Rename per-sample ML allelic fractions and counts so that they don't have the same name as the per-site INFO fields, and clarify wording in VCF header

This commit is contained in:
Guillermo del Angel 2012-08-09 09:58:20 -04:00
parent 3e2752667c
commit 71ee8d87b3
4 changed files with 12 additions and 10 deletions

View File

@ -681,8 +681,8 @@ public class GeneralPloidyExactAFCalculationModel extends AlleleFrequencyCalcula
}
// per-pool logging of AC and AF
gb.attribute(VCFConstants.MLE_ALLELE_COUNT_KEY, alleleCounts.size() == 1 ? alleleCounts.get(0) : alleleCounts);
gb.attribute(VCFConstants.MLE_ALLELE_FREQUENCY_KEY, alleleFreqs.size() == 1 ? alleleFreqs.get(0) : alleleFreqs);
gb.attribute(VCFConstants.MLE_PER_SAMPLE_ALLELE_COUNT_KEY, alleleCounts.size() == 1 ? alleleCounts.get(0) : alleleCounts);
gb.attribute(VCFConstants.MLE_PER_SAMPLE_ALLELE_FRACTION_KEY, alleleFreqs.size() == 1 ? alleleFreqs.get(0) : alleleFreqs);
// remove PLs if necessary
if (newLikelihoods.length > MAX_LENGTH_FOR_POOL_PL_LOGGING)

View File

@ -46,33 +46,33 @@ public class UnifiedGenotyperGeneralPloidyIntegrationTest extends WalkerTest {
@Test
public void testBOTH_GGA_Pools() {
PC_LSV_Test(String.format(" -maxAltAlleles 2 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_BOTH_GGA","BOTH","d76e3b910259da819f1e1b2adc68ba8d");
PC_LSV_Test(String.format(" -maxAltAlleles 2 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_BOTH_GGA","BOTH","0934f72865388999efec64bd9d4a9b93");
}
@Test
public void testINDEL_GGA_Pools() {
PC_LSV_Test(String.format(" -maxAltAlleles 1 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_INDEL_GGA","INDEL","ffadcdaee613dab975197bed0fc78da3");
PC_LSV_Test(String.format(" -maxAltAlleles 1 -ploidy 24 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",LSV_ALLELES),"LSV_INDEL_GGA","INDEL","126581c72d287722437274d41b6fed7b");
}
@Test
public void testINDEL_maxAltAlleles2_ploidy3_Pools_noRef() {
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","96087fe9240e3656cc2a4e0ff0174d5b");
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 3","LSV_INDEL_DISC_NOREF_p3","INDEL","b543aa1c3efedb301e525c1d6c50ed8d");
}
@Test
public void testINDEL_maxAltAlleles2_ploidy1_Pools_noRef() {
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1","LSV_INDEL_DISC_NOREF_p1","INDEL","6fdae7093831ecfc82a06dd707d62fe9");
PC_LSV_Test_NoRef(" -maxAltAlleles 2 -ploidy 1","LSV_INDEL_DISC_NOREF_p1","INDEL","55b20557a836bb92688e68f12d7f5dc4");
}
@Test
public void testMT_SNP_DISCOVERY_sp4() {
PC_MT_Test(CEUTRIO_BAM, " -maxAltAlleles 1 -ploidy 8", "MT_SNP_DISCOVERY_sp4","6b27634214530d379db70391a9cfc2d7");
PC_MT_Test(CEUTRIO_BAM, " -maxAltAlleles 1 -ploidy 8", "MT_SNP_DISCOVERY_sp4","7eb889e8e07182f4c3d64609591f9459");
}
@Test
public void testMT_SNP_GGA_sp10() {
PC_MT_Test(CEUTRIO_BAM, String.format(" -maxAltAlleles 1 -ploidy 20 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",NA12891_CALLS), "MT_SNP_GGA_sp10", "e74d4c73ece45d7fb676b99364df4f1a");
PC_MT_Test(CEUTRIO_BAM, String.format(" -maxAltAlleles 1 -ploidy 20 -gt_mode GENOTYPE_GIVEN_ALLELES -out_mode EMIT_ALL_SITES -alleles %s",NA12891_CALLS), "MT_SNP_GGA_sp10", "db8114877b99b14f7180fdcd24b040a7");
}
}

View File

@ -312,8 +312,8 @@ public class UnifiedGenotyper extends LocusWalker<List<VariantCallContext>, Unif
// add the pool values for each genotype
if (UAC.samplePloidy != VariantContextUtils.DEFAULT_PLOIDY) {
headerInfo.add(new VCFFormatHeaderLine(VCFConstants.MLE_ALLELE_COUNT_KEY, VCFHeaderLineCount.A, VCFHeaderLineType.Integer, "Maximum likelihood expectation (MLE) for the allele counts (not necessarily the same as the AC), for each ALT allele, in the same order as listed, for this pool"));
headerInfo.add(new VCFFormatHeaderLine(VCFConstants.MLE_ALLELE_FREQUENCY_KEY, VCFHeaderLineCount.A, VCFHeaderLineType.Float, "Maximum likelihood expectation (MLE) for the allele frequency (not necessarily the same as the AF), for each ALT allele, in the same order as listed, for this pool"));
headerInfo.add(new VCFFormatHeaderLine(VCFConstants.MLE_PER_SAMPLE_ALLELE_COUNT_KEY, VCFHeaderLineCount.A, VCFHeaderLineType.Integer, "Maximum likelihood expectation (MLE) for the alternate allele count, in the same order as listed, for each individual sample"));
headerInfo.add(new VCFFormatHeaderLine(VCFConstants.MLE_PER_SAMPLE_ALLELE_FRACTION_KEY, VCFHeaderLineCount.A, VCFHeaderLineType.Float, "Maximum likelihood expectation (MLE) for the alternate allele fraction, in the same order as listed, for each individual sample"));
}
if (UAC.referenceSampleName != null) {
headerInfo.add(new VCFInfoHeaderLine(VCFConstants.REFSAMPLE_DEPTH_KEY, 1, VCFHeaderLineType.Integer, "Total reference sample depth"));

View File

@ -36,6 +36,8 @@ public final class VCFConstants {
public static final String MLE_ALLELE_COUNT_KEY = "MLEAC";
public static final String ALLELE_FREQUENCY_KEY = "AF";
public static final String MLE_ALLELE_FREQUENCY_KEY = "MLEAF";
public static final String MLE_PER_SAMPLE_ALLELE_COUNT_KEY = "MLPSAC";
public static final String MLE_PER_SAMPLE_ALLELE_FRACTION_KEY = "MLPSAF";
public static final String ALLELE_NUMBER_KEY = "AN";
public static final String RMS_BASE_QUALITY_KEY = "BQ";
public static final String CIGAR_KEY = "CIGAR";