public static BaseUtils.BaseSubstitutionType getSNPSubstitutionType(VariantContext context) {
if (!context.isSNP() || !context.isBiallelic())
throw new IllegalStateException(
"Requested SNP substitution type for bialleic non-SNP " + context);
return BaseUtils.SNPSubstitutionType(
context.getReference().getBases()[0], context.getAlternateAllele(0).getBases()[0]);
}
@Requires({"eval != null", "comp != null"})
private EvalCompMatchType doEvalAndCompMatch(
final VariantContext eval, final VariantContext comp, boolean requireStrictAlleleMatch) {
// find all of the matching comps
if (comp.getType() != eval.getType()) return EvalCompMatchType.NO_MATCH;
// find the comp which matches both the reference allele and alternate allele from eval
final Allele altEval =
eval.getAlternateAlleles().size() == 0 ? null : eval.getAlternateAllele(0);
final Allele altComp =
comp.getAlternateAlleles().size() == 0 ? null : comp.getAlternateAllele(0);
if ((altEval == null && altComp == null)
|| (altEval != null
&& altEval.equals(altComp)
&& eval.getReference().equals(comp.getReference()))) return EvalCompMatchType.STRICT;
else return requireStrictAlleleMatch ? EvalCompMatchType.NO_MATCH : EvalCompMatchType.LENIENT;
}
private Map<String, Object> annotateSNP(AlignmentContext stratifiedContext, VariantContext vc) {
if (!stratifiedContext.hasBasePileup()) return null;
HashMap<Byte, Integer> alleleCounts = new HashMap<Byte, Integer>();
for (Allele allele : vc.getAlternateAlleles()) alleleCounts.put(allele.getBases()[0], 0);
ReadBackedPileup pileup = stratifiedContext.getBasePileup();
int totalDepth = pileup.size();
Map<String, Object> map = new HashMap<String, Object>();
map.put(getKeyNames().get(0), totalDepth); // put total depth in right away
if (totalDepth == 0) return map; // done, can not compute FA at 0 coverage!!
int mq0 = 0; // number of "ref" reads that are acually mq0
for (PileupElement p : pileup) {
if (p.getMappingQual() == 0) {
mq0++;
continue;
}
if (alleleCounts.containsKey(p.getBase())) // non-mq0 read and it's an alt
alleleCounts.put(p.getBase(), alleleCounts.get(p.getBase()) + 1);
}
if (mq0 == totalDepth) return map; // if all reads are mq0, there is nothing left to do
// we need to add counts in the correct order
String[] fracs = new String[alleleCounts.size()];
for (int i = 0; i < vc.getAlternateAlleles().size(); i++) {
fracs[i] =
String.format(
"%.3f",
((float) alleleCounts.get(vc.getAlternateAllele(i).getBases()[0]))
/ (totalDepth - mq0));
}
map.put(getKeyNames().get(1), fracs);
return map;
}
public String update1(
VariantContext vc1,
RefMetaDataTracker tracker,
ReferenceContext ref,
AlignmentContext context) {
nCalledLoci++;
// Note from Eric:
// This is really not correct. What we really want here is a polymorphic vs. monomorphic count
// (i.e. on the Genotypes).
// So in order to maintain consistency with the previous implementation (and the intention of
// the original author), I've
// added in a proxy check for monomorphic status here.
// Protect against case when vc only as no-calls too - can happen if we strafity by sample and
// sample as a single no-call.
