예제 #1
0
  @Override
  public void execute() {
    VariantContextWriter vcw =
        new VariantContextWriterBuilder()
            .setOutputFile(out)
            .clearIndexCreator()
            .setReferenceDictionary(VCF.getFileHeader().getSequenceDictionary())
            .build();

    VCFHeader vh = VCF.getFileHeader();
    vh.addMetaDataLine(
        new VCFInfoHeaderLine(
            "region", 1, VCFHeaderLineType.String, "region type in which the variant occurs"));

    vcw.writeHeader(VCF.getFileHeader());

    IntervalTreeMap<String> itm = loadRegions();

    int num = 0;

    for (VariantContext vc : VCF) {
      if (!vc.isFiltered()
          && vc.getGenotype(CHILD).isCalled()
          && vc.getGenotype(MOTHER).isCalled()
          && vc.getGenotype(FATHER).isCalled()
          && vc.getGenotype(CHILD).getType().equals(GenotypeType.HET)
          && vc.getGenotype(MOTHER).getType().equals(GenotypeType.HOM_REF)
          && vc.getGenotype(FATHER).getType().equals(GenotypeType.HOM_REF)
          && (!vc.getGenotype(CHILD).hasAnyAttribute("GQ")
              || (vc.getGenotype(CHILD).getGQ() > GQ_THRESHOLD
                  && vc.getGenotype(MOTHER).getGQ() > GQ_THRESHOLD
                  && vc.getGenotype(FATHER).getGQ() > GQ_THRESHOLD))
          && (!vc.getGenotype(CHILD).hasAnyAttribute("DP")
              || (vc.getGenotype(CHILD).getDP() > DP_THRESHOLD
                  && vc.getGenotype(MOTHER).getDP() > DP_THRESHOLD
                  && vc.getGenotype(FATHER).getDP() > DP_THRESHOLD))) {
        vcw.add(vc);

        num++;
      }
    }

    vcw.close();

    log.info("num: {}", num);
  }
예제 #2
0
  @Override
  protected Object doWork() {
    IOUtil.assertFileIsReadable(INPUT);
    IOUtil.assertFileIsReadable(REFERENCE_SEQUENCE);
    IOUtil.assertFileIsReadable(CHAIN);
    IOUtil.assertFileIsWritable(OUTPUT);
    IOUtil.assertFileIsWritable(REJECT);

    ////////////////////////////////////////////////////////////////////////
    // Setup the inputs
    ////////////////////////////////////////////////////////////////////////
    final LiftOver liftOver = new LiftOver(CHAIN);
    final VCFFileReader in = new VCFFileReader(INPUT, false);

    logger.info("Loading up the target reference genome.");
    final ReferenceSequenceFileWalker walker = new ReferenceSequenceFileWalker(REFERENCE_SEQUENCE);
    final Map<String, byte[]> refSeqs = new HashMap<>();
    for (final SAMSequenceRecord rec : walker.getSequenceDictionary().getSequences()) {
      refSeqs.put(rec.getSequenceName(), walker.get(rec.getSequenceIndex()).getBases());
    }
    CloserUtil.close(walker);

    ////////////////////////////////////////////////////////////////////////
    // Setup the outputs
    ////////////////////////////////////////////////////////////////////////
    final VCFHeader inHeader = in.getFileHeader();
    final VCFHeader outHeader = new VCFHeader(inHeader);
    outHeader.setSequenceDictionary(walker.getSequenceDictionary());
    final VariantContextWriter out =
        new VariantContextWriterBuilder()
            .setOption(Options.INDEX_ON_THE_FLY)
            .setOutputFile(OUTPUT)
            .setReferenceDictionary(walker.getSequenceDictionary())
            .build();
    out.writeHeader(outHeader);

    final VariantContextWriter rejects =
        new VariantContextWriterBuilder()
            .setOutputFile(REJECT)
            .unsetOption(Options.INDEX_ON_THE_FLY)
            .build();
    final VCFHeader rejectHeader = new VCFHeader(in.getFileHeader());
    for (final VCFFilterHeaderLine line : FILTERS) rejectHeader.addMetaDataLine(line);
    rejects.writeHeader(rejectHeader);

