From e48f7544785437d52d2149ac82d5750e9c6746ac Mon Sep 17 00:00:00 2001 From: Eric Banks Date: Thu, 13 Jun 2013 19:29:08 -0400 Subject: [PATCH] Fixes to several of the annotations for reduced reads (and other issues). 1. Have the RMSMappingQuality annotation take into account the fact that reduced reads represent multiple reads. 2. The rank sume tests should not be using reduced reads (because they do not represent distinct observations). 3. Fixed a massive bug in the BaseQualityRankSumTest annotation! It was not using the base qualities but rather the read likelihoods?! Added a unit test for Rank Sum Tests to prove that the distributions are correctly getting assigned appropriate p-values. Also, and just as importantly, the test shows that using reduced reads in the rank sum tests skews the results and makes insignificant distributions look significant (so it can falsely cause the filtering of good sites). Also included in this commit is a massive refactor of the RankSumTest class as requested by the reviewer. --- .../annotator/BaseQualityRankSumTest.java | 42 +---- .../annotator/ClippingRankSumTest.java | 31 +--- .../gatk/walkers/annotator/FisherStrand.java | 16 +- .../annotator/MappingQualityRankSumTest.java | 42 +---- .../walkers/annotator/RMSMappingQuality.java | 55 +++---- .../gatk/walkers/annotator/RankSumTest.java | 139 +++++++++++----- .../walkers/annotator/ReadPosRankSumTest.java | 95 ++++------- .../walkers/annotator/RankSumUnitTest.java | 151 ++++++++++++++++++ .../VariantAnnotatorIntegrationTest.java | 6 +- ...perGeneralPloidySuite1IntegrationTest.java | 2 +- ...perGeneralPloidySuite2IntegrationTest.java | 2 +- ...dGenotyperIndelCallingIntegrationTest.java | 16 +- ...GenotyperNormalCallingIntegrationTest.java | 8 +- ...dGenotyperReducedReadsIntegrationTest.java | 6 +- ...lexAndSymbolicVariantsIntegrationTest.java | 6 +- .../HaplotypeCallerIntegrationTest.java | 18 +-- 16 files changed, 373 insertions(+), 262 deletions(-) create mode 100644 protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/RankSumUnitTest.java diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/BaseQualityRankSumTest.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/BaseQualityRankSumTest.java index a3a9e50e9..534834d0e 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/BaseQualityRankSumTest.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/BaseQualityRankSumTest.java @@ -47,13 +47,11 @@ package org.broadinstitute.sting.gatk.walkers.annotator; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.StandardAnnotation; -import org.broadinstitute.sting.utils.genotyper.MostLikelyAllele; -import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; +import org.broadinstitute.sting.utils.sam.GATKSAMRecord; +import org.broadinstitute.sting.utils.sam.ReadUtils; import org.broadinstitute.variant.vcf.VCFHeaderLineType; import org.broadinstitute.variant.vcf.VCFInfoHeaderLine; import org.broadinstitute.sting.utils.pileup.PileupElement; -import org.broadinstitute.sting.utils.pileup.ReadBackedPileup; -import org.broadinstitute.variant.variantcontext.Allele; import java.util.*; @@ -71,37 +69,11 @@ public class BaseQualityRankSumTest extends RankSumTest implements StandardAnnot public List getDescriptions() { return Arrays.asList(new VCFInfoHeaderLine("BaseQRankSum", 1, VCFHeaderLineType.Float, "Z-score from Wilcoxon rank sum test of Alt Vs. Ref base qualities")); } - protected void fillQualsFromPileup(final List allAlleles, final int refLoc, - final ReadBackedPileup pileup, - final PerReadAlleleLikelihoodMap alleleLikelihoodMap, - final List refQuals, final List altQuals){ - - if (alleleLikelihoodMap == null) { - // use fast SNP-based version if we don't have per-read allele likelihoods - for ( final PileupElement p : pileup ) { - if ( isUsableBase(p) ) { - if ( allAlleles.get(0).equals(Allele.create(p.getBase(),true)) ) { - refQuals.add((double)p.getQual()); - } else if ( allAlleles.contains(Allele.create(p.getBase()))) { - altQuals.add((double)p.getQual()); - } - } - } - return; - } - - for (Map el : alleleLikelihoodMap.getLikelihoodMapValues()) { - final MostLikelyAllele a = PerReadAlleleLikelihoodMap.getMostLikelyAllele(el); - if (! a.isInformative()) - continue; // read is non-informative - if (a.getMostLikelyAllele().isReference()) - refQuals.add(-10.0*(double)el.get(a.getMostLikelyAllele())); - else if (allAlleles.contains(a.getMostLikelyAllele())) - altQuals.add(-10.0*(double)el.get(a.getMostLikelyAllele())); - - - } + protected Double getElementForRead(final GATKSAMRecord read, final int refLoc) { + return (double)read.getBaseQualities()[ReadUtils.getReadCoordinateForReferenceCoordinateUpToEndOfRead(read, refLoc, ReadUtils.ClippingTail.RIGHT_TAIL)]; } - + protected Double getElementForPileupElement(final PileupElement p) { + return (double)p.getQual(); + } } \ No newline at end of file diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ClippingRankSumTest.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ClippingRankSumTest.java index 366512119..68e983bb8 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ClippingRankSumTest.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ClippingRankSumTest.java @@ -46,14 +46,11 @@ package org.broadinstitute.sting.gatk.walkers.annotator; -import org.broadinstitute.sting.utils.genotyper.MostLikelyAllele; -import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; +import org.broadinstitute.sting.utils.pileup.PileupElement; import org.broadinstitute.variant.vcf.VCFHeaderLineType; import org.broadinstitute.variant.vcf.VCFInfoHeaderLine; -import org.broadinstitute.sting.utils.pileup.ReadBackedPileup; import org.broadinstitute.sting.utils.sam.AlignmentUtils; import org.broadinstitute.sting.utils.sam.GATKSAMRecord; -import org.broadinstitute.variant.variantcontext.Allele; import java.util.