if (vc1.isMonomorphicInSamples()) {
nRefLoci++;
} else {
switch (vc1.getType()) {
case NO_VARIATION:
// shouldn't get here
break;
case SNP:
nVariantLoci++;
nSNPs++;
if (vc1.getAttributeAsBoolean("ISSINGLETON", false)) nSingletons++;
break;
case MNP:
nVariantLoci++;
nMNPs++;
if (vc1.getAttributeAsBoolean("ISSINGLETON", false)) nSingletons++;
break;
case INDEL:
nVariantLoci++;
if (vc1.isSimpleInsertion()) nInsertions++;
else if (vc1.isSimpleDeletion()) nDeletions++;
else nComplex++;
break;
case MIXED:
nVariantLoci++;
nMixed++;
break;
case SYMBOLIC:
nSymbolic++;
break;
default:
throw new ReviewedStingException("Unexpected VariantContext type " + vc1.getType());
}
}
String refStr = vc1.getReference().getBaseString().toUpperCase();
String aaStr =
vc1.hasAttribute("ANCESTRALALLELE")
? vc1.getAttributeAsString("ANCESTRALALLELE", null).toUpperCase()
: null;
// if (aaStr.equals(".")) {
// aaStr = refStr;
// }
// ref aa alt class
// A C A der homozygote
// A C C anc homozygote
// A A A ref homozygote
// A A C
// A C A
// A C C
for (final Genotype g : vc1.getGenotypes()) {
final String altStr =
vc1.getAlternateAlleles().size() > 0
? vc1.getAlternateAllele(0).getBaseString().toUpperCase()
: null;
switch (g.getType()) {
case NO_CALL:
nNoCalls++;
break;
case HOM_REF:
nHomRef++;
if (aaStr != null && altStr != null && !refStr.equalsIgnoreCase(aaStr)) {
nHomDerived++;
}
break;
case HET:
nHets++;
break;
case HOM_VAR:
nHomVar++;
if (aaStr != null && altStr != null && !altStr.equalsIgnoreCase(aaStr)) {
nHomDerived++;
}
break;
case MIXED:
break;
default:
throw new ReviewedStingException("BUG: Unexpected genotype type: " + g);
}
}
return null; // we don't capture any interesting sites
}
/**
* 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());
}
}
/**
* 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;
}
private Map<String, Object> annotateIndel(AlignmentContext stratifiedContext, VariantContext vc) {
if (!stratifiedContext.hasExtendedEventPileup()) {
return null;
}
ReadBackedExtendedEventPileup pileup = stratifiedContext.getExtendedEventPileup();
if (pileup == null) return null;
int totalDepth = pileup.size();
Map<String, Object> map = new HashMap<String, Object>();
map.put(getKeyNames().get(0), totalDepth); // put total depth in right away
if (totalDepth == 0) return map;
int mq0 = 0; // number of "ref" reads that are acually mq0
HashMap<String, Integer> alleleCounts = new HashMap<String, Integer>();
Allele refAllele = vc.getReference();
for (Allele allele : vc.getAlternateAlleles()) {
if (allele.isNoCall()) {
continue; // this does not look so good, should we die???
}
alleleCounts.put(getAlleleRepresentation(allele), 0);
}
for (ExtendedEventPileupElement e : pileup.toExtendedIterable()) {
if (e.getMappingQual() == 0) {
mq0++;
continue;
}
if (e.isInsertion()) {
final String b = e.getEventBases();
if (alleleCounts.containsKey(b)) {
alleleCounts.put(b, alleleCounts.get(b) + 1);
}
} else {
if (e.isDeletion()) {
if (e.getEventLength() == refAllele.length()) {
// this is indeed the deletion allele recorded in VC
final String b = DEL;
if (alleleCounts.containsKey(b)) {
alleleCounts.put(b, alleleCounts.get(b) + 1);
}
}
// else {
// System.out.print(" deletion of WRONG length found");
// }
}
}
}
if (mq0 == totalDepth) return map;
String[] fracs = new String[alleleCounts.size()];
for (int i = 0; i < vc.getAlternateAlleles().size(); i++)
fracs[i] =
String.format(
"%.3f",
((float) alleleCounts.get(getAlleleRepresentation(vc.getAlternateAllele(i))))
/ (totalDepth - mq0));
map.put(getKeyNames().get(1), fracs);
// map.put(getKeyNames().get(0), counts);
return map;
}