@Override
public void update2(
VariantContext eval,
VariantContext comp,
RefMetaDataTracker tracker,
ReferenceContext ref,
AlignmentContext context) {
if (comp != null) { // we only need to consider sites in comp
if (REQUIRE_IDENTICAL_ALLELES && (eval != null && haveDifferentAltAlleles(eval, comp)))
nDifferentAlleleSites++;
else {
SiteStatus evalStatus = calcSiteStatus(eval);
final Set<String> evalSamples = getWalker().getSampleNamesForEvaluation();
if (comp.hasGenotypes() && !evalSamples.isEmpty() && comp.hasGenotypes(evalSamples))
// if we have genotypes in both eval and comp, subset comp down just the samples in eval
comp = comp.subContextFromSamples(evalSamples, false);
SiteStatus compStatus = calcSiteStatus(comp);
counts[compStatus.ordinal()][evalStatus.ordinal()]++;
}
}
}
/**
* Update the attributes of the attributes map given the VariantContext to reflect the proper
* chromosome-based VCF tags
*
* @param vc the VariantContext
* @param attributes the attributes map to populate; must not be null; may contain old values
* @param removeStaleValues should we remove stale values from the mapping?
* @return the attributes map provided as input, returned for programming convenience
*/
public static Map<String, Object> calculateChromosomeCounts(
VariantContext vc, Map<String, Object> attributes, boolean removeStaleValues) {
final int AN = vc.getCalledChrCount();
// if everyone is a no-call, remove the old attributes if requested
if (AN == 0 && removeStaleValues) {
if (attributes.containsKey(VCFConstants.ALLELE_COUNT_KEY))
attributes.remove(VCFConstants.ALLELE_COUNT_KEY);
if (attributes.containsKey(VCFConstants.ALLELE_FREQUENCY_KEY))
attributes.remove(VCFConstants.ALLELE_FREQUENCY_KEY);
if (attributes.containsKey(VCFConstants.ALLELE_NUMBER_KEY))
attributes.remove(VCFConstants.ALLELE_NUMBER_KEY);
return attributes;
}
if (vc.hasGenotypes()) {
attributes.put(VCFConstants.ALLELE_NUMBER_KEY, AN);
// if there are alternate alleles, record the relevant tags
if (vc.getAlternateAlleles().size() > 0) {
final ArrayList<String> alleleFreqs = new ArrayList<String>();
final ArrayList<Integer> alleleCounts = new ArrayList<Integer>();
for (Allele allele : vc.getAlternateAlleles()) {
int altChromosomes = vc.getCalledChrCount(allele);
alleleCounts.add(altChromosomes);
if (AN == 0) {
alleleFreqs.add("0.0");
} else {
// todo -- this is a performance problem
final String freq =
String.format(
makePrecisionFormatStringFromDenominatorValue((double) AN),
((double) altChromosomes / (double) AN));
alleleFreqs.add(freq);
}
}
attributes.put(
VCFConstants.ALLELE_COUNT_KEY,
alleleCounts.size() == 1 ? alleleCounts.get(0) : alleleCounts);
attributes.put(
VCFConstants.ALLELE_FREQUENCY_KEY,
alleleFreqs.size() == 1 ? alleleFreqs.get(0) : alleleFreqs);
} else {
attributes.put(VCFConstants.ALLELE_COUNT_KEY, 0);
attributes.put(VCFConstants.ALLELE_FREQUENCY_KEY, 0.0);
}
}
return attributes;
}
//
// helper routines
//
private SiteStatus calcSiteStatus(VariantContext vc) {
if (vc == null) return SiteStatus.NO_CALL;
if (vc.isFiltered()) return SiteStatus.FILTERED;
if (vc.isMonomorphicInSamples()) return SiteStatus.MONO;
if (vc.hasGenotypes())
return SiteStatus
.POLY; // must be polymorphic if isMonomorphicInSamples was false and there are genotypes
if (vc.hasAttribute(VCFConstants.ALLELE_COUNT_KEY)) {
int ac = 0;
if (vc.getNAlleles() > 2) {
return SiteStatus.POLY;
} else ac = vc.getAttributeAsInt(VCFConstants.ALLELE_COUNT_KEY, 0);
return ac > 0 ? SiteStatus.POLY : SiteStatus.MONO;
} else {
return TREAT_ALL_SITES_IN_EVAL_VCF_AS_CALLED
? SiteStatus.POLY
: SiteStatus.NO_CALL; // we can't figure out what to do
}
}
public static boolean canBeOutputToBeagle(VariantContext v) {
return v != null && !v.isFiltered() && v.isBiallelic() && v.hasGenotypes();
}
/**
* add a record to the file
*
* @param vc the Variant Context object
* @param refBase the ref base used for indels
* @param refBaseShouldBeAppliedToEndOfAlleles *** THIS SHOULD BE FALSE EXCEPT FOR AN INDEL AT THE
