zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Alleles are now returned in a consistent order, so we can deal with tri-allelic sites 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@4002 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
ebanks 2010-08-10 15:21:10 +00:00
parent 67063deb16
commit 3ff6e3404e
4 changed files with 41 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
java/src/org/broadinstitute/sting/gatk/walkers/annotator/DepthPerAlleleBySample.java
View File

@ -2,8 +2,10 @@ package org.broadinstitute.sting.gatk.walkers.annotator;
import org.broad.tribble.util.variantcontext.Genotype;
import org.broad.tribble.util.variantcontext.VariantContext;
import org.broad.tribble.util.variantcontext.Allele;
import org.broad.tribble.vcf.VCFFormatHeaderLine;
import org.broad.tribble.vcf.VCFHeaderLineType;
import org.broad.tribble.vcf.VCFCompoundHeaderLine;
import org.broadinstitute.sting.gatk.contexts.*;
import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker;
import org.broadinstitute.sting.gatk.walkers.annotator.interfaces.*;
@ -18,9 +20,6 @@ public class DepthPerAlleleBySample implements GenotypeAnnotation, StandardAnnot
if ( g == null || !g.isCalled() )
return null;
if ( !vc.isBiallelic() )
return null;
if ( vc.isSNP() )
return annotateSNP(stratifiedContext, vc);
if ( vc.isIndel() )
@ -32,8 +31,8 @@ public class DepthPerAlleleBySample implements GenotypeAnnotation, StandardAnnot
private Map<String,Object> annotateSNP(StratifiedAlignmentContext stratifiedContext, VariantContext vc) {
HashMap<Byte, Integer> alleleCounts = new HashMap<Byte, Integer>();
alleleCounts.put(vc.getReference().getBases()[0], 0);
alleleCounts.put(vc.getAlternateAllele(0).getBases()[0], 0);
for ( Allele allele : vc.getAlleles() )
alleleCounts.put(allele.getBases()[0], 0);
ReadBackedPileup pileup = stratifiedContext.getContext(StratifiedAlignmentContext.StratifiedContextType.COMPLETE).getBasePileup();
for ( PileupElement p : pileup ) {
@ -42,10 +41,11 @@ public class DepthPerAlleleBySample implements GenotypeAnnotation, StandardAnnot
}
// we need to add counts in the correct order
Integer[] counts = new Integer[2];
Integer[] counts = new Integer[alleleCounts.size()];
counts[0] = alleleCounts.get(vc.getReference().getBases()[0]);
counts[1] = alleleCounts.get(vc.getAlternateAllele(0).getBases()[0]);
for (int i = 0; i < vc.getAlternateAlleles().size(); i++)
counts[i+1] = alleleCounts.get(vc.getAlternateAllele(i).getBases()[0]);
Map<String, Object> map = new HashMap<String, Object>();
map.put(getKeyNames().get(0), counts);
return map;
@ -76,5 +76,5 @@ public class DepthPerAlleleBySample implements GenotypeAnnotation, StandardAnnot
public List<String> getKeyNames() { return Arrays.asList("AD"); }
public List<VCFFormatHeaderLine> getDescriptions() { return Arrays.asList(new VCFFormatHeaderLine(getKeyNames().get(0), 2, VCFHeaderLineType.Integer, "Allelic depths for the ref and alt alleles; currently only works for bi-allelic sites")); }
public List<VCFFormatHeaderLine> getDescriptions() { return Arrays.asList(new VCFFormatHeaderLine(getKeyNames().get(0), VCFCompoundHeaderLine.UNBOUNDED, VCFHeaderLineType.Integer, "Allelic depths for the ref and alt alleles in the order listed")); }
}

16
java/test/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorIntegrationTest.java
View File

