Examples with SamRecordFilter htsjdk.samtools.filter.SamRecordFilter used on opensource projects

Example 6 with SamRecordFilter

use of htsjdk.samtools.filter.SamRecordFilter in project gatk by broadinstitute.

the class CountingPairedFilter method doWork.

@Override
protected Object doWork() {
    IOUtil.assertFileIsReadable(INPUT);
    IOUtil.assertFileIsWritable(OUTPUT);
    IOUtil.assertFileIsReadable(REFERENCE_SEQUENCE);
    // Setup all the inputs
    final ProgressLogger progress = new ProgressLogger(logger, 10000000, "Processed", "loci");
    final ReferenceSequenceFileWalker refWalker = new ReferenceSequenceFileWalker(REFERENCE_SEQUENCE);
    final SamReader in = SamReaderFactory.makeDefault().referenceSequence(REFERENCE_SEQUENCE).validationStringency(VALIDATION_STRINGENCY).open(INPUT);
    // Load up the reference sequence and double check sequence dictionaries
    if (!in.getFileHeader().getSequenceDictionary().isEmpty()) {
        SequenceUtil.assertSequenceDictionariesEqual(in.getFileHeader().getSequenceDictionary(), refWalker.getSequenceDictionary());
    }
    final SamLocusIterator iterator = new SamLocusIterator(in);
    final List<SamRecordFilter> filters = new ArrayList<>();
    final CountingFilter dupeFilter = new CountingDuplicateFilter();
    final CountingFilter mapqFilter = new CountingMapQFilter(MINIMUM_MAPPING_QUALITY);
    final CountingPairedFilter pairFilter = new CountingPairedFilter();
    filters.add(mapqFilter);
    filters.add(dupeFilter);
    filters.add(pairFilter);
    // Not a counting filter because we never want to count reads twice
    filters.add(new SecondaryAlignmentFilter());
    iterator.setSamFilters(filters);
    iterator.setEmitUncoveredLoci(true);
    // Handled separately because we want to count bases
    iterator.setMappingQualityScoreCutoff(0);
    // Handled separately because we want to count bases
    iterator.setQualityScoreCutoff(0);
    iterator.setIncludeNonPfReads(false);
    final int max = COVERAGE_CAP;
    final long[] HistogramArray = new long[max + 1];
    final long[] baseQHistogramArray = new long[Byte.MAX_VALUE];
    final boolean usingStopAfter = STOP_AFTER > 0;
    final long stopAfter = STOP_AFTER - 1;
    long counter = 0;
    long basesExcludedByBaseq = 0;
    long basesExcludedByOverlap = 0;
    long basesExcludedByCapping = 0;
    // Loop through all the loci
    while (iterator.hasNext()) {
        final SamLocusIterator.LocusInfo info = iterator.next();
        // Check that the reference is not N
        final ReferenceSequence ref = refWalker.get(info.getSequenceIndex());
        final byte base = ref.getBases()[info.getPosition() - 1];
        if (base == 'N')
            continue;
        // Figure out the coverage while not counting overlapping reads twice, and excluding various things
        final Set<String> readNames = new HashSet<>(info.getRecordAndOffsets().size());
        int pileupSize = 0;
        for (final SamLocusIterator.RecordAndOffset recs : info.getRecordAndOffsets()) {
            if (recs.getBaseQuality() < MINIMUM_BASE_QUALITY) {
                ++basesExcludedByBaseq;
                continue;
            }
            if (!readNames.add(recs.getRecord().getReadName())) {
                ++basesExcludedByOverlap;
                continue;
            }
            pileupSize++;
            if (pileupSize <= max) {
                baseQHistogramArray[recs.getRecord().getBaseQualities()[recs.getOffset()]]++;
            }
        }
        final int depth = Math.min(readNames.size(), max);
