ref,
final GenomeLoc refLoc,
final GenomeLoc activeRegionWindow,
final GenomeLocParser genomeLocParser,
final RefMetaDataTracker tracker,
final List<VariantContext> activeAllelesToGenotype,
final boolean emitReferenceConfidence) {
// sanity check input arguments
if (haplotypes == null || haplotypes.isEmpty()) throw new IllegalArgumentException("haplotypes input should be non-empty and non-null, got "+haplotypes);
if (readLikelihoods == null || readLikelihoods.sampleCount() == 0) throw new IllegalArgumentException("readLikelihoods input should be non-empty and non-null, got "+readLikelihoods);
if (ref == null || ref.length == 0 ) throw new IllegalArgumentException("ref bytes input should be non-empty and non-null, got " + Arrays.toString(ref));
if (refLoc == null || refLoc.size() != ref.length) throw new IllegalArgumentException(" refLoc must be non-null and length must match ref bytes, got "+refLoc);
if (activeRegionWindow == null ) throw new IllegalArgumentException("activeRegionWindow must be non-null");
if (activeAllelesToGenotype == null ) throw new IllegalArgumentException("activeAllelesToGenotype must be non-null");
if (genomeLocParser == null ) throw new IllegalArgumentException("genomeLocParser must be non-null");
// update the haplotypes so we're ready to call, getting the ordered list of positions on the reference
// that carry events among the haplotypes
final TreeSet<Integer> startPosKeySet = decomposeHaplotypesIntoVariantContexts(haplotypes, readLikelihoods, ref, refLoc, activeAllelesToGenotype);
// Walk along each position in the key set and create each event to be outputted
final Set<Haplotype> calledHaplotypes = new HashSet<>();
final List<VariantContext> returnCalls = new ArrayList<>();
final int ploidy = configuration.genotypeArgs.samplePloidy;
final List<Allele> noCallAlleles = GATKVariantContextUtils.noCallAlleles(ploidy);
for( final int loc : startPosKeySet ) {
if( loc >= activeRegionWindow.getStart() && loc <= activeRegionWindow.getStop() ) { // genotyping an event inside this active region
final List<VariantContext> eventsAtThisLoc = getVCsAtThisLocation(haplotypes, loc, activeAllelesToGenotype);
if( eventsAtThisLoc.isEmpty() ) { continue; }
// Create the event mapping object which maps the original haplotype events to the events present at just this locus
final Map<Event, List<Haplotype>> eventMapper = createEventMapper(loc, eventsAtThisLoc, haplotypes);
// Sanity check the priority list for mistakes
final List<String> priorityList = makePriorityList(eventsAtThisLoc);
// Merge the event to find a common reference representation
VariantContext mergedVC = GATKVariantContextUtils.simpleMerge(eventsAtThisLoc, priorityList,
GATKVariantContextUtils.FilteredRecordMergeType.KEEP_IF_ANY_UNFILTERED,
GATKVariantContextUtils.GenotypeMergeType.PRIORITIZE, false, false, null, false, false);
if( mergedVC == null )
continue;
final GenotypeLikelihoodsCalculationModel.Model calculationModel = mergedVC.isSNP()
? GenotypeLikelihoodsCalculationModel.Model.SNP : GenotypeLikelihoodsCalculationModel.Model.INDEL;
if (emitReferenceConfidence)
mergedVC = addNonRefSymbolicAllele(mergedVC);
final Map<VariantContext, Allele> mergeMap = new LinkedHashMap<>();
mergeMap.put(null, mergedVC.getReference()); // the reference event (null) --> the reference allele
for(int iii = 0; iii < eventsAtThisLoc.size(); iii++) {
mergeMap.put(eventsAtThisLoc.get(iii), mergedVC.getAlternateAllele(iii)); // BUGBUG: This is assuming that the order of alleles is the same as the priority list given to simpleMerge function
}
final Map<Allele, List<Haplotype>> alleleMapper = createAlleleMapper(mergeMap, eventMapper);
if( configuration.DEBUG && logger != null ) {
if (logger != null) logger.info("Genotyping event at " + loc + " with alleles = " + mergedVC.getAlleles());
}
ReadLikelihoods<Allele> readAlleleLikelihoods = readLikelihoods.marginalize(alleleMapper, genomeLocParser.createPaddedGenomeLoc(genomeLocParser.createGenomeLoc(mergedVC), ALLELE_EXTENSION));
