@Override
protected void finish() {
final MetricsFile<PreAdapterSummaryMetrics, Integer> preAdapterSummaryMetricsFile =
getMetricsFile();
final MetricsFile<PreAdapterDetailMetrics, Integer> preAdapterDetailMetricsFile =
getMetricsFile();
final MetricsFile<BaitBiasSummaryMetrics, Integer> baitBiasSummaryMetricsFile =
getMetricsFile();
final MetricsFile<BaitBiasDetailMetrics, Integer> baitBiasDetailMetricsFile = getMetricsFile();
for (final ArtifactCounter counter : artifactCounters.values()) {
// build metrics
counter.finish();
// write metrics
preAdapterSummaryMetricsFile.addAllMetrics(counter.getPreAdapterSummaryMetrics());
baitBiasSummaryMetricsFile.addAllMetrics(counter.getBaitBiasSummaryMetrics());
for (final PreAdapterDetailMetrics preAdapterDetailMetrics :
counter.getPreAdapterDetailMetrics()) {
if (CONTEXTS_TO_PRINT.isEmpty()
|| CONTEXTS_TO_PRINT.contains(preAdapterDetailMetrics.CONTEXT)) {
preAdapterDetailMetricsFile.addMetric(preAdapterDetailMetrics);
}
}
for (final BaitBiasDetailMetrics baitBiasDetailMetrics : counter.getBaitBiasDetailMetrics()) {
if (CONTEXTS_TO_PRINT.isEmpty()
|| CONTEXTS_TO_PRINT.contains(baitBiasDetailMetrics.CONTEXT)) {
baitBiasDetailMetricsFile.addMetric(baitBiasDetailMetrics);
}
}
}
preAdapterDetailMetricsFile.write(preAdapterDetailsOut);
preAdapterSummaryMetricsFile.write(preAdapterSummaryOut);
baitBiasDetailMetricsFile.write(baitBiasDetailsOut);
baitBiasSummaryMetricsFile.write(baitBiasSummaryOut);
}
@Override
protected void acceptRead(final SAMRecord rec, final ReferenceSequence ref) {
// see if the whole read should be skipped
if (recordFilter.filterOut(rec)) return;
// check read group + library
final String library =
(rec.getReadGroup() == null)
? UNKNOWN_LIBRARY
: getOrElse(rec.getReadGroup().getLibrary(), UNKNOWN_LIBRARY);
if (!libraries.contains(library)) {
// should never happen if SAM is valid
throw new PicardException("Record contains library that is missing from header: " + library);
}
// set up some constants that don't change in the loop below
final int contextFullLength = 2 * CONTEXT_SIZE + 1;
final ArtifactCounter counter = artifactCounters.get(library);
final byte[] readBases = rec.getReadBases();
final byte[] readQuals;
if (USE_OQ) {
final byte[] tmp = rec.getOriginalBaseQualities();
readQuals = tmp == null ? rec.getBaseQualities() : tmp;
} else {
readQuals = rec.getBaseQualities();
}
// iterate over aligned positions
for (final AlignmentBlock block : rec.getAlignmentBlocks()) {
for (int offset = 0; offset < block.getLength(); offset++) {
// remember, these are 1-based!
final int readPos = block.getReadStart() + offset;
final int refPos = block.getReferenceStart() + offset;
// skip low BQ sites
final byte qual = readQuals[readPos - 1];
if (qual < MINIMUM_QUALITY_SCORE) continue;
// skip N bases in read
final char readBase = Character.toUpperCase((char) readBases[readPos - 1]);
if (readBase == 'N') continue;
/**
* Skip regions outside of intervals.
*
* <p>NB: IntervalListReferenceSequenceMask.get() has side-effects which assume that
* successive ReferenceSequence's passed to this method will be in-order (e.g. it will break
* if you call acceptRead() with chr1, then chr2, then chr1 again). So this only works if
* the underlying iteration is coordinate-sorted.
*/
if (intervalMask != null && !intervalMask.get(ref.getContigIndex(), refPos)) continue;
// skip dbSNP sites
if (dbSnpMask != null && dbSnpMask.isDbSnpSite(ref.getName(), refPos)) continue;
// skip the ends of the reference
final int contextStartIndex = refPos - CONTEXT_SIZE - 1;
if (contextStartIndex < 0 || contextStartIndex + contextFullLength > ref.length()) continue;
// skip contexts with N bases
final String context = getRefContext(ref, contextStartIndex, contextFullLength);
if (context.contains("N")) continue;
// count the base!
counter.countRecord(context, readBase, rec);
}
}
}
