@Override
public void initialize(
AnnotatorCompatible walker, GenomeAnalysisEngine toolkit, Set<VCFHeaderLine> headerLines) {
// Make sure that we actually have a valid SnpEff rod binding (just in case the user specified
// -A SnpEff
// without providing a SnpEff rod via --snpEffFile):
if (!isValidRodBinding(walker.getSnpEffRodBinding())) {
canAnnotate = false;
return;
}
RodBinding<VariantContext> snpEffRodBinding = walker.getSnpEffRodBinding();
// Make sure that the SnpEff version number and command-line header lines are present in the VCF
// header of
// the SnpEff rod, and that the file was generated by a supported version of SnpEff:
VCFHeader snpEffVCFHeader =
GATKVCFUtils.getVCFHeadersFromRods(toolkit, Arrays.asList(snpEffRodBinding.getName()))
.get(snpEffRodBinding.getName());
VCFHeaderLine snpEffVersionLine =
snpEffVCFHeader.getOtherHeaderLine(SNPEFF_VCF_HEADER_VERSION_LINE_KEY);
VCFHeaderLine snpEffCommandLine =
snpEffVCFHeader.getOtherHeaderLine(SNPEFF_VCF_HEADER_COMMAND_LINE_KEY);
if (!isValidSnpEffVersionAndCommandLine(snpEffVersionLine, snpEffCommandLine)) {
canAnnotate = false;
return;
}
// If everything looks ok, add the SnpEff version number and command-line header lines to the
// header of the VCF output file, changing the key names so that our output file won't be
// mistaken in the future for a SnpEff output file:
headerLines.add(
new VCFHeaderLine(OUTPUT_VCF_HEADER_VERSION_LINE_KEY, snpEffVersionLine.getValue()));
headerLines.add(
new VCFHeaderLine(OUTPUT_VCF_HEADER_COMMAND_LINE_KEY, snpEffCommandLine.getValue()));
// Can only be called from VariantAnnotator
if (!(walker instanceof VariantAnnotator)) {
if (walker != null)
logger.warn(
"Annotation will not be calculated, must be called from VariantAnnotator, not "
+ walker.getClass().getName());
else logger.warn("Annotation will not be calculated, must be called from VariantAnnotator");
canAnnotate = false;
return;
}
}
private void writeRecord(VariantContext vc, RefMetaDataTracker tracker, GenomeLoc loc) {
if (!wroteHeader) {
wroteHeader = true;
// setup the header fields
Set<VCFHeaderLine> hInfo = new HashSet<VCFHeaderLine>();
hInfo.addAll(GATKVCFUtils.getHeaderFields(getToolkit(), Arrays.asList(variants.getName())));
hInfo.add(VCFStandardHeaderLines.getFormatLine(VCFConstants.GENOTYPE_KEY));
allowedGenotypeFormatStrings.add(VCFConstants.GENOTYPE_KEY);
for (VCFHeaderLine field : hInfo) {
if (field instanceof VCFFormatHeaderLine) {
allowedGenotypeFormatStrings.add(((VCFFormatHeaderLine) field).getID());
}
}
samples = new LinkedHashSet<String>();
if (sampleName != null) {
samples.add(sampleName);
} else {
// try VCF first
samples =
SampleUtils.getSampleListWithVCFHeader(getToolkit(), Arrays.asList(variants.getName()));
if (samples.isEmpty()) {
List<Feature> features = tracker.getValues(variants, loc);
if (features.size() == 0)
throw new IllegalStateException(
"No rod data is present, but we just created a VariantContext");
Feature f = features.get(0);
if (f instanceof RawHapMapFeature)
samples.addAll(Arrays.asList(((RawHapMapFeature) f).getSampleIDs()));
else samples.addAll(vc.getSampleNames());
}
}
vcfwriter.writeHeader(new VCFHeader(hInfo, samples));
}
vc = GATKVariantContextUtils.purgeUnallowedGenotypeAttributes(vc, allowedGenotypeFormatStrings);
vcfwriter.add(vc);
}
/**
* Gets the header lines for the VCF writer
*
* @return A set of VCF header lines
*/
private static Set<VCFHeaderLine> getHeaderInfo() {
Set<VCFHeaderLine> headerLines = new HashSet<>();
// INFO fields for overall data
headerLines.add(VCFStandardHeaderLines.getInfoLine(VCFConstants.END_KEY));
headerLines.add(GATKVCFHeaderLines.getInfoLine(GATKVCFConstants.AVG_INTERVAL_DP_KEY));
headerLines.add(GATKVCFHeaderLines.getInfoLine(GATKVCFConstants.INTERVAL_GC_CONTENT_KEY));
headerLines.add(
new VCFInfoHeaderLine(
"Diagnose Targets", 0, VCFHeaderLineType.Flag, "DiagnoseTargets mode"));
// FORMAT fields for each genotype
headerLines.add(VCFStandardHeaderLines.getFormatLine(VCFConstants.GENOTYPE_FILTER_KEY));
headerLines.add(
GATKVCFHeaderLines.getFormatLine(GATKVCFConstants.AVG_INTERVAL_DP_BY_SAMPLE_KEY));
headerLines.add(GATKVCFHeaderLines.getFormatLine(GATKVCFConstants.LOW_COVERAGE_LOCI));
headerLines.add(GATKVCFHeaderLines.getFormatLine(GATKVCFConstants.ZERO_COVERAGE_LOCI));
// FILTER fields
for (CallableStatus stat : CallableStatus.values())
headerLines.add(new VCFFilterHeaderLine(stat.name(), stat.description));