if (configuration.isSampleContaminationPresent())
readAlleleLikelihoods.contaminationDownsampling(configuration.getSampleContamination());
if (emitReferenceConfidence)
readAlleleLikelihoods.addNonReferenceAllele(GATKVCFConstants.NON_REF_SYMBOLIC_ALLELE);
final GenotypesContext genotypes = calculateGLsForThisEvent( readAlleleLikelihoods, mergedVC, noCallAlleles );
final VariantContext call = calculateGenotypes(new VariantContextBuilder(mergedVC).genotypes(genotypes).make(), calculationModel);
if( call != null ) {
readAlleleLikelihoods = prepareReadAlleleLikelihoodsForAnnotation(readLikelihoods, perSampleFilteredReadList,
genomeLocParser, emitReferenceConfidence, alleleMapper, readAlleleLikelihoods, call);
ReferenceContext referenceContext = new ReferenceContext(genomeLocParser, genomeLocParser.createGenomeLoc(mergedVC.getChr(), mergedVC.getStart(), mergedVC.getEnd()), refLoc, ref);
VariantContext annotatedCall = annotationEngine.annotateContextForActiveRegion(referenceContext, tracker,readAlleleLikelihoods, call);
if( call.getAlleles().size() != mergedVC.getAlleles().size() )
annotatedCall = GATKVariantContextUtils.reverseTrimAlleles(annotatedCall);
// maintain the set of all called haplotypes
for ( final Allele calledAllele : call.getAlleles() ) {
final List<Haplotype> haplotypeList = alleleMapper.get(calledAllele);
if (haplotypeList == null) continue;
calledHaplotypes.addAll(haplotypeList);
}
returnCalls.add( annotatedCall );
}
}
}
final List<VariantContext> phasedCalls = doPhysicalPhasing ? phaseCalls(returnCalls, calledHaplotypes) : returnCalls;
return new CalledHaplotypes(phasedCalls, calledHaplotypes);
}
/**
* Tries to phase the individual alleles based on pairwise comparisons to the other alleles based on all called haplotypes
*
* @param calls the list of called alleles
* @param calledHaplotypes the set of haplotypes used for calling
* @return a non-null list which represents the possibly phased version of the calls
*/
protected List<VariantContext> phaseCalls(final List<VariantContext> calls, final Set<Haplotype> calledHaplotypes) {
// construct a mapping from alternate allele to the set of haplotypes that contain that allele
final Map<VariantContext, Set<Haplotype>> haplotypeMap = constructHaplotypeMapping(calls, calledHaplotypes);
// construct a mapping from call to phase set ID
final Map<VariantContext, Pair<Integer, String>> phaseSetMapping = new HashMap<>();
final int uniqueCounterEndValue = constructPhaseSetMapping(calls, haplotypeMap, calledHaplotypes.size() - 1, phaseSetMapping);
// we want to establish (potential) *groups* of phased variants, so we need to be smart when looking at pairwise phasing partners
return constructPhaseGroups(calls, phaseSetMapping, uniqueCounterEndValue);
}
/**
* Construct the mapping from alternate allele to the set of haplotypes that contain that allele
*
* @param originalCalls the original unphased calls
* @param calledHaplotypes the set of haplotypes used for calling
* @return non-null Map
*/
protected static Map<VariantContext, Set<Haplotype>> constructHaplotypeMapping(final List<VariantContext> originalCalls,
final Set<Haplotype> calledHaplotypes) {
final Map<VariantContext, Set<Haplotype>> haplotypeMap = new HashMap<>(originalCalls.size());
for ( final VariantContext call : originalCalls ) {
// don't try to phase if there is not exactly 1 alternate allele
if ( ! isBiallelic(call) ) {
haplotypeMap.put(call, Collections.<Haplotype>emptySet());
continue;
}
// keep track of the haplotypes that contain this particular alternate allele
final Set<Haplotype> hapsWithAllele = new HashSet<>();
final Allele alt = call.getAlternateAllele(0);
for ( final Haplotype h : calledHaplotypes ) {
for ( final VariantContext event : h.getEventMap().getVariantContexts() ) {
if ( event.getStart() == call.getStart() && event.getAlternateAlleles().contains(alt) )
hapsWithAllele.add(h);
}
