void BedIntersectPE::IntersectBamPE(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB->loadBedFileIntoMap();
// open the BAM file
BamReader reader;
BamWriter writer;
reader.Open(bamFile);
// get header & reference information
string bamHeader = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// open a BAM output to stdout if we are writing BAM
if (_bamOutput == true) {
// set compression mode
BamWriter::CompressionMode compressionMode = BamWriter::Compressed;
if ( _isUncompressedBam ) compressionMode = BamWriter::Uncompressed;
writer.SetCompressionMode(compressionMode);
// open our BAM writer
writer.Open("stdout", bamHeader, refs);
}
// track the previous and current sequence
// names so that we can identify blocks of
// alignments for a given read ID.
string prevName, currName;
prevName = currName = "";
vector<BamAlignment> alignments; // vector of BAM alignments for a given ID in a BAM file.
alignments.reserve(100);
_bedA->bedType = 10; // it's a full BEDPE given it's BAM
// rip through the BAM file and convert each mapped entry to BEDPE
BamAlignment bam1, bam2;
while (reader.GetNextAlignment(bam1)) {
// the alignment must be paired
if (bam1.IsPaired() == true) {
// grab the second alignment for the pair.
reader.GetNextAlignment(bam2);
// require that the alignments are from the same query
if (bam1.Name == bam2.Name) {
ProcessBamBlock(bam1, bam2, refs, writer);
}
else {
cerr << "*****ERROR: -bedpe requires BAM to be sorted or grouped by query name. " << endl;
exit(1);
}
}
}
// close up
reader.Close();
if (_bamOutput == true) {
writer.Close();
}
}
bool RevertTool::RevertToolPrivate::Run(void) {
// opens the BAM file without checking for indexes
BamReader reader;
if ( !reader.Open(m_settings->InputFilename) ) {
cerr << "Could not open input BAM file... quitting." << endl;
return false;
}
// get BAM file metadata
const string& headerText = reader.GetHeaderText();
const RefVector& references = reader.GetReferenceData();
// open writer
BamWriter writer;
bool writeUncompressed = ( m_settings->OutputFilename == Options::StandardOut() && !m_settings->IsForceCompression );
if ( !writer.Open(m_settings->OutputFilename, headerText, references, writeUncompressed) ) {
cerr << "Could not open " << m_settings->OutputFilename << " for writing." << endl;
return false;
}
// plow through file, reverting alignments
BamAlignment al;
while ( reader.GetNextAlignment(al) ) {
RevertAlignment(al);
writer.SaveAlignment(al);
}
// clean and exit
reader.Close();
writer.Close();
return true;
}
// 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;
}
}
}
void BedCoverage::CollectCoverageBam(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB->loadBedCovFileIntoMap();
// open the BAM file
BamReader reader;
reader.Open(bamFile);
// get header & reference information
string header = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// convert each aligned BAM entry to BED
// and compute coverage on B
BamAlignment bam;
while (reader.GetNextAlignment(bam)) {
if (bam.IsMapped()) {
// treat the BAM alignment as a single "block"
if (_obeySplits == false) {
// construct a new BED entry from the current BAM alignment.
BED a;
a.chrom = refs.at(bam.RefID).RefName;
a.start = bam.Position;
a.end = bam.GetEndPosition(false, false);
a.strand = "+";
if (bam.IsReverseStrand()) a.strand = "-";
_bedB->countHits(a, _sameStrand, _diffStrand, _countsOnly);
}
// split the BAM alignment into discrete blocks and
// look for overlaps only within each block.
else {
// vec to store the discrete BED "blocks" from a
bedVector bedBlocks;
// since we are counting coverage, we do want to split blocks when a
// deletion (D) CIGAR op is encountered (hence the true for the last parm)
GetBamBlocks(bam, refs.at(bam.RefID).RefName, bedBlocks, false, true);
// use countSplitHits to avoid over-counting each split chunk
// as distinct read coverage.
_bedB->countSplitHits(bedBlocks, _sameStrand, _diffStrand, _countsOnly);
}
}
}
// report the coverage (summary or histogram) for BED B.
if (_countsOnly == true)
ReportCounts();
else
ReportCoverage();
// close the BAM file
reader.Close();
}
void parser::fetchLines(vector<BamAlignment>& result, uint32_t n,
const std::string& file) {
BamReader bam;
BamAlignment read;
Guarded<FileNotGood> g(!(bam.Open(file)), file.c_str());
const RefVector refvec = bam.GetReferenceData();
while (bam.GetNextAlignment(read) && n) {
result.push_back(read);
// cout << "read " << n << "\t" << read << "\n";
n--;
}
}
bool CoverageTool::CoverageToolPrivate::Run(void) {
// if output filename given
ofstream outFile;
if ( m_settings->HasOutputFile ) {
// open output file stream
outFile.open(m_settings->OutputFilename.c_str());
if ( !outFile ) {
cerr << "bamtools coverage ERROR: could not open " << m_settings->OutputFilename
<< " for output" << endl;
return false;
}
// set m_out to file's streambuf
m_out.rdbuf(outFile.rdbuf());
}
//open our BAM reader
BamReader reader;
if ( !reader.Open(m_settings->InputBamFilename) ) {
cerr << "bamtools coverage ERROR: could not open input BAM file: " << m_settings->InputBamFilename << endl;
return false;
}
// retrieve references
m_references = reader.GetReferenceData();
// set up our output 'visitor'
CoverageVisitor* cv = new CoverageVisitor(m_references, &m_out);
// set up pileup engine with 'visitor'
PileupEngine pileup;
pileup.AddVisitor(cv);
// process input data
BamAlignment al;
while ( reader.GetNextAlignment(al) )
pileup.AddAlignment(al);
// clean up
reader.Close();
if ( m_settings->HasOutputFile )
outFile.close();
delete cv;
cv = 0;
// return success
return true;
}
int main (int argc, char *argv[]) {
if( (argc== 1) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cout<<"Usage:setAsUnpaired [in bam] [outbam]"<<endl<<"this program takes flags all paired sequences as singles"<<endl;