return headerLines;
}
/** Prepare the output file and the list of available features. */
public void initialize() {
if (LIST) {
AnnotationHelpUtils.listAnnotations();
System.exit(0);
}
// get the list of all sample names from the variant VCF input rod, if applicable
final List<String> rodName = Arrays.asList(variantCollection.variants.getName());
final Set<String> samples = SampleUtils.getUniqueSamplesFromRods(getToolkit(), rodName);
if (USE_ALL_ANNOTATIONS)
engine = new VariantAnnotatorEngine(annotationsToExclude, this, getToolkit());
else
engine =
new VariantAnnotatorEngine(
annotationGroupsToUse, annotationsToUse, annotationsToExclude, this, getToolkit());
engine.initializeExpressions(expressionsToUse);
engine.setExpressionAlleleConcordance(expressionAlleleConcordance);
// setup the header fields
// note that if any of the definitions conflict with our new ones, then we want to overwrite the
// old ones
final Set<VCFHeaderLine> hInfo = new HashSet<>();
hInfo.addAll(engine.getVCFAnnotationDescriptions());
for (final VCFHeaderLine line : GATKVCFUtils.getHeaderFields(getToolkit(), rodName)) {
if (isUniqueHeaderLine(line, hInfo)) hInfo.add(line);
}
// for the expressions, pull the info header line from the header of the resource rod
for (final VariantAnnotatorEngine.VAExpression expression : engine.getRequestedExpressions()) {
// special case the ID field
if (expression.fieldName.equals("ID")) {
hInfo.add(
new VCFInfoHeaderLine(
expression.fullName,
1,
VCFHeaderLineType.String,
"ID field transferred from external VCF resource"));
continue;
}
VCFInfoHeaderLine targetHeaderLine = null;
for (final VCFHeaderLine line :
GATKVCFUtils.getHeaderFields(getToolkit(), Arrays.asList(expression.binding.getName()))) {
if (line instanceof VCFInfoHeaderLine) {
final VCFInfoHeaderLine infoline = (VCFInfoHeaderLine) line;
if (infoline.getID().equals(expression.fieldName)) {
targetHeaderLine = infoline;
break;
}
}
}
if (targetHeaderLine != null) {
if (targetHeaderLine.getCountType() == VCFHeaderLineCount.INTEGER)
hInfo.add(
new VCFInfoHeaderLine(
expression.fullName,
targetHeaderLine.getCount(),
targetHeaderLine.getType(),
targetHeaderLine.getDescription()));
else
hInfo.add(
new VCFInfoHeaderLine(
expression.fullName,
targetHeaderLine.getCountType(),
targetHeaderLine.getType(),
targetHeaderLine.getDescription()));
} else {
hInfo.add(
new VCFInfoHeaderLine(
expression.fullName,
VCFHeaderLineCount.UNBOUNDED,
VCFHeaderLineType.String,
"Value transferred from another external VCF resource"));
}
}
engine.makeHeaderInfoMap(hInfo);
engine.invokeAnnotationInitializationMethods(hInfo);
VCFHeader vcfHeader = new VCFHeader(hInfo, samples);
vcfWriter.writeHeader(vcfHeader);
}
/** Parse the familial relationship specification, and initialize VCF writer */
public void initialize() {
List<String> rodNames = new ArrayList<String>();
rodNames.add(variantCollection.variants.getName());
Map<String, VCFHeader> vcfRods = GATKVCFUtils.getVCFHeadersFromRods(getToolkit(), rodNames);
Set<String> vcfSamples =
SampleUtils.getSampleList(
vcfRods, GATKVariantContextUtils.GenotypeMergeType.REQUIRE_UNIQUE);
// set up tumor and normal samples
if (!vcfSamples.contains(normalSample))
throw new UserException.BadArgumentValue(
"--normalSample",
"the normal sample " + normalSample + " doesn't match any sample from the input VCF");
if (!vcfSamples.contains(tumorSample))
throw new UserException.BadArgumentValue(
"--tumorSample",
"the tumor sample " + tumorSample + " doesn't match any sample from the input VCF");
logger.info("Normal sample: " + normalSample);
logger.info("Tumor sample: " + tumorSample);
Set<VCFHeaderLine> headerLines = new HashSet<VCFHeaderLine>();
headerLines.addAll(GATKVCFUtils.getHeaderFields(this.getToolkit()));
headerLines.add(
new VCFInfoHeaderLine(
VCFConstants.SOMATIC_KEY,
0,
VCFHeaderLineType.Flag,
"Is this a confidently called somatic mutation"));
headerLines.add(
new VCFInfoHeaderLine(
SOMATIC_LOD_TAG_NAME,
1,
VCFHeaderLineType.Float,
"log10 probability that the site is a somatic mutation"));
headerLines.add(
new VCFInfoHeaderLine(
SOMATIC_AC_TAG_NAME,
1,
VCFHeaderLineType.Integer,
"Allele count of samples with somatic event"));
headerLines.add(
new VCFInfoHeaderLine(
SOMATIC_NONREF_TAG_NAME,
1,
VCFHeaderLineType.Integer,
"Number of samples with somatic event"));
samples.add(normalSample);
samples.add(tumorSample);
vcfWriter.writeHeader(new VCFHeader(headerLines, samples));
}