}
haplotypeMap.put(call, hapsWithAllele);
}
return haplotypeMap;
}
/**
* Construct the mapping from call (variant context) to phase set ID
*
* @param originalCalls the original unphased calls
* @param haplotypeMap mapping from alternate allele to the set of haplotypes that contain that allele
* @param totalAvailableHaplotypes the total number of possible haplotypes used in calling
* @param phaseSetMapping the map to populate in this method;
* note that it is okay for this method NOT to populate the phaseSetMapping at all (e.g. in an impossible-to-phase situation)
* @return the next incremental unique index
*/
protected static int constructPhaseSetMapping(final List<VariantContext> originalCalls,
final Map<VariantContext, Set<Haplotype>> haplotypeMap,
final int totalAvailableHaplotypes,
final Map<VariantContext, Pair<Integer, String>> phaseSetMapping) {
final int numCalls = originalCalls.size();
int uniqueCounter = 0;
// use the haplotype mapping to connect variants that are always/never present on the same haplotypes
for ( int i = 0; i < numCalls - 1; i++ ) {
final VariantContext call = originalCalls.get(i);
final Set<Haplotype> haplotypesWithCall = haplotypeMap.get(call);
if ( haplotypesWithCall.isEmpty() )
continue;
final boolean callIsOnAllHaps = haplotypesWithCall.size() == totalAvailableHaplotypes;
for ( int j = i+1; j < numCalls; j++ ) {
final VariantContext comp = originalCalls.get(j);
final Set<Haplotype> haplotypesWithComp = haplotypeMap.get(comp);
if ( haplotypesWithComp.isEmpty() )
continue;
// if the variants are together on all haplotypes, record that fact.
// another possibility is that one of the variants is on all possible haplotypes (i.e. it is homozygous).
final boolean compIsOnAllHaps = haplotypesWithComp.size() == totalAvailableHaplotypes;
if ( (haplotypesWithCall.size() == haplotypesWithComp.size() && haplotypesWithCall.containsAll(haplotypesWithComp)) || callIsOnAllHaps || compIsOnAllHaps ) {
// create a new group if these are the first entries
if ( ! phaseSetMapping.containsKey(call) ) {
// note that if the comp is already in the map then that is very bad because it means that there is
// another variant that is in phase with the comp but not with the call. Since that's an un-phasable
// situation, we should abort if we encounter it.
if ( phaseSetMapping.containsKey(comp) ) {
phaseSetMapping.clear();
return 0;
}
// An important note: even for homozygous variants we are setting the phase as "0|1" here.
// We do this because we cannot possibly know for sure at this time that the genotype for this
// sample will actually be homozygous downstream: there are steps in the pipeline that are liable
// to change the genotypes. Because we can't make those assumptions here, we have decided to output
// the phase as if the call is heterozygous and then "fix" it downstream as needed.
phaseSetMapping.put(call, new Pair<>(uniqueCounter, phase01));
phaseSetMapping.put(comp, new Pair<>(uniqueCounter, phase01));
uniqueCounter++;
}
// otherwise it's part of an existing group so use that group's unique ID
else if ( ! phaseSetMapping.containsKey(comp) ) {
final Pair<Integer, String> callPhase = phaseSetMapping.get(call);
phaseSetMapping.put(comp, new Pair<>(callPhase.first, callPhase.second));
}
}
// if the variants are apart on *all* haplotypes, record that fact
else if ( haplotypesWithCall.size() + haplotypesWithComp.size() == totalAvailableHaplotypes ) {
final Set<Haplotype> intersection = new HashSet<>();
intersection.addAll(haplotypesWithCall);
intersection.retainAll(haplotypesWithComp);
if ( intersection.isEmpty() ) {
// create a new group if these are the first entries
if ( ! phaseSetMapping.containsKey(call) ) {
// note that if the comp is already in the map then that is very bad because it means that there is
// another variant that is in phase with the comp but not with the call. Since that's an un-phasable
// situation, we should abort if we encounter it.