*; @@ -74,26 +71,12 @@ public class ClippingRankSumTest extends RankSumTest { public List getDescriptions() { return Arrays.asList(new VCFInfoHeaderLine("ClippingRankSum", 1, VCFHeaderLineType.Float, "Z-score From Wilcoxon rank sum test of Alt vs. Ref number of hard clipped bases")); } - - protected void fillQualsFromPileup(final List allAlleles, - final int refLoc, - final ReadBackedPileup pileup, - final PerReadAlleleLikelihoodMap likelihoodMap, final List refQuals, final List altQuals) { - // todo - only support non-pileup case for now, e.g. active-region based version - if (pileup != null || likelihoodMap == null) - return; - - for (Map.Entry> el : likelihoodMap.getLikelihoodReadMap().entrySet()) { - - final MostLikelyAllele a = PerReadAlleleLikelihoodMap.getMostLikelyAllele(el.getValue()); - if (! a.isInformative()) - continue; // read is non-informative - if (a.getMostLikelyAllele().isReference()) - refQuals.add((double)AlignmentUtils.getNumHardClippedBases(el.getKey())); - else if (allAlleles.contains(a.getMostLikelyAllele())) - altQuals.add((double)AlignmentUtils.getNumHardClippedBases(el.getKey())); - - } + protected Double getElementForRead(final GATKSAMRecord read, final int refLoc) { + return (double)AlignmentUtils.getNumHardClippedBases(read); } + protected Double getElementForPileupElement(final PileupElement p) { + // TODO - we only support the non-pileup case for now, e.g. an active-region based version + return null; + } } diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/FisherStrand.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/FisherStrand.java index 957eb1aba..876dbf039 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/FisherStrand.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/FisherStrand.java @@ -300,7 +300,7 @@ public class FisherStrand extends InfoFieldAnnotation implements StandardAnnotat for ( Map.Entry sample : stratifiedContexts.entrySet() ) { for (PileupElement p : sample.getValue().getBasePileup()) { - if ( ! RankSumTest.isUsableBase(p, false) ) // ignore deletions + if ( ! isUsableBase(p) ) // ignore deletions and bad MQ continue; if ( p.getQual() < minQScoreToConsider || p.getMappingQual() < minQScoreToConsider ) @@ -313,6 +313,20 @@ public class FisherStrand extends InfoFieldAnnotation implements StandardAnnotat return table; } + /** + * Can the base in this pileup element be used in comparative tests? + * + * @param p the pileup element to consider + * + * @return true if this base is part of a meaningful read for comparison, false otherwise + */ + private static boolean isUsableBase(final PileupElement p) { + return !( p.isDeletion() || + p.getMappingQual() == 0 || + p.getMappingQual() == QualityUtils.MAPPING_QUALITY_UNAVAILABLE || + ((int) p.getQual()) < QualityUtils.MIN_USABLE_Q_SCORE); + } + private static void updateTable(final int[][] table, final Allele allele, final GATKSAMRecord read, final Allele ref, final Allele alt, final int representativeCount) { final boolean matchesRef = allele.equals(ref, true); diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/MappingQualityRankSumTest.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/MappingQualityRankSumTest.java index 3873138a2..0ebb09961 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/MappingQualityRankSumTest.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/MappingQualityRankSumTest.java @@ -47,14 +47,10 @@ package org.broadinstitute.sting.gatk.walkers.annotator; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.StandardAnnotation; -import org.broadinstitute.sting.utils.genotyper.MostLikelyAllele; -import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; +import org.broadinstitute.sting.utils.pileup.PileupElement; import org.broadinstitute.variant.vcf.VCFHeaderLineType; import org.broadinstitute.variant.vcf.VCFInfoHeaderLine; -import org.broadinstitute.sting.utils.pileup.PileupElement; -import org.broadinstitute.sting.utils.pileup.ReadBackedPileup; import org.broadinstitute.sting.utils.sam.GATKSAMRecord; -import org.broadinstitute.variant.variantcontext.Allele; import java.util.*; @@ -73,35 +69,11 @@ public class MappingQualityRankSumTest extends RankSumTest implements StandardAn public List getDescriptions() { return Arrays.asList(new VCFInfoHeaderLine("MQRankSum", 1, VCFHeaderLineType.Float, "Z-score From Wilcoxon rank sum test of Alt vs. Ref read mapping qualities")); } - protected void fillQualsFromPileup(final List allAlleles, - final int refLoc, - final ReadBackedPileup pileup, - final PerReadAlleleLikelihoodMap likelihoodMap, - final List refQuals, final List altQuals) { - - if (pileup != null && likelihoodMap == null) { - // old UG snp-only path through the annotations - for ( final PileupElement p : pileup ) { - if ( isUsableBase(p) ) { - if ( allAlleles.get(0).equals(Allele.create(p.getBase(), true)) ) { - refQuals.add((double)p.getMappingQual()); - } else if ( allAlleles.contains(Allele.create(p.getBase()))) { - altQuals.add((double)p.getMappingQual()); - } - } - } - return; - } - for (Map.Entry> el : likelihoodMap.getLikelihoodReadMap().entrySet()) { - final MostLikelyAllele a = PerReadAlleleLikelihoodMap.getMostLikelyAllele(el.getValue()); - // BUGBUG: There needs to be a comparable isUsableBase check here - if (! a.isInformative()) - continue; // read is non-informative - if (a.getMostLikelyAllele().isReference()) - refQuals.add((double)el.getKey().getMappingQuality()); - else if (allAlleles.contains(a.getMostLikelyAllele())) - altQuals.add((double)el.getKey().getMappingQuality()); - } + protected Double getElementForRead(final GATKSAMRecord read, final int refLoc) { + return (double)read.getMappingQuality(); } - } \ No newline at end of file + protected Double getElementForPileupElement(final PileupElement p) { + return (double)p.getRead().getMappingQuality(); + } +} \ No newline at end of file diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RMSMappingQuality.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RMSMappingQuality.java index 18348162e..d9bc5966c 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RMSMappingQuality.