return 1;
}
string bamfiletopen = string(argv[1]);
string bamFileOUT = string(argv[2]);
BamReader reader;
BamWriter writer;
if ( !reader.Open(bamfiletopen) ) {
cerr << "Could not open input BAM files." << endl;
return 1;
}
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
if ( !writer.Open(bamFileOUT,header,references) ) {
cerr << "Could not open output BAM file "<<bamFileOUT << endl;
return 1;
}
BamAlignment al;
while ( reader.GetNextAlignment(al) ) {
if(al.IsMapped()){
cerr << "Cannot yet handle mapped reads " << endl;
return 1;
}
al.SetIsPaired (false);
writer.SaveAlignment(al);
} //while al
reader.Close();
writer.Close();
return 0;
}
int main(const int argc, char* const argv[]) {
int c, min_mapQ=0, seed=chrono::system_clock::now().time_since_epoch().count();
unsigned int flag_on=0, flag_off=0;
string fn_tgt, fn_in, fn_out="", out_format="b";
while ((c = getopt(argc, argv, "SbBcCt:h1Ho:q:f:F:ul:r:?T:R:L:s:@:m:x:U:")) >= 0) {
switch (c) {
case 's': seed = atoi(optarg); break;
case 'm': break;
case 'c': break;
case 'S': break;
case 'b': break;
case 'C': break;
case 'h': break;
case 'H': break;
case 'o': fn_out = optarg; break;
case 'U': break;
case 'f': flag_on |= strtol(optarg, 0, 0); break;
case 'F': flag_off |= strtol(optarg, 0, 0); break;
case 'q': min_mapQ = atoi(optarg); break;
case 'u': out_format = "u"; break;
case '1': break;
case 'l': break;
case 'r': break;
case 't': fn_tgt = optarg; break;
case 'R': break;
case '?': return usage();
case 'T': break;
case 'B': break;
case '@': break;
case 'x': break;
default: return usage();
}
}
if (fn_tgt.compare("") == 0) return usage();
if (argc == optind) return usage();
fn_in = argv[optind];
BamReader reader;
if (!reader.Open(fn_in)) {
cerr << "ERROR: cannot open [" << fn_in << "] for reading\n";
return 1;
}
if (!reader.LocateIndex()) {
cerr << "ERROR: cannot find BAM index for [" << fn_in << "]\n";
return 1;
}
const SamHeader header = reader.GetHeader();
if (header.SortOrder.compare("coordinate") != 0) {
cerr << "ERROR: [" << fn_in << "] not sorted by coordinate\n";
return 1;
}
const RefVector refseq = reader.GetReferenceData();
vector<BamRegion> regions;
vector<unsigned int> src_depths, tgt_depths;
if (read_region_depth(fn_tgt.c_str(), reader, regions, src_depths, tgt_depths) != 0) return 1;
BamWriter writer;
if (!writer.Open(fn_out, header, refseq)) {
cerr << "ERROR: cannot open [" << fn_out << "] for writing\n";
return 1;
}
BamAlignment aln;
vector<BamAlignment> reads;
vector<string> paired, unpaired;
unordered_map<int, int> kept;
unordered_map<string, unsigned int> seen, sampled;
unordered_map<string, vector<int> > pool;
for (size_t i=0; i<regions.size(); ++i) {
reads.clear();
paired.clear();
unpaired.clear();
kept.clear();
pool.clear();
char region_string[256];
sprintf(region_string, "%s:%d-%d", refseq[regions[i].LeftRefID].RefName.c_str(), regions[i].LeftPosition, regions[i].RightPosition);
if (!reader.SetRegion(regions[i])) {
cerr << "WARNING: failed to locate [" << region_string << "]\n";
//cerr << "WARNING: failed to locate [" << refseq[regions[i].LeftRefID].RefName << ':' << regions[i].LeftPosition << '-' << regions[i].RightPosition << "]\n";
continue;
}
while (reader.GetNextAlignment(aln)) {
if ((aln.AlignmentFlag & flag_on) == flag_on && !(aln.AlignmentFlag & flag_off) && aln.MapQuality >= min_mapQ)
reads.push_back(aln);
}
if (reads.size() == 0) continue;
unsigned int depth = 0;
for (size_t k=0; k<reads.size(); ++k) {
aln = reads[k];
string rn = aln.Name;
if (seen.find(rn) != seen.end()) { // if seen in previous regions
if (sampled.find(rn) != sampled.end()) { // if self or mate sampled before, sample it
if (sampled[rn] != aln.AlignmentFlag) kept[k] = 1; // if mate sampled before, keep it
depth += get_overlap(aln, regions[i]);
}
if (seen[rn] != aln.AlignmentFlag) seen[rn] = aln.AlignmentFlag;
}
else { // if not seen in previous regions
pool[rn].push_back(k);
}
if (depth > tgt_depths[i]) break;
}
if (depth < tgt_depths[i]) {
for (auto it=pool.begin(); it!=pool.end(); ++it) {
if (it->second.size()>1)
paired.push_back(it->first);
else
unpaired.push_back(it->first);
}
shuffle(paired.begin(), paired.end(), default_random_engine(seed));
shuffle(unpaired.begin(), unpaired.end(), default_random_engine(seed));
int n1=paired.size(), n2=unpaired.size(), k1, k2, k3;
while (depth < tgt_depths[i] && n1+n2 > 0) {
if (n1>0) {
k1 = pool[paired[--n1]][0];
k2 = pool[paired[n1]][1];
depth += get_overlap(reads[k1], regions[i]);
depth += get_overlap(reads[k2], regions[i]);
kept[k1] = 1; kept[k2] = 1;
continue;
}
if (n2>0) {
k3 = pool[unpaired[--n2]][0];
depth += get_overlap(reads[k3], regions[i]);
kept[k3] = 1;
continue;
}
}
}
for (auto it=pool.begin(); it!=pool.end(); ++it) {
string rn = it->first;
seen[rn] = reads[pool[rn].back()].AlignmentFlag;
}
for (auto it=kept.begin(); it!=kept.end(); ++it) {
int k = it->first;
string rn = reads[k].Name;
sampled[rn] = reads[k].AlignmentFlag;
writer.SaveAlignment(reads[k]);
}
cerr << "INFO: target=[" << tgt_depths[i] << "], actual=[" << depth << "], N(reads)=[" << reads.size() << "], N(kept)=[" << kept.size() << "] at [" << region_string << "]\n";
}
reader.Close();
return 0;
}
void BedIntersect::IntersectBam(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB->loadBedFileIntoMap();
// open the BAM file
BamReader reader;
BamWriter writer;
reader.Open(bamFile);
// get header & reference information
string header = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// open a BAM output to stdout if we are writing BAM
if (_bamOutput == true) {
// open our BAM writer
writer.Open("stdout", header, refs, _isUncompressedBam);
}
vector<BED> hits;
// reserve some space
hits.reserve(100);
_bedA->bedType = 6;
BamAlignment bam;
// get each set of alignments for each pair.
