// ValidateReaders checks that all the readers point to BAM files representing
// alignments against the same set of reference sequences, and that the
// sequences are identically ordered. If these checks fail the operation of
// the multireader is undefined, so we force program exit.
void BamMultiReader::ValidateReaders(void) const {
int firstRefCount = readers.front().first->GetReferenceCount();
BamTools::RefVector firstRefData = readers.front().first->GetReferenceData();
for (vector<pair<BamReader*, BamAlignment*> >::const_iterator it = readers.begin(); it != readers.end(); ++it) {
BamReader* reader = it->first;
BamTools::RefVector currentRefData = reader->GetReferenceData();
BamTools::RefVector::const_iterator f = firstRefData.begin();
BamTools::RefVector::const_iterator c = currentRefData.begin();
if (reader->GetReferenceCount() != firstRefCount || firstRefData.size() != currentRefData.size()) {
cerr << "ERROR: mismatched number of references in " << reader->GetFilename()
<< " expected " << firstRefCount
<< " reference sequences but only found " << reader->GetReferenceCount() << endl;
exit(1);
}
// this will be ok; we just checked above that we have identically-sized sets of references
// here we simply check if they are all, in fact, equal in content
while (f != firstRefData.end()) {
if (f->RefName != c->RefName || f->RefLength != c->RefLength) {
cerr << "ERROR: mismatched references found in " << reader->GetFilename()
<< " expected: " << endl;
for (BamTools::RefVector::const_iterator a = firstRefData.begin(); a != firstRefData.end(); ++a)
cerr << a->RefName << " " << a->RefLength << endl;
cerr << "but found: " << endl;
for (BamTools::RefVector::const_iterator a = currentRefData.begin(); a != currentRefData.end(); ++a)
cerr << a->RefName << " " << a->RefLength << endl;
exit(1);
}
++f; ++c;
}
}
}
NewGenomeFile::NewGenomeFile(const BamTools::RefVector &refVector)
: _maxId(-1)
{
for (size_t i = 0; i < refVector.size(); ++i) {
QuickString chrom = refVector[i].RefName;
CHRPOS length = refVector[i].RefLength;
_maxId++;
_chromSizeIds[chrom] = pair<CHRPOS, CHRPOS>(length, _maxId);
}
}
NewGenomeFile::NewGenomeFile(const BamTools::RefVector &refVector)
: _maxId(-1)
{
size_t i = 0;
for (; i < refVector.size(); ++i) {
string chrom = refVector[i].RefName;
CHRPOS length = refVector[i].RefLength;
_maxId++;
_chromSizeIds[chrom] = pair<CHRPOS, CHRPOS>(length, _maxId);
_chromList.push_back(chrom);
}
// Special: BAM files can have unmapped reads, which show as no chromosome, or an empty chrom string.
// Add in an empty chrom so these don't error.
_maxId++;
_chromSizeIds[""] = pair<CHRPOS, int>(0, _maxId);
_chromList.push_back("");
}
RegionCoverage::RegionCoverage( const BamTools::RefVector& references )
: m_contigList(NULL)
, m_bcovRegion(NULL)
, m_rcovRegion(NULL)
, m_lastRegionAssigned(NULL)
, m_numAuxFields(0)
, m_ncovDepths(0)
{
m_numRefContigs = references.size();
if( m_numRefContigs ) {
m_contigList = new TargetContig*[m_numRefContigs];
for( size_t i = 0; i < m_numRefContigs; ++i ) {
m_contigList[i] = new TargetContig( references[i].RefName, references[i].RefLength );
m_contigIdx[ references[i].RefName ] = i;
}
}
// initial values force set up on first call to 'iterator'
m_bcovContigIdx = m_rcovContigIdx = m_numRefContigs;
m_bcovRegionPos = 0;
}
int bamToLociMeth(std::string bamFile1,
std::string bamFile2,
const char * outFile,
std::string sample,
int d, int freq, int methdiff)
{
std::map <std::string, int> allpatterns;
std::map <std::string, int> patternMeth;
getAllPatterns(allpatterns, patternMeth);
BamTools::BamReader reader1;
BamTools::BamReader reader2;
bamCheck(bamFile1, reader1);
bamCheck(bamFile2, reader2);
// count reads number of bam files for normalization.
