ReadLikelihoods class introduction final changes before merging

Stories:

        https://www.pivotaltracker.com/story/show/70222086
        https://www.pivotaltracker.com/story/show/67961652

Changes:

  Done some changes that I missed in relation with making sure that all PairHMM implentations use the same interface; as a consequence we were running always the standard PairHMM.
  Fixed some additional bugs detected when running it on full wgs single sample and exom multi sample data set.
  Updated some integration test md5s.
This commit is contained in:
Valentin Ruano-Rubio 2014-08-11 17:47:25 -04:00
parent 9a9a68409e
commit b39508cd15
5 changed files with 15 additions and 15 deletions

View File

@ -447,8 +447,8 @@ public class PairHMMIndelErrorModel {
final List<Haplotype> distinctHaplotypesList = Arrays.asList(distinctHaplotypesSet.toArray(new Haplotype[distinctHaplotypesSet.size()])); final List<Haplotype> distinctHaplotypesList = Arrays.asList(distinctHaplotypesSet.toArray(new Haplotype[distinctHaplotypesSet.size()]));
// Get the likelihoods for our clipped read against each of our trimmed haplotypes. // Get the likelihoods for our clipped read against each of our trimmed haplotypes.
final ReadLikelihoods<Haplotype> rl = new ReadLikelihoods<>( final ReadLikelihoods<Haplotype> rl = new ReadLikelihoods<>(
Collections.singletonList("DUMMY_SAMPLE"),distinctHaplotypesList,Collections.singletonMap("DUMMY_SAMPLE",Collections.singletonList(processedRead))); Collections.singletonList("DUMMY_SAMPLE"),distinctHaplotypesList,Collections.singletonMap("DUMMY_SAMPLE",Collections.singletonList(processedRead)));
final ReadLikelihoods.Matrix<Haplotype> dummySampleLikelihoods = rl.sampleMatrix(0); final ReadLikelihoods.Matrix<Haplotype> dummySampleLikelihoods = rl.sampleMatrix(0);
pairHMM.computeLikelihoods(rl.sampleMatrix(0), Collections.singletonList(processedRead), readGCPArrayMap); pairHMM.computeLikelihoods(rl.sampleMatrix(0), Collections.singletonList(processedRead), readGCPArrayMap);
@ -458,7 +458,7 @@ public class PairHMMIndelErrorModel {
final int hIndex = rl.alleleIndex(h); final int hIndex = rl.alleleIndex(h);
final double readLikelihood = dummySampleLikelihoods.get(hIndex,0); final double readLikelihood = dummySampleLikelihoods.get(hIndex,0);
readLikelihoods[readIdx][j++] = readLikelihood; readLikelihoods[readIdx][j++] = readLikelihood;
perReadAlleleLikelihoodMap.add(read,a,readLikelihood); perReadAlleleLikelihoodMap.add(p,a,readLikelihood);
} }
} }
} }

View File

@ -79,6 +79,6 @@ public class UnifiedGenotyperGeneralPloidySuite1IntegrationTest extends WalkerTe
@Test(enabled = true) @Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() { public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() {
executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1", "LSV_INDEL_DISC_NOREF_p1", "INDEL", "a9730f883a2e01e93d56b11321c9fd52"); executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1", "LSV_INDEL_DISC_NOREF_p1", "INDEL", "ae90d6be28c19e455083a47bc95c3b1b");
} }
} }

View File

@ -58,7 +58,7 @@ public class UnifiedGenotyperGeneralPloidySuite2IntegrationTest extends WalkerTe
@Test(enabled = true) @Test(enabled = true)
public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() { public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() {
executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","93cf7ed2645599d0ddd319c6989eadcc"); executor.PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","9fc4f04105bde31bafc93548745cb67e");
} }
@Test(enabled = true) @Test(enabled = true)

