public Map<String, Object> annotate(
RefMetaDataTracker tracker,
AnnotatorCompatible walker,
ReferenceContext ref,
Map<String, AlignmentContext> stratifiedContexts,
VariantContext vc) {
if (mendelianViolation == null) {
if (checkAndSetSamples(((Walker) walker).getSampleDB())) {
mendelianViolation =
new MendelianViolation(((VariantAnnotator) walker).minGenotypeQualityP);
} else {
throw new UserException(
"Mendelian violation annotation can only be used from the Variant Annotator, and must be provided a valid PED file (-ped) from the command line containing only 1 trio.");
}
}
Map<String, Object> toRet = new HashMap<String, Object>(1);
boolean hasAppropriateGenotypes =
vc.hasGenotype(motherId)
&& vc.getGenotype(motherId).hasLikelihoods()
&& vc.hasGenotype(fatherId)
&& vc.getGenotype(fatherId).hasLikelihoods()
&& vc.hasGenotype(childId)
&& vc.getGenotype(childId).hasLikelihoods();
if (hasAppropriateGenotypes)
toRet.put(
"MVLR", mendelianViolation.violationLikelihoodRatio(vc, motherId, fatherId, childId));
return toRet;
}
private boolean isConcordant(VariantContext vc, Collection<VariantContext> compVCs) {
if (vc == null || compVCs == null || compVCs.isEmpty()) return false;
// if we're not looking for specific samples then the fact that we have both VCs is enough to
// call it concordant.
if (NO_SAMPLES_SPECIFIED) return true;
// make a list of all samples contained in this variant VC that are being tracked by the user
// command line arguments.
Set<String> variantSamples = vc.getSampleNames();
variantSamples.retainAll(samples);
// check if we can find all samples from the variant rod in the comp rod.
for (String sample : variantSamples) {
boolean foundSample = false;
for (VariantContext compVC : compVCs) {
Genotype varG = vc.getGenotype(sample);
Genotype compG = compVC.getGenotype(sample);
if (haveSameGenotypes(varG, compG)) {
foundSample = true;
break;
}
}
// if at least one sample doesn't have the same genotype, we don't have concordance
if (!foundSample) {
return false;
}
}
return true;
}
/**
* Checks if vc has a variant call for (at least one of) the samples.
*
* @param vc the variant rod VariantContext. Here, the variant is the dataset you're looking for
* discordances to (e.g. HapMap)
* @param compVCs the comparison VariantContext (discordance
* @return
*/
private boolean isDiscordant(VariantContext vc, Collection<VariantContext> compVCs) {
if (vc == null) return false;
// if we're not looking at specific samples then the absence of a compVC means discordance
if (NO_SAMPLES_SPECIFIED) return (compVCs == null || compVCs.isEmpty());
// check if we find it in the variant rod
Map<String, Genotype> genotypes = vc.getGenotypes(samples);
for (Genotype g : genotypes.values()) {
if (sampleHasVariant(g)) {
// There is a variant called (or filtered with not exclude filtered option set) that is not
// HomRef for at least one of the samples.
if (compVCs == null) return true;
// Look for this sample in the all vcs of the comp ROD track.
boolean foundVariant = false;
for (VariantContext compVC : compVCs) {
if (sampleHasVariant(compVC.getGenotype(g.getSampleName()))) {
foundVariant = true;
break;
}
}
// if (at least one sample) was not found in all VCs of the comp ROD, we have discordance
if (!foundVariant) return true;
}
}
return false; // we only get here if all samples have a variant in the comp rod.
}
static boolean allSamplesAreMergeable(VariantContext vc1, VariantContext vc2) {
// Check that each sample's genotype in vc2 is uniquely appendable onto its genotype in vc1:
for (final Genotype gt1 : vc1.getGenotypes()) {
Genotype gt2 = vc2.getGenotype(gt1.getSampleName());
if (!alleleSegregationIsKnown(gt1, gt2)) // can merge if: phased, or if either is a hom
return false;
}
return true;
}
/**
* Helper method to subset a VC record, modifying some metadata stored in the INFO field (i.e. AN,
* AC, AF).
