position BamAlignmentReader::GetLastPositionInBam(const std::string& bamPath, Region::SharedPtr regionPtr)
{
BamTools::BamReader bamReader;
if (!bamReader.Open(bamPath))
{
throw "Unable to open bam file";
}
bamReader.LocateIndex();
int refID = bamReader.GetReferenceID(regionPtr->getReferenceID());
auto referenceData = bamReader.GetReferenceData();
bamReader.Close();
return referenceData[refID].RefLength;
}
void Config::InitializationClustering() {
struct stat st;
if(stat(Workspace.c_str(),&st) == 0 and st.st_mode and S_IFDIR != 0) Log("[Warning] Workspace directory already present");
else if (mkdir(Workspace.c_str(), 0755) != 0) {
Log("[Error] Could not create workspace directory: " + Workspace);
exit(1);
}
RunningTasksFile = Workspace + "/" + FilePrefix + "running.tasks";
StatsFile = Workspace + "/" + FilePrefix + "stats";
BinClusterFile = Workspace + "/" + FilePrefix + "bpc";
clusterFile = new ClusterFile(BinClusterFile);
clusterDir = Workspace + "/clusters/";
if(stat(clusterDir.c_str(),&st) == 0 and st.st_mode and S_IFDIR != 0) Log("[Warning] Cluster directory already present");
else if (mkdir(clusterDir.c_str(), 0755) != 0) {
Log("[Error] Could not create cluster directory: " + clusterDir);
exit(1);
}
insertsizeDir = Workspace + "/insertsize/";
if(stat(insertsizeDir.c_str(),&st) == 0 and st.st_mode and S_IFDIR != 0) Log("[Warning] Insertsize directory already present");
else if (mkdir(insertsizeDir.c_str(), 0755) != 0) {
Log("[Error] Could not create insertsize directory: " + insertsizeDir);
exit(1);
}
coverageDir = Workspace + "/coverage/";
if(stat(coverageDir.c_str(),&st) == 0 and st.st_mode and S_IFDIR != 0) Log("[Warning] Coverage directory already present");
else if (mkdir(coverageDir.c_str(), 0755) != 0) {
Log("[Error] Could not create coverage directory: " + coverageDir);
exit(1);
}
if (!ForwardBam.empty() && !ReverseBam.empty() && PairedBam.empty()) {
UsePairedBam = false;
} else if (ForwardBam.empty() && ReverseBam.empty() && !PairedBam.empty()) {
UsePairedBam = true;
} else {
Log("[Error] No correct bam file(s)");
exit(1);
}
BamTools::BamAlignment alignment;
BamTools::BamReader BamReader;
if (UsePairedBam) {
BamReader.Open(PairedBam);
if (not BamReader.IsOpen()) {
Log("[Error] Could not open paired bam");
exit(1);
}
if (PairedIndex.empty()) {
if (not BamReader.LocateIndex(BamTools::BamIndex::STANDARD)) {
PairedIndex = PairedBam.substr(0,PairedBam.find_last_of(".bam")-3) + ".bai";
BamReader.OpenIndex(PairedIndex);
}
if (not BamReader.HasIndex()) {
Log("[Error] No index for bamfile");
exit(1);
}
}
BamTools::SamHeader header = BamReader.GetHeader();
for (BamTools::SamReadGroupIterator it = header.ReadGroups.Begin(); it != header.ReadGroups.End(); it++) {
BamTools::SamReadGroup* readgroup = &*it;
readNameConverter.TrimName(readgroup->ID);
readNameConverter.AddReadGroup(readgroup->ID);
}
long int count = 0;
while (BamReader.GetNextAlignment(alignment)) {
string RG;
if (alignment.GetTag("RG", RG)) {
if (not NameTrim.empty()) readNameConverter.TrimName(RG);
if (readNameConverter.AddReadGroup(RG)) {
Log("[Warning] Readgroup '" + RG + "' found in reads but not in header");
count = 0;
}