if ( phaseSetMapping.containsKey(comp) ) {
phaseSetMapping.clear();
return 0;
}
phaseSetMapping.put(call, new Pair<>(uniqueCounter, phase01));
phaseSetMapping.put(comp, new Pair<>(uniqueCounter, phase10));
uniqueCounter++;
}
// otherwise it's part of an existing group so use that group's unique ID
else if ( ! phaseSetMapping.containsKey(comp) ){
final Pair<Integer, String> callPhase = phaseSetMapping.get(call);
phaseSetMapping.put(comp, new Pair<>(callPhase.first, callPhase.second.equals(phase01) ? phase10 : phase01));
}
}
}
}
}
return uniqueCounter;
}
/**
* Assemble the phase groups together and update the original calls accordingly
*
* @param originalCalls the original unphased calls
* @param phaseSetMapping mapping from call (variant context) to phase group ID
* @param indexTo last index (exclusive) of phase group IDs
* @return a non-null list which represents the possibly phased version of the calls
*/
protected static List<VariantContext> constructPhaseGroups(final List<VariantContext> originalCalls,
final Map<VariantContext, Pair<Integer, String>> phaseSetMapping,
final int indexTo) {
final List<VariantContext> phasedCalls = new ArrayList<>(originalCalls);
// if we managed to find any phased groups, update the VariantContexts
for ( int count = 0; count < indexTo; count++ ) {
// get all of the (indexes of the) calls that belong in this group (keeping them in the original order)
final List<Integer> indexes = new ArrayList<>();
for ( int index = 0; index < originalCalls.size(); index++ ) {
final VariantContext call = originalCalls.get(index);
if ( phaseSetMapping.containsKey(call) && phaseSetMapping.get(call).first == count )
indexes.add(index);
}
if ( indexes.size() < 2 )
throw new IllegalStateException("Somehow we have a group of phased variants that has fewer than 2 members");
// create a unique ID based on the leftmost one
final String uniqueID = createUniqueID(originalCalls.get(indexes.get(0)));
// update the VCs
for ( final int index : indexes ) {
final VariantContext originalCall = originalCalls.get(index);
final VariantContext phasedCall = phaseVC(originalCall, uniqueID, phaseSetMapping.get(originalCall).second);
phasedCalls.set(index, phasedCall);
}
}
return phasedCalls;
}
/**
* Is this variant bi-allelic? This implementation is very much specific to this class so shouldn't be pulled out into a generalized place.
*
* @param vc the variant context
* @return true if this variant context is bi-allelic, ignoring the NON-REF symbolic allele, false otherwise
*/
private static boolean isBiallelic(final VariantContext vc) {
return vc.isBiallelic() || (vc.getNAlleles() == 3 && vc.getAlternateAlleles().contains(GATKVCFConstants.NON_REF_SYMBOLIC_ALLELE));
}
/**
* Create a unique identifier given the variant context
*
* @param vc the variant context
* @return non-null String
*/
private static String createUniqueID(final VariantContext vc) {
return String.format("%d_%s_%s", vc.getStart(), vc.getReference().getDisplayString(), vc.getAlternateAllele(0).getDisplayString());
// return base + "_0," + base + "_1";
}
/**
* Add physical phase information to the provided variant context
*
* @param vc the variant context
* @param ID the ID to use
* @param phaseGT the phase GT string to use
* @return phased non-null variant context
*/
private static VariantContext phaseVC(final VariantContext vc, final String ID, final String phaseGT) {
final List<Genotype> phasedGenotypes = new ArrayList<>();
for ( final Genotype g : vc.getGenotypes() )
phasedGenotypes.add(new GenotypeBuilder(g).attribute(GATKVCFConstants.HAPLOTYPE_CALLER_PHASING_ID_KEY, ID).attribute(GATKVCFConstants.HAPLOTYPE_CALLER_PHASING_GT_KEY, phaseGT).make());
return new VariantContextBuilder(vc).genotypes(phasedGenotypes).make();
}
private VariantContext addNonRefSymbolicAllele(final VariantContext mergedVC) {
final VariantContextBuilder vcb = new VariantContextBuilder(mergedVC);
final List<Allele> originalList = mergedVC.getAlleles();
final List<Allele> alleleList = new ArrayList<>(originalList.size() + 1);
alleleList.addAll(mergedVC.getAlleles());
alleleList.add(GATKVCFConstants.NON_REF_SYMBOLIC_ALLELE);
vcb.alleles(alleleList);
return vcb.make();
}
// Builds the read-likelihoods collection to use for annotation considering user arguments and the collection
// used for genotyping.