while (reader.GetNextAlignment(bam)) {
if (bam.IsMapped()) {
BED a;
a.chrom = refs.at(bam.RefID).RefName;
a.start = bam.Position;
a.end = bam.GetEndPosition(false);
// build the name field from the BAM alignment.
a.name = bam.Name;
if (bam.IsFirstMate()) a.name += "/1";
if (bam.IsSecondMate()) a.name += "/2";
a.score = ToString(bam.MapQuality);
a.strand = "+";
if (bam.IsReverseStrand()) a.strand = "-";
if (_bamOutput == true) {
bool overlapsFound = false;
// treat the BAM alignment as a single "block"
if (_obeySplits == false) {
overlapsFound = FindOneOrMoreOverlap(a);
}
// split the BAM alignment into discrete blocks and
// look for overlaps only within each block.
else {
bool overlapFoundForBlock;
bedVector bedBlocks; // vec to store the discrete BED "blocks" from a
// we don't want to split on "D" ops, hence the "false"
getBamBlocks(bam, refs, bedBlocks, false);
vector<BED>::const_iterator bedItr = bedBlocks.begin();
vector<BED>::const_iterator bedEnd = bedBlocks.end();
for (; bedItr != bedEnd; ++bedItr) {
overlapFoundForBlock = FindOneOrMoreOverlap(a);
if (overlapFoundForBlock == true)
overlapsFound = true;
}
}
if (overlapsFound == true) {
if (_noHit == false)
writer.SaveAlignment(bam);
}
else {
if (_noHit == true) {
writer.SaveAlignment(bam);
}
}
}
else {
// treat the BAM alignment as a single BED "block"
if (_obeySplits == false) {
FindOverlaps(a, hits);
hits.clear();
}
// split the BAM alignment into discrete BED blocks and
// look for overlaps only within each block.
else {
bedVector bedBlocks; // vec to store the discrete BED "blocks" from a
getBamBlocks(bam, refs, bedBlocks, false);
vector<BED>::const_iterator bedItr = bedBlocks.begin();
vector<BED>::const_iterator bedEnd = bedBlocks.end();
for (; bedItr != bedEnd; ++bedItr) {
FindOverlaps(*bedItr, hits);
hits.clear();
}
}
}
}
}
// close the relevant BAM files.
reader.Close();
if (_bamOutput == true) {
writer.Close();
}
}
void BedIntersectPE::IntersectBamPE(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB->loadBedFileIntoMap();
// open the BAM file
BamReader reader;
BamWriter writer;
reader.Open(bamFile);
// get header & reference information
string bamHeader = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// open a BAM output to stdout if we are writing BAM
if (_bamOutput == true) {
// set compression mode
BamWriter::CompressionMode compressionMode = BamWriter::Compressed;
if ( _isUncompressedBam ) compressionMode = BamWriter::Uncompressed;
writer.SetCompressionMode(compressionMode);
// open our BAM writer
writer.Open("stdout", bamHeader, refs);
}
// track the previous and current sequence
// names so that we can identify blocks of
// alignments for a given read ID.
string prevName, currName;
prevName = currName = "";
vector<BamAlignment> alignments; // vector of BAM alignments for a given ID in a BAM file.
alignments.reserve(100);
_bedA->bedType = 10; // it's a full BEDPE given it's BAM
// rip through the BAM file and convert each mapped entry to BEDPE
BamAlignment bam1, bam2;
while (reader.GetNextAlignment(bam1)) {
reader.GetNextAlignment(bam2);
if (bam1.Name != bam2.Name) {
while (bam1.Name != bam2.Name)
{
if (bam1.IsPaired())
{
cerr << "*****WARNING: Query " << bam1.Name
<< " is marked as paired, but it's mate does not occur"
<< " next to it in your BAM file. Skipping. " << endl;
}
bam1 = bam2;
reader.GetNextAlignment(bam2);
}
}
else if (bam1.IsPaired() && bam1.IsPaired()) {
ProcessBamBlock(bam1, bam2, refs, writer);
}
}
// close up
reader.Close();
if (_bamOutput == true) {
writer.Close();
}
}
int main (int argc, const char *argv[])
{
printf ("------------- bamrealignment --------------\n");
OptArgs opts;
opts.ParseCmdLine(argc, argv);
vector<int> score_vals(4);
string input_bam = opts.GetFirstString ('i', "input", "");
string output_bam = opts.GetFirstString ('o', "output", "");
opts.GetOption(score_vals, "4,-6,-5,-2", 's', "scores");
int clipping = opts.GetFirstInt ('c', "clipping", 2);
bool anchors = opts.GetFirstBoolean ('a', "anchors", true);
int bandwidth = opts.GetFirstInt ('b', "bandwidth", 10);
bool verbose = opts.GetFirstBoolean ('v', "verbose", false);
bool debug = opts.GetFirstBoolean ('d', "debug", false);
int format = opts.GetFirstInt ('f', "format", 1);
int num_threads = opts.GetFirstInt ('t', "threads", 8);
string log_fname = opts.GetFirstString ('l', "log", "");
if (input_bam.empty() or output_bam.empty())
return PrintHelp();
opts.CheckNoLeftovers();
std::ofstream logf;
if (log_fname.size ())
{
logf.open (log_fname.c_str ());
if (!logf.is_open ())
{
fprintf (stderr, "bamrealignment: Failed to open log file %s\n", log_fname.c_str());
return 1;
}
}
BamReader reader;
if (!reader.Open(input_bam)) {
fprintf(stderr, "bamrealignment: Failed to open input file %s\n", input_bam.c_str());
return 1;
}
SamHeader header = reader.GetHeader();
RefVector refs = reader.GetReferenceData();
BamWriter writer;
writer.SetNumThreads(num_threads);
if (format == 1)
writer.SetCompressionMode(BamWriter::Uncompressed);
else
writer.SetCompressionMode(BamWriter::Compressed);
if (!writer.Open(output_bam, header, refs)) {
fprintf(stderr, "bamrealignment: Failed to open output file %s\n", output_bam.c_str());
return 1;
}
// The meat starts here ------------------------------------
if (verbose)
cout << "Verbose option is activated, each alignment will print to screen." << endl
<< " After a read hit RETURN to continue to the next one," << endl