@ -31,7 +31,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testHasAnnotsAsking1() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample2.vcf -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -L 1:10,020,000-10,021,000", 1,
Arrays.asList("c2a52baffc7254f222697f0f274bddfa"));
Arrays.asList("b48199f9dc6f91c61418536151afa8fd"));
executeTest("test file has annotations, asking for annotations, #1", spec);
}
@ -39,7 +39,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testHasAnnotsAsking2() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample3.vcf -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -L 1:10,000,000-10,050,000", 1,
Arrays.asList("d9ec651bf6fa225f684719038f1c48c9"));
Arrays.asList("dcc7c522c4178b4fd9a0e5439bcdaebc"));
executeTest("test file has annotations, asking for annotations, #2", spec);
}
@ -63,7 +63,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testNoAnnotsAsking1() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample2empty.vcf -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -L 1:10,020,000-10,021,000", 1,
Arrays.asList("05658e763022cbfdcef4d2e7c0b02b4e"));
Arrays.asList("133275d150a8100ba4dc756d17b23ef1"));
executeTest("test file doesn't have annotations, asking for annotations, #1", spec);
}
@ -71,7 +71,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testNoAnnotsAsking2() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample3empty.vcf -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -L 1:10,000,000-10,050,000", 1,
Arrays.asList("d11ff0ac631f214b6023ab0232be4b14"));
Arrays.asList("120fc0d4af1d370f1b306700258464b9"));
executeTest("test file doesn't have annotations, asking for annotations, #2", spec);
}
@ -79,7 +79,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testOverwritingHeader() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample4.vcf -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -L 1:10,001,292", 1,
Arrays.asList("20cb9e8b7d883ed945e13845e84153ad"));
Arrays.asList("10c31f8ab903843538a7604ed1e5405c"));
executeTest("test overwriting header", spec);
}
@ -87,7 +87,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testNoReads() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample3empty.vcf -BTI variant", 1,
Arrays.asList("edde543d79202c3cc5a76a654f88c7de"));
Arrays.asList("a6ec667a656e7ba368de9dbae781eef3"));
executeTest("not passing it any reads", spec);
}
@ -95,7 +95,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testDBTagWithDbsnp() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -D " + GATKDataLocation + "dbsnp_129_b36.rod -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample3empty.vcf -BTI variant", 1,
Arrays.asList("dee4c3065c832004201104a314637d18"));
Arrays.asList("8b4c309fe4900b5c8ae720c649715c7d"));
executeTest("getting DB tag with dbSNP", spec);
}
@ -103,7 +103,7 @@ public class VariantAnnotatorIntegrationTest extends WalkerTest {
public void testDBTagWithHapMap() {
WalkerTestSpec spec = new WalkerTestSpec(
baseTestString() + " -B compH3,VCF," + validationDataLocation + "fakeHM3.vcf -G \"Standard\" -B variant,VCF," + validationDataLocation + "vcfexample3empty.vcf -BTI variant", 1,
Arrays.asList("8eefa78f0137d6a1f247e336655e6d4c"));
Arrays.asList("ff6b3468f21b262de671e823349cbb3c"));
executeTest("getting DB tag with HM3", spec);
}
}

32
java/test/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperIntegrationTest.java
View File