protected ReadLikelihoods<Allele> prepareReadAlleleLikelihoodsForAnnotation(
final ReadLikelihoods<Haplotype> readHaplotypeLikelihoods,
final Map<String, List<GATKSAMRecord>> perSampleFilteredReadList,
final GenomeLocParser genomeLocParser,
final boolean emitReferenceConfidence,
final Map<Allele, List<Haplotype>> alleleMapper,
final ReadLikelihoods<Allele> readAlleleLikelihoodsForGenotyping,
final VariantContext call) {
final ReadLikelihoods<Allele> readAlleleLikelihoodsForAnnotations;
final GenomeLoc loc = genomeLocParser.createGenomeLoc(call);
// We can reuse for annotation the likelihood for genotyping as long as there is no contamination filtering
// or the user want to use the contamination filtered set for annotations.
// Otherwise (else part) we need to do it again.
if (configuration.USE_FILTERED_READ_MAP_FOR_ANNOTATIONS || !configuration.isSampleContaminationPresent()) {
readAlleleLikelihoodsForAnnotations = readAlleleLikelihoodsForGenotyping;
readAlleleLikelihoodsForAnnotations.filterToOnlyOverlappingUnclippedReads(loc);
} else {
readAlleleLikelihoodsForAnnotations = readHaplotypeLikelihoods.marginalize(alleleMapper, loc);
if (emitReferenceConfidence)
readAlleleLikelihoodsForAnnotations.addNonReferenceAllele(
GATKVCFConstants.NON_REF_SYMBOLIC_ALLELE);
}
// Skim the filtered map based on the location so that we do not add filtered read that are going to be removed
// right after a few lines of code bellow.
final Map<String, List<GATKSAMRecord>> overlappingFilteredReads = overlappingFilteredReads(perSampleFilteredReadList, loc);
readAlleleLikelihoodsForAnnotations.addReads(overlappingFilteredReads,0);
return readAlleleLikelihoodsForAnnotations;
}
private Map<String, List<GATKSAMRecord>> overlappingFilteredReads(final Map<String, List<GATKSAMRecord>> perSampleFilteredReadList, final GenomeLoc loc) {
final Map<String,List<GATKSAMRecord>> overlappingFilteredReads = new HashMap<>(perSampleFilteredReadList.size());
for (final Map.Entry<String,List<GATKSAMRecord>> sampleEntry : perSampleFilteredReadList.entrySet()) {
final List<GATKSAMRecord> originalList = sampleEntry.getValue();
final String sample = sampleEntry.getKey();
if (originalList == null || originalList.size() == 0)
continue;
final List<GATKSAMRecord> newList = new ArrayList<>(originalList.size());
for (final GATKSAMRecord read : originalList) {
if (ReadLikelihoods.unclippedReadOverlapsRegion(read, loc))
newList.add(read);
}
if (newList.size() == 0)
continue;
overlappingFilteredReads.put(sample,newList);
}
return overlappingFilteredReads;
}
/**
* Go through the haplotypes we assembled, and decompose them into their constituent variant contexts
*
* @param haplotypes the list of haplotypes we're working with
* @param readLikelihoods map from samples -> the per read allele likelihoods
* @param ref the reference bases (over the same interval as the haplotypes)
* @param refLoc the span of the reference bases
* @param activeAllelesToGenotype alleles we want to ensure are scheduled for genotyping (GGA mode)
* @return never {@code null} but perhaps an empty list if there is no variants to report.
