/**
   * For each variant in the file, determine the phasing for the child and replace the child's
   * genotype with the trio's genotype
   *
   * @param tracker the reference meta-data tracker
   * @param ref the reference context
   * @param context the alignment context
   * @return null
   */
  @Override
  public Integer map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
    if (tracker != null) {
      for (VariantContext vc :
          tracker.getValues(variantCollection.variants, context.getLocation())) {
        vc = vc.subContextFromSamples(samples);
        if (!vc.isPolymorphicInSamples()) continue;

        double log10pSomatic = calcLog10pSomatic(vc);

        // write in the somatic status probability
        Map<String, Object> attrs = new HashMap<String, Object>(); // vc.getAttributes());
        if (!minimalVCF) attrs.putAll(vc.getAttributes());
        attrs.put(SOMATIC_LOD_TAG_NAME, log10pSomatic);
        if (log10pSomatic > somaticMinLOD) {
          attrs.put(VCFConstants.SOMATIC_KEY, true);
          attrs.put(SOMATIC_NONREF_TAG_NAME, calculateTumorNNR(vc));
          attrs.put(SOMATIC_AC_TAG_NAME, calculateTumorAC(vc));
        }
        final VariantContextBuilder builder = new VariantContextBuilder(vc).attributes(attrs);
        VariantContextUtils.calculateChromosomeCounts(builder, false);
        VariantContext newvc = builder.make();

        vcfWriter.add(newvc);
      }

      return null;
    }

    return null;
  }
示例#2
0
  @Override
  public CallableBaseState map(
      RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
    CalledState state;

    if (BaseUtils.isNBase(ref.getBase())) {
      state = CalledState.REF_N;
    } else {
      // count up the depths of all and QC+ bases
      int rawDepth = 0, QCDepth = 0, lowMAPQDepth = 0;
      for (PileupElement e : context.getBasePileup()) {
        rawDepth++;

        if (e.getMappingQual() <= maxLowMAPQ) lowMAPQDepth++;

        if (e.getMappingQual() >= minMappingQuality
            && (e.getQual() >= minBaseQuality || e.isDeletion())) {
          QCDepth++;
        }
      }

      // System.out.printf("%s rawdepth = %d QCDepth = %d lowMAPQ = %d%n", context.getLocation(),
      // rawDepth, QCDepth, lowMAPQDepth);
      if (rawDepth == 0) {
        state = CalledState.NO_COVERAGE;
      } else if (rawDepth >= minDepthLowMAPQ
          && MathUtils.ratio(lowMAPQDepth, rawDepth) >= maxLowMAPQFraction) {
        state = CalledState.POOR_MAPPING_QUALITY;
      } else if (QCDepth < minDepth) {
        state = CalledState.LOW_COVERAGE;
      } else if (rawDepth >= maxDepth && maxDepth != -1) {
        state = CalledState.EXCESSIVE_COVERAGE;
      } else {
        state = CalledState.CALLABLE;
      }
    }

    return new CallableBaseState(getToolkit().getGenomeLocParser(), context.getLocation(), state);
  }
示例#3
0
  /**
   * For each site of interest, annotate based on the requested annotation types
   *
   * @param tracker the meta-data tracker
   * @param ref the reference base
   * @param context the context for the given locus
   * @return 1 if the locus was successfully processed, 0 if otherwise
   */
  public Integer map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
    if (tracker == null) return 0;

    // get the variant contexts for all the variants at the location
    Collection<VariantContext> VCs =
        tracker.getValues(variantCollection.variants, context.getLocation());
    if (VCs.isEmpty()) return 0;

    Collection<VariantContext> annotatedVCs = VCs;

    // if the reference base is not ambiguous, we can annotate
    Map<String, AlignmentContext> stratifiedContexts;
    if (BaseUtils.simpleBaseToBaseIndex(ref.getBase()) != -1) {
      stratifiedContexts = AlignmentContextUtils.splitContextBySampleName(context.getBasePileup());
      annotatedVCs = new ArrayList<>(VCs.size());
      for (VariantContext vc : VCs)
        annotatedVCs.add(engine.annotateContext(tracker, ref, stratifiedContexts, vc));
    }

    for (VariantContext annotatedVC : annotatedVCs) vcfWriter.add(annotatedVC);

    return 1;
  }
  public Integer map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
    if (tracker == null) return 0;

    final Collection<VariantContext> VCs = tracker.getValues(variants, context.getLocation());
    if (VCs.size() == 0) return 0;

    final VCcontext vc = new VCcontext(VariantContextUtils.sitesOnlyVariantContexts(VCs), ref);

