/**
* Main work method. Reads the BAM file once and collects sorted information about
* the 5' ends of both ends of each read (or just one end in the case of pairs).
* Then makes a pass through those determining duplicates before re-reading the
* input file and writing it out with duplication flags set correctly.
*/
int MarkDuplicates::runInternal() {
ogeNameThread("am_MarkDuplicates");
if(verbose)
cerr << "Reading input file and constructing read end information." << endl;
buildSortedReadEndLists();
generateDuplicateIndexes();
if(verbose)
cerr << "Marking " << numDuplicateIndices << " records as duplicates." << endl;
BamReader in;
in.Open(getBufferFileName());
// Now copy over the file while marking all the necessary indexes as duplicates
long recordInFileIndex = 0;
long written = 0;
while (true) {
BamAlignment * prec = in.GetNextAlignment();
if(!prec)
break;
if (prec->IsPrimaryAlignment()) {
if (duplicateIndexes.count(recordInFileIndex) == 1)
prec->SetIsDuplicate(true);
else
prec->SetIsDuplicate(false);
}
recordInFileIndex++;
if (removeDuplicates && prec->IsDuplicate()) {
// do nothing
}
else {
putOutputAlignment(prec);
if (verbose && read_count && ++written % 100000 == 0) {
cerr << "\rWritten " << written << " records (" << written * 100 / read_count <<"%)." << std::flush;
}
}
}
if (verbose && read_count)
cerr << "\rWritten " << written << " records (" << written * 100 / read_count <<"%)." << endl;
in.Close();
remove(getBufferFileName().c_str());
return 0;
}
int DataStatisticsTool::Execute()
{
// iterate over reads in BAM file(s)
BamAlignment alignObj;
while(bamReader.GetNextAlignment(alignObj))
{
if (alignObj.IsDuplicate()) continue;
if (alignObj.IsFailedQC()) continue;
if (!alignObj.IsMapped()) continue;
if (!alignObj.IsPrimaryAlignment()) continue;
if (alignObj.IsPaired() && !alignObj.IsProperPair()) continue;
if (alignObj.IsPaired() && !alignObj.IsMateMapped()) continue;
if (!alignObj.HasTag("MD")) continue;
// // debug
// GenericBamAlignmentTools::printBamAlignmentCigar(alignObj);
// GenericBamAlignmentTools::printBamAlignmentMD(alignObj);
// shift InDel
GenericBamAlignmentTools::leftShiftInDel(alignObj);
// // debug
// GenericBamAlignmentTools::printBamAlignmentCigar(alignObj);
// GenericBamAlignmentTools::printBamAlignmentMD(alignObj);
// get the alignment sequences
string alignRead;
string alignGenome;
GenericBamAlignmentTools::getAlignmentSequences(alignObj, alignRead, alignGenome);
// update the statistics
statistics.update(alignRead, alignGenome);
}
// print to screen
cout << statistics << endl;
// statistics.printMatchMismatch();
// close BAM reader
bamReader.Close();
// close Fasta
genomeFasta.Close();
return 1;
}
bool check(const PropertyFilter& filter, const BamAlignment& al) {
bool keepAlignment = true;
const PropertyMap& properties = filter.Properties;
PropertyMap::const_iterator propertyIter = properties.begin();
PropertyMap::const_iterator propertyEnd = properties.end();
for ( ; propertyIter != propertyEnd; ++propertyIter ) {
// check alignment data field depending on propertyName
const string& propertyName = (*propertyIter).first;
const PropertyFilterValue& valueFilter = (*propertyIter).second;
if ( propertyName == ALIGNMENTFLAG_PROPERTY ) keepAlignment &= valueFilter.check(al.AlignmentFlag);
else if ( propertyName == CIGAR_PROPERTY ) {
stringstream cigarSs;
const vector<CigarOp>& cigarData = al.CigarData;
if ( !cigarData.empty() ) {
vector<CigarOp>::const_iterator cigarBegin = cigarData.begin();
vector<CigarOp>::const_iterator cigarIter = cigarBegin;
vector<CigarOp>::const_iterator cigarEnd = cigarData.end();
for ( ; cigarIter != cigarEnd; ++cigarIter ) {
const CigarOp& op = (*cigarIter);
cigarSs << op.Length << op.Type;
}
keepAlignment &= valueFilter.check(cigarSs.str());
}
}
else if ( propertyName == INSERTSIZE_PROPERTY ) keepAlignment &= valueFilter.check(al.InsertSize);
else if ( propertyName == ISDUPLICATE_PROPERTY ) keepAlignment &= valueFilter.check(al.IsDuplicate());
else if ( propertyName == ISFAILEDQC_PROPERTY ) keepAlignment &= valueFilter.check(al.IsFailedQC());
else if ( propertyName == ISFIRSTMATE_PROPERTY ) keepAlignment &= valueFilter.check(al.IsFirstMate());