View File

@ -73,7 +73,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
" -o %s" + " -o %s" +
" -L 1:10,000,000-10,500,000", " -L 1:10,000,000-10,500,000",
1, 1,
Arrays.asList("24d5f65860507ec7f8d9441bfee8653a")); Arrays.asList("8a4de9e1f59cffe80a4372cf02fe809e"));
executeTest(String.format("test indel caller in SLX"), spec); executeTest(String.format("test indel caller in SLX"), spec);
} }
@ -100,7 +100,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
" -o %s" + " -o %s" +
" -L 1:10,000,000-10,500,000", " -L 1:10,000,000-10,500,000",
1, 1,
Arrays.asList("188dc4d70dc5f4de09f645270cc3f4e1")); Arrays.asList("2b92df91a9337b9d9f03db5699bb41f2"));
executeTest(String.format("test indel calling, multiple technologies"), spec); executeTest(String.format("test indel calling, multiple technologies"), spec);
} }
@ -110,7 +110,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation + 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, "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("31aaad9adbf12bb7fb12d53080e5c2f5")); Arrays.asList("591c9a7bd7d64974b4a2dd932fdef138"));
executeTest("test MultiSample Pilot2 indels with alleles passed in", spec); executeTest("test MultiSample Pilot2 indels with alleles passed in", spec);
} }
@ -120,7 +120,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
baseCommandIndels + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " baseCommandIndels + " --output_mode EMIT_ALL_SITES --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles "
+ privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation + + privateTestDir + "indelAllelesForUG.vcf -I " + validationDataLocation +
"pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1, "pilot2_daughters.chr20.10k-11k.bam -o %s -L 20:10,000,000-10,100,000", 1,
Arrays.asList("fecb8ae4e8f5ea2b2bf3826d674298f7")); Arrays.asList("e784b6d1b3725f72a2c42b19e8d24cb1"));
executeTest("test MultiSample Pilot2 indels with alleles passed in and emitting all sites", spec); executeTest("test MultiSample Pilot2 indels with alleles passed in and emitting all sites", spec);
} }
@ -135,7 +135,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec2 = new WalkerTest.WalkerTestSpec(
baseCommandIndels + " --genotyping_mode GENOTYPE_GIVEN_ALLELES -alleles " + result.get(0).getAbsolutePath() + " -I " + validationDataLocation + 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, "low_coverage_CEU.chr1.10k-11k.bam -o %s -L " + result.get(0).getAbsolutePath(), 1,
Arrays.asList("0662ec6fe6d6e9807f758b1c82adf056")); Arrays.asList("d101beb3d1f71a14a31438cce9786881"));
executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2); executeTest("test MultiSample Pilot1 CEU indels using GENOTYPE_GIVEN_ALLELES", spec2);
} }
@ -176,7 +176,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
public void testMinIndelFraction0() { public void testMinIndelFraction0() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.0", 1, assessMinIndelFraction + " -minIndelFrac 0.0", 1,
Arrays.asList("3d5140ffc02a125ee26c63ec55449192")); Arrays.asList("96c2ab043f523c7676e4bc36dd3c9d0b"));
executeTest("test minIndelFraction 0.0", spec); executeTest("test minIndelFraction 0.0", spec);
} }
@ -184,7 +184,7 @@ public class UnifiedGenotyperIndelCallingIntegrationTest extends WalkerTest {
public void testMinIndelFraction25() { public void testMinIndelFraction25() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
assessMinIndelFraction + " -minIndelFrac 0.25", 1, assessMinIndelFraction + " -minIndelFrac 0.25", 1,
Arrays.asList("7c62eb9142d9124e49c123dcc052ce57")); Arrays.asList("53ce3aa9ee16302ab2ec21f742b5e157"));
executeTest("test minIndelFraction 0.25", spec); executeTest("test minIndelFraction 0.25", spec);
} }

View File

@ -65,7 +65,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{
public void testMultiSamplePilot1() { public void testMultiSamplePilot1() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( 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, baseCommand + " -I " + validationDataLocation + "low_coverage_CEU.chr1.10k-11k.bam -o %s -L 1:10,022,000-10,025,000", 1,
Arrays.asList("c6462121e5a419546f50d53646274fd3")); Arrays.asList("058b54cff14527485c7b1ab035ebb6c4"));
executeTest("test MultiSample Pilot1", spec); executeTest("test MultiSample Pilot1", spec);
} }
@ -97,7 +97,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{
public void testMultipleSNPAlleles() { public void testMultipleSNPAlleles() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm BOTH --dbsnp " + b37dbSNP129 + " -I " + privateTestDir + "multiallelic.snps.bam -o %s -L " + privateTestDir + "multiallelic.snps.intervals", 1, "-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --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("85e8ba3cfeef5e3b72418bcfaec11668")); Arrays.asList("120270ddf48dfd461f756c89ea1ab074"));
executeTest("test Multiple SNP alleles", spec); executeTest("test Multiple SNP alleles", spec);
} }
@ -113,7 +113,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{
public void testReverseTrim() { public void testReverseTrim() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --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, "-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --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("bd867fb341e75a9f1fc78d25bda93f63")); Arrays.asList("bc5a143868e3ad3acc9bb7c09798cdf2"));
executeTest("test reverse trim", spec); executeTest("test reverse trim", spec);
} }
@ -121,7 +121,7 @@ public class UnifiedGenotyperNormalCallingIntegrationTest extends WalkerTest{
public void testMismatchedPLs() { public void testMismatchedPLs() {
WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec( WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
"-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + privateTestDir + "mismatchedPLs.bam -o %s -L 1:24020341", 1, "-T UnifiedGenotyper --contamination_fraction_to_filter 0.05 --disableDithering -R " + b37KGReference + " --no_cmdline_in_header -glm INDEL -I " + privateTestDir + "mismatchedPLs.bam -o %s -L 1:24020341", 1,
Arrays.asList("e87419cd42fb9684ee3674ee54abe567")); Arrays.asList("02fd6357bd1082115822c8a931cbb6a3"));
executeTest("test mismatched PLs", spec); executeTest("test mismatched PLs", spec);
} }
} }