@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;
}
}
@Override
public Map<String, Object> annotate(
final RefMetaDataTracker tracker,
final AnnotatorCompatible walker,
final ReferenceContext ref,
final Map<String, AlignmentContext> stratifiedContexts,
final VariantContext vc,
final Map<String, PerReadAlleleLikelihoodMap> stratifiedPerReadAlleleLikelihoodMap) {
// Can not annotate if failed initialization conditions
if (!canAnnotate) return null;
RodBinding<VariantContext> snpEffRodBinding = walker.getSnpEffRodBinding();
// Get only SnpEff records that start at this locus, not merely span it:
List<VariantContext> snpEffRecords = tracker.getValues(snpEffRodBinding, ref.getLocus());
// Within this set, look for a SnpEff record whose ref/alt alleles match the record to annotate.
// If there is more than one such record, we only need to pick the first one, since the
// biological
// effects will be the same across all such records:
VariantContext matchingRecord = getMatchingSnpEffRecord(snpEffRecords, vc);
if (matchingRecord == null) {
return null;
}
// Parse the SnpEff INFO field annotation from the matching record into individual effect
// objects:
List<SnpEffEffect> effects = parseSnpEffRecord(matchingRecord);
if (effects.isEmpty()) {
return null;
}
// Add only annotations for one of the most biologically-significant effects from this set:
SnpEffEffect mostSignificantEffect = getMostSignificantEffect(effects);
return mostSignificantEffect.getAnnotations();
}