<< " or press q RETURN to quit the program," << endl
<< " or press s Return to silence verbose," << endl
<< " or press c RETURN to continue printing without further prompt." << endl << endl;
unsigned int readcounter = 0;
unsigned int mapped_readcounter = 0;
unsigned int realigned_readcounter = 0;
unsigned int modified_alignment_readcounter = 0;
unsigned int pos_update_readcounter = 0;
unsigned int failed_clip_realigned_readcount = 0;
unsigned int already_perfect_readcount = 0;
unsigned int bad_md_tag_readcount = 0;
unsigned int error_recreate_ref_readcount = 0;
unsigned int error_clip_anchor_readcount = 0;
unsigned int error_sw_readcount = 0;
unsigned int error_unclip_readcount = 0;
unsigned int start_position_shift;
int orig_position;
int new_position;
string md_tag, new_md_tag, input = "x";
vector<CigarOp> new_cigar_data;
vector<MDelement> new_md_data;
bool position_shift = false;
time_t start_time = time(NULL);
Realigner aligner;
aligner.verbose_ = verbose;
aligner.debug_ = debug;
if (!aligner.SetScores(score_vals))
cout << "bamrealignment: Four scores need to be provided: match, mismatch, gap open, gap extend score!" << endl;
aligner.SetAlignmentBandwidth(bandwidth);
BamAlignment alignment;
while(reader.GetNextAlignment(alignment)){
readcounter ++;
position_shift = false;
if ( (readcounter % 100000) == 0 )
cout << "Processed " << readcounter << " reads. Elapsed time: " << (time(NULL) - start_time) << endl;
if (alignment.IsMapped()) {
orig_position = alignment.Position;
mapped_readcounter++;
aligner.SetClipping(clipping, !alignment.IsReverseStrand());
if (aligner.verbose_) {
cout << endl;
if (alignment.IsReverseStrand())
cout << "The read is from the reverse strand." << endl;
else
cout << "The read is from the forward strand." << endl;
}
if (!alignment.GetTag("MD", md_tag)) {
if (aligner.verbose_)
cout << "Warning: Skipping read " << alignment.Name << ". It is mapped but missing MD tag." << endl;
if (logf.is_open ())
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "MISSMD" << '\n';
bad_md_tag_readcount++;
} else if (aligner.CreateRefFromQueryBases(alignment.QueryBases, alignment.CigarData, md_tag, anchors)) {
bool clipfail = false;
if (Realigner::CR_ERR_CLIP_ANCHOR == aligner.GetCreateRefError ())
{
clipfail = true;
failed_clip_realigned_readcount ++;
}
if (!aligner.computeSWalignment(new_cigar_data, new_md_data, start_position_shift)) {
if (aligner.verbose_)
cout << "Error in the alignment! Not updating read information." << endl;
if (logf.is_open ())
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "SWERR" << '\n';
error_sw_readcount++;
writer.SaveAlignment(alignment); // Write alignment unchanged
continue;
}
if (!aligner.addClippedBasesToTags(new_cigar_data, new_md_data, alignment.QueryBases.size())) {
if (aligner.verbose_)
cout << "Error when adding clipped anchors back to tags! Not updating read information." << endl;
if (logf.is_open ())
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "UNCLIPERR" << '\n';
writer.SaveAlignment(alignment); // Write alignment unchanged
error_unclip_readcount ++;
continue;
}
new_md_tag = aligner.GetMDstring(new_md_data);
realigned_readcounter++;
// adjust start position of read
if (!aligner.LeftAnchorClipped() and start_position_shift != 0) {
new_position = aligner.updateReadPosition(alignment.CigarData, (int)start_position_shift, alignment.Position);
if (new_position != alignment.Position) {
pos_update_readcounter++;
position_shift = true;
alignment.Position = new_position;
}
}
if (position_shift || alignment.CigarData.size () != new_cigar_data.size () || md_tag != new_md_tag)
{
if (logf.is_open ())
{
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "MOD";
if (position_shift)
logf << "-SHIFT";
if (clipfail)
logf << " NOCLIP";
logf << '\n';
}
modified_alignment_readcounter++;
}
else
{
if (logf.is_open ())
{
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "UNMOD";
if (clipfail)
logf << " NOCLIP";
logf << '\n';
}
}
if (aligner.verbose_){
cout << alignment.Name << endl;
cout << "------------------------------------------" << endl;
// Wait for input to continue or quit program
if (input.size() == 0)
input = 'x';
else if (input[0] != 'c' and input[0] != 'C')
getline(cin, input);
if (input.size()>0){
if (input[0] == 'q' or input[0] == 'Q')
return 1;
else if (input[0] == 's' or input[0] == 'S')
aligner.verbose_ = false;
}
}
// Finally update alignment information
alignment.CigarData = new_cigar_data;
alignment.EditTag("MD", "Z" , new_md_tag);
} // end of CreateRef else if
else {
switch (aligner.GetCreateRefError ())
{
case Realigner::CR_ERR_RECREATE_REF:
if (logf.is_open ())
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "RECRERR" << '\n';
error_recreate_ref_readcount++;
break;
case Realigner::CR_ERR_CLIP_ANCHOR:
if (logf.is_open ())
logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "CLIPERR" << '\n';
error_clip_anchor_readcount++;
break;
default:
// On a good run this writes way too many reads to the log file - don't want to create a too large txt file
// if (logf.is_open ())
//logf << alignment.Name << '\t' << alignment.IsReverseStrand() << '\t' << alignment.RefID << '\t' << setfill ('0') << setw (8) << orig_position << '\t' << "PERFECT" << '\n';
already_perfect_readcount++;
break;
}
if (aligner.verbose_) {
cout << alignment.Name << endl;
cout << "------------------------------------------" << endl;
// Wait for input to continue or quit program