*
* @param vc the VariantContext record to subset
* @param samples the samples to extract
* @return the subsetted VariantContext
*/
private VariantContext subsetRecord(VariantContext vc, Set<String> samples) {
if (samples == null || samples.isEmpty()) return vc;
ArrayList<Genotype> genotypes = new ArrayList<Genotype>();
for (Map.Entry<String, Genotype> genotypePair : vc.getGenotypes().entrySet()) {
if (samples.contains(genotypePair.getKey())) genotypes.add(genotypePair.getValue());
}
VariantContext sub = vc.subContextFromGenotypes(genotypes, vc.getAlleles());
// if we have fewer alternate alleles in the selected VC than in the original VC, we need to
// strip out the GL/PLs (because they are no longer accurate)
if (vc.getAlleles().size() != sub.getAlleles().size())
sub = VariantContext.modifyGenotypes(sub, VariantContextUtils.stripPLs(vc.getGenotypes()));
HashMap<String, Object> attributes = new HashMap<String, Object>(sub.getAttributes());
int depth = 0;
for (String sample : sub.getSampleNames()) {
Genotype g = sub.getGenotype(sample);
if (g.isNotFiltered() && g.isCalled()) {
String dp = (String) g.getAttribute("DP");
if (dp != null
&& !dp.equals(VCFConstants.MISSING_DEPTH_v3)
&& !dp.equals(VCFConstants.MISSING_VALUE_v4)) {
depth += Integer.valueOf(dp);
}
}
}
if (KEEP_ORIGINAL_CHR_COUNTS) {
if (attributes.containsKey(VCFConstants.ALLELE_COUNT_KEY))
attributes.put("AC_Orig", attributes.get(VCFConstants.ALLELE_COUNT_KEY));
if (attributes.containsKey(VCFConstants.ALLELE_FREQUENCY_KEY))
attributes.put("AF_Orig", attributes.get(VCFConstants.ALLELE_FREQUENCY_KEY));
if (attributes.containsKey(VCFConstants.ALLELE_NUMBER_KEY))
attributes.put("AN_Orig", attributes.get(VCFConstants.ALLELE_NUMBER_KEY));
}
VariantContextUtils.calculateChromosomeCounts(sub, attributes, false);
attributes.put("DP", depth);
sub = VariantContext.modifyAttributes(sub, attributes);
return sub;
}
static boolean someSampleHasDoubleNonReferenceAllele(VariantContext vc1, VariantContext vc2) {
for (final Genotype gt1 : vc1.getGenotypes()) {
Genotype gt2 = vc2.getGenotype(gt1.getSampleName());
List<Allele> site1Alleles = gt1.getAlleles();
List<Allele> site2Alleles = gt2.getAlleles();
Iterator<Allele> all2It = site2Alleles.iterator();
for (Allele all1 : site1Alleles) {
Allele all2 = all2It.next(); // this is OK, since allSamplesAreMergeable()
if (all1.isNonReference() && all2.isNonReference()) // corresponding alleles are alternate
return true;
}
}
return false;
}
static boolean doubleAllelesSegregatePerfectlyAmongSamples(
VariantContext vc1, VariantContext vc2) {
// Check that Alleles at vc1 and at vc2 always segregate together in all samples (including
// reference):
Map<Allele, Allele> allele1ToAllele2 = new HashMap<Allele, Allele>();
Map<Allele, Allele> allele2ToAllele1 = new HashMap<Allele, Allele>();
// Note the segregation of the alleles for the reference genome:
allele1ToAllele2.put(vc1.getReference(), vc2.getReference());
allele2ToAllele1.put(vc2.getReference(), vc1.getReference());
// Note the segregation of the alleles for each sample (and check that it is consistent with the
// reference and all previous samples).