/*int n1=0;
int n2=0;
BamTools::BamAlignment al1;
BamTools::BamAlignment al2;
while(reader1.GetNextAlignment(al1)){
n1=n1+1;
}
while(reader2.GetNextAlignment(al2)){
n2=n2+1;
}
std::cout << bamFile1 << " read count: " << n1 << ".\n" << bamFile2 << " read count: " << n2 << "." << std::endl;
double fnorm1k2=double(n1)/double(n2);
*/
// get reference name from first bam
const BamTools::RefVector refs = reader1.GetReferenceData();
reader1.LocateIndex();
reader2.LocateIndex();
if (reader1.HasIndex() & reader2.HasIndex())
{
std::cerr << "Output file: " << outFile << std::endl;
std::cerr << "===============================" << std::endl;
ogzstream myfile;
myfile.open (outFile);
header(myfile, allpatterns);
for(BamTools::RefVector::const_iterator i = refs.begin(); i != refs.end(); ++i)
{
std::map<std::string, std::vector<std::string> > lociMeth1;
std::map<std::string, std::vector<std::string> > lociMeth2;
// iterate by chromosome for bamFile1 and bamFile2
readerToMeth(reader1, reader2, lociMeth1, lociMeth2, i, d, refs, sample);
//entropy analysis
interLoci(lociMeth1, lociMeth2, allpatterns, patternMeth, freq, methdiff, myfile);
}
reader1.Close();
reader2.Close();
myfile.close();
finished();
return 0;
} else {
std::cerr << "Could not load index data for all input BAM file(s)... Aborting." << std::endl;
return false;
}
}
inline int
run(Config const& c, TSingleHit)
{
// Create library objects
typedef std::map<std::string, LibraryInfo> TLibraryMap;
typedef std::map<std::string, TLibraryMap> TSampleLibrary;
TSampleLibrary sampleLib;
// Scan libraries
for(unsigned int file_c = 0; file_c < c.files.size(); ++file_c) {
// Get a sample name
std::string sampleName(c.files[file_c].stem().string());
// Check that all input bam files exist
BamTools::BamReader reader;
if ( ! reader.Open(c.files[file_c].string()) ) {
std::cerr << "Could not open input bam file: " << c.files[file_c].string() << std::endl;
reader.Close();
return -1;
}
// Check that all input bam files are indexed
reader.LocateIndex();
if ( !reader.HasIndex() ) {
std::cerr << "Missing bam index file: " << c.files[file_c].string() << std::endl;
reader.Close();
return -1;
}
// Get library parameters and overall maximum insert size
TLibraryMap libInfo;
getLibraryParams(c.files[file_c], libInfo, 0, 5);
sampleLib.insert(std::make_pair(sampleName, libInfo));
}
// Read all SV intervals
typedef std::vector<StructuralVariantRecord> TSVs;
TSVs svs;
std::map<unsigned int, std::string> idToName;
unsigned int intervalCount=1;
if (boost::filesystem::exists(c.int_file) && boost::filesystem::is_regular_file(c.int_file) && boost::filesystem::file_size(c.int_file)) {
Memory_mapped_file interval_file(c.int_file.string().c_str());
char interval_buffer[Memory_mapped_file::MAX_LINE_LENGTH];
while (interval_file.left_bytes() > 0) {
interval_file.read_line(interval_buffer);
// Read single interval line
StructuralVariantRecord sv;
Tokenizer token(interval_buffer, Memory_mapped_file::MAX_LINE_LENGTH);
std::string interval_rname;
token.getString(sv.chr);
sv.svStart = token.getUInt();
sv.svEnd = token.getUInt() + 1;
std::string svName;
token.getString(svName);
idToName.insert(std::make_pair(intervalCount, svName));
sv.id = intervalCount++;
svs.push_back(sv);
}
interval_file.close();
} else {
// Create artificial intervals
BamTools::BamReader readerRef;
if ( ! readerRef.Open(c.files[0].string()) ) return -1;
BamTools::RefVector references = readerRef.GetReferenceData();
typename BamTools::RefVector::const_iterator itRef = references.begin();
for(int refIndex=0;itRef!=references.end();++itRef, ++refIndex) {
int32_t pos = 0;
while (pos < references[refIndex].RefLength) {
int32_t window_len = pos+c.window_size;
if (window_len > references[refIndex].RefLength) window_len = references[refIndex].RefLength;
StructuralVariantRecord sv;
sv.chr = references[refIndex].RefName;
sv.svStart = pos;
sv.svEnd = window_len;
std::stringstream s;
s << sv.chr << ":" << sv.svStart << "-" << sv.svEnd;
idToName.insert(std::make_pair(intervalCount, s.str()));
sv.id = intervalCount++;
svs.push_back(sv);
pos += c.window_offset;
}
}
}
// Output data types