if (input.size() == 0)
input = 'x';
else if (input[0] != 'c' and input[0] != 'C')
getline(cin, input);
if (input.size()>0){
if (input[0] == 'q' or input[0] == 'Q')
return 1;
else if (input[0] == 's' or input[0] == 'S')
aligner.verbose_ = false;
}
}
}
// --- Debug output for Rajesh ---
if (debug && aligner.invalid_cigar_in_input) {
aligner.verbose_ = true;
cout << "Invalid cigar string / md tag pair in read " << alignment.Name << endl;
// Rerun reference generation to display error
aligner.CreateRefFromQueryBases(alignment.QueryBases, alignment.CigarData, md_tag, anchors);
aligner.verbose_ = verbose;
aligner.invalid_cigar_in_input = false;
}
// --- --- ---
} // end of if isMapped
writer.SaveAlignment(alignment);
} // end while loop over reads
if (aligner.invalid_cigar_in_input)
cerr << "WARNING bamrealignment: There were invalid cigar string / md tag pairs in the input bam file." << endl;
// ----------------------------------------------------------------
// program end -- output summary information
cout << " File: " << input_bam << endl
<< " Total reads: " << readcounter << endl
<< " Mapped reads: " << mapped_readcounter << endl;
if (bad_md_tag_readcount)
cout << " Skipped: bad MD tags: " << bad_md_tag_readcount << endl;
if (error_recreate_ref_readcount)
cout << " Skipped: unable to recreate ref: " << error_recreate_ref_readcount << endl;
if (error_clip_anchor_readcount)
cout << " Skipped: error clipping anchor: " << error_clip_anchor_readcount << endl;
cout << " Skipped: already perfect: " << already_perfect_readcount << endl
<< " Total reads realigned: " << mapped_readcounter - already_perfect_readcount - bad_md_tag_readcount - error_recreate_ref_readcount - error_clip_anchor_readcount << endl;
if (failed_clip_realigned_readcount)
cout << " (including " << failed_clip_realigned_readcount << " that failed to clip)" << endl;
if (error_sw_readcount)
cout << " Failed to complete SW alignment: " << error_sw_readcount << endl;
if (error_unclip_readcount)
cout << " Failed to unclip anchor: " << error_unclip_readcount << endl;
cout << " Succesfully realigned: " << realigned_readcounter << endl
<< " Modified alignments: " << modified_alignment_readcounter << endl
<< " Shifted position: " << pos_update_readcounter << endl;
cout << "Processing time: " << (time(NULL)-start_time) << " seconds." << endl;
cout << "INFO: The output BAM file may be unsorted." << endl;
cout << "------------------------------------------" << endl;
return 0;
}
void BedWindow::WindowIntersectBam(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB->loadBedFileIntoMap();
// open the BAM file
BamReader reader;
BamWriter writer;
reader.Open(bamFile);
// get header & reference information
string bamHeader = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// open a BAM output to stdout if we are writing BAM
if (_bamOutput == true) {
// set compression mode
BamWriter::CompressionMode compressionMode = BamWriter::Compressed;
if ( _isUncompressedBam ) compressionMode = BamWriter::Uncompressed;
writer.SetCompressionMode(compressionMode);
// open our BAM writer
writer.Open("stdout", bamHeader, refs);
}
vector<BED> hits; // vector of potential hits
// reserve some space
hits.reserve(100);
_bedA->bedType = 6;
BamAlignment bam;
bool overlapsFound;
// get each set of alignments for each pair.
while (reader.GetNextAlignment(bam)) {
if (bam.IsMapped()) {
BED a;
a.chrom = refs.at(bam.RefID).RefName;
a.start = bam.Position;
a.end = bam.GetEndPosition(false, false);
// build the name field from the BAM alignment.
a.name = bam.Name;
if (bam.IsFirstMate()) a.name += "/1";
if (bam.IsSecondMate()) a.name += "/2";
a.score = ToString(bam.MapQuality);
a.strand = "+"; if (bam.IsReverseStrand()) a.strand = "-";
if (_bamOutput == true) {
overlapsFound = FindOneOrMoreWindowOverlaps(a);
if (overlapsFound == true) {
if (_noHit == false)
writer.SaveAlignment(bam);
}
else {
if (_noHit == true)
writer.SaveAlignment(bam);
}
}
else {
FindWindowOverlaps(a, hits);
hits.clear();
}
}
// BAM IsMapped() is false
else if (_noHit == true) {
writer.SaveAlignment(bam);
}
}
// close the relevant BAM files.
reader.Close();
if (_bamOutput == true) {
writer.Close();
}
}
int main(int argc, char** argv) {
FILE *fsp = 0, *fep = 0;
int flag; // 0
char chrom[255]; // 1
int pos; // 2
int quality; // 3
char *cigar; // 4
size_t len = 0;
char chromcopy[255] = { 0 };
char fn[255];
int i;
int forward, reverse;
bool direction; // 0
bool leftflag;
const char *prefix = "";
int cnt = 0;
BamReader reader;
//if (argc < 2 || argc > 4) usage();
for (i=1; i<argc; i++) {
if ((strcmp(argv[i], "-p") == 0) && (i != argc-1)) {
prefix = argv[++i];
} else {
if (!reader.IsOpen()) {
try {
reader.Open(argv[i]);
} catch (exception& e) {
cout << e.what();
throw;
}
} else {
usage();
}
}
}
if (!reader.IsOpen()) usage();
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
BamAlignment al;
while(reader.GetNextAlignmentCore(al)) {
flag = al.AlignmentFlag;
if ((flag & 256) || (flag & 4) || (flag & 512) || (flag & 1024)) continue;
direction = ( (flag & 16) == 16);
strcpy(chrom, references[al.RefID].RefName.c_str());
pos = al.Position+1;
quality = al.MapQuality;
if (quality == 0) continue;
if (strcmp(chromcopy, chrom) != 0) {
if (fsp) fclose(fsp);
if (fep) fclose(fep);
fn[0] = 0;
if (prefix != "") {
strcat(fn, prefix);
}
strcat(fn, chrom);
strcat(fn, "_forward.txt");
fsp = fopen(fn, "w");
fn[0] = 0;
if (prefix != "") {
strcat(fn, prefix);