for (final Genotype gt1 : vc1.getGenotypes()) {
Genotype gt2 = vc2.getGenotype(gt1.getSampleName());
List<Allele> site1Alleles = gt1.getAlleles();
List<Allele> site2Alleles = gt2.getAlleles();
Iterator<Allele> all2It = site2Alleles.iterator();
for (Allele all1 : site1Alleles) {
Allele all2 = all2It.next();
Allele all1To2 = allele1ToAllele2.get(all1);
if (all1To2 == null) allele1ToAllele2.put(all1, all2);
else if (!all1To2.equals(all2)) // all1 segregates with two different alleles at site 2
return false;
Allele all2To1 = allele2ToAllele1.get(all2);
if (all2To1 == null) allele2ToAllele1.put(all2, all1);
else if (!all2To1.equals(all1)) // all2 segregates with two different alleles at site 1
return false;
}
}
return true;
}
/**
* add the genotype data
*
* @param vc the variant context
* @param genotypeFormatKeys Genotype formatting string
* @param alleleMap alleles for this context
* @throws IOException for writer
*/
private void addGenotypeData(
VariantContext vc, Map<Allele, String> alleleMap, List<String> genotypeFormatKeys)
throws IOException {
for (String sample : mHeader.getGenotypeSamples()) {
mWriter.write(VCFConstants.FIELD_SEPARATOR);
Genotype g = vc.getGenotype(sample);
if (g == null) {
// TODO -- The VariantContext needs to know what the general ploidy is of the samples
// TODO -- We shouldn't be assuming diploid genotypes here!
mWriter.write(VCFConstants.EMPTY_GENOTYPE);
continue;
}
List<String> attrs = new ArrayList<String>(genotypeFormatKeys.size());
for (String key : genotypeFormatKeys) {
if (key.equals(VCFConstants.GENOTYPE_KEY)) {
if (!g.isAvailable()) {
throw new ReviewedStingException(
"GTs cannot be missing for some samples if they are available for others in the record");
}
writeAllele(g.getAllele(0), alleleMap);
for (int i = 1; i < g.getPloidy(); i++) {
mWriter.write(g.isPhased() ? VCFConstants.PHASED : VCFConstants.UNPHASED);
writeAllele(g.getAllele(i), alleleMap);
}
continue;
}
Object val = g.hasAttribute(key) ? g.getAttribute(key) : VCFConstants.MISSING_VALUE_v4;
// some exceptions
if (key.equals(VCFConstants.GENOTYPE_QUALITY_KEY)) {
if (Math.abs(g.getNegLog10PError() - Genotype.NO_NEG_LOG_10PERROR) < 1e-6)
val = VCFConstants.MISSING_VALUE_v4;
else {
val = getQualValue(Math.min(g.getPhredScaledQual(), VCFConstants.MAX_GENOTYPE_QUAL));
}
} else if (key.equals(VCFConstants.GENOTYPE_FILTER_KEY)) {
val =
g.isFiltered()
? ParsingUtils.join(";", ParsingUtils.sortList(g.getFilters()))
: (g.filtersWereApplied()
? VCFConstants.PASSES_FILTERS_v4
: VCFConstants.UNFILTERED);
}
VCFFormatHeaderLine metaData = mHeader.getFormatHeaderLine(key);
if (metaData != null) {
int numInFormatField = metaData.getCount(vc.getAlternateAlleles().size());
if (numInFormatField > 1 && val.equals(VCFConstants.MISSING_VALUE_v4)) {
// If we have a missing field but multiple values are expected, we need to construct a
// new string with all fields.
// For example, if Number=2, the string has to be ".,."