@ -24,7 +24,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
public void testMultiSamplePilot1Joint() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -varout %s -L 1:10,022,000-10,025,000", 1,
Arrays.asList("73d514f53f1630832b3bed65c67fe869"));
Arrays.asList("99ff7d7031cc5038ec95f7872311c6b5"));
executeTest("testMultiSamplePilot1 - Joint Estimate", spec);
}
@ -32,7 +32,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
public void testMultiSamplePilot2Joint() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "pilot2_daughters.chr20.10k-11k.bam -varout %s -L 20:10,000,000-10,050,000", 1,
Arrays.asList("e21ff78ca74bdaacaea562795a90e979"));
Arrays.asList("497ac35ab217d5c0c21ee7733219ef5d"));
executeTest("testMultiSamplePilot2 - Joint Estimate", spec);
}
@ -40,7 +40,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
public void testSingleSamplePilot2Joint() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -varout %s -L 1:10,000,000-10,100,000", 1,
Arrays.asList("20878adf2f56687ef03f32e6bac8fb30"));
Arrays.asList("29670c1c6ae4a1e7fc27c4d78c434a72"));
executeTest("testSingleSamplePilot2 - Joint Estimate", spec);
}
@ -52,7 +52,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
@Test
public void testParallelization() {
String md5 = "664f02681a0ff09bdf3269f97284009c";
String md5 = "c6a4347807f624c2a31fb95080ed68b8";
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -varout %s -L 1:10,000,000-10,075,000", 1,
@ -79,11 +79,11 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
@Test
public void testParameter() {
HashMap<String, String> e = new HashMap<String, String>();
e.put( "-genotype", "5d3f4a25039959e1758a035b8dc595b7" );
e.put( "-all_bases", "402740e2e1ef23a246120dde91a42dcd" );
e.put( "--min_base_quality_score 26", "a6fe67665b1ffe1d53194199b1d28b9a" );
e.put( "--min_mapping_quality_score 26", "3fcb922f5939221ee73661f83d604050" );
e.put( "--max_mismatches_in_40bp_window 5", "93a0d181625c2c98c35acd03599f6f8b" );
e.put( "-genotype", "fefaff399971f080c1a5393ea7f6d5d1" );
e.put( "-all_bases", "300decbc3ebd4022da1045cb6c1bdbbb" );
e.put( "--min_base_quality_score 26", "3fd8b652f2fca1dbd9156aec5978a90c" );
e.put( "--min_mapping_quality_score 26", "7d50c7cd55b3a9dee6aca3980c4612a8" );
e.put( "--max_mismatches_in_40bp_window 5", "3e347f056ed9cf2e64f2060771244a6c" );
for ( Map.Entry<String, String> entry : e.entrySet() ) {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
@ -97,12 +97,12 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
public void testConfidence() {
WalkerTest.WalkerTestSpec spec1 = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -varout %s -L 1:10,000,000-10,010,000 -stand_call_conf 10 ", 1,
Arrays.asList("292ecd223494a2cff64a9cb72b0850d9"));
Arrays.asList("0937f45888cab1aacfa129d45d82384f"));
executeTest("testConfidence1", spec1);
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommand + " -I " + validationDataLocation + "NA12878.1kg.p2.chr1_10mb_11_mb.SLX.bam -varout %s -L 1:10,000,000-10,010,000 -stand_emit_conf 10 ", 1,
Arrays.asList("65e29fad296e3d03312087e7e9168095"));
Arrays.asList("99419b1852b744092bc7418747dc78f3"));
executeTest("testConfidence2", spec2);
}
@ -114,8 +114,8 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
@Test
public void testHeterozyosity() {
HashMap<Double, String> e = new HashMap<Double, String>();
e.put( 0.01, "79be222f9c5fa7091723b8172c236a22" );
e.put( 1.0 / 1850, "e217d213a7ebaf9c4bd41f79234055cc" );
e.put( 0.01, "f64f88e8c085ff21db1570bb5cd74f8a" );
e.put( 1.0 / 1850, "0804d411ed7cf77c79e1ff540b8b85f1" );
for ( Map.Entry<Double, String> entry : e.entrySet() ) {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
@ -134,8 +134,8 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
@Test
public void testOtherBaseCallModel() {
HashMap<String, String> e = new HashMap<String, String>();
e.put( "one_state", "8e7a18b71c08d655df15b3e5204ae924" );
e.put( "three_state", "eacf6c82bf81d20f3fc4a74051869375" );
e.put( "one_state", "a2b9f07c406ff60b81b42a8d5da45f10" );
e.put( "three_state", "5a1b3c215e18686a4254849efcb45c40" );
for ( Map.Entry<String, String> entry : e.entrySet() ) {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
@ -158,7 +158,7 @@ public class UnifiedGenotyperIntegrationTest extends WalkerTest {
" -varout %s" +
" -L 1:10,000,000-10,100,000",
1,
Arrays.asList("99102fb35f0f33f582da8dbde64d80cd"));
Arrays.asList("ce22e33e550ea62d8786deb38a70049a"));
executeTest(String.format("testMultiTechnologies"), spec);
}

16
java/test/org/broadinstitute/sting/gatk/walkers/variantutils/CombineVariantsIntegrationTest.java
View File