    // TODO -- what should we do about filtered records?

    if (!queue.isEmpty()) {

      final VCcontext previous = queue.getLast();
      if (!previous.loc.onSameContig(vc.loc)
          || previous.loc.distance(vc.loc) > MAX_DISTANCE_BETWEEN_MERGED_RECORDS
          || queue.getFirst().loc.distance(vc.loc) > MAX_HAPLOTYPE_TO_CONSIDER) {
        purgeQueue();
      }
    }

    queue.addLast(vc);
    return 0;
  }
  public Event map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {

    boolean hasIndel = false;
    boolean hasInsertion = false;
    boolean hasPointEvent = false;

    int furthestStopPos = -1;

    // look at the rods for indels or SNPs
    if (tracker != null) {
      for (VariantContext vc : tracker.getValues(known)) {
        switch (vc.getType()) {
          case INDEL:
            hasIndel = true;
            if (vc.isSimpleInsertion()) hasInsertion = true;
            break;
          case SNP:
            hasPointEvent = true;
            break;
          case MIXED:
            hasPointEvent = true;
            hasIndel = true;
            if (vc.isSimpleInsertion()) hasInsertion = true;
            break;
          default:
            break;
        }
        if (hasIndel) furthestStopPos = vc.getEnd();
      }
    }

    // look at the normal context to get deletions and positions with high entropy
    final ReadBackedPileup pileup = context.getBasePileup();

    int mismatchQualities = 0, totalQualities = 0;
    final byte refBase = ref.getBase();
    for (PileupElement p : pileup) {

      // check the ends of the reads to see how far they extend
      furthestStopPos = Math.max(furthestStopPos, p.getRead().getAlignmentEnd());

      // is it a deletion or insertion?
      if (p.isDeletion() || p.isBeforeInsertion()) {
        hasIndel = true;
        if (p.isBeforeInsertion()) hasInsertion = true;
      }

      // look for mismatches
      else if (lookForMismatchEntropy) {
        if (p.getBase() != refBase) mismatchQualities += p.getQual();
        totalQualities += p.getQual();
      }
    }

    // make sure we're supposed to look for high entropy
    if (lookForMismatchEntropy
        && pileup.getNumberOfElements() >= minReadsAtLocus
        && (double) mismatchQualities / (double) totalQualities >= mismatchThreshold)
      hasPointEvent = true;

    // return null if no event occurred
    if (!hasIndel && !hasPointEvent) return null;

    // return null if we didn't find any usable reads/rods associated with the event
    if (furthestStopPos == -1) return null;

    GenomeLoc eventLoc = context.getLocation();
    if (hasInsertion)
      eventLoc =
          getToolkit()
              .getGenomeLocParser()
              .createGenomeLoc(eventLoc.getContig(), eventLoc.getStart(), eventLoc.getStart() + 1);

    EVENT_TYPE eventType =
        (hasIndel
            ? (hasPointEvent ? EVENT_TYPE.BOTH : EVENT_TYPE.INDEL_EVENT)
            : EVENT_TYPE.POINT_EVENT);

    return new Event(eventLoc, furthestStopPos, eventType);
  }
  public CountedData map(
      RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {

    final CountedData counter = new CountedData();

    // For some reason RodWalkers get map calls with null trackers
    if (tracker == null) return counter;

    VariantContext vcComp = tracker.getFirstValue(alleles);
    if (vcComp == null) return counter;

    // todo - not sure I want this, may be misleading to filter extended indel events.
    if (isInsideExtendedIndel(vcComp, ref)) return counter;