    ////////////////////////////////////////////////////////////////////////
    // Read the input VCF, lift the records over and write to the sorting
    // collection.
    ////////////////////////////////////////////////////////////////////////
    long failedLiftover = 0, failedAlleleCheck = 0, total = 0;
    logger.info("Lifting variants over and sorting.");

    final SortingCollection<VariantContext> sorter =
        SortingCollection.newInstance(
            VariantContext.class,
            new VCFRecordCodec(outHeader),
            outHeader.getVCFRecordComparator(),
            MAX_RECORDS_IN_RAM,
            TMP_DIR);

    ProgressLogger progress = new ProgressLogger(logger, 1000000, "read");

    for (final VariantContext ctx : in) {
      ++total;
      final Interval source =
          new Interval(
              ctx.getContig(),
              ctx.getStart(),
              ctx.getEnd(),
              false,
              ctx.getContig() + ":" + ctx.getStart() + "-" + ctx.getEnd());
      final Interval target = liftOver.liftOver(source, 1.0);

      if (target == null) {
        rejects.add(new VariantContextBuilder(ctx).filter(FILTER_CANNOT_LIFTOVER).make());
        failedLiftover++;
      } else {
        // Fix the alleles if we went from positive to negative strand
        final List<Allele> alleles = new ArrayList<>();
        for (final Allele oldAllele : ctx.getAlleles()) {
          if (target.isPositiveStrand() || oldAllele.isSymbolic()) {
            alleles.add(oldAllele);
          } else {
            alleles.add(
                Allele.create(
                    SequenceUtil.reverseComplement(oldAllele.getBaseString()),
                    oldAllele.isReference()));
          }
        }

        // Build the new variant context
        final VariantContextBuilder builder =
            new VariantContextBuilder(
                ctx.getSource(), target.getContig(), target.getStart(), target.getEnd(), alleles);

        builder.id(ctx.getID());
        builder.attributes(ctx.getAttributes());
        builder.genotypes(ctx.getGenotypes());
        builder.filters(ctx.getFilters());
        builder.log10PError(ctx.getLog10PError());

        // Check that the reference allele still agrees with the reference sequence
        boolean mismatchesReference = false;
        for (final Allele allele : builder.getAlleles()) {
          if (allele.isReference()) {
            final byte[] ref = refSeqs.get(target.getContig());
            final String refString =
                StringUtil.bytesToString(ref, target.getStart() - 1, target.length());

            if (!refString.equalsIgnoreCase(allele.getBaseString())) {
              mismatchesReference = true;
            }

            break;
          }
        }

        if (mismatchesReference) {
          rejects.add(new VariantContextBuilder(ctx).filter(FILTER_MISMATCHING_REF_ALLELE).make());
          failedAlleleCheck++;
        } else {
          sorter.add(builder.make());
        }
      }

      progress.record(ctx.getContig(), ctx.getStart());
    }

    final NumberFormat pfmt = new DecimalFormat("0.0000%");
    final String pct = pfmt.format((failedLiftover + failedAlleleCheck) / (double) total);
    logger.info("Processed ", total, " variants.");
    logger.info(Long.toString(failedLiftover), " variants failed to liftover.");
    logger.info(
        Long.toString(failedAlleleCheck),
        " variants lifted over but had mismatching reference alleles after lift over.");
    logger.info(pct, " of variants were not successfully lifted over and written to the output.");

    rejects.close();
    in.close();

    ////////////////////////////////////////////////////////////////////////
    // Write the sorted outputs to the final output file
    ////////////////////////////////////////////////////////////////////////
    sorter.doneAdding();
    progress = new ProgressLogger(logger, 1000000, "written");
    logger.info("Writing out sorted records to final VCF.");

    for (final VariantContext ctx : sorter) {
      out.add(ctx);
      progress.record(ctx.getContig(), ctx.getStart());
    }
    out.close();
    sorter.cleanup();

    return null;
  }
  @Override
  public void runCommand() {

    logger.info("MergeVCFColumnsCommand");
    /*
     * Assumptions
     * (1) Only two vcfs that are sorted with the same contig order
     * (2) if contigs on it same order, then we will just skip that contig
     * (3) No overlapping samples allowed
     *
     * Output:
     * A vcf where intersecting sites are merged together and will only return biallelic markers
     * the info field will be cleared
     * the only GT FORMAT field will be there
     */