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RMSMappingQuality.java @@ -56,6 +56,7 @@ import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.StandardAnnota import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; import org.broadinstitute.sting.utils.MathUtils; import org.broadinstitute.sting.utils.QualityUtils; +import org.broadinstitute.sting.utils.sam.ReadUtils; import org.broadinstitute.variant.vcf.VCFConstants; import org.broadinstitute.variant.vcf.VCFInfoHeaderLine; import org.broadinstitute.variant.vcf.VCFStandardHeaderLines; @@ -77,55 +78,41 @@ public class RMSMappingQuality extends InfoFieldAnnotation implements StandardAn final Map stratifiedContexts, final VariantContext vc, final Map perReadAlleleLikelihoodMap ) { - int totalSize = 0, index = 0; - int qualities[]; - if (stratifiedContexts != null) { + + final List qualities = new ArrayList<>(); + if ( stratifiedContexts != null ) { if ( stratifiedContexts.size() == 0 ) return null; - for ( AlignmentContext context : stratifiedContexts.values() ) - totalSize += context.size(); - - qualities = new int[totalSize]; - - for ( Map.Entry sample : stratifiedContexts.entrySet() ) { - AlignmentContext context = sample.getValue(); - for (PileupElement p : context.getBasePileup() ) - index = fillMappingQualitiesFromPileupAndUpdateIndex(p.getRead(), index, qualities); + for ( final Map.Entry sample : stratifiedContexts.entrySet() ) { + final AlignmentContext context = sample.getValue(); + for ( final PileupElement p : context.getBasePileup() ) + fillMappingQualitiesFromPileup(p.getRead().getMappingQuality(), p.getRepresentativeCount(), qualities); } } else if (perReadAlleleLikelihoodMap != null) { if ( perReadAlleleLikelihoodMap.size() == 0 ) return null; - for ( PerReadAlleleLikelihoodMap perReadLikelihoods : perReadAlleleLikelihoodMap.values() ) - totalSize += perReadLikelihoods.size(); - - qualities = new int[totalSize]; - for ( PerReadAlleleLikelihoodMap perReadLikelihoods : perReadAlleleLikelihoodMap.values() ) { - for (GATKSAMRecord read : perReadLikelihoods.getStoredElements()) - index = fillMappingQualitiesFromPileupAndUpdateIndex(read, index, qualities); - - - } + for ( final PerReadAlleleLikelihoodMap perReadLikelihoods : perReadAlleleLikelihoodMap.values() ) { + for ( final GATKSAMRecord read : perReadLikelihoods.getStoredElements() ) + fillMappingQualitiesFromPileup(read.getMappingQuality(), (read.isReducedRead() ? read.getReducedCount(ReadUtils.getReadCoordinateForReferenceCoordinateUpToEndOfRead(read, vc.getStart(), ReadUtils.ClippingTail.RIGHT_TAIL)) : 1), qualities); + } } else return null; - - - double rms = MathUtils.rms(qualities); - Map map = new HashMap(); - map.put(getKeyNames().get(0), String.format("%.2f", rms)); - return map; + final double rms = MathUtils.rms(qualities); + return Collections.singletonMap(getKeyNames().get(0), (Object)String.format("%.2f", rms)); } - private static int fillMappingQualitiesFromPileupAndUpdateIndex(final GATKSAMRecord read, final int inputIdx, final int[] qualities) { - int outputIdx = inputIdx; - if ( read.getMappingQuality() != QualityUtils.MAPPING_QUALITY_UNAVAILABLE ) - qualities[outputIdx++] = read.getMappingQuality(); - - return outputIdx; + private static void fillMappingQualitiesFromPileup(final int mq, final int representativeCount, final List qualities) { + if ( mq != QualityUtils.MAPPING_QUALITY_UNAVAILABLE ) { + if ( representativeCount == 1 ) + qualities.add(mq); + else + qualities.addAll(Collections.nCopies(representativeCount, mq)); + } } public List getKeyNames() { return Arrays.asList(VCFConstants.RMS_MAPPING_QUALITY_KEY); } diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RankSumTest.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RankSumTest.java index ef456824e..37508fc06 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RankSumTest.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/RankSumTest.java @@ -53,9 +53,11 @@ import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.ActiveRegionBasedAnnotation; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.AnnotatorCompatible; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.InfoFieldAnnotation; +import org.broadinstitute.sting.utils.genotyper.MostLikelyAllele; import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; import org.broadinstitute.sting.utils.MannWhitneyU; import org.broadinstitute.sting.utils.QualityUtils; +import org.broadinstitute.sting.utils.sam.GATKSAMRecord; import org.broadinstitute.variant.vcf.VCFHeaderLine; import org.broadinstitute.sting.utils.collections.Pair; import org.broadinstitute.sting.utils.pileup.PileupElement; @@ -87,31 +89,33 @@ public abstract class RankSumTest extends InfoFieldAnnotation implements ActiveR if (genotypes == null || genotypes.size() == 0) return null; - final ArrayList refQuals = new ArrayList(); - final ArrayList altQuals = new ArrayList(); + final ArrayList refQuals = new ArrayList<>(); + final ArrayList altQuals = new ArrayList<>(); for ( final Genotype genotype : genotypes.iterateInSampleNameOrder() ) { - PerReadAlleleLikelihoodMap indelLikelihoodMap = null; - ReadBackedPileup pileup = null; + boolean usePileup = true; - if (stratifiedContexts != null) { // the old UG SNP-only path through the annotations - final AlignmentContext context = stratifiedContexts.get(genotype.getSampleName()); - if ( context != null ) - pileup = context.getBasePileup(); + if ( stratifiedPerReadAlleleLikelihoodMap != null ) { + final PerReadAlleleLikelihoodMap likelihoodMap = stratifiedPerReadAlleleLikelihoodMap.get(genotype.getSampleName()); + if ( likelihoodMap != null && !likelihoodMap.isEmpty() ) { + fillQualsFromLikelihoodMap(vc.getAlleles(), vc.getStart(), likelihoodMap, refQuals, altQuals); + usePileup = false; + } } - if (stratifiedPerReadAlleleLikelihoodMap != null ) - indelLikelihoodMap = stratifiedPerReadAlleleLikelihoodMap.get(genotype.getSampleName()); - if (indelLikelihoodMap != null && indelLikelihoodMap.isEmpty()) - indelLikelihoodMap = null; - // treat an empty likelihood map as a null reference - will simplify contract with fillQualsFromPileup - if (indelLikelihoodMap == null && pileup == null) - continue; - - fillQualsFromPileup(vc.getAlleles(), vc.getStart(), pileup, indelLikelihoodMap, refQuals, altQuals ); + // the old UG SNP-only path through the annotations + if ( usePileup && stratifiedContexts != null ) { + final AlignmentContext context = stratifiedContexts.get(genotype.getSampleName()); + if ( context != null ) { + final ReadBackedPileup pileup = context.getBasePileup(); + if ( pileup != null ) + fillQualsFromPileup(vc.getAlleles(), pileup, refQuals, altQuals); + } + } } - if (refQuals.isEmpty() && altQuals.isEmpty()) + + if ( refQuals.isEmpty() && altQuals.isEmpty() ) return null; final MannWhitneyU mannWhitneyU = new MannWhitneyU(useDithering); @@ -136,18 +140,72 @@ public abstract class RankSumTest extends InfoFieldAnnotation implements ActiveR // we are testing that set1 (the alt bases) have lower quality scores than set2 (the ref bases) final Pair testResults = mannWhitneyU.runOneSidedTest(MannWhitneyU.USet.SET1); - final Map map = new HashMap(); + final Map map = new HashMap<>(); if (!Double.isNaN(testResults.first)) map.put(getKeyNames().get(0), String.format("%.3f", testResults.first)); return map; } - protected abstract void fillQualsFromPileup(final List alleles, - final int refLoc, - final ReadBackedPileup readBackedPileup, - final PerReadAlleleLikelihoodMap alleleLikelihoodMap, - final List refQuals, - final List altQuals); + private void fillQualsFromPileup(final List alleles, + final ReadBackedPileup pileup, + final List refQuals, + final List altQuals) { + for ( final PileupElement p : pileup ) { + if ( isUsableBase(p) ) { + final Double value = getElementForPileupElement(p); + if ( value == null ) + continue; + + if ( alleles.get(0).equals(Allele.create(p.getBase(), true)) ) + refQuals.add(value); + else if ( alleles.contains(Allele.create(p.getBase())) ) + altQuals.add(value); + } + } + } + + private void fillQualsFromLikelihoodMap(final List alleles, + final int refLoc, + final PerReadAlleleLikelihoodMap likelihoodMap, + final List refQuals, + final List altQuals) { + for ( final Map.Entry> el : likelihoodMap.getLikelihoodReadMap().entrySet() ) { + final MostLikelyAllele a = PerReadAlleleLikelihoodMap.getMostLikelyAllele(el.getValue()); + if ( ! a.isInformative() ) + continue; // read is non-informative + + final GATKSAMRecord read = el.getKey(); + if ( isUsableRead(read, refLoc) ) { + final Double value = getElementForRead(read, refLoc); + if ( value == null ) + continue; + + if ( a.getMostLikelyAllele().isReference() ) + refQuals.add(value); + else if ( alleles.contains(a.getMostLikelyAllele()) ) + altQuals.add(value); + } + } + } + + /** + * Get the element for the given read at the given reference position + * + * @param read the read + * @param refLoc the reference position + * @return a Double representing the element to be used in the rank sum test, or null if it should not be used + */ + protected abstract Double getElementForRead(final GATKSAMRecord read, final int refLoc); + + // TODO -- until the ReadPosRankSumTest stops treating these differently, we need to have separate methods for GATKSAMRecords and PileupElements. Yuck. + + /** + * Get the element for the given read at the given reference position + * + * @param p the pileup element + * @return a Double representing the element to be used in the rank sum test, or null if it should not be used + */ + protected abstract Double getElementForPileupElement(final PileupElement p); /** * Can the base in this pileup element be used in comparative tests between ref / alt bases? @@ -157,30 +215,33 @@ public abstract class RankSumTest extends InfoFieldAnnotation implements ActiveR * @param p the pileup element to consider * @return true if this base is part of a meaningful read for comparison, false otherwise */ - public static boolean isUsableBase(final PileupElement p) { - return isUsableBase(p, false); + protected boolean isUsableBase(final PileupElement p) { + return !(p.isDeletion() || + p.getMappingQual() == 0 || + p.getMappingQual() == QualityUtils.MAPPING_QUALITY_UNAVAILABLE || + ((int) p.getQual()) < QualityUtils.MIN_USABLE_Q_SCORE || // need the unBAQed quality score here + p.getRead().isReducedRead() ); } /** - * Can the base in this pileup element be used in comparative tests between ref / alt bases? + * Can the read be used in comparative tests between ref / alt bases? * - * @param p the pileup element to consider - * @param allowDeletions if true, allow p to be a deletion base - * @return true if this base is part of a meaningful read for comparison, false otherwise + * @param read the read to consider + * @param refLoc the reference location + * @return true if this read is meaningful for comparison, false otherwise */ - public static boolean isUsableBase(final PileupElement p, final boolean allowDeletions) { - return !((! allowDeletions && p.isDeletion()) || - p.getMappingQual() == 0 || - p.getMappingQual() == QualityUtils.MAPPING_QUALITY_UNAVAILABLE || - ((int) p.getQual()) < QualityUtils.MIN_USABLE_Q_SCORE); // need the unBAQed quality score here + protected boolean isUsableRead(final GATKSAMRecord read, final int refLoc) { + return !( read.getMappingQuality() == 0 || + read.getMappingQuality() == QualityUtils.MAPPING_QUALITY_UNAVAILABLE || + read.isReducedRead() ); } /** * Initialize the rank sum test annotation using walker and engine information. Right now this checks to see if * engine randomization is turned off, and if so does not dither. - * @param walker - * @param toolkit - * @param headerLines + * @param walker the walker + * @param toolkit the GATK engine + * @param headerLines the header lines */ public void initialize ( AnnotatorCompatible walker, GenomeAnalysisEngine toolkit, Set headerLines ) { useDithering = ! toolkit.getArguments().disableDithering; diff --git a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ReadPosRankSumTest.java b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ReadPosRankSumTest.java index 6ce4aab49..37faaed22 100644 --- a/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ReadPosRankSumTest.java +++ b/protected/java/src/org/broadinstitute/sting/gatk/walkers/annotator/ReadPosRankSumTest.java @@ -51,17 +51,13 @@ import net.sf.samtools.CigarElement; import net.sf.samtools.CigarOperator; import net.sf.samtools.SAMRecord; import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.StandardAnnotation; -import org.broadinstitute.sting.utils.genotyper.MostLikelyAllele; -import org.broadinstitute.sting.utils.genotyper.PerReadAlleleLikelihoodMap; import org.broadinstitute.sting.gatk.walkers.indels.PairHMMIndelErrorModel; import org.broadinstitute.variant.vcf.VCFHeaderLineType; import org.broadinstitute.variant.vcf.VCFInfoHeaderLine; import org.broadinstitute.sting.utils.pileup.PileupElement; -import org.broadinstitute.sting.utils.pileup.ReadBackedPileup; import org.broadinstitute.sting.utils.sam.AlignmentUtils; import org.broadinstitute.sting.utils.sam.GATKSAMRecord; import org.broadinstitute.sting.utils.sam.ReadUtils; -import org.broadinstitute.variant.variantcontext.Allele; import java.util.