        if (depth < readNames.size())
            basesExcludedByCapping += readNames.size() - max;
        HistogramArray[depth]++;
        // Record progress and perhaps stop
        progress.record(info.getSequenceName(), info.getPosition());
        if (usingStopAfter && ++counter > stopAfter)
            break;
    }
    // Construct and write the outputs
    final Histogram<Integer> histo = new Histogram<>("coverage", "count");
    for (int i = 0; i < HistogramArray.length; ++i) {
        histo.increment(i, HistogramArray[i]);
    }
    // Construct and write the outputs
    final Histogram<Integer> baseQHisto = new Histogram<>("value", "baseq_count");
    for (int i = 0; i < baseQHistogramArray.length; ++i) {
        baseQHisto.increment(i, baseQHistogramArray[i]);
    }
    final WgsMetrics metrics = generateWgsMetrics();
    metrics.GENOME_TERRITORY = (long) histo.getSumOfValues();
    metrics.MEAN_COVERAGE = histo.getMean();
    metrics.SD_COVERAGE = histo.getStandardDeviation();
    metrics.MEDIAN_COVERAGE = histo.getMedian();
    metrics.MAD_COVERAGE = histo.getMedianAbsoluteDeviation();
    final long basesExcludedByDupes = dupeFilter.getFilteredBases();
    final long basesExcludedByMapq = mapqFilter.getFilteredBases();
    final long basesExcludedByPairing = pairFilter.getFilteredBases();
    final double total = histo.getSum();
    final double totalWithExcludes = total + basesExcludedByDupes + basesExcludedByMapq + basesExcludedByPairing + basesExcludedByBaseq + basesExcludedByOverlap + basesExcludedByCapping;
    metrics.PCT_EXC_DUPE = basesExcludedByDupes / totalWithExcludes;
    metrics.PCT_EXC_MAPQ = basesExcludedByMapq / totalWithExcludes;
    metrics.PCT_EXC_UNPAIRED = basesExcludedByPairing / totalWithExcludes;
    metrics.PCT_EXC_BASEQ = basesExcludedByBaseq / totalWithExcludes;
    metrics.PCT_EXC_OVERLAP = basesExcludedByOverlap / totalWithExcludes;
    metrics.PCT_EXC_CAPPED = basesExcludedByCapping / totalWithExcludes;
    metrics.PCT_EXC_TOTAL = (totalWithExcludes - total) / totalWithExcludes;
    metrics.PCT_5X = MathUtils.sum(HistogramArray, 5, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_10X = MathUtils.sum(HistogramArray, 10, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_15X = MathUtils.sum(HistogramArray, 15, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_20X = MathUtils.sum(HistogramArray, 20, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_25X = MathUtils.sum(HistogramArray, 25, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_30X = MathUtils.sum(HistogramArray, 30, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_40X = MathUtils.sum(HistogramArray, 40, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_50X = MathUtils.sum(HistogramArray, 50, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_60X = MathUtils.sum(HistogramArray, 60, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_70X = MathUtils.sum(HistogramArray, 70, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_80X = MathUtils.sum(HistogramArray, 80, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_90X = MathUtils.sum(HistogramArray, 90, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    metrics.PCT_100X = MathUtils.sum(HistogramArray, 100, HistogramArray.length) / (double) metrics.GENOME_TERRITORY;
    final MetricsFile<WgsMetrics, Integer> out = getMetricsFile();
    out.addMetric(metrics);
    out.addHistogram(histo);
    if (INCLUDE_BQ_HISTOGRAM) {
        out.addHistogram(baseQHisto);
    }
    out.write(OUTPUT);
    return null;
}