typedef std::pair<std::string, int> TSampleSVPair;
typedef std::pair<int, int> TBpRead;
typedef std::map<TSampleSVPair, TBpRead> TCountMap;
TCountMap countMap;
// Annotate coverage
annotateCoverage(c.files, c.minMapQual, c.inclCigar, sampleLib, svs, countMap, TSingleHit());
// Output library statistics
std::cout << "Library statistics" << std::endl;
TSampleLibrary::const_iterator sampleIt=sampleLib.begin();
for(;sampleIt!=sampleLib.end();++sampleIt) {
std::cout << "Sample: " << sampleIt->first << std::endl;
TLibraryMap::const_iterator libIt=sampleIt->second.begin();
for(;libIt!=sampleIt->second.end();++libIt) {
std::cout << "RG: ID=" << libIt->first << ",Median=" << libIt->second.median << ",MAD=" << libIt->second.mad << ",Orientation=" << (int) libIt->second.defaultOrient << ",MappedReads=" << libIt->second.mappedReads << ",DuplicatePairs=" << libIt->second.non_unique_pairs << ",UniquePairs=" << libIt->second.unique_pairs << std::endl;
}
}
// Output file
boost::iostreams::filtering_ostream dataOut;
dataOut.push(boost::iostreams::gzip_compressor());
dataOut.push(boost::iostreams::file_sink(c.outfile.string().c_str(), std::ios_base::out | std::ios_base::binary));
// Iterate all SVs
typename TSVs::const_iterator itSV = svs.begin();
typename TSVs::const_iterator itSVEnd = svs.end();
for(;itSV!=itSVEnd;++itSV) {
dataOut << itSV->chr << "\t" << itSV->svStart << "\t" << itSV->svEnd << "\t" << idToName.find(itSV->id)->second;
// Iterate all samples
for(unsigned int file_c = 0; file_c < c.files.size(); ++file_c) {
// Get the sample name
std::string sampleName(c.files[file_c].stem().string());
TSampleSVPair sampleSVPair = std::make_pair(sampleName, itSV->id);
typename TCountMap::iterator countMapIt=countMap.find(sampleSVPair);
dataOut << "\t";
if (c.avg_flag) dataOut << ( (countMapIt->second.first) / (double) (itSV->svEnd - itSV->svStart)) << "\t";
if (c.bp_flag) dataOut << countMapIt->second.first << "\t";
dataOut << countMapIt->second.second;
}
dataOut << std::endl;
}
// End
boost::posix_time::ptime now = boost::posix_time::second_clock::local_time();
std::cout << '[' << boost::posix_time::to_simple_string(now) << "] Done." << std::endl;;
return 0;
}
inline int
run(Config const& c, TCoverageType covType)
{
// Create library objects
typedef boost::unordered_map<std::string, LibraryInfo> TLibraryMap;
typedef boost::unordered_map<std::string, TLibraryMap> TSampleLibrary;
TSampleLibrary sampleLib;
// Scan libraries
for(unsigned int file_c = 0; file_c < c.files.size(); ++file_c) {
// Get a sample name
std::string sampleName(c.files[file_c].stem().string());
// Check that all input bam files exist
BamTools::BamReader reader;
if ( ! reader.Open(c.files[file_c].string()) ) {
std::cerr << "Could not open input bam file: " << c.files[file_c].string() << std::endl;
reader.Close();
return -1;
}
// Check that all input bam files are indexed
reader.LocateIndex();
if ( !reader.HasIndex() ) {
std::cerr << "Missing bam index file: " << c.files[file_c].string() << std::endl;
reader.Close();
return -1;
}
// Get library parameters and overall maximum insert size
TLibraryMap libInfo;
getLibraryParams(c.files[file_c], libInfo, 0, 5);
sampleLib.insert(std::make_pair(sampleName, libInfo));
}
// Get references
BamTools::BamReader readerRef;
if ( ! readerRef.Open(c.files[0].string()) ) return -1;
BamTools::RefVector references = readerRef.GetReferenceData();
// Read all SV intervals
typedef std::vector<CovRecord> TSVs;
TSVs svs;
std::map<unsigned int, std::string> idToName;
unsigned int intervalCount=1;
if (boost::filesystem::exists(c.int_file) && boost::filesystem::is_regular_file(c.int_file) && boost::filesystem::file_size(c.int_file)) {
typedef boost::unordered_map<std::string, unsigned int> TMapChr;
TMapChr mapChr;
typename BamTools::RefVector::const_iterator itRef = references.begin();
for(unsigned int i = 0;itRef!=references.end();++itRef, ++i) mapChr[ itRef->RefName ] = i;
std::ifstream interval_file(c.int_file.string().c_str(), std::ifstream::in);
if (interval_file.is_open()) {
while (interval_file.good()) {
std::string intervalLine;
getline(interval_file, intervalLine);