*/
protected TreeSet<Integer> decomposeHaplotypesIntoVariantContexts(final List<Haplotype> haplotypes,
final ReadLikelihoods readLikelihoods,
final byte[] ref,
final GenomeLoc refLoc,
final List<VariantContext> activeAllelesToGenotype) {
final boolean in_GGA_mode = !activeAllelesToGenotype.isEmpty();
// Using the cigar from each called haplotype figure out what events need to be written out in a VCF file
final TreeSet<Integer> startPosKeySet = EventMap.buildEventMapsForHaplotypes(haplotypes, ref, refLoc, configuration.DEBUG);
if ( !in_GGA_mode ) {
// run the event merger if we're not in GGA mode
if (crossHaplotypeEventMerger == null)
throw new IllegalStateException(" no variant merger was provided at set-up when needed in GGA mode");
final boolean mergedAnything = crossHaplotypeEventMerger.merge(haplotypes, readLikelihoods, startPosKeySet, ref, refLoc);
if ( mergedAnything )
cleanUpSymbolicUnassembledEvents( haplotypes ); // the newly created merged events could be overlapping the unassembled events
} else {
startPosKeySet.clear();
for( final VariantContext compVC : activeAllelesToGenotype ) {
startPosKeySet.add( compVC.getStart() );
}
}
return startPosKeySet;
}
/**
* Get the priority list (just the list of sources for these variant context) used to merge overlapping events into common reference view
* @param vcs a list of variant contexts
* @return the list of the sources of vcs in the same order
*/
protected List<String> makePriorityList(final List<VariantContext> vcs) {
final List<String> priorityList = new LinkedList<>();
for ( final VariantContext vc : vcs ) priorityList.add(vc.getSource());
return priorityList;
}
protected List<VariantContext> getVCsAtThisLocation(final List<Haplotype> haplotypes,
final int loc,
final List<VariantContext> activeAllelesToGenotype) {
// the overlapping events to merge into a common reference view
final List<VariantContext> eventsAtThisLoc = new ArrayList<>();
if( activeAllelesToGenotype.isEmpty() ) {
for( final Haplotype h : haplotypes ) {
final EventMap eventMap = h.getEventMap();
final VariantContext vc = eventMap.get(loc);
if( vc != null && !containsVCWithMatchingAlleles(eventsAtThisLoc, vc) ) {
eventsAtThisLoc.add(vc);
}
}
} else { // we are in GGA mode!
int compCount = 0;
for( final VariantContext compVC : activeAllelesToGenotype ) {
if( compVC.getStart() == loc ) {
int alleleCount = 0;
for( final Allele compAltAllele : compVC.getAlternateAlleles() ) {
List<Allele> alleleSet = new ArrayList<>(2);
alleleSet.add(compVC.getReference());
alleleSet.add(compAltAllele);
final String vcSourceName = "Comp" + compCount + "Allele" + alleleCount;
// check if this event is already in the list of events due to a repeat in the input alleles track
final VariantContext candidateEventToAdd = new VariantContextBuilder(compVC).alleles(alleleSet).source(vcSourceName).make();
boolean alreadyExists = false;
for( final VariantContext eventToTest : eventsAtThisLoc ) {
if( eventToTest.hasSameAllelesAs(candidateEventToAdd) ) {
alreadyExists = true;
}
}
if( !alreadyExists ) {
eventsAtThisLoc.add(candidateEventToAdd);
}
alleleCount++;
}
}
compCount++;
}
}
return eventsAtThisLoc;
}
/**
* For a particular event described in inputVC, form PL vector for each sample by looking into allele read map and filling likelihood matrix for each allele
* @param readLikelihoods Allele map describing mapping from reads to alleles and corresponding likelihoods
* @param mergedVC Input VC with event to genotype
* @return GenotypesContext object wrapping genotype objects with PLs
*/
@Requires({"readLikelihoods!= null", "mergedVC != null"})
@Ensures("result != null")
protected GenotypesContext calculateGLsForThisEvent( final ReadLikelihoods<Allele> readLikelihoods, final VariantContext mergedVC, final List<Allele> noCallAlleles ) {
final List<Allele> vcAlleles = mergedVC.getAlleles();
final AlleleList<Allele> alleleList = readLikelihoods.alleleCount() == vcAlleles.size() ? readLikelihoods : new IndexedAlleleList<>(vcAlleles);