*; @@ -83,55 +79,34 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio return Arrays.asList(new VCFInfoHeaderLine("ReadPosRankSum", 1, VCFHeaderLineType.Float, "Z-score from Wilcoxon rank sum test of Alt vs. Ref read position bias")); } - protected void fillQualsFromPileup(final List allAlleles, - final int refLoc, - final ReadBackedPileup pileup, - final PerReadAlleleLikelihoodMap alleleLikelihoodMap, - final List refQuals, final List altQuals) { + protected Double getElementForRead(final GATKSAMRecord read, final int refLoc) { + final int offset = ReadUtils.getReadCoordinateForReferenceCoordinate( read.getSoftStart(), read.getCigar(), refLoc, ReadUtils.ClippingTail.RIGHT_TAIL, true ); + if ( offset == ReadUtils.CLIPPING_GOAL_NOT_REACHED ) + return null; - if (alleleLikelihoodMap == null) { - // use old UG SNP-based version if we don't have per-read allele likelihoods - for ( final PileupElement p : pileup ) { - if ( isUsableBase(p) && p.getRead().getCigar() != null ) { - int readPos = AlignmentUtils.calcAlignmentByteArrayOffset(p.getRead().getCigar(), p, 0, 0); - - readPos = getFinalReadPosition(p.getRead(),readPos); - - if ( allAlleles.get(0).equals(Allele.create(p.getBase(), true)) ) { - refQuals.add((double)readPos); - } else if ( allAlleles.contains(Allele.create(p.getBase()))) { - altQuals.add((double)readPos); - } - } - } - return; - } - - for (Map.Entry> el : alleleLikelihoodMap.getLikelihoodReadMap().entrySet()) { - final MostLikelyAllele a = PerReadAlleleLikelihoodMap.getMostLikelyAllele(el.getValue()); - if (! a.isInformative() ) - continue; // read is non-informative - - final GATKSAMRecord read = el.getKey(); - if ( read.getSoftStart() + read.getCigar().getReadLength() <= refLoc ) { // make sure the read actually covers the requested ref loc - continue; - } - final int offset = ReadUtils.getReadCoordinateForReferenceCoordinate( read.getSoftStart(), read.getCigar(), refLoc, ReadUtils.ClippingTail.RIGHT_TAIL, true ); - if ( offset == ReadUtils.CLIPPING_GOAL_NOT_REACHED || read.getCigar() == null ) - continue; - int readPos = AlignmentUtils.calcAlignmentByteArrayOffset( read.getCigar(), offset, false, 0, 0 ); - final int numAlignedBases = AlignmentUtils.getNumAlignedBasesCountingSoftClips( read ); - if (readPos > numAlignedBases / 2) - readPos = numAlignedBases - (readPos + 1); - - if (a.getMostLikelyAllele().isReference()) - refQuals.add((double)readPos); - else if (allAlleles.contains(a.getMostLikelyAllele())) - altQuals.add((double)readPos); - } + int readPos = AlignmentUtils.calcAlignmentByteArrayOffset( read.getCigar(), offset, false, 0, 0 ); + final int numAlignedBases = AlignmentUtils.getNumAlignedBasesCountingSoftClips( read ); + if (readPos > numAlignedBases / 2) + readPos = numAlignedBases - (readPos + 1); + return (double)readPos; } - int getFinalReadPosition(GATKSAMRecord read, int initialReadPosition) { + protected Double getElementForPileupElement(final PileupElement p) { + final int offset = AlignmentUtils.calcAlignmentByteArrayOffset(p.getRead().getCigar(), p, 0, 0); + return (double)getFinalReadPosition(p.getRead(), offset); + } + + @Override + protected boolean isUsableBase(final PileupElement p) { + return super.isUsableBase(p) && p.getRead().getCigar() != null; + } + + @Override + protected boolean isUsableRead(final GATKSAMRecord read, final int refLoc) { + return super.isUsableRead(read, refLoc) && read.getSoftStart() + read.getCigar().getReadLength() > refLoc; + } + + private int getFinalReadPosition(final GATKSAMRecord read, final int initialReadPosition) { final int numAlignedBases = getNumAlignedBases(read); int readPos = initialReadPosition; @@ -141,7 +116,8 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio return readPos; } - int getNumClippedBasesAtStart(SAMRecord read) { + + private int getNumClippedBasesAtStart(final SAMRecord read) { // compute total number of clipped bases (soft or hard clipped) // check for hard clips (never consider these bases): final Cigar c = read.getCigar(); @@ -151,8 +127,8 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio if (first.getOperator() == CigarOperator.H) { numStartClippedBases = first.getLength(); } - byte[] unclippedReadBases = read.getReadBases(); - byte[] unclippedReadQuals = read.getBaseQualities(); + final byte[] unclippedReadBases = read.getReadBases(); + final byte[] unclippedReadQuals = read.getBaseQualities(); // Do a stricter base clipping than provided by CIGAR string, since this one may be too conservative, // and may leave a string of Q2 bases still hanging off the reads. @@ -167,11 +143,11 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio return numStartClippedBases; } - int getNumAlignedBases(SAMRecord read) { + private int getNumAlignedBases(final GATKSAMRecord read) { return read.getReadLength() - getNumClippedBasesAtStart(read) - getNumClippedBasesAtEnd(read); } - int getNumClippedBasesAtEnd(SAMRecord read) { + private int getNumClippedBasesAtEnd(final GATKSAMRecord read) { // compute total number of clipped bases (soft or hard clipped) // check for hard clips (never consider these bases): final Cigar c = read.getCigar(); @@ -181,8 +157,8 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio if (last.getOperator() == CigarOperator.H) { numEndClippedBases = last.getLength(); } - byte[] unclippedReadBases = read.getReadBases(); - byte[] unclippedReadQuals = read.getBaseQualities(); + final byte[] unclippedReadBases = read.getReadBases(); + final byte[] unclippedReadQuals = read.getBaseQualities(); // Do a stricter base clipping than provided by CIGAR string, since this one may be too conservative, // and may leave a string of Q2 bases still hanging off the reads. @@ -193,11 +169,6 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio break; } - return numEndClippedBases; } - - int getOffsetFromClippedReadStart(SAMRecord read, int offset) { - return offset - getNumClippedBasesAtStart(read); - } } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/RankSumUnitTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/RankSumUnitTest.java new file mode 100644 index 000000000..fec83e1a8 --- /dev/null +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/RankSumUnitTest.java @@ -0,0 +1,151 @@ +/* +* 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 7 Cambridge Center, 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 GATK2 and related documentation, if any, as they exist on the EFFECTIVE DATE and can be downloaded from http://www.broadinstitute/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. +* 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. 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 Broad Institute, Inc. +* Notice of attribution: The GATK2 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. +* +* 4. 