}
count++;
if (count > 10000) break;
}
BamReader.Close();
} else {
BamReader.Open(ForwardBam);
if (not BamReader.IsOpen()) {
Log("[Error] Could not open first/forward bam");
exit(1);
}
if (ForwardIndex.empty()) {
if (not BamReader.LocateIndex(BamTools::BamIndex::STANDARD)) {
ForwardIndex = ForwardBam.substr(0,ForwardBam.find_last_of(".bam")-3) + ".bai";
BamReader.OpenIndex(ForwardIndex);
}
if (not BamReader.HasIndex()) {
Log("[Error] No index for forward bamfile");
exit(1);
}
}
BamTools::SamHeader forwardheader = BamReader.GetHeader();
for (BamTools::SamReadGroupIterator it = forwardheader.ReadGroups.Begin(); it != forwardheader.ReadGroups.End(); it++) {
BamTools::SamReadGroup* readgroup = &*it;
readNameConverter.TrimName(readgroup->ID);
readNameConverter.AddReadGroup(readgroup->ID);
}
long int count = 0;
while (BamReader.GetNextAlignment(alignment)) {
string RG;
if (alignment.GetTag("RG", RG)) {
if (!NameTrim.empty()) readNameConverter.TrimName(RG);
if (readNameConverter.AddReadGroup(RG)) {
Log("[Warning] Readgroup '" + RG + "' found in forward reads but not in header");
count = 0;
}
}
count++;
if (count > 10000) break;
}
BamReader.Close();
BamReader.Open(ReverseBam);
if (not BamReader.IsOpen()) {
Log("[Error] Could not open second/reverse bam");
exit(1);
}
if (ReverseIndex.empty()) {
if (not BamReader.LocateIndex(BamTools::BamIndex::STANDARD)) {
ReverseIndex = ReverseBam.substr(0,ReverseBam.find_last_of(".bam")-3) + ".bai";
BamReader.OpenIndex(ReverseIndex);
}
if (not BamReader.HasIndex()) {
Log("[Error] No index for reverse bamfile");
exit(1);
}
}
BamTools::SamHeader reverseheader = BamReader.GetHeader();
for (BamTools::SamReadGroupIterator it = reverseheader.ReadGroups.Begin(); it != reverseheader.ReadGroups.End(); it++) {
BamTools::SamReadGroup* readgroup = &*it;
readNameConverter.TrimName(readgroup->ID);
if (readNameConverter.AddReadGroup(readgroup->ID)) {
Log("[Warning] Readgroup '" + readgroup->ID + "' found in reverse but not in forward");
}
}
count = 0;
while (BamReader.GetNextAlignment(alignment)) {
string RG;
if (alignment.GetTag("RG", RG)) {
if (!NameTrim.empty()) readNameConverter.TrimName(RG);
if (readNameConverter.AddReadGroup(RG)) {
Log("[Warning] Readgroup '" + RG + "' found in reverse reads but not in header");
count = 0;
}
}
count++;
if (count > 10000) break;
}
BamReader.Close();
}
for(map<string, int>::iterator it = readNameConverter.ReadGroups.begin(); it!=readNameConverter.ReadGroups.end(); ++it) {
ostringstream logBuffer;
logBuffer << "Readgroup found: " << it->second << " - " << it->first;
Log(logBuffer.str());
}
writeConfigFile(Workspace + FilePrefix + "config");
}
//
// Main
//
int somaticVariantFiltersMain(int argc, char** argv)
{
parseSomaticVariantFiltersOptions(argc, argv);
Timer* pTimer = new Timer(PROGRAM_IDENT);
// Load Reference
ReadTable refTable(opt::referenceFile, SRF_NO_VALIDATION);
refTable.indexReadsByID();
// Load BAMs
BamTools::BamReader* pTumorBamReader = new BamTools::BamReader;
pTumorBamReader->Open(opt::tumorBamFile);
pTumorBamReader->LocateIndex();
assert(pTumorBamReader->HasIndex());
BamTools::BamReader* pNormalBamReader = new BamTools::BamReader;