}
strcat(fn, chrom);
strcat(fn, "_reverse.txt");
fep = fopen(fn, "w");
strcpy(chromcopy, chrom);
}
vector<CigarOp> cigars = al.CigarData;
forward = pos;
reverse = pos;
leftflag = true;
for (i = 0; i < cigars.size(); i++) {
if (cigars[i].Type == 'S' || cigars[i].Type == 'H') {
if (leftflag == false) break;
continue;
} else {
if (cigars[i].Type != 'I') reverse += cigars[i].Length;
leftflag = false;
}
}
if (direction == 0)
fprintf(fsp, "%d\n", forward);
else
fprintf(fep, "%d\n", reverse-1);
cnt++;
if (cnt % 500000 == 0)
printf("%d ", cnt);
}
reader.Close();
printf("\n");
if (fsp) fclose(fsp);
if (fep) fclose(fep);
}
int main (int argc, char *argv[]) {
if( (argc!= 4 && argc !=5 && argc !=6) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cerr<<"Usage:splitByRG [in bam] [rg Tally] [out prefix] (optional target)"<<endl<<"this program will subsample a BAM file per read group for a certain target\nFor example splitByRG in.bam tally.txt out will create\nout.rg1.bam\nout.rg2.bam\n"<<endl;
return 1;
}
string bamfiletopen = string(argv[1]);
string rgTally = string(argv[2]);
string bamDirOutPrefix = string(argv[3]);
int target = 200000;
int maxTarget = 1000000;
if(argc==5){
target = destringify<int> ( string(argv[4]) );
}
if(argc==6){
target = destringify<int> ( string(argv[4]) );
maxTarget = destringify<int> ( string(argv[5]) );
}
cerr<<"minimum fragments:\t"<<target<<endl;
cerr<<"target fragments:\t"<<maxTarget<<endl;
string line;
ifstream myFileTally;
map<string,double> rg2Fraction;
myFileTally.open(rgTally.c_str(), ios::in);
cerr<<"Retained groups:\n"<<endl;
cerr<<"RG\t#mapped\tfraction retained"<<endl;
cerr<<"-----------------------------------"<<endl;
if (myFileTally.is_open()){
while ( getline (myFileTally,line)){
vector<string> tokens = allTokens(line,'\t');
if(tokens.size() > 6)
if( tokens[1] == "pass" &&
(tokens[0] != "\"\"" &&
tokens[0] != "control" &&
tokens[0] != "TOTAL") ){
//cout<<tokens[0]<<"\t"<<tokens[5]<<endl;
int count = destringify<int>(tokens[5]);
if(count>target){
if(count>=maxTarget){
rg2Fraction[ tokens[0] ] = double(maxTarget)/double(count);
cout<<tokens[0]<<"\t"<<count<<"\t"<<double(maxTarget)/double(count)<<endl;
}else{
cout<<tokens[0]<<"\t"<<count<<"\t"<<1.0<<endl;
rg2Fraction[ tokens[0] ] = 1.0;
}
}
}
}
myFileTally.close();
}else{
cerr << "Unable to open file "<<rgTally<<endl;
return 1;
}
map<string,BamWriter *> rg2BamWriter;
// if(!isDirectory(bamDirOut)){
// cerr<<"ERROR: the out directory does not exist"<<endl;
// return 1;
// }
BamReader reader;
if ( !reader.Open(bamfiletopen) ) {
cerr << "Could not open input BAM files." << endl;
return 1;
}
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
vector<RefData> refData=reader.GetReferenceData();
SamReadGroupDictionary srgd=header.ReadGroups;
for(SamReadGroupConstIterator srgci=srgd.ConstBegin();
srgci<srgd.ConstEnd();
srgci++){
//cout<<*srgci<<endl;
const SamReadGroup rg = (*srgci);
//cout<<rg.ID<<endl;
if( rg2Fraction.find(rg.ID) != rg2Fraction.end() ){
rg2BamWriter[rg.ID] = new BamWriter();
rg2BamWriter[rg.ID]->Open(bamDirOutPrefix+"."+rg.ID+".bam",header,references);
}
//cout<<bamDirOutPrefix+"."+rg.ID+".bam"<<endl;
}
// return 1;
// BamWriter unmapped;
// cout<<header.ToString()<<endl;
// return 1;
// if ( !unmapped.Open(bamDirOutPrefix+".unmapped.bam",header,references) ) {
// cerr << "Could not open output BAM file "<< bamDirOutPrefix+".unmapped.bam" << endl;
// return 1;
// }
// cout<<"reading"<<endl;
BamAlignment al;
unsigned int total=0;
while ( reader.GetNextAlignment(al) ) {
if(al.HasTag("RG") &&
al.IsMapped() ){
string rgTag;
al.GetTag("RG",rgTag);
//cout<<rgTag<<endl;
if(rg2BamWriter.find(rgTag) == rg2BamWriter.end()){ //new: ignore completely
}else{
if( randomProb() <= rg2Fraction[ rgTag ] ){
rg2BamWriter[rgTag]->SaveAlignment(al);
//cout<<"wrote "<<rgTag<<endl;
} else{
//cout<<"skipped "<<rgTag<<endl;
}
}
}// else{
// string rgTag="unknown";
// //cout<<rgTag<<endl;
// if(rg2BamWriter.find(rgTag) == rg2BamWriter.end()){ //new
// cerr<<"Found new RG "<<rgTag<<endl;
// rg2BamWriter[rgTag] = new BamWriter();
// if ( !rg2BamWriter[rgTag]->Open(bamDirOutPrefix+"."+rgTag+".bam",header,references) ) {
// cerr << "Could not open output BAM file "<< bamDirOutPrefix<<"."<<rgTag<<".bam" << endl;
// return 1;
// }
// rg2BamWriter[rgTag]->SaveAlignment(al);
// }else{
// rg2BamWriter[rgTag]->SaveAlignment(al);
// }
// // cerr << "Cannot get RG tag for " << al.Name<<endl;
// // return 1;
// }
total++;
} //while al
reader.Close();
// writer.Close();
// unmapped.Close();
map<string,BamWriter *>::iterator rg2BamWriterIt;
for (rg2BamWriterIt =rg2BamWriter.begin();
rg2BamWriterIt!=rg2BamWriter.end();
rg2BamWriterIt++){
rg2BamWriterIt->second->Close();
}
cerr<<"Wrote succesfully "<<total<<" reads"<<endl;
return 0;
}
int GenericIndividualSnpCall::call(Fasta &fastaObj, BamReader &bamObj, BamRegion &roi, GenericProbabilisticAlignment &probAligner, VariantCallSetting& snpCallSettings, vector<GenericVariant> &variantSet)
{
RefVector chromosomes = bamObj.GetReferenceData();
// set up genome blocks
vector<int> BlockChrID, BlockLeftPos, BlockRightPos;
int BlockNumber=setupGenomeBlock(chromosomes, roi, BlockChrID, BlockLeftPos, BlockRightPos);
int numSNP = 0;
// iterate throught blocks
for (int i=0; i<BlockNumber; ++i)
{
if (m_verbosity>=1)
{
cout << "processing " << chromosomes[BlockChrID[i]].RefName << ":" << BlockLeftPos[i]+1 << "-" << BlockRightPos[i] << endl;