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. +* +* 5. 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. +* +* 6. 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. +* +* 7. MISCELLANEOUS +* 7.1 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. +* 7.2 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. +* 7.3 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. +* 7.4 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. +* 7.5 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. +* 7.6 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. +* 7.7 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.sting.gatk.walkers.annotator; + +import org.broadinstitute.sting.gatk.walkers.compression.reducereads.*; +import org.broadinstitute.sting.gatk.walkers.compression.reducereads.BaseCounts; +import org.broadinstitute.sting.utils.MannWhitneyU; +import org.testng.Assert; +import org.testng.annotations.BeforeClass; +import org.testng.annotations.DataProvider; +import org.testng.annotations.Test; + +import java.util.ArrayList; +import java.util.Arrays; +import java.util.Collections; +import java.util.List; + +public class RankSumUnitTest { + + List distribution20, distribution30, distribution20_40; + static final int observations = 100; + + @BeforeClass + public void init() { + distribution20 = new ArrayList<>(observations); + distribution30 = new ArrayList<>(observations); + distribution20_40 = new ArrayList<>(observations); + + final int skew = 3; + makeDistribution(distribution20, 20, skew, observations); + makeDistribution(distribution30, 30, skew, observations); + makeDistribution(distribution20_40, 20, skew, observations/2); + makeDistribution(distribution20_40, 40, skew, observations/2); + + // shuffle the observations + Collections.shuffle(distribution20); + Collections.shuffle(distribution30); + Collections.shuffle(distribution20_40); + } + + private static void makeDistribution(final List result, final int target, final int skew, final int numObservations) { + final int rangeStart = target - skew; + final int rangeEnd = target + skew; + + int current = rangeStart; + for ( int i = 0; i < numObservations; i++ ) { + result.add(current++); + if ( current > rangeEnd ) + current = rangeStart; + } + } + + @DataProvider(name = "DistributionData") + public Object[][] makeDistributionData() { + List tests = new ArrayList(); + + for ( final int numToReduce : Arrays.asList(0, 10, 50, 100) ) { + tests.add(new Object[]{distribution20, distribution20, numToReduce, true, "20-20"}); + tests.add(new Object[]{distribution30, distribution30, numToReduce, true, "30-30"}); + tests.add(new Object[]{distribution20_40, distribution20_40, numToReduce, true, "20/40-20/40"}); + + tests.add(new Object[]{distribution20, distribution30, numToReduce, false, "20-30"}); + tests.add(new Object[]{distribution30, distribution20, numToReduce, false, "30-20"}); + + tests.add(new Object[]{distribution20, distribution20_40, numToReduce, false, "20-20/40"}); + tests.add(new Object[]{distribution30, distribution20_40, numToReduce, true, "30-20/40"}); + } + + return tests.toArray(new Object[][]{}); + } + + @Test(enabled = true, dataProvider = "DistributionData") + public void testDistribution(final List distribution1, final List distribution2, final int numToReduceIn2, final boolean distributionsShouldBeEqual, final String debugString) { + final MannWhitneyU mannWhitneyU = new MannWhitneyU(true); + + for ( final Integer num : distribution1 ) + mannWhitneyU.add(num, MannWhitneyU.USet.SET1); + + final List dist2 = new ArrayList<>(distribution2); + if ( numToReduceIn2 > 0 ) { + final org.broadinstitute.sting.gatk.walkers.compression.reducereads.BaseCounts counts = new BaseCounts(); + for ( int i = 0; i < numToReduceIn2; i++ ) { + final int value = dist2.remove(0); + counts.incr(BaseIndex.A, (byte)value, 0, false); + } + + final int qual = (int)counts.averageQualsOfBase(BaseIndex.A); + for ( int i = 0; i < numToReduceIn2; i++ ) + dist2.add(qual); + } + + for ( final Integer num : dist2 ) + mannWhitneyU.add(num, MannWhitneyU.USet.SET2); + + final Double result = mannWhitneyU.runTwoSidedTest().second; + Assert.assertFalse(Double.isNaN(result)); + + if ( distributionsShouldBeEqual ) { + // TODO -- THIS IS THE FAILURE POINT OF USING REDUCED READS WITH RANK SUM TESTS + if ( numToReduceIn2 >= observations / 2 ) + return; + Assert.assertTrue(result > 0.1, String.format("%f %d %d", result, numToReduceIn2, dist2.get(0))); + } else { + Assert.assertTrue(result < 0.01, String.format("%f %d %d", result, numToReduceIn2, dist2.get(0))); + } + } +} diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorIntegrationTest.java index 961a28bcf..e7d7300ae 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorIntegrationTest.java @@ -78,7 +78,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest { public void testHasAnnotsAsking1() { WalkerTestSpec spec = new WalkerTestSpec( baseTestString() + " -G Standard --variant " + privateTestDir + "vcfexample2.vcf -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -L 1:10,020,000-10,021,000", 1, - Arrays.asList("42889072698af972f2004ccfe8eae15e")); + Arrays.asList("823868a4b5b5ec2cdf080c059d04d31a")); executeTest("test file has annotations, asking for annotations, #1", spec); } @@ -112,7 +112,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest { public void testNoAnnotsAsking1() { WalkerTestSpec spec = new WalkerTestSpec( baseTestString() + " -G Standard --variant " + privateTestDir + "vcfexample2empty.vcf -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -L 1:10,020,000-10,021,000", 1, - Arrays.asList("7e755bb09169699b76850e76b71a5f5a")); + Arrays.asList("6f873b3152db291e18e3a04fbce2e117")); executeTest("test file doesn't have annotations, asking for annotations, #1", spec); } @@ -128,7 +128,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest { public void testExcludeAnnotations() { WalkerTestSpec spec = new WalkerTestSpec( baseTestString() + " -G Standard -XA FisherStrand -XA ReadPosRankSumTest --variant " + privateTestDir + "vcfexample2empty.vcf -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -L 1:10,020,000-10,021,000", 1, - Arrays.asList("e17596007d0db7673d138a9ae4890e82")); + Arrays.asList("552c2ad9dbfaa85d51d2def93c8229c6")); executeTest("test exclude annotations", spec); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite1IntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite1IntegrationTest.java