@ -60,21 +60,21 @@ public class CombineVariantsIntegrationTest extends WalkerTest {
}
@Test public void test1SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "bbeb813ff559b630570725419e4e1adc"); }
@Test public void test2SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "5039478bd5bb05b4c4e8309cdd887e5f", " -setKey foo"); }
@Test public void test3SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "8502338872d241857379fd25878a7700", " -setKey null"); }
@Test public void test1SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "f203232c2fa51862e911940ad9d60387"); }
@Test public void test2SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "29aa605c3f08f3ad58f5eea64dc709c1", " -setKey foo"); }
@Test public void test3SNP() { test1InOut("pilot2.snps.vcf4.genotypes.vcf", "445bff7b13b31c900492f1ccaed62a80", " -setKey null"); }
@Test public void testOfficialCEUPilotCalls() { test1InOut("CEU.trio.2010_03.genotypes.vcf.gz", "38b7e64b91c726867a604cf95b9cb10a"); } // official project VCF files in tabix format
@Test public void test1Indel1() { test1InOut("CEU.dindel.vcf4.trio.2010_06.indel.genotypes.vcf", "4c667935099544a1863e70ae88ddd685"); }
@Test public void test1Indel1() { test1InOut("CEU.dindel.vcf4.trio.2010_06.indel.genotypes.vcf", "cdf66ad481b7f98204e368b968d6d8ec"); }
@Test public void test1Indel2() { test1InOut("CEU.dindel.vcf4.low_coverage.2010_06.indel.genotypes.vcf", "17c8468b1b963c9abc49dff17fd811ba"); }
@Test public void combineTrioCalls() { combine2("CEU.trio.2010_03.genotypes.vcf.gz", "YRI.trio.2010_03.genotypes.vcf.gz", "", "f3fce9ae729548e7e7c378a8282df235"); } // official project VCF files in tabix format
@Test public void combineTrioCalls() { combine2("CEU.trio.2010_03.genotypes.vcf.gz", "YRI.trio.2010_03.genotypes.vcf.gz", "", "27b509c5c64c5b7520d348bceaca67f5"); } // official project VCF files in tabix format
@Test public void combineTrioCallsMin() { combine2("CEU.trio.2010_03.genotypes.vcf.gz", "YRI.trio.2010_03.genotypes.vcf.gz", " -minimalVCF", "7ad58764b855ec7ad61075dda63567b3"); } // official project VCF files in tabix format
@Test public void combine2Indels() { combine2("CEU.dindel.vcf4.trio.2010_06.indel.genotypes.vcf", "CEU.dindel.vcf4.low_coverage.2010_06.indel.genotypes.vcf", "", "7fce3e9539ce4eb0a839d9236c928dfc"); }
@Test public void combine2Indels() { combine2("CEU.dindel.vcf4.trio.2010_06.indel.genotypes.vcf", "CEU.dindel.vcf4.low_coverage.2010_06.indel.genotypes.vcf", "", "a438947372f5e748931d9c5e2ba5fc3a"); }
@Test public void combineSNPsAndIndels() { combine2("CEU.trio.2010_03.genotypes.vcf.gz", "CEU.dindel.vcf4.low_coverage.2010_06.indel.genotypes.vcf", "", "528cff763b95b82717008d2c05e8c393"); }
@Test public void combineSNPsAndIndels() { combine2("CEU.trio.2010_03.genotypes.vcf.gz", "CEU.dindel.vcf4.low_coverage.2010_06.indel.genotypes.vcf", "", "19f8c7ed0ed3b59c53ac76164679b7f5"); }
@Test public void uniqueSNPs() { combine2("pilot2.snps.vcf4.genotypes.vcf", "yri.trio.gatk_glftrio.intersection.annotated.filtered.chr1.vcf", "", "a9126d1cbe1fdf741236763fb3e3461f"); }
@Test public void uniqueSNPs() { combine2("pilot2.snps.vcf4.genotypes.vcf", "yri.trio.gatk_glftrio.intersection.annotated.filtered.chr1.vcf", "", "96383d62d6b9f0e7ee2d3637a985af28"); }
@Test public void threeWayWithRefs() {
WalkerTestSpec spec = new WalkerTestSpec(