Example 7 with SamRecordFilter

use of htsjdk.samtools.filter.SamRecordFilter in project gatk-protected by broadinstitute.

the class AllelicCountCollector method collect.

/**
     * Returns an {@link AllelicCountCollection} based on the pileup at sites (specified by an interval list)
     * in a sorted BAM file.  Reads and bases below the specified mapping quality and base quality, respectively,
     * are filtered out of the pileup.  The alt count is defined as the total count minus the ref count, and the
     * alt nucleotide is defined as the non-ref base with the highest count, with ties broken by the order of the
     * bases in {@link AllelicCountCollector#BASES}.
     * @param bamFile           sorted BAM file
     * @param siteIntervals     interval list of sites
     * @param minMappingQuality minimum mapping quality required for reads to be included in pileup
     * @param minBaseQuality    minimum base quality required for bases to be included in pileup
     * @return                  AllelicCountCollection of ref/alt counts at sites in BAM file
     */
public AllelicCountCollection collect(final File bamFile, final IntervalList siteIntervals, final int minMappingQuality, final int minBaseQuality) {
    try (final SamReader reader = readerFactory.open(bamFile)) {
        ParamUtils.isPositiveOrZero(minMappingQuality, "Minimum mapping quality must be nonnegative.");
        ParamUtils.isPositiveOrZero(minBaseQuality, "Minimum base quality must be nonnegative.");
        if (reader.getFileHeader().getSortOrder() != SAMFileHeader.SortOrder.coordinate) {
            throw new UserException.BadInput("BAM file " + bamFile.toString() + " must be coordinate sorted.");
        }
        final int numberOfSites = siteIntervals.size();
        final boolean useIndex = numberOfSites < MAX_INTERVALS_FOR_INDEX;
        final SamLocusIterator locusIterator = new SamLocusIterator(reader, siteIntervals, useIndex);
        //set read and locus filters [note: read counts match IGV, but off by a few from pysam.mpileup]
        final List<SamRecordFilter> samFilters = Arrays.asList(new NotPrimaryAlignmentFilter(), new DuplicateReadFilter());
        locusIterator.setSamFilters(samFilters);
        locusIterator.setEmitUncoveredLoci(true);
        locusIterator.setIncludeNonPfReads(false);
        locusIterator.setMappingQualityScoreCutoff(minMappingQuality);
        locusIterator.setQualityScoreCutoff(minBaseQuality);
        logger.info("Examining " + numberOfSites + " sites in total...");
        int locusCount = 0;
        final AllelicCountCollection counts = new AllelicCountCollection();
        for (final SamLocusIterator.LocusInfo locus : locusIterator) {
            if (locusCount % NUMBER_OF_SITES_PER_LOGGED_STATUS_UPDATE == 0) {
                logger.info("Examined " + locusCount + " sites.");
            }
            locusCount++;
            final Nucleotide refBase = Nucleotide.valueOf(referenceWalker.get(locus.getSequenceIndex()).getBases()[locus.getPosition() - 1]);
            if (!BASES.contains(refBase)) {
                logger.warn(String.format("The reference position at %d has an unknown base call (value: %s). Skipping...", locus.getPosition(), refBase.toString()));
                continue;
            }
            final Nucleotide.Counter baseCounts = getPileupBaseCounts(locus);
            //only include total ACGT counts in binomial test (exclude N, etc.)
            final int totalBaseCount = BASES.stream().mapToInt(b -> (int) baseCounts.get(b)).sum();
            final int refReadCount = (int) baseCounts.get(refBase);
            //we take alt = total - ref instead of the actual alt count
            final int altReadCount = totalBaseCount - refReadCount;
            final Nucleotide altBase = inferAltFromPileupBaseCounts(baseCounts, refBase);
            counts.add(new AllelicCount(new SimpleInterval(locus.getSequenceName(), locus.getPosition(), locus.getPosition()), refReadCount, altReadCount, refBase, altBase));
        }
        logger.info(locusCount + " sites out of " + numberOfSites + " total sites were examined.");
        return counts;
    } catch (final IOException | SAMFormatException e) {
        throw new UserException("Unable to collect allelic counts from " + bamFile);
    }
}

Example 8 with SamRecordFilter

use of htsjdk.samtools.filter.SamRecordFilter in project gatk by broadinstitute.

the class BayesianHetPulldownCalculator method getSamLocusIteratorWithDefaultFilters.

/**
     * Returns a {@link SamLocusIterator} object for a given {@link SamReader} and {@link IntervalList} with filters
     * on minimum base quality and minimum mapping quality
     *
     * @param samReader a SamReader object
     * @return a SamLocusIterator object
     */
private SamLocusIterator getSamLocusIteratorWithDefaultFilters(final SamReader samReader) {
    final SamLocusIterator locusIterator = new SamLocusIterator(samReader, snpIntervals, false);
    /* set read and locus filters */
    final List<SamRecordFilter> samFilters = Arrays.asList(new NotPrimaryAlignmentFilter(), new DuplicateReadFilter());
    locusIterator.setSamFilters(samFilters);
    locusIterator.setEmitUncoveredLoci(false);
    locusIterator.setIncludeNonPfReads(false);
    locusIterator.setMappingQualityScoreCutoff(minMappingQuality);
    locusIterator.setQualityScoreCutoff(minBaseQuality);
    return locusIterator;
}

Example 9 with SamRecordFilter

use of htsjdk.samtools.filter.SamRecordFilter in project gatk by broadinstitute.

the class HetPulldownCalculator method getHetPulldown.