    Collection<File> vcfs = applicationOptions.getVcfs();
    String outfile = applicationOptions.getOutFile();

    if (vcfs.size() != 2) {
      throw new IllegalArgumentException("This function requires exactly two vcfs");
    }

    Iterator<File> vcfFileIter = vcfs.iterator();

    File vcf1 = vcfFileIter.next();
    File vcf2 = vcfFileIter.next();
    VCFFileReader reader1 = new VCFFileReader(vcf1, false);
    VCFFileReader reader2 = new VCFFileReader(vcf2, false);

    Iterator<VariantContext> iter1 = reader1.iterator();
    Iterator<VariantContext> iter2 = reader2.iterator();

    VariantContextComparator comparator = new VariantContextComparator();

    /*
     * Merge headers
     */
    VCFHeader header1 = reader1.getFileHeader();

    VCFHeader header2 = reader2.getFileHeader();

    List<String> samples1 = header1.getGenotypeSamples();
    List<String> samples2 = header2.getGenotypeSamples();

    List<String> mergedSamples = new ArrayList<>(samples1.size() + samples2.size());
    mergedSamples.addAll(samples1);
    mergedSamples.addAll(samples2);

    // Validate that there are no duplicates
    HashSet<String> sampleSet = new HashSet<String>();
    for (String id : mergedSamples) {
      if (sampleSet.contains(id)) {
        throw new IllegalArgumentException("Duplicate id found: " + id);
      } else {
        sampleSet.add(id);
      }
    }

    HashSet<VCFHeaderLine> meta = new HashSet<>();
    meta.add(new VCFFormatHeaderLine("GT", 1, VCFHeaderLineType.String, "GT"));
    meta.addAll(header1.getContigLines());
    VCFHeader mergedHeader = new VCFHeader(meta, mergedSamples);

    /*
     * Create encoder
     */
    VCFEncoder encoder = new VCFEncoder(mergedHeader, false, false);

    BufferedWriter writer = null;
    try {
      if (outfile.endsWith(".gz")) {
        BlockCompressedOutputStream outstream = new BlockCompressedOutputStream(new File(outfile));
        writer = new BufferedWriter(new OutputStreamWriter(outstream));
      } else {
        writer =
            Files.newBufferedWriter(
                Paths.get(outfile),
                Charset.defaultCharset(),
                StandardOpenOption.CREATE,
                StandardOpenOption.TRUNCATE_EXISTING);
      }

      /*
       * Write header
       */
      VCFHeaderWriter.writeHeader(writer, mergedHeader);
      logger.info("Wrote header");

      VariantContext previous1 = null;
      VariantContext previous2 = null;

      int count = 0;

      int countFile1 = 0;
      int countFile2 = 0;

      boolean usePrevious1 = false;
      boolean usePrevious2 = false;
      while (iter1.hasNext() || iter2.hasNext()) {

        if ((iter1.hasNext() || usePrevious1) && (iter2.hasNext() || usePrevious2)) {

          VariantContext variant1 = null;

          VariantContext variant2 = null;

          //					if(usePrevious1 == true && usePrevious2 == true &&
          // comparator.compare(previous1,previous2) != 0) {
          //						//then skip both
          //						usePrevious1 = false;
          //						usePrevious2 = false;
          //					}

          if (usePrevious1) {
            variant1 = previous1;
          } else {
            variant1 = iter1.next();
            countFile1++;
          }

          if (usePrevious2) {
            variant2 = previous2;
          } else {
            variant2 = iter2.next();
            countFile2++;
          }