}
clock_t startTime = clock();
// genome
string BlockGenome;
fastaObj.GetSequence(BlockChrID[i], BlockLeftPos[i], BlockRightPos[i], BlockGenome);
map<int,list<tuple<char,int,int,double>>> BlockBamData;
AlleleSet BlockSnpAlleleCandidates;
// profile SNP sites by the simple method
simpleSnpCall(BlockGenome, bamObj, BlockChrID[i], BlockLeftPos[i], BlockRightPos[i], BlockSnpAlleleCandidates, BlockBamData);
// merge SNP sites to SNP blocks
vector<tuple<int,int,list<Allele>>> BlockSnpLoci;
mergeSnpSitesToBlocks(BlockSnpAlleleCandidates, BlockSnpLoci);
// iterate through Snp locus
for (int j=0; j<BlockSnpLoci.size(); j++)
{
int BlockSnpLeftPos = get<0>(BlockSnpLoci[j]);
int BlockSnpRightPos = get<1>(BlockSnpLoci[j]);
// it is a SNP site
if (BlockSnpRightPos==BlockSnpLeftPos+1)
{
simpleBayesianSnpCall(fastaObj, bamObj, BlockChrID[i], BlockSnpLeftPos, BlockSnpRightPos, get<2>(BlockSnpLoci[j]), BlockBamData[BlockSnpLeftPos], snpCallSettings, variantSet);
}else if (BlockSnpRightPos==BlockSnpLeftPos+2)
{
for (int pos=BlockSnpLeftPos; pos<BlockSnpRightPos; pos++)
{
list<Allele> fAlleles = get<2>(BlockSnpLoci[j]);
list<Allele> tAlleles;
for (list<Allele>::iterator faIter=fAlleles.begin(); faIter!=fAlleles.end(); faIter++)
{
if (faIter->m_chrPosition==pos)
tAlleles.emplace_back(*faIter);
}
if (!tAlleles.empty())
simpleBayesianSnpCall(fastaObj, bamObj, BlockChrID[i], pos, pos+1, tAlleles, BlockBamData[pos], snpCallSettings, variantSet);
}
}
else // it is a MNP site
{
PyroHMMsnp(fastaObj, bamObj, BlockChrID[i], BlockSnpLeftPos, BlockSnpRightPos, probAligner, get<2>(BlockSnpLoci[j]), snpCallSettings, variantSet);
}
}
clock_t endTime = clock();
if (m_verbosity>=1)
{
cout << "time elapsed " << ((endTime-startTime)/(double)CLOCKS_PER_SEC/60.) << " minutes";
cout << ", ";
cout << "call " << variantSet.size()-numSNP << " SNPs" << endl;
}
numSNP = variantSet.size();
}
return variantSet.size();
}
void BedGenomeCoverage::CoverageBam(string bamFile) {
ResetChromCoverage();
// open the BAM file
BamReader reader;
if (!reader.Open(bamFile)) {
cerr << "Failed to open BAM file " << bamFile << endl;
exit(1);
}
// get header & reference information
string header = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// load the BAM header references into a BEDTools "genome file"
_genome = new GenomeFile(refs);
// convert each aligned BAM entry to BED
// and compute coverage on B
BamAlignment bam;
while (reader.GetNextAlignment(bam)) {
// skip if the read is unaligned
if (bam.IsMapped() == false)
continue;
bool _isReverseStrand = bam.IsReverseStrand();
//changing second mate's strand to opposite
if( _dUTP && bam.IsPaired() && bam.IsMateMapped() && bam.IsSecondMate())
_isReverseStrand = !bam.IsReverseStrand();
// skip if we care about strands and the strand isn't what
// the user wanted
if ( (_filterByStrand == true) &&
((_requestedStrand == "-") != _isReverseStrand) )
continue;
// extract the chrom, start and end from the BAM alignment
string chrom(refs.at(bam.RefID).RefName);
CHRPOS start = bam.Position;
CHRPOS end = bam.GetEndPosition(false, false) - 1;
// are we on a new chromosome?
if ( chrom != _currChromName )
StartNewChrom(chrom);
if(_pair_chip_) {
// Skip if not a proper pair
if (bam.IsPaired() && (!bam.IsProperPair() or !bam.IsMateMapped()) )
continue;
// Skip if wrong coordinates
if( ( (bam.Position<bam.MatePosition) && bam.IsReverseStrand() ) ||
( (bam.MatePosition < bam.Position) && bam.IsMateReverseStrand() ) ) {
//chemically designed: left on positive strand, right on reverse one
continue;
}
/*if(_haveSize) {
if (bam.IsFirstMate() && bam.IsReverseStrand()) { //put fragmentSize in to the middle of pair end_fragment
int mid = bam.MatePosition+abs(bam.InsertSize)/2;
if(mid<_fragmentSize/2)
AddCoverage(0, mid+_fragmentSize/2);
else
AddCoverage(mid-_fragmentSize/2, mid+_fragmentSize/2);
}
else if (bam.IsFirstMate() && bam.IsMateReverseStrand()) { //put fragmentSize in to the middle of pair end_fragment
int mid = start+abs(bam.InsertSize)/2;
if(mid<_fragmentSize/2)
AddCoverage(0, mid+_fragmentSize/2);
else
AddCoverage(mid-_fragmentSize/2, mid+_fragmentSize/2);
}
} else */
if (bam.IsFirstMate() && bam.IsReverseStrand()) { //prolong to the mate to the left
AddCoverage(bam.MatePosition, end);
}
else if (bam.IsFirstMate() && bam.IsMateReverseStrand()) { //prolong to the mate to the right
AddCoverage(start, start + abs(bam.InsertSize) - 1);
}
} else if (_haveSize) {
if(bam.IsReverseStrand()) {
if(end<_fragmentSize) { //sometimes fragmentSize is bigger :(
AddCoverage(0, end);
} else {
AddCoverage(end + 1 - _fragmentSize, end );
}
} else {
AddCoverage(start,start+_fragmentSize - 1);
}
} else
// add coverage accordingly.
if (!_only_5p_end && !_only_3p_end) {
bedVector bedBlocks;
// we always want to split blocks when a D CIGAR op is found.
// if the user invokes -split, we want to also split on N ops.
if (_obeySplits) { // "D" true, "N" true
GetBamBlocks(bam, refs.at(bam.RefID).RefName, bedBlocks, true, true);
}
else { // "D" true, "N" false
GetBamBlocks(bam, refs.at(bam.RefID).RefName, bedBlocks, true, false);
}
AddBlockedCoverage(bedBlocks);
}
else if (_only_5p_end) {
CHRPOS pos = ( !bam.IsReverseStrand() ) ? start : end;
AddCoverage(pos,pos);
}
else if (_only_3p_end) {
CHRPOS pos = ( bam.IsReverseStrand() ) ? start : end;
AddCoverage(pos,pos);
}
}
// close the BAM
reader.Close();
// process the results of the last chromosome.