/**
     * For a normal or tumor sample, returns a data structure giving (intervals, reference counts, alternate counts),
     * where intervals give positions of likely heterozygous SNP sites.
     *
     * <p>
     *     For a normal sample:
     *     <ul>
     *         The IntervalList snpIntervals gives common SNP sites in 1-based format.
     *     </ul>
     *     <ul>
     *         The p-value threshold must be specified for a two-sided binomial test,
     *         which is used to determine SNP sites from snpIntervals that are
     *         compatible with a heterozygous SNP, given the sample.  Only these sites are output.
     *     </ul>
     * </p>
     * <p>
     *     For a tumor sample:
     *     <ul>
     *         The IntervalList snpIntervals gives heterozygous SNP sites likely to be present in the normal sample.
     *         This should be from {@link HetPulldownCalculator#getNormal} in 1-based format.
     *         Only these sites are output.
     *     </ul>
     * </p>
     * @param bamFile           sorted BAM file for sample
     * @param snpIntervals      IntervalList of SNP sites
     * @param sampleType        flag indicating type of sample (SampleType.NORMAL or SampleType.TUMOR)
     *                          (determines whether to perform binomial test)
     * @param pvalThreshold     p-value threshold for two-sided binomial test, used for normal sample
     * @param minimumRawReads   minimum number of total reads that must be present at a het site
     * @return                  Pulldown of heterozygous SNP sites in 1-based format
     */
private Pulldown getHetPulldown(final File bamFile, final IntervalList snpIntervals, final SampleType sampleType, final double pvalThreshold, final int minimumRawReads) {
    try (final SamReader bamReader = SamReaderFactory.makeDefault().validationStringency(validationStringency).referenceSequence(refFile).open(bamFile);
        final ReferenceSequenceFileWalker refWalker = new ReferenceSequenceFileWalker(refFile)) {
        if (bamReader.getFileHeader().getSortOrder() != SAMFileHeader.SortOrder.coordinate) {
            throw new UserException.BadInput("BAM file " + bamFile.toString() + " must be coordinate sorted.");
        }
        final Pulldown hetPulldown = new Pulldown(bamReader.getFileHeader());
        final int totalNumberOfSNPs = snpIntervals.size();
        final SamLocusIterator locusIterator = new SamLocusIterator(bamReader, snpIntervals, totalNumberOfSNPs < MAX_INTERVALS_FOR_INDEX);
        //set read and locus filters [note: read counts match IGV, but off by a few from pysam.mpileup]
        final List<SamRecordFilter> samFilters = Arrays.asList(new NotPrimaryAlignmentFilter(), new DuplicateReadFilter());
        locusIterator.setSamFilters(samFilters);
        locusIterator.setEmitUncoveredLoci(false);
        locusIterator.setIncludeNonPfReads(false);
        locusIterator.setMappingQualityScoreCutoff(minMappingQuality);
        locusIterator.setQualityScoreCutoff(minBaseQuality);
        logger.info("Examining " + totalNumberOfSNPs + " sites in total...");
        int locusCount = 0;
        for (final SamLocusIterator.LocusInfo locus : locusIterator) {
            if (locusCount % NUMBER_OF_SITES_PER_LOGGED_STATUS_UPDATE == 0) {
                logger.info("Examined " + locusCount + " covered sites.");
            }
            locusCount++;
            //include N, etc. reads here
            final int totalReadCount = locus.getRecordAndOffsets().size();
            if (totalReadCount < minimumRawReads) {
                continue;
            }
            final Nucleotide.Counter baseCounts = getPileupBaseCounts(locus);
            //only include total ACGT counts in binomial test (exclude N, etc.)
            final int totalBaseCount = Arrays.stream(BASES).mapToInt(b -> (int) baseCounts.get(b)).sum();
            if (sampleType == SampleType.NORMAL && !isPileupHetCompatible(baseCounts, totalBaseCount, pvalThreshold)) {
                continue;
            }
            final Nucleotide refBase = Nucleotide.valueOf(refWalker.get(locus.getSequenceIndex()).getBases()[locus.getPosition() - 1]);
            final int refReadCount = (int) baseCounts.get(refBase);
            final int altReadCount = totalBaseCount - refReadCount;
            hetPulldown.add(new AllelicCount(new SimpleInterval(locus.getSequenceName(), locus.getPosition(), locus.getPosition()), refReadCount, altReadCount));
        }
        logger.info(locusCount + " covered sites out of " + totalNumberOfSNPs + " total sites were examined.");
        return hetPulldown;
    } catch (final IOException | SAMFormatException e) {
        throw new UserException(e.getMessage());
    }
}