typedef boost::tokenizer< boost::char_separator<char> > Tokenizer;
boost::char_separator<char> sep(" \t,;");
Tokenizer tokens(intervalLine, sep);
Tokenizer::iterator tokIter = tokens.begin();
if (tokIter!=tokens.end()) {
std::string chrName=*tokIter++;
TMapChr::const_iterator mapChrIt = mapChr.find(chrName);
if (mapChrIt != mapChr.end()) {
if (tokIter!=tokens.end()) {
CovRecord sv;
sv.chr = mapChrIt->second;
sv.svStart = boost::lexical_cast<int32_t>(*tokIter++);
sv.svEnd = boost::lexical_cast<int32_t>(*tokIter++) + 1;
std::string svName = *tokIter;
idToName.insert(std::make_pair(intervalCount, svName));
sv.id = intervalCount++;
svs.push_back(sv);
}
}
}
}
interval_file.close();
}
} else {
// Create artificial intervals
typename BamTools::RefVector::const_iterator itRef = references.begin();
for(int refIndex=0;itRef!=references.end();++itRef, ++refIndex) {
int32_t pos = 0;
unsigned int wSize = c.window_size;
unsigned int wOffset = c.window_offset;
if (c.window_num>0) {
wSize=(itRef->RefLength / c.window_num) + 1;
wOffset=wSize;
}
while (pos < references[refIndex].RefLength) {
int32_t window_len = pos+wSize;
if (window_len > references[refIndex].RefLength) window_len = references[refIndex].RefLength;
CovRecord sv;
sv.chr = refIndex;
sv.svStart = pos;
sv.svEnd = window_len;
std::stringstream s;
s << references[sv.chr].RefName << ":" << sv.svStart << "-" << sv.svEnd;
idToName.insert(std::make_pair(intervalCount, s.str()));
sv.id = intervalCount++;
svs.push_back(sv);
pos += wOffset;
}
}
}
// Output data types
typedef std::pair<std::string, int> TSampleSVPair;
typedef std::pair<int, int> TBpRead;
typedef std::map<TSampleSVPair, TBpRead> TCountMap;
TCountMap countMap;
// Annotate coverage
if (c.inclCigar) annotateCoverage(c.files, c.minGenoQual, sampleLib, svs, countMap, BpLevelType<BpLevelCount>(), covType);
else annotateCoverage(c.files, c.minGenoQual, sampleLib, svs, countMap, BpLevelType<NoBpLevelCount>(), covType);
// Output library statistics
std::cout << "Library statistics" << std::endl;
TSampleLibrary::const_iterator sampleIt=sampleLib.begin();
for(;sampleIt!=sampleLib.end();++sampleIt) {
std::cout << "Sample: " << sampleIt->first << std::endl;
TLibraryMap::const_iterator libIt=sampleIt->second.begin();
for(;libIt!=sampleIt->second.end();++libIt) {
std::cout << "RG: ID=" << libIt->first << ",Median=" << libIt->second.median << ",MAD=" << libIt->second.mad << ",Orientation=" << (int) libIt->second.defaultOrient << std::endl;
}
}
// Output file
boost::iostreams::filtering_ostream dataOut;
dataOut.push(boost::iostreams::gzip_compressor());
dataOut.push(boost::iostreams::file_sink(c.outfile.string().c_str(), std::ios_base::out | std::ios_base::binary));
// Print header
dataOut << "#chr\tstart\tend\tid";
for(unsigned int file_c = 0; file_c < c.files.size(); ++file_c) {
std::string sampleName(c.files[file_c].stem().string());
dataOut << "\t";
if (c.avg_flag) dataOut << sampleName << "_avgcov" << "\t";
if (c.bp_flag) dataOut << sampleName << "_bpcount" << "\t";
if ((c.bp_flag) || (c.avg_flag)) dataOut << sampleName << "_readcount";
else dataOut << sampleName;
}
dataOut << std::endl;
// Iterate all SVs
typename TSVs::const_iterator itSV = svs.begin();
typename TSVs::const_iterator itSVEnd = svs.end();
for(;itSV!=itSVEnd;++itSV) {
dataOut << references[itSV->chr].RefName << "\t" << itSV->svStart << "\t" << itSV->svEnd << "\t" << idToName.find(itSV->id)->second;
// Iterate all samples
for(unsigned int file_c = 0; file_c < c.files.size(); ++file_c) {
// Get the sample name
std::string sampleName(c.files[file_c].stem().string());
TSampleSVPair sampleSVPair = std::make_pair(sampleName, itSV->id);
typename TCountMap::iterator countMapIt=countMap.find(sampleSVPair);
dataOut << "\t";
if (c.avg_flag) dataOut << ( (countMapIt->second.first) / (double) (itSV->svEnd - itSV->svStart)) << "\t";
if (c.bp_flag) dataOut << countMapIt->second.first << "\t";
dataOut << countMapIt->second.second;
}
dataOut << std::endl;
}
// End
boost::posix_time::ptime now = boost::posix_time::second_clock::local_time();
std::cout << '[' << boost::posix_time::to_simple_string(now) << "] Done." << std::endl;;
return 0;
}