pNormalBamReader->Open(opt::normalBamFile);
pNormalBamReader->LocateIndex();
assert(pNormalBamReader->HasIndex());
// Track duplicated variants
HashSet<std::string> duplicateHash;
std::ifstream input(opt::vcfFile.c_str());
std::string line;
while(getline(input, line))
{
if(line.empty())
continue;
if(line[0] == '#')
{
std::cout << line << "\n";
continue;
}
// parse record
VCFRecord record(line);
if(record.isMultiAllelic())
{
std::cerr << "Error: multi-allelic VCF found, please run vcfbreakmulti\n";
exit(EXIT_FAILURE);
}
// Check if we've seen this variant already
std::string key = makeVariantKey(record);
if(duplicateHash.find(key) != duplicateHash.end())
continue;
else
duplicateHash.insert(key);
if(opt::verbose > 0)
{
std::stringstream ss;
ss << "Variant: " << record << "\n";
fprintf(stderr, "===============================================\n%s", ss.str().c_str());
}
StringStringHash tagHash;
makeTagHash(record, tagHash);
StringVector fail_reasons;
int hplen = 0;
if(!getTagValue(tagHash, "HPLen", hplen))
hplen = calculateHomopolymerLength(record, &refTable);
if(hplen > opt::maxHPLen)
fail_reasons.push_back("Homopolymer");
double dust = 0.0f;
if(!getTagValue(tagHash, "Dust", dust))
dust = HapgenUtil::calculateDustScoreAtPosition(record.refName,
record.refPosition,
&refTable);
if(dust > opt::maxDust)
fail_reasons.push_back("LowComplexity");
double af;
if(getTagValue(tagHash, "AF", af) && af < opt::minAF)
fail_reasons.push_back("LowAlleleFrequency");
int varDP;
if(getTagValue(tagHash, "VarDP", varDP) && varDP < opt::minVarDP)
fail_reasons.push_back("LowVarDP");
double strandBias;
if(getTagValue(tagHash, "SB", strandBias) && strandBias >= opt::maxStrandBias)
fail_reasons.push_back("StrandBias");
CoverageStats tumor_stats = getVariantCoverage(pTumorBamReader, record, &refTable);
CoverageStats normal_stats = getVariantCoverage(pNormalBamReader, record, &refTable);
if(opt::verbose > 0)
{
fprintf(stderr, "Tumor: [%zu %zu]\n", tumor_stats.n_total_reads, tumor_stats.n_evidence_reads);
fprintf(stderr, "Normal: [%zu %zu]\n", normal_stats.n_total_reads, normal_stats.n_evidence_reads);
}
if(normal_stats.n_evidence_reads > opt::maxNormalReads)
fail_reasons.push_back("NormalEvidence");
if(normal_stats.n_total_reads < opt::minNormalDepth)
fail_reasons.push_back("LowNormalDepth");
if(!tumor_stats.snv_evidence_quals.empty())
{
double median_quality = median(tumor_stats.snv_evidence_quals);
if(median_quality < opt::minMedianQuality)
fail_reasons.push_back("LowQuality");
}
if(tumor_stats.median_mapping_quality < opt::minMedianQuality)
fail_reasons.push_back("LowMappingQuality");
if(!fail_reasons.empty())
{
if(record.passStr != "PASS" && record.passStr != ".")
fail_reasons.insert(fail_reasons.begin(), record.passStr);
std::stringstream strss;
std::copy(fail_reasons.begin(), fail_reasons.end(), std::ostream_iterator<std::string>(strss, ";"));
record.passStr = strss.str();
record.passStr.erase(record.passStr.size() - 1); // erase trailing ;
}
std::cout << record << "\n";
}
// Cleanup
delete pTumorBamReader;
delete pNormalBamReader;
delete pTimer;
return 0;
}
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;
}