          // check that variants are ordered correctly
          if (previous1 != null
              && previous1 != variant1
              && comparator.compare(previous1, variant1) > 0) {
            throw new IllegalStateException(
                previous1.getContig()
                    + ":"
                    + previous1.getStart()
                    + " > "
                    + variant1.getContig()
                    + ":"
                    + variant1.getStart());
          }

          if (previous2 != null
              && previous2 != variant2
              && comparator.compare(previous2, variant2) > 0) {
            throw new IllegalStateException(
                previous2.getContig()
                    + ":"
                    + previous2.getStart()
                    + " > "
                    + variant2.getContig()
                    + ":"
                    + variant2.getStart());
          }

          int cmp = comparator.compare(variant1, variant2);

          if (cmp < 0) {
            // logger.info("Skipping VCF1: " + variant1.getContig() + ":" + variant1.getStart() +
            // "\t" + variant1.getReference().toString() +  "\t" + variant1.getAlternateAlleles());

            if (usePrevious1 == true && usePrevious2 == true) {
              // variant1 < variant2
              // we need to go to next variant in vcf1
              usePrevious1 = false;
            }

            usePrevious2 = true;
          } else if (cmp > 0) {

            if (usePrevious1 == true && usePrevious2 == true) {
              // variant1 > variant2
              // we need to go to next variant in vcf2
              usePrevious2 = false;
            }

            usePrevious1 = true;
            // logger.info("Skipping VCF2: " + variant2.getContig() + ":" + variant2.getStart() +
            // "\t" + variant2.getReference().toString() +  "\t" + variant2.getAlternateAlleles());
          } else {
            // they equal position
            usePrevious1 = false;
            usePrevious2 = false;

            if (variant1.isBiallelic()
                && variant2.isBiallelic()
                && variant1.getReference().equals(variant2.getReference())
                && variant1.getAlternateAllele(0).equals(variant2.getAlternateAllele(0))) {

              // TODO: Finish merging
              // both variants are bialleleic and the reference and alternative alleles match

              count++;
              if (count % 10000 == 0) {
                logger.info(count + " mergeable variants found");
              }

              VariantContext merged = VariantContextMerger.merge(variant1, variant2);

              writer.write(encoder.encode(merged));
              writer.write("\n");

            } else {
              // skip if they do not equal

              //			logger.info("Skipping: " + variant1.getContig() + ":" + variant1.getStart() +
              // "\t" + variant1.getReference().toString() +  "\t" +
              // variant1.getAlternateAlleles());
              //			logger.info("Skipping: " + variant2.getContig() + ":" + variant2.getStart() +
              // "\t" + variant2.getReference().toString() +  "\t" +
              // variant2.getAlternateAlleles());
            }
          }

          previous1 = variant1;
          previous2 = variant2;
        } else if (iter1.hasNext()) {
          // just skip remaining variants
          VariantContext current = iter1.next();
          countFile1++;

          if (previous1 != null && current != null && comparator.compare(previous1, current) > 0) {
            throw new IllegalStateException(
                previous1.getContig()
                    + ":"
                    + previous1.getStart()
                    + " > "
                    + current.getContig()
                    + ":"
                    + current.getStart());
          }

          previous1 = current;

          // logger.info("Skipping: " + previous1.getContig() + ":" + previous1.getStart() + "\t" +
          // previous1.getReference().toString() +  "\t" + previous1.getAlternateAlleles());

        } else if (iter2.hasNext()) {
          // just skip remaining variants
          // fixed bug/ was iter1 changed to iter2
          VariantContext current = iter2.next();
          countFile2++;

          if (previous2 != null && current != null && comparator.compare(previous2, current) > 0) {
            throw new IllegalStateException(
                previous2.getContig()
                    + ":"
                    + previous2.getStart()
                    + " > "
                    + current.getContig()
                    + ":"
                    + current.getStart());
          }

          previous2 = current;

          // logger.info("Skipping: " + previous2.getContig() + ":" + previous2.getStart() + "\t" +
          // previous2.getReference().toString() +  "\t" + previous2.getAlternateAlleles());
        } else {
          throw new IllegalStateException("Error should not of reached this point");
        }
      }

      reader1.close();
      reader2.close();

      logger.info(count + "  merged variants");
      logger.info(countFile1 + "  variants in " + vcf1.getAbsolutePath());
      logger.info(countFile2 + "  variants in " + vcf2.getAbsolutePath());

    } catch (Exception e) {
      e.printStackTrace();
    } finally {
      if (writer != null) {
        try {
          logger.info("Flushing writer");
          writer.close();
        } catch (IOException e) {
          // TODO Auto-generated catch block
          e.printStackTrace();
        }
      }
    }

    logger.info("finished merging vcfs");
  }