StringBuilder sb = new StringBuilder(VCFConstants.MISSING_VALUE_v4);
for (int i = 1; i < numInFormatField; i++) {
sb.append(",");
sb.append(VCFConstants.MISSING_VALUE_v4);
}
val = sb.toString();
}
}
// assume that if key is absent, then the given string encoding suffices
String outputValue = formatVCFField(val);
if (outputValue != null) attrs.add(outputValue);
}
// strip off trailing missing values
for (int i = attrs.size() - 1; i >= 0; i--) {
if (isMissingValue(attrs.get(i))) attrs.remove(i);
else break;
}
for (int i = 0; i < attrs.size(); i++) {
if (i > 0 || genotypeFormatKeys.contains(VCFConstants.GENOTYPE_KEY))
mWriter.write(VCFConstants.GENOTYPE_FIELD_SEPARATOR);
mWriter.write(attrs.get(i));
}
}
}
/**
* Subset VC record if necessary and emit the modified record (provided it satisfies criteria for
* printing)
*
* @param tracker the ROD tracker
* @param ref reference information
* @param context alignment info
* @return 1 if the record was printed to the output file, 0 if otherwise
*/
@Override
public Integer map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
if (tracker == null) return 0;
Collection<VariantContext> vcs =
tracker.getValues(variantCollection.variants, context.getLocation());
if (vcs == null || vcs.size() == 0) {
return 0;
}
for (VariantContext vc : vcs) {
if (MENDELIAN_VIOLATIONS) {
boolean foundMV = false;
for (MendelianViolation mv : mvSet) {
if (mv.isViolation(vc)) {
foundMV = true;
// System.out.println(vc.toString());
if (outMVFile != null)
outMVFileStream.format(
"MV@%s:%d. REF=%s, ALT=%s, AC=%d, momID=%s, dadID=%s, childID=%s, momG=%s, momGL=%s, dadG=%s, dadGL=%s, "
+ "childG=%s childGL=%s\n",
vc.getChr(),
vc.getStart(),
vc.getReference().getDisplayString(),
vc.getAlternateAllele(0).getDisplayString(),
vc.getChromosomeCount(vc.getAlternateAllele(0)),
mv.getSampleMom(),
mv.getSampleDad(),
mv.getSampleChild(),
vc.getGenotype(mv.getSampleMom()).toBriefString(),
vc.getGenotype(mv.getSampleMom()).getLikelihoods().getAsString(),
vc.getGenotype(mv.getSampleDad()).toBriefString(),
vc.getGenotype(mv.getSampleMom()).getLikelihoods().getAsString(),
vc.getGenotype(mv.getSampleChild()).toBriefString(),
vc.getGenotype(mv.getSampleChild()).getLikelihoods().getAsString());
}
}
if (!foundMV) break;
}
if (DISCORDANCE_ONLY) {
Collection<VariantContext> compVCs =
tracker.getValues(discordanceTrack, context.getLocation());
if (!isDiscordant(vc, compVCs)) return 0;
}
if (CONCORDANCE_ONLY) {
Collection<VariantContext> compVCs =
tracker.getValues(concordanceTrack, context.getLocation());
if (!isConcordant(vc, compVCs)) return 0;
}
if (alleleRestriction.equals(NumberAlleleRestriction.BIALLELIC) && !vc.isBiallelic())
continue;
if (alleleRestriction.equals(NumberAlleleRestriction.MULTIALLELIC) && vc.isBiallelic())
continue;
if (!selectedTypes.contains(vc.getType())) continue;
VariantContext sub = subsetRecord(vc, samples);
if ((sub.isPolymorphic() || !EXCLUDE_NON_VARIANTS)
&& (!sub.isFiltered() || !EXCLUDE_FILTERED)) {
for (VariantContextUtils.JexlVCMatchExp jexl : jexls) {
if (!VariantContextUtils.match(sub, jexl)) {
return 0;
}
}
if (SELECT_RANDOM_NUMBER) {