    // Do not operate on variants that are not covered to the optional minimum depth
    if (!context.hasReads()
        || (minDepth > 0 && context.getBasePileup().getBases().length < minDepth)) {
      counter.nUncovered = 1L;
      final GVstatus status = getGVstatus(vcComp);
      if (status == GVstatus.T) counter.nAltNotCalled = 1L;
      else if (status == GVstatus.F) counter.nRefNotCalled = 1L;
      else counter.nNoStatusNotCalled = 1L;

      return counter;
    }

    VariantCallContext call;
    if (vcComp.isSNP()) {
      call = snpEngine.calculateLikelihoodsAndGenotypes(tracker, ref, context).get(0);
    } else if (vcComp.isIndel()) {
      call = indelEngine.calculateLikelihoodsAndGenotypes(tracker, ref, context).get(0);
    } else if (bamIsTruth) {
      // assume it's a SNP if no variation is present; this is necessary so that we can test
      // supposed monomorphic sites against the truth bam
      call = snpEngine.calculateLikelihoodsAndGenotypes(tracker, ref, context).get(0);
    } else {
      logger.info(
          "Not SNP or INDEL "
              + vcComp.getChr()
              + ":"
              + vcComp.getStart()
              + " "
              + vcComp.getAlleles());
      return counter;
    }

    boolean writeVariant = true;

    if (bamIsTruth) {
      if (call.confidentlyCalled) {
        // If truth is a confident REF call
        if (call.isVariant()) {
          if (vcComp.isVariant()) counter.nAltCalledAlt = 1L;
          else {
            counter.nAltCalledRef = 1L;
            if (printInterestingSites)
              System.out.println("Truth=ALT Call=REF at " + call.getChr() + ":" + call.getStart());
          }
        }
        // If truth is a confident ALT call
        else {
          if (vcComp.isVariant()) {
            counter.nRefCalledAlt = 1L;
            if (printInterestingSites)
              System.out.println("Truth=REF Call=ALT at " + call.getChr() + ":" + call.getStart());
          } else counter.nRefCalledRef = 1L;
        }
      } else {
        counter.nNotConfidentCalls = 1L;
        if (printInterestingSites)
          System.out.println("Truth is not confident at " + call.getChr() + ":" + call.getStart());
        writeVariant = false;
      }
    } else {
      //            if (!vcComp.hasExtendedAttribute("GV"))
      //                throw new UserException.BadInput("Variant has no GV annotation in the INFO
      // field. " + vcComp.getChr() + ":" + vcComp.getStart());

      final GVstatus status = getGVstatus(vcComp);
      if (call.isCalledAlt(callConf)) {
        if (status == GVstatus.T) counter.nAltCalledAlt = 1L;
        else if (status == GVstatus.F) {
          counter.nRefCalledAlt = 1L;
          if (printInterestingSites)
            System.out.println("Truth=REF Call=ALT at " + call.getChr() + ":" + call.getStart());
        } else counter.nNoStatusCalledAlt = 1L;
      } else if (call.isCalledRef(callConf)) {
        if (status == GVstatus.T) {
          counter.nAltCalledRef = 1L;
          if (printInterestingSites)
            System.out.println("Truth=ALT Call=REF at " + call.getChr() + ":" + call.getStart());
        } else if (status == GVstatus.F) counter.nRefCalledRef = 1L;
        else counter.nNoStatusCalledRef = 1L;
      } else {
        counter.nNotConfidentCalls = 1L;
        if (status == GVstatus.T) counter.nAltNotCalled = 1L;
        else if (status == GVstatus.F) counter.nRefNotCalled = 1L;
        else counter.nNoStatusNotCalled = 1L;

        if (printInterestingSites)
          System.out.println("Truth is not confident at " + call.getChr() + ":" + call.getStart());
        writeVariant = false;
      }
    }

    if (vcfWriter != null && writeVariant) {
      if (!vcComp.hasAttribute(GATKVCFConstants.GENOTYPE_AND_VALIDATE_STATUS_KEY)) {
        vcfWriter.add(
            new VariantContextBuilder(vcComp)
                .attribute(
                    GATKVCFConstants.GENOTYPE_AND_VALIDATE_STATUS_KEY,
                    call.isCalledAlt(callConf) ? "ALT" : "REF")
                .make());
      } else vcfWriter.add(vcComp);
    }
    return counter;
  }