index c791d08ae..2d36a27d1 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite1IntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite1IntegrationTest.java @@ -79,6 +79,6 @@ public class UnifiedGenotyperGeneralPloidySuite1IntegrationTest extends WalkerTe @Test(enabled = true) public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() { - executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1", "LSV_INDEL_DISC_NOREF_p1", "INDEL", "353c97bfb05a939b3838dc8eee50326b"); + executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1", "LSV_INDEL_DISC_NOREF_p1", "INDEL", "dd28b14d732852bffbba4f22f7697227"); } } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite2IntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite2IntegrationTest.java index 1022b6e15..117e54ef8 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite2IntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperGeneralPloidySuite2IntegrationTest.java @@ -58,7 +58,7 @@ public class UnifiedGenotyperGeneralPloidySuite2IntegrationTest extends WalkerTe @Test(enabled = true) public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() { - executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","7e4e1397d5cff68aeba3595e671574fc"); + executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","369ad0ff28bb9ce7974dc2c7343c8737"); } @Test(enabled = true) diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java index 64a27c4c3..49d429c0d 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIndelCallingIntegrationTest.java @@ -73,7 +73,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { " -o %s" + " -L 1:10,000,000-10,500,000", 1, - Arrays.asList("14ad6eeed46e9b6f4757370267b1a1cc")); + Arrays.asList("ef8151aa699da3272c1ae0986d16ca21")); executeTest(String.format("test indel caller in SLX"), spec); } @@ -88,7 +88,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { " -minIndelCnt 1" + " -L 1:10,000,000-10,100,000", 1, - Arrays.asList("d9572a227ccb13a6baa6dc4fb65bc1e5")); + Arrays.asList("7f88229ccefb74513efb199b61183cb8")); executeTest(String.format("test indel caller in SLX with low min allele count"), spec); } @@ -101,7 +101,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { " -o %s" + " -L 1:10,000,000-10,500,000", 1, - Arrays.asList("cd184a2a5a1932dcf3e8f0424652176b")); + Arrays.asList("1928ad48bcd0ca180e046bc235cfb3f4")); executeTest(String.format("test indel calling, multiple technologies"), spec); } @@ -111,7 +111,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1, - Arrays.asList("e8d98996eb81ece8cfb52437920ae2e0")); + Arrays.asList("6663e434a7b549f23bfd52db90e53a1a")); executeTest("test MultiSample Pilot2 indels with alleles passed in", spec); } @@ -121,7 +121,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { baseCommandIndels + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1, - Arrays.asList("23a78c16f64bffe1dea3a5587fcabdad")); + Arrays.asList("581c552664e536df6d0f102fb0d10e5a")); executeTest("test MultiSample Pilot2 indels with alleles passed in and emitting all sites", spec); } @@ -136,7 +136,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec( baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + result.get(0).getAbsolutePath() + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L " + result.get(0).getAbsolutePath(), 1, - Arrays.asList("facac578891a4f2be63ddd5ba6b9096b")); + Arrays.asList("587bf6bad368ed81189747a84f913ab2")); executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2); } @@ -176,7 +176,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { public void testMinIndelFraction0() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( assessMinIndelFraction + " -minIndelFrac 0.0", 1, - Arrays.asList("e90256acfc360fc4bf377094732a673a")); + Arrays.asList("862d82c8aa35f1da4f9e67b5b48dfe52")); executeTest("test minIndelFraction 0.0", spec); } @@ -184,7 +184,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest { public void testMinIndelFraction25() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( assessMinIndelFraction + " -minIndelFrac 0.25", 1, - Arrays.asList("98abcfb0a008050eba8b9c285a25b2a0")); + Arrays.asList("8d9fc96be07db791737ac18135de4d63")); executeTest("test minIndelFraction 0.25", spec); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java index f7c5e6fd5..439039f9b 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperNormalCallingIntegrationTest.java @@ -64,7 +64,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{ public void testMultiSamplePilot1() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( baseCommand + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10,022,000-10,025,000", 1, - Arrays.asList("474dfb943a307c86cabe2043970c58f3")); + Arrays.asList("a9466c1e3ce1fc4bac83086b25a6df54")); executeTest("test MultiSample Pilot1", spec); } @@ -96,7 +96,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{ public void testMultipleSNPAlleles() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T UnifiedGenotyper --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH --dbsnp " + b37dbSNP129 + " -I " + privateTestDir + "multiallelic.snps.bam -o %s -L " + privateTestDir + "multiallelic.snps.intervals", 1, - Arrays.asList("f576d86656cc37c0a869c7ac911f4c7c")); + Arrays.asList("70a21812d4dd6b72c44f60c74d508d5b")); executeTest("test Multiple SNP alleles", spec); } @@ -112,7 +112,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{ public void testReverseTrim() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T UnifiedGenotyper --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + validationDataLocation + "CEUTrio.HiSeq.b37.chr20.10_11mb.bam -o %s -L 20:10289124 -L 20:10090289", 1, - Arrays.asList("94d7a907fdca7e8c9fd6bb8a87b2bab2")); + Arrays.asList("f3da1ff1e49a831af055ca52d6d07dd7")); executeTest("test reverse trim", spec); } @@ -120,7 +120,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{ public void testMismatchedPLs() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T UnifiedGenotyper --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + privateTestDir + "mismatchedPLs.bam -o %s -L 1:24020341", 1, - Arrays.asList("94bfccbd06043e90ae1b1c66fc3afe07")); + Arrays.asList("20ff311f363c51b7385a76f6f296759c")); executeTest("test mismatched PLs", spec); } } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperReducedReadsIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperReducedReadsIntegrationTest.java index b9830de8e..33810e255 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperReducedReadsIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperReducedReadsIntegrationTest.java @@ -63,18 +63,18 @@ public class UnifiedGenotyperReducedReadsIntegrationTest extends WalkerTest { public void testReducedBam() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T UnifiedGenotyper --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "bamExample.ReducedRead.ADAnnotation.bam -o %s -L 1:67,225,396-67,288,518", 1, - Arrays.asList("e6565060b44a7804935973efcd56e596")); + Arrays.asList("ffde0d5e23523e4bd9e7e18f62d37d0f")); executeTest("test calling on a ReducedRead BAM", spec); } @Test public void testReducedBamSNPs() { - testReducedCalling("SNP", "ab776d74c41ce2b859e2b2466a76204a"); + testReducedCalling("SNP", "e8de8c523751ad2fa2ee20185ba5dea7"); } @Test public void testReducedBamINDELs() { - testReducedCalling("INDEL", "22110b001e2d3dd45d7872334086b2b9"); + testReducedCalling("INDEL", "4b4902327fb132f9aaab3dd5ace934e1"); } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java index c1b8f8a70..0636d7c1b 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java @@ -64,7 +64,7 @@ public class HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest extends Wa @Test public void testHaplotypeCallerMultiSampleComplex1() { - HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "03944bbedb012e2ac2026a84baa0560c"); + HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "4a3479fc4ad387d381593b328f737a1b"); } private void HCTestSymbolicVariants(String bam, String args, String md5) { @@ -88,12 +88,12 @@ public class HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest extends Wa @Test public void testHaplotypeCallerMultiSampleGGAComplex() { HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:119673-119823 -L 20:121408-121538", - "321dc9f3d330790bac7981ffae00cb0c"); + "b7a01525c00d02b3373513a668a43c6a"); } @Test public void testHaplotypeCallerMultiSampleGGAMultiAllelic() { HCTestComplexGGA(NA12878_CHR20_BAM, "-L 20:133041-133161 -L 20:300207-300337", - "7e9f99d4cba8087dac66ea871b910d7e"); + "a2a42055b068334f415efb07d6bb9acd"); } } diff --git a/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerIntegrationTest.java b/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerIntegrationTest.java index da92f39fc..aca1172d4 100644 --- a/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerIntegrationTest.java +++ b/protected/java/test/org/broadinstitute/sting/gatk/walkers/haplotypecaller/HaplotypeCallerIntegrationTest.java @@ -78,12 +78,12 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { @Test public void testHaplotypeCallerMultiSample() { - HCTest(CEUTRIO_BAM, "", "09d84bc1aef2dd9c185934752172b794"); + HCTest(CEUTRIO_BAM, "", "baa5a2eedc8f06ce9f8f98411ee09f8a"); } @Test public void testHaplotypeCallerSingleSample() { - HCTest(NA12878_BAM, "", "5c074930b27d1f5c942fe755c2a8be27"); + HCTest(NA12878_BAM, "", "f09e03d41238697b23f95716a12667cb"); } @Test(enabled = false) // can't annotate the rsID's yet @@ -94,7 +94,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { @Test public void testHaplotypeCallerMultiSampleGGA() { HCTest(CEUTRIO_BAM, "--max_alternate_alleles 3 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles " + validationDataLocation + "combined.phase1.chr20.raw.indels.sites.vcf", - "005a6d1933913a5d96fc56d01303fa95"); + "130d36448faeb7b8d4bce4be12dacd3a"); } @Test @@ -110,7 +110,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { @Test public void testHaplotypeCallerSingleSampleIndelQualityScores() { - HCTestIndelQualityScores(NA12878_RECALIBRATED_BAM, "", "9b6f667ad87e19c38d16fefe63c37484"); + HCTestIndelQualityScores(NA12878_RECALIBRATED_BAM, "", "7c20aa62633f4ce8ebf12950fbf05ec0"); } private void HCTestNearbySmallIntervals(String bam, String args, String md5) { @@ -147,7 +147,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { @Test public void testHaplotypeCallerNearbySmallIntervals() { - HCTestNearbySmallIntervals(NA12878_BAM, "", "6e170d03047caefc2fba3f1c1f8de132"); + HCTestNearbySmallIntervals(NA12878_BAM, "", "0ddc56f0a0fbcfefda79aa20b2ecf603"); } // This problem bam came from a user on the forum and it spotted a problem where the ReadClipper @@ -186,7 +186,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { public void HCTestReducedBam() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T HaplotypeCaller --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "bamExample.ReducedRead.ADAnnotation.bam -o %s -L 1:67,225,396-67,288,518", 1, - Arrays.asList("a47ef09a8701128cfb301a83b7bb0728")); + Arrays.asList("5fe9310addf881bed4fde2354e59ce34")); executeTest("HC calling on a ReducedRead BAM", spec); } @@ -194,7 +194,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { public void testReducedBamWithReadsNotFullySpanningDeletion() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T HaplotypeCaller --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -I " + privateTestDir + "reduced.readNotFullySpanningDeletion.bam -o %s -L 1:167871297", 1, - Arrays.asList("0cb99f6bb3e630add4b3486c496fa508")); + Arrays.asList("26a9917f6707536636451266de0116c3")); executeTest("test calling on a ReducedRead BAM where the reads do not fully span a deletion", spec); } @@ -208,7 +208,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { public void HCTestDBSNPAnnotationWGS() { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T HaplotypeCaller --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -I " + NA12878_PCRFREE + " -o %s -L 20:10,000,000-10,100,000 -D " + b37dbSNP132, 1, - Arrays.asList("92f947cc89e4f50cf2ef3121d2fe308d")); + Arrays.asList("cc6f2a76ee97ecc14a5f956ffbb21d88")); executeTest("HC calling with dbSNP ID annotation on WGS intervals", spec); } @@ -217,7 +217,7 @@ public class HaplotypeCallerIntegrationTest extends WalkerTest { WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( "-T HaplotypeCaller --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -I " + NA12878_PCRFREE + " -o %s -L 20:10,000,000-11,000,000 -D " + b37dbSNP132 + " -L " + hg19Intervals + " -isr INTERSECTION", 1, - Arrays.asList("91877c8ea3eb0e0316d9ad11fdcc1a87")); + Arrays.asList("51e91c8af61a6b47807165906baefb00")); executeTest("HC calling with dbSNP ID annotation on WEx intervals", spec); } }