randomlyAddVariant(++variantNumber, sub, ref.getBase());
} else if (!SELECT_RANDOM_FRACTION
|| (GenomeAnalysisEngine.getRandomGenerator().nextDouble() < fractionRandom)) {
vcfWriter.add(sub);
}
}
}
return 1;
}
static VariantContext reallyMergeIntoMNP(
VariantContext vc1, VariantContext vc2, ReferenceSequenceFile referenceFile) {
int startInter = vc1.getEnd() + 1;
int endInter = vc2.getStart() - 1;
byte[] intermediateBases = null;
if (startInter <= endInter) {
intermediateBases =
referenceFile.getSubsequenceAt(vc1.getChr(), startInter, endInter).getBases();
StringUtil.toUpperCase(intermediateBases);
}
MergedAllelesData mergeData =
new MergedAllelesData(
intermediateBases, vc1, vc2); // ensures that the reference allele is added
GenotypesContext mergedGenotypes = GenotypesContext.create();
for (final Genotype gt1 : vc1.getGenotypes()) {
Genotype gt2 = vc2.getGenotype(gt1.getSampleName());
List<Allele> site1Alleles = gt1.getAlleles();
List<Allele> site2Alleles = gt2.getAlleles();
List<Allele> mergedAllelesForSample = new LinkedList<Allele>();
/* NOTE: Since merged alleles are added to mergedAllelesForSample in the SAME order as in the input VC records,
we preserve phase information (if any) relative to whatever precedes vc1:
*/
Iterator<Allele> all2It = site2Alleles.iterator();
for (Allele all1 : site1Alleles) {
Allele all2 = all2It.next(); // this is OK, since allSamplesAreMergeable()
Allele mergedAllele = mergeData.ensureMergedAllele(all1, all2);
mergedAllelesForSample.add(mergedAllele);
}
double mergedGQ = Math.max(gt1.getLog10PError(), gt2.getLog10PError());
Set<String> mergedGtFilters =
new HashSet<
String>(); // Since gt1 and gt2 were unfiltered, the Genotype remains unfiltered
Map<String, Object> mergedGtAttribs = new HashMap<String, Object>();
PhaseAndQuality phaseQual = calcPhaseForMergedGenotypes(gt1, gt2);
if (phaseQual.PQ != null) mergedGtAttribs.put(ReadBackedPhasingWalker.PQ_KEY, phaseQual.PQ);
Genotype mergedGt =
new Genotype(
gt1.getSampleName(),
mergedAllelesForSample,
mergedGQ,
mergedGtFilters,
mergedGtAttribs,
phaseQual.isPhased);
mergedGenotypes.add(mergedGt);
}
String mergedName = mergeVariantContextNames(vc1.getSource(), vc2.getSource());
double mergedLog10PError = Math.min(vc1.getLog10PError(), vc2.getLog10PError());
Set<String> mergedFilters =
new HashSet<
String>(); // Since vc1 and vc2 were unfiltered, the merged record remains unfiltered
Map<String, Object> mergedAttribs = mergeVariantContextAttributes(vc1, vc2);
// ids
List<String> mergedIDs = new ArrayList<String>();
if (vc1.hasID()) mergedIDs.add(vc1.getID());
if (vc2.hasID()) mergedIDs.add(vc2.getID());
String mergedID =
mergedIDs.isEmpty()
? VCFConstants.EMPTY_ID_FIELD
: Utils.join(VCFConstants.ID_FIELD_SEPARATOR, mergedIDs);
VariantContextBuilder mergedBuilder =
new VariantContextBuilder(
mergedName,
vc1.getChr(),
vc1.getStart(),
vc2.getEnd(),
mergeData.getAllMergedAlleles())
.id(mergedID)
.genotypes(mergedGenotypes)
.log10PError(mergedLog10PError)
.filters(mergedFilters)
.attributes(mergedAttribs);
VariantContextUtils.calculateChromosomeCounts(mergedBuilder, true);
return mergedBuilder.make();
}