final GenotypingLikelihoods<Allele> likelihoods = genotypingModel.calculateLikelihoods(alleleList,new GenotypingData<>(ploidyModel,readLikelihoods));
final int sampleCount = samples.sampleCount();
final GenotypesContext result = GenotypesContext.create(sampleCount);
for (int s = 0; s < sampleCount; s++)
result.add(new GenotypeBuilder(samples.sampleAt(s)).alleles(noCallAlleles).PL(likelihoods.sampleLikelihoods(s).getAsPLs()).make());
return result;
}
/**
* Removes symbolic events from list of haplotypes
* @param haplotypes Input/output list of haplotypes, before/after removal
*/
// TODO - split into input haplotypes and output haplotypes as not to share I/O arguments
@Requires("haplotypes != null")
protected static void cleanUpSymbolicUnassembledEvents( final List<Haplotype> haplotypes ) {
final List<Haplotype> haplotypesToRemove = new ArrayList<>();
for( final Haplotype h : haplotypes ) {
for( final VariantContext vc : h.getEventMap().getVariantContexts() ) {
if( vc.isSymbolic() ) {
for( final Haplotype h2 : haplotypes ) {
for( final VariantContext vc2 : h2.getEventMap().getVariantContexts() ) {
if( vc.getStart() == vc2.getStart() && (vc2.isIndel() || vc2.isMNP()) ) { // unfortunately symbolic alleles can't currently be combined with non-point events
haplotypesToRemove.add(h);
break;
}
}
}
}
}
}
haplotypes.removeAll(haplotypesToRemove);
}
protected static Map<Allele, List<Haplotype>> createAlleleMapper( final Map<VariantContext, Allele> mergeMap, final Map<Event, List<Haplotype>> eventMap ) {
final Map<Allele, List<Haplotype>> alleleMapper = new LinkedHashMap<>();
for( final Map.Entry<VariantContext, Allele> entry : mergeMap.entrySet() ) {
alleleMapper.put(entry.getValue(), eventMap.get(new Event(entry.getKey())));
}
return alleleMapper;
}
@Requires({"haplotypes.size() >= eventsAtThisLoc.size() + 1"})
@Ensures({"result.size() == eventsAtThisLoc.size() + 1"})
protected static Map<Event, List<Haplotype>> createEventMapper( final int loc, final List<VariantContext> eventsAtThisLoc, final List<Haplotype> haplotypes) {
final Map<Event, List<Haplotype>> eventMapper = new LinkedHashMap<>(eventsAtThisLoc.size()+1);
final Event refEvent = new Event(null);
eventMapper.put(refEvent, new ArrayList<Haplotype>());
for( final VariantContext vc : eventsAtThisLoc ) {
eventMapper.put(new Event(vc), new ArrayList<Haplotype>());
}
for( final Haplotype h : haplotypes ) {
if( h.getEventMap().get(loc) == null ) {
eventMapper.get(refEvent).add(h);
} else {
for( final VariantContext vcAtThisLoc : eventsAtThisLoc ) {
if( h.getEventMap().get(loc).hasSameAllelesAs(vcAtThisLoc) ) {
eventMapper.get(new Event(vcAtThisLoc)).add(h);
break;
}
}
}
}
return eventMapper;
}
@Deprecated
protected static Map<Integer,VariantContext> generateVCsFromAlignment( final Haplotype haplotype, final byte[] ref, final GenomeLoc refLoc, final String sourceNameToAdd ) {
return new EventMap(haplotype, ref, refLoc, sourceNameToAdd);
}
protected static boolean containsVCWithMatchingAlleles( final List<VariantContext> list, final VariantContext vcToTest ) {
for( final VariantContext vc : list ) {
if( vc.hasSameAllelesAs(vcToTest) ) {
return true;
}
}
return false;
}
protected static class Event {
public VariantContext vc;
public Event( final VariantContext vc ) {
this.vc = vc;
}
@Override
public boolean equals( final Object obj ) {
return obj instanceof Event && ((((Event) obj).vc == null && vc == null) || (((Event) obj).vc != null && vc != null && ((Event) obj).vc.hasSameAllelesAs(vc))) ;
}
@Override
public int hashCode() {
return (vc == null ? -1 : vc.getAlleles().hashCode());
}
}
/**
* Returns the ploidy-model used by this genotyping engine.
*
* @return never {@code null}.
*/
public PloidyModel getPloidyModel() {
return ploidyModel;
}
/**
* Returns the genotyping-model used by this genotyping engine.
*
* @return never {@code null}.
*/
public GenotypingModel getGenotypingModel() {
return genotypingModel;
}
}