CArchiveMerge::CArchiveMerge (
const vector <string>& inputFilenames,
const string& outputFilename,
uint64_t *readNo,
const bool& isSolid,
const string& commandLine,
const string& paired_end_ann_file,
const string& single_end_ann_file,
const unsigned int& fragmentLength,
const unsigned int& localAlignmentSearchRadius,
const bool& hasSpecial,
const unsigned char& statMappingQuality)
: _inputFilenames(inputFilenames)
, _outputFilename(outputFilename)
, _readNo (readNo)
, _isSolid (isSolid)
, _expectedFragmentLength (fragmentLength)
, _localAlignmentSearchRadius(localAlignmentSearchRadius)
, _refIndex()
, _referenceSequences()
, _referenceSequencesWoSpecial()
, _readGroups()
, _alignmentStatus(0)
, _isPairedEnd(false)
, _hasSpecial (hasSpecial)
, _statMappingQuality (statMappingQuality)
, _sequencingTechnologies()
, _readGroupsMap()
, _specialArchiveName()
, _specialCode1()
, _specialCode2()
, _specialReader()
, _specialAl()
, _special_owner(0)
, _specialArchiveEmpty(true)
, _specialReferenceSequences()
, _counters()
, _statisticsMaps()
, _sHeader()
, _rHeader()
, _sBam()
, _rBam()
, za1()
, za2()
, _entropy()
, _mqCalculator()
{
//_statisticsMaps = mStatisticsMaps;
MosaikReadFormat::CAlignmentReader reader;
reader.Open( _inputFilenames[0] );
reader.GetReadGroups(_readGroups);
_alignmentStatus = reader.GetStatus();
reader.Close();
_isPairedEnd = ( ( _alignmentStatus & AS_PAIRED_END_READ ) != 0 ) ? true : false;
_refIndex.resize(_inputFilenames.size(), 0);
for ( vector<MosaikReadFormat::ReadGroup>::iterator ite = _readGroups.begin(); ite != _readGroups.end(); ++ite ) {
ite->ReadGroupCode = ite->GetCode( *ite );
_readGroupsMap[ ite->ReadGroupCode ] = *ite ;
}
_sequencingTechnologies = _readGroups[0].SequencingTechnology;
// the last archive is the one containing alignments located at special references
for ( unsigned int i = 0; i < _inputFilenames.size(); i++ ) {
reader.Open( _inputFilenames[i] );
// grab reference info from the archive
vector<ReferenceSequence> referenceSequences;
reader.GetReferenceSequences(referenceSequences);
CopyReferenceString( referenceSequences );
_referenceSequences.insert( _referenceSequences.end(), referenceSequences.begin(), referenceSequences.end() );
_refIndex[i] = ( i == 0 ) ? referenceSequences.size() : referenceSequences.size() + _refIndex[i-1];
// don't include the special references
if ((_hasSpecial) && (i != ( _inputFilenames.size() - 1)))
_referenceSequencesWoSpecial.insert( _referenceSequencesWoSpecial.end(), referenceSequences.begin(), referenceSequences.end() );
// includes the special references
if ((_hasSpecial) && (i == ( _inputFilenames.size() - 1)))
_specialReferenceSequences.insert( _specialReferenceSequences.end(), referenceSequences.begin(), referenceSequences.end() );
referenceSequences.clear();
reader.Close();
}
if (!_hasSpecial)
_referenceSequencesWoSpecial = _referenceSequences;
/*
if (_hasSpecial) {
for (unsigned int i = 0; i < _specialReferenceSequences.size(); ++i) {
printf("%s\n",_specialReferenceSequences[i].Species.c_str());
}
}
*/
_sHeader.SortOrder = SORTORDER_UNSORTED;
//_uHeader.SortOrder = SORTORDER_UNSORTED;
_rHeader.SortOrder = SORTORDER_UNSORTED;
_sHeader.pReferenceSequences = &_referenceSequences;
//_uHeader.pReferenceSequences = &_referenceSequencesWoSpecial;
