@Override
public List<VCFInfoHeaderLine> getDescriptions() {
return Arrays.asList(
new VCFInfoHeaderLine(
InfoFieldKey.EFFECT_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"The highest-impact effect resulting from the current variant (or one of the highest-impact effects, if there is a tie)"),
new VCFInfoHeaderLine(
InfoFieldKey.IMPACT_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Impact of the highest-impact effect resulting from the current variant "
+ Arrays.toString(EffectImpact.values())),
new VCFInfoHeaderLine(
InfoFieldKey.FUNCTIONAL_CLASS_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Functional class of the highest-impact effect resulting from the current variant: "
+ Arrays.toString(EffectFunctionalClass.values())),
new VCFInfoHeaderLine(
InfoFieldKey.CODON_CHANGE_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Old/New codon for the highest-impact effect resulting from the current variant"),
new VCFInfoHeaderLine(
InfoFieldKey.AMINO_ACID_CHANGE_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Old/New amino acid for the highest-impact effect resulting from the current variant (in HGVS style)"),
new VCFInfoHeaderLine(
InfoFieldKey.GENE_NAME_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Gene name for the highest-impact effect resulting from the current variant"),
new VCFInfoHeaderLine(
InfoFieldKey.GENE_BIOTYPE_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Gene biotype for the highest-impact effect resulting from the current variant"),
new VCFInfoHeaderLine(
InfoFieldKey.TRANSCRIPT_ID_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Transcript ID for the highest-impact effect resulting from the current variant"),
new VCFInfoHeaderLine(
InfoFieldKey.EXON_ID_KEY.getKeyName(),
1,
VCFHeaderLineType.String,
"Exon ID for the highest-impact effect resulting from the current variant"));
}
private void outputMissingInterval(final IntervalStratification stats) {
final GenomeLoc interval = stats.getInterval();
final boolean missing[] = new boolean[interval.size()];
Arrays.fill(missing, true);
for (AbstractStratification sample : stats.getElements()) {
if (hasMissingStatuses(sample)) {
int pos = 0;
for (AbstractStratification locus : sample.getElements()) {
if (locus.callableStatuses().isEmpty()) {
missing[pos] = false;
}
pos++;
}
}
}
int start = -1;
boolean insideMissing = false;
for (int i = 0; i < missing.length; i++) {
if (missing[i] && !insideMissing) {
start = interval.getStart() + i;
insideMissing = true;
} else if (!missing[i] && insideMissing) {
final int stop = interval.getStart() + i - 1;
outputMissingInterval(interval.getContig(), start, stop);
insideMissing = false;
}
}
if (insideMissing) {
outputMissingInterval(interval.getContig(), start, interval.getStop());
}
}
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);
}
@Override
public List<String> getKeyNames() {
return Arrays.asList(
InfoFieldKey.EFFECT_KEY.getKeyName(),
InfoFieldKey.IMPACT_KEY.getKeyName(),
InfoFieldKey.FUNCTIONAL_CLASS_KEY.getKeyName(),
InfoFieldKey.CODON_CHANGE_KEY.getKeyName(),
InfoFieldKey.AMINO_ACID_CHANGE_KEY.getKeyName(),
InfoFieldKey.GENE_NAME_KEY.getKeyName(),
InfoFieldKey.GENE_BIOTYPE_KEY.getKeyName(),
InfoFieldKey.TRANSCRIPT_ID_KEY.getKeyName(),
InfoFieldKey.EXON_ID_KEY.getKeyName());
}
private List<SnpEffEffect> parseSnpEffRecord(VariantContext snpEffRecord) {
List<SnpEffEffect> parsedEffects = new ArrayList<SnpEffEffect>();
Object effectFieldValue = snpEffRecord.getAttribute(SNPEFF_INFO_FIELD_KEY);
if (effectFieldValue == null) {
return parsedEffects;
}
// The VCF codec stores multi-valued fields as a List<String>, and single-valued fields as a
// String.
// We can have either in the case of SnpEff, since there may be one or more than one effect in
// this record.
List<String> individualEffects;
if (effectFieldValue instanceof List) {
individualEffects = (List<String>) effectFieldValue;
} else {
individualEffects = Arrays.asList((String) effectFieldValue);
}
for (String effectString : individualEffects) {
String[] effectNameAndMetadata = effectString.split(SNPEFF_EFFECT_METADATA_DELIMITER);
if (effectNameAndMetadata.length != 2) {
logger.warn(
String.format(
"Malformed SnpEff effect field at %s:%d, skipping: %s",
snpEffRecord.getChr(), snpEffRecord.getStart(), effectString));
continue;
}
String effectName = effectNameAndMetadata[0];
String[] effectMetadata =
effectNameAndMetadata[1].split(SNPEFF_EFFECT_METADATA_SUBFIELD_DELIMITER, -1);
SnpEffEffect parsedEffect = new SnpEffEffect(effectName, effectMetadata);
if (parsedEffect.isWellFormed()) {
parsedEffects.add(parsedEffect);
} else {
logger.warn(
String.format(
"Skipping malformed SnpEff effect field at %s:%d. Error was: \"%s\". Field was: \"%s\"",
snpEffRecord.getChr(),
snpEffRecord.getStart(),
parsedEffect.getParseError(),
effectString));
}
}
return parsedEffects;
}
@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 String supportedSnpEffVersionsString() {
return String.format(
"%s, as well as later versions when run with the option %s",
Arrays.toString(SUPPORTED_SNPEFF_VERSIONS), SNPEFF_GATK_COMPATIBILITY_ARGUMENT);
}
/** 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);
}