bool GetNextAlignment(BamAlignment& al, BamMultiReader& reader, int32_t refID)
{
bool good = reader.GetNextAlignmentCore(al);
if(not good or al.RefID != refID)
return false;
else if(not al.IsMapped() or al.IsDuplicate() or al.IsFailedQC())
return GetNextAlignment(al, reader, refID);
else
return true;
}
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;
}
// use current input alignment to update BAM file alignment stats
void StatsTool::StatsToolPrivate::ProcessAlignment(const BamAlignment& al) {
// increment total alignment counter
++numReads;
// check the paired-independent flags
if ( al.IsDuplicate() ) ++numDuplicates;
if ( al.IsFailedQC() ) ++numFailedQC;
if ( al.IsMapped() ) ++numMapped;
// check forward/reverse strand
if ( al.IsReverseStrand() )
++numReverseStrand;
else
++numForwardStrand;
// if alignment is paired-end
if ( al.IsPaired() ) {
// increment PE counter
++numPaired;
// increment first mate/second mate counters
if ( al.IsFirstMate() ) ++numFirstMate;
if ( al.IsSecondMate() ) ++numSecondMate;
// if alignment is mapped, check mate status
if ( al.IsMapped() ) {
// if mate mapped
if ( al.IsMateMapped() )
++numBothMatesMapped;
// else singleton
else
++numSingletons;
}
// check for explicit proper pair flag
if ( al.IsProperPair() ) ++numProperPair;
// store insert size for first mate
if ( settings->IsShowingInsertSizeSummary && al.IsFirstMate() && (al.InsertSize != 0) ) {
int insertSize = abs(al.InsertSize);
insertSizes.push_back( insertSize );
}
}
}
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;
}