_rHeader.pReferenceSequences = &_referenceSequencesWoSpecial;
_sHeader.pReadGroups = &_readGroups;
//_uHeader.pReadGroups = &_readGroups;
_rHeader.pReadGroups = &_readGroups;
ProgramGroup pg;
pg.ID = "MosaikAligner";
stringstream ss;
ss << (int)MOSAIK_MAJOR_VERSION << "." << (int)MOSAIK_MINOR_VERSION << "." << (int)MOSAIK_BUILD_VERSION;
pg.VN = ss.str();
pg.CL = commandLine;
_sHeader.pg.ID = "MosaikAligner";
_rHeader.pg.ID = "MosaikAligner";
_sHeader.pg.VN = ss.str();
_rHeader.pg.VN = ss.str();
_sHeader.pg.CL = commandLine;
_rHeader.pg.CL = commandLine;
_mqCalculator.Open(paired_end_ann_file, single_end_ann_file);
}
void CMosaikAligner::MergeArchives(void) {
// set active threads
unsigned int nThread = ( mSettings.NumThreads < outputFilenames.size() ) ? mSettings.NumThreads : outputFilenames.size();
vector< string > temporaryFiles;
temporaryFiles.resize( outputFilenames.size() );
for ( unsigned int i = 0; i < outputFilenames.size(); i++ ) {
string tempFilename;
CFileUtilities::GetTempFilename(tempFilename);
temporaryFiles[i] = tempFilename;
}
// calculate total # of reads
unsigned int nReads = 0;
for ( unsigned int i = 0 ; i < outputFilenames.size(); i++ ) {
MosaikReadFormat::CAlignmentReader reader;
reader.Open( outputFilenames[i] );
nReads += reader.GetNumReads();
reader.Close();
}
// if nThread is too large, it'll open too many files at the same time.
// Then, we'll get an error since system doesn't allow us to open any file.
if ( nThread > 7 )
nThread = 7;
// prepare reference offset vector for SOLiD
//vector<unsigned int> refOffsets;
//refOffsets.resize(referenceGroups.size());
//for ( unsigned int i = 0; i < referenceGroups.size(); i++ ) {
// unsigned int startRef = referenceGroups[i].first;
// refOffsets[i] = referenceSequences[startRef].Begin;
//}
CConsole::Heading();
cout << endl << "Sorting alignment archive:" << endl;
CConsole::Reset();
SortThread sThread ( outputFilenames, temporaryFiles, nThread, nReads, mSettings.MedianFragmentLength );
sThread.Start();
CConsole::Heading();
cout << "Merging alignment archive:" << endl;
CConsole::Reset();
unsigned int readNo = 0;
unsigned int nMaxAlignment = 1000;
CProgressBar<unsigned int>::StartThread(&readNo, 0, nReads, "reads");
CArchiveMerge merger( temporaryFiles, mSettings.OutputReadArchiveFilename, nMaxAlignment, &readNo );
merger.Merge();
CProgressBar<unsigned int>::WaitThread();
for ( unsigned int i = 0; i < outputFilenames.size(); i++ )
rm(outputFilenames[i].c_str());
for ( unsigned int i = 0; i < temporaryFiles.size(); i++ )
rm(temporaryFiles[i].c_str());
// get statistics information
CArchiveMerge::StatisticsCounters mergeCounters;
merger.GetStatisticsCounters( mergeCounters );
mStatisticsCounters.AlignedReads = mergeCounters.AlignedReads;
mStatisticsCounters.BothNonUniqueReads = mergeCounters.BothNonUniqueReads;
mStatisticsCounters.BothUniqueReads = mergeCounters.BothUniqueReads;
mStatisticsCounters.OneNonUniqueReads = mergeCounters.OneNonUniqueReads;
mStatisticsCounters.OrphanedReads = mergeCounters.OrphanedReads;
mStatisticsCounters.FilteredOutMates = mergeCounters.FilteredOutMates;
mStatisticsCounters.NonUniqueMates = mergeCounters.NonUniqueMates;
mStatisticsCounters.UniqueMates = mergeCounters.UniqueMates;
}