public void onComplete() {
// summarize read data
if (metrics.TOTAL_READS > 0) {
metrics.PCT_PF_READS = (double) metrics.PF_READS / (double) metrics.TOTAL_READS;
metrics.PCT_ADAPTER = this.adapterReads / (double) metrics.PF_READS;
metrics.MEAN_READ_LENGTH = readLengthHistogram.getMean();
// Calculate BAD_CYCLES
metrics.BAD_CYCLES = 0;
for (final Histogram<Integer>.Bin cycleBin : badCycleHistogram.values()) {
final double badCyclePercentage = cycleBin.getValue() / metrics.TOTAL_READS;
if (badCyclePercentage >= .8) {
metrics.BAD_CYCLES++;
}
}
if (doRefMetrics) {
if (metrics.PF_READS > 0)
metrics.PCT_PF_READS_ALIGNED =
(double) metrics.PF_READS_ALIGNED / (double) metrics.PF_READS;
if (metrics.PF_READS_ALIGNED > 0)
metrics.PCT_READS_ALIGNED_IN_PAIRS =
(double) metrics.READS_ALIGNED_IN_PAIRS / (double) metrics.PF_READS_ALIGNED;
if (metrics.PF_READS_ALIGNED > 0)
metrics.STRAND_BALANCE = numPositiveStrand / (double) metrics.PF_READS_ALIGNED;
if (this.chimerasDenominator > 0)
metrics.PCT_CHIMERAS = this.chimeras / (double) this.chimerasDenominator;
if (nonBisulfiteAlignedBases > 0)
metrics.PF_MISMATCH_RATE =
mismatchHistogram.getSum() / (double) nonBisulfiteAlignedBases;
metrics.PF_HQ_MEDIAN_MISMATCHES = hqMismatchHistogram.getMedian();
if (hqNonBisulfiteAlignedBases > 0)
metrics.PF_HQ_ERROR_RATE =
hqMismatchHistogram.getSum() / (double) hqNonBisulfiteAlignedBases;
if (metrics.PF_ALIGNED_BASES > 0)
metrics.PF_INDEL_RATE = this.indels / (double) metrics.PF_ALIGNED_BASES;
}
}
}
private void collectReadData(final SAMRecord record, final ReferenceSequence ref) {
metrics.TOTAL_READS++;
readLengthHistogram.increment(record.getReadBases().length);
if (!record.getReadFailsVendorQualityCheckFlag()) {
metrics.PF_READS++;
if (isNoiseRead(record)) metrics.PF_NOISE_READS++;
if (record.getReadUnmappedFlag()) {
// If the read is unmapped see if it's adapter sequence
final byte[] readBases = record.getReadBases();
if (!(record instanceof BAMRecord)) StringUtil.toUpperCase(readBases);
if (isAdapterSequence(readBases)) {
this.adapterReads++;
}
} else if (doRefMetrics) {
metrics.PF_READS_ALIGNED++;
if (!record.getReadNegativeStrandFlag()) numPositiveStrand++;
if (record.getReadPairedFlag() && !record.getMateUnmappedFlag()) {
metrics.READS_ALIGNED_IN_PAIRS++;
// Check that both ends have mapq > minimum
final Integer mateMq = record.getIntegerAttribute("MQ");
if (mateMq == null
|| mateMq >= MAPPING_QUALITY_THRESOLD
&& record.getMappingQuality() >= MAPPING_QUALITY_THRESOLD) {
++this.chimerasDenominator;
// With both reads mapped we can see if this pair is chimeric
if (Math.abs(record.getInferredInsertSize()) > maxInsertSize
|| !record.getReferenceIndex().equals(record.getMateReferenceIndex())) {
++this.chimeras;
}
}
}
}
}
}
private void collectQualityData(final SAMRecord record, final ReferenceSequence reference) {
// If the read isnt an aligned PF read then look at the read for no-calls
if (record.getReadUnmappedFlag()
|| record.getReadFailsVendorQualityCheckFlag()
|| !doRefMetrics) {
final byte[] readBases = record.getReadBases();
for (int i = 0; i < readBases.length; i++) {
if (SequenceUtil.isNoCall(readBases[i])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
} else if (!record.getReadFailsVendorQualityCheckFlag()) {
final boolean highQualityMapping = isHighQualityMapping(record);
if (highQualityMapping) metrics.PF_HQ_ALIGNED_READS++;
final byte[] readBases = record.getReadBases();
final byte[] refBases = reference.getBases();
final byte[] qualities = record.getBaseQualities();
final int refLength = refBases.length;
long mismatchCount = 0;
long hqMismatchCount = 0;
for (final AlignmentBlock alignmentBlock : record.getAlignmentBlocks()) {
final int readIndex = alignmentBlock.getReadStart() - 1;
final int refIndex = alignmentBlock.getReferenceStart() - 1;
final int length = alignmentBlock.getLength();
for (int i = 0; i < length && refIndex + i < refLength; ++i) {
final int readBaseIndex = readIndex + i;
boolean mismatch =
!SequenceUtil.basesEqual(readBases[readBaseIndex], refBases[refIndex + i]);
boolean bisulfiteBase = false;
if (mismatch && isBisulfiteSequenced) {
if ((record.getReadNegativeStrandFlag()
&& (refBases[refIndex + i] == 'G' || refBases[refIndex + i] == 'g')
&& (readBases[readBaseIndex] == 'A' || readBases[readBaseIndex] == 'a'))
|| ((!record.getReadNegativeStrandFlag())
&& (refBases[refIndex + i] == 'C' || refBases[refIndex + i] == 'c')
&& (readBases[readBaseIndex] == 'T')
|| readBases[readBaseIndex] == 't')) {
bisulfiteBase = true;
mismatch = false;
}
}
if (mismatch) mismatchCount++;
metrics.PF_ALIGNED_BASES++;
if (!bisulfiteBase) nonBisulfiteAlignedBases++;
if (highQualityMapping) {
metrics.PF_HQ_ALIGNED_BASES++;
if (!bisulfiteBase) hqNonBisulfiteAlignedBases++;
if (qualities[readBaseIndex] >= BASE_QUALITY_THRESHOLD)
metrics.PF_HQ_ALIGNED_Q20_BASES++;
if (mismatch) hqMismatchCount++;
}
if (mismatch || SequenceUtil.isNoCall(readBases[readBaseIndex])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
}
mismatchHistogram.increment(mismatchCount);
hqMismatchHistogram.increment(hqMismatchCount);
// Add any insertions and/or deletions to the global count
for (final CigarElement elem : record.getCigar().getCigarElements()) {
final CigarOperator op = elem.getOperator();
if (op == CigarOperator.INSERTION || op == CigarOperator.DELETION) ++this.indels;
}
}
}