* EXTREME BEGINNING OF A CONTIG (WHERE THERE IS NO PREVIOUS BASE, SO WE USE THE BASE AFTER
* THE EVENT INSTEAD)
*/
public void add(VariantContext vc, byte refBase, boolean refBaseShouldBeAppliedToEndOfAlleles) {
if (mHeader == null)
throw new IllegalStateException(
"The VCF Header must be written before records can be added: " + locationString());
if (doNotWriteGenotypes) vc = VariantContext.modifyGenotypes(vc, null);
try {
vc =
VariantContext.createVariantContextWithPaddedAlleles(
vc, refBase, refBaseShouldBeAppliedToEndOfAlleles);
// if we are doing on the fly indexing, add the record ***before*** we write any bytes
if (indexer != null) indexer.addFeature(vc, positionalStream.getPosition());
Map<Allele, String> alleleMap = new HashMap<Allele, String>(vc.getAlleles().size());
alleleMap.put(Allele.NO_CALL, VCFConstants.EMPTY_ALLELE); // convenience for lookup
// CHROM
mWriter.write(vc.getChr());
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// POS
mWriter.write(String.valueOf(vc.getStart()));
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// ID
String ID = vc.hasID() ? vc.getID() : VCFConstants.EMPTY_ID_FIELD;
mWriter.write(ID);
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// REF
alleleMap.put(vc.getReference(), "0");
String refString = vc.getReference().getDisplayString();
mWriter.write(refString);
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// ALT
if (vc.isVariant()) {
Allele altAllele = vc.getAlternateAllele(0);
alleleMap.put(altAllele, "1");
String alt = altAllele.getDisplayString();
mWriter.write(alt);
for (int i = 1; i < vc.getAlternateAlleles().size(); i++) {
altAllele = vc.getAlternateAllele(i);
alleleMap.put(altAllele, String.valueOf(i + 1));
alt = altAllele.getDisplayString();
mWriter.write(",");
mWriter.write(alt);
}
} else {
mWriter.write(VCFConstants.EMPTY_ALTERNATE_ALLELE_FIELD);
}
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// QUAL
if (!vc.hasNegLog10PError()) mWriter.write(VCFConstants.MISSING_VALUE_v4);
else mWriter.write(getQualValue(vc.getPhredScaledQual()));
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// FILTER
String filters =
vc.isFiltered()
? ParsingUtils.join(";", ParsingUtils.sortList(vc.getFilters()))
: (filtersWereAppliedToContext || vc.filtersWereApplied()
? VCFConstants.PASSES_FILTERS_v4
: VCFConstants.UNFILTERED);
mWriter.write(filters);
mWriter.write(VCFConstants.FIELD_SEPARATOR);
// INFO
Map<String, String> infoFields = new TreeMap<String, String>();
for (Map.Entry<String, Object> field : vc.getAttributes().entrySet()) {
String key = field.getKey();
if (key.equals(VariantContext.ID_KEY)
|| key.equals(VariantContext.REFERENCE_BASE_FOR_INDEL_KEY)
|| key.equals(VariantContext.UNPARSED_GENOTYPE_MAP_KEY)
|| key.equals(VariantContext.UNPARSED_GENOTYPE_PARSER_KEY)) continue;
String outputValue = formatVCFField(field.getValue());
if (outputValue != null) infoFields.put(key, outputValue);
}
writeInfoString(infoFields);
// FORMAT
if (vc.hasAttribute(VariantContext.UNPARSED_GENOTYPE_MAP_KEY)) {
mWriter.write(VCFConstants.FIELD_SEPARATOR);
mWriter.write(vc.getAttributeAsString(VariantContext.UNPARSED_GENOTYPE_MAP_KEY, ""));
} else {
List<String> genotypeAttributeKeys = new ArrayList<String>();
if (vc.hasGenotypes()) {
genotypeAttributeKeys.addAll(calcVCFGenotypeKeys(vc));
} else if (mHeader.hasGenotypingData()) {
// this needs to be done in case all samples are no-calls
genotypeAttributeKeys.add(VCFConstants.GENOTYPE_KEY);
}
if (genotypeAttributeKeys.size() > 0) {
String genotypeFormatString =
ParsingUtils.join(VCFConstants.GENOTYPE_FIELD_SEPARATOR, genotypeAttributeKeys);
mWriter.write(VCFConstants.FIELD_SEPARATOR);
mWriter.write(genotypeFormatString);
addGenotypeData(vc, alleleMap, genotypeAttributeKeys);
}
}
mWriter.write("\n");
mWriter.flush(); // necessary so that writing to an output stream will work
} catch (IOException e) {
throw new RuntimeException("Unable to write the VCF object to " + locationString());
}
}