else if ( propertyName == ISMAPPED_PROPERTY ) keepAlignment &= valueFilter.check(al.IsMapped());
else if ( propertyName == ISMATEMAPPED_PROPERTY ) keepAlignment &= valueFilter.check(al.IsMateMapped());
else if ( propertyName == ISMATEREVERSESTRAND_PROPERTY ) keepAlignment &= valueFilter.check(al.IsMateReverseStrand());
else if ( propertyName == ISPAIRED_PROPERTY ) keepAlignment &= valueFilter.check(al.IsPaired());
else if ( propertyName == ISPRIMARYALIGNMENT_PROPERTY ) keepAlignment &= valueFilter.check(al.IsPrimaryAlignment());
else if ( propertyName == ISPROPERPAIR_PROPERTY ) keepAlignment &= valueFilter.check(al.IsProperPair());
else if ( propertyName == ISREVERSESTRAND_PROPERTY ) keepAlignment &= valueFilter.check(al.IsReverseStrand());
else if ( propertyName == ISSECONDMATE_PROPERTY ) keepAlignment &= valueFilter.check(al.IsSecondMate());
else if ( propertyName == ISSINGLETON_PROPERTY ) {
const bool isSingleton = al.IsPaired() && al.IsMapped() && !al.IsMateMapped();
keepAlignment &= valueFilter.check(isSingleton);
}
else if ( propertyName == MAPQUALITY_PROPERTY ) keepAlignment &= valueFilter.check(al.MapQuality);
else if ( propertyName == MATEPOSITION_PROPERTY ) keepAlignment &= ( al.IsPaired() && al.IsMateMapped() && valueFilter.check(al.MateRefID) );
else if ( propertyName == MATEREFERENCE_PROPERTY ) {
if ( !al.IsPaired() || !al.IsMateMapped() ) return false;
BAMTOOLS_ASSERT_MESSAGE( (al.MateRefID>=0 && (al.MateRefID<(int)filterToolReferences.size())), "Invalid MateRefID");
const string& refName = filterToolReferences.at(al.MateRefID).RefName;
keepAlignment &= valueFilter.check(refName);
}
else if ( propertyName == NAME_PROPERTY ) keepAlignment &= valueFilter.check(al.Name);
else if ( propertyName == POSITION_PROPERTY ) keepAlignment &= valueFilter.check(al.Position);
else if ( propertyName == QUERYBASES_PROPERTY ) keepAlignment &= valueFilter.check(al.QueryBases);
else if ( propertyName == REFERENCE_PROPERTY ) {
BAMTOOLS_ASSERT_MESSAGE( (al.RefID>=0 && (al.RefID<(int)filterToolReferences.size())), "Invalid RefID");
const string& refName = filterToolReferences.at(al.RefID).RefName;
keepAlignment &= valueFilter.check(refName);
}
else if ( propertyName == TAG_PROPERTY ) keepAlignment &= checkAlignmentTag(valueFilter, al);
else BAMTOOLS_ASSERT_UNREACHABLE;
// if alignment fails at ANY point, just quit and return false
if ( !keepAlignment ) return false;
}
BAMTOOLS_ASSERT_MESSAGE( keepAlignment, "Error in BamAlignmentChecker... keepAlignment should be true here");
return keepAlignment;
}
virtual void main()
{
//init
QTextStream out(stdout);
BamReader reader;
NGSHelper::openBAM(reader, getInfile("in"));
FastqOutfileStream out1(getOutfile("out1"), false);
FastqOutfileStream out2(getOutfile("out2"), false);
long long c_unpaired = 0;
long long c_paired = 0;
int max_cached = 0;
//iterate through reads
BamAlignment al;
QHash<QByteArray, BamAlignment> al_cache;
while (reader.GetNextAlignment(al))
{
//skip secondary alinments
if(!al.IsPrimaryAlignment()) continue;
//skip unpaired
if(!al.IsPaired())
{
++c_unpaired;
continue;
}
QByteArray name(al.Name.data()); //TODO use QByteArray::fromStdString (when upgraded to Qt5.4)
//store cached read when we encounter the mate
if (al_cache.contains(name))
{
BamAlignment mate = al_cache.take(name);
//out << name << " [AL] First: " << al.IsFirstMate() << " Reverse: " << al.IsReverseStrand() << " Seq: " << al.QueryBases.data() << endl;
//out << name << " [MA] First: " << mate.IsFirstMate() << " Reverse: " << mate.IsReverseStrand() << " Seq: " << mate.QueryBases.data() << endl;
if (al.IsFirstMate())
{
write(out1, al, al.IsReverseStrand());
write(out2, mate, mate.IsReverseStrand());
}
else
{
write(out1, mate, mate.IsReverseStrand());
write(out2, al, al.IsReverseStrand());
}
++c_paired;
}
//cache read for later retrieval
else
{
al_cache.insert(name, al);
}
max_cached = std::max(max_cached, al_cache.size());
}
reader.Close();
out1.close();
out2.close();
//write debug output
out << "Pair reads (written) : " << c_paired << endl;
out << "Unpaired reads (skipped) : " << c_unpaired << endl;
out << "Unmatched paired reads (skipped): " << al_cache.size() << endl;
out << endl;
out << "Maximum cached reads : " << max_cached << endl;
}