ReportChromCoverage(_currChromCoverage, _currChromSize,
_currChromName, _currChromDepthHist);
// report all empty chromsomes
PrintEmptyChromosomes();
// report the overall coverage if asked.
PrintFinalCoverage();
}
void TagBam::Tag() {
// open the annotations files for processing;
OpenAnnoFiles();
// open the BAM file
BamReader reader;
BamWriter writer;
if (!reader.Open(_bamFile)) {
cerr << "Failed to open BAM file " << _bamFile << endl;
exit(1);
}
// get header & reference information
string bamHeader = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// set compression mode
BamWriter::CompressionMode compressionMode = BamWriter::Compressed;
// if ( _isUncompressedBam ) compressionMode = BamWriter::Uncompressed;
writer.SetCompressionMode(compressionMode);
// open our BAM writer
writer.Open("stdout", bamHeader, refs);
// rip through the BAM file and test for overlaps with each annotation file.
BamAlignment al;
vector<BED> hits;
while (reader.GetNextAlignment(al)) {
if (al.IsMapped() == true) {
BED a;
a.chrom = refs.at(al.RefID).RefName;
a.start = al.Position;
a.end = al.GetEndPosition(false, false);
a.strand = "+";
if (al.IsReverseStrand()) a.strand = "-";
ostringstream annotations;
// annotate the BAM file based on overlaps with the annotation files.
for (size_t i = 0; i < _annoFiles.size(); ++i)
{
// grab the current annotation file.
BedFile *anno = _annoFiles[i];
if (!_useNames && !_useScores && !_useIntervals) {
// add the label for this annotation file to tag if there is overlap
if (anno->anyHits(a.chrom, a.start, a.end, a.strand,
_sameStrand, _diffStrand, _overlapFraction, false))
{
annotations << _annoLabels[i] << ";";
}
}
// use the score field
else if (!_useNames && _useScores && !_useIntervals) {
anno->allHits(a.chrom, a.start, a.end, a.strand,
hits, _sameStrand, _diffStrand, 0.0, false);
for (size_t i = 0; i < hits.size(); ++i) {
annotations << hits[i].score;
if (i < hits.size() - 1) annotations << ",";
}
if (hits.size() > 0) annotations << ";";
hits.clear();
}
// use the name field from the annotation files to populate tag
else if (_useNames && !_useScores && !_useIntervals) {
anno->allHits(a.chrom, a.start, a.end, a.strand,
hits, _sameStrand, _diffStrand, 0.0, false);
for (size_t j = 0; j < hits.size(); ++j) {
annotations << hits[j].name;
if (j < hits.size() - 1) annotations << ",";
}
if (hits.size() > 0) annotations << ";";
hits.clear();
}
// use the full interval information annotation files to populate tag
else if (!_useNames && !_useScores && _useIntervals) {
anno->allHits(a.chrom, a.start, a.end, a.strand,
hits, _sameStrand, _diffStrand, 0.0, false);
for (size_t j = 0; j < hits.size(); ++j) {
annotations << _annoLabels[i] << ":" <<
hits[j].chrom << ":" <<
hits[j].start << "-" <<
hits[j].end << "," <<
hits[j].name << "," <<
hits[j].score << "," <<
hits[j].strand;
if (j < hits.size() - 1) annotations << ",";
}
if (hits.size() > 0) annotations << ";";
hits.clear();
}
}
// were there any overlaps with which to make a tag?
if (annotations.str().size() > 0) {
al.AddTag(_tag, "Z", annotations.str().substr(0, annotations.str().size() - 1)); // get rid of the last ";"
}
}
writer.SaveAlignment(al);
}
reader.Close();
writer.Close();
// close the annotations files;
CloseAnnoFiles();
}
int
main(int argc, char* argv[]) {
// validate argument count
if( argc != 2 ) {
cerr << "USAGE: " << argv[0] << " <input BAM file> " << endl;
return EXIT_FAILURE;
}
string filename = argv[1];
//cerr << "Printing alignments from file: " << filename << endl;
BamReader reader;
if (!reader.Open(filename)) {
cerr << "could not open filename " << filename << endl;
return EXIT_FAILURE;
}
cerr << filename << ": Done opening" << endl;
// Header can't be used to accurately determine sort order because samtools never
// changes it; instead, check after loading each read as is done with "samtools index"
// We don't need to load an index (right?)
// if (!reader.LocateIndex()) {
// const string index_filename = filename + ".bai";
// if (!reader.OpenIndex(index_filename)) {
// cerr << "could not open index" << endl;
// }
// }
const SamHeader header = reader.GetHeader();
cerr << filename << ": Done getting header" << endl;
const RefVector refs = reader.GetReferenceData();
cerr << filename << ": Done getting reference data" << endl;
BamWriter writer;
if (! output_bam_filename.empty()) {
if (! writer.Open(output_bam_filename, header, refs)) {
cerr << "Could not open BAM output file " << output_bam_filename << endl;
return EXIT_FAILURE;
}
cerr << filename << ": Done opening BAM output file " << output_bam_filename << endl;
}
alignmentMap read1Map; // a single map, for all reads awaiting their mate
typedef map<string,int32_t> stringMap;
typedef stringMap::iterator stringMapI;
stringMap ref_mates;
// alignmentMap read1Map, read2Map;
BamAlignment full_al;
int32_t count = 0;
uint32_t max_reads_in_map = 0;
int32_t n_reads_skipped_unmapped = 0;
int32_t n_reads_skipped_mate_unmapped = 0;
int32_t n_reads_skipped_wont_see_mate = 0;
int32_t n_reads_skipped_mate_tail_est = 0;
int32_t n_reads_skipped_ref_mate = 0;
int32_t n_reads = 0;
int32_t n_singleton_reads = 0;
int32_t last_RefID = -1;
int32_t last_Position = -1;
cerr << filename << ": Looking for up to " << pairs_to_process << " link pairs,"
<< " total tail = " << link_pair_total_tail
<< " critical tail = " << link_pair_crit_tail
<< ", must be on diff chromosome = " << link_pair_diff_chrom << endl;
while (reader.GetNextAlignment(full_al)
&& (! pairs_to_process || count < pairs_to_process)) {
BamAlignment al = full_al;
//printAlignmentInfo(al, refs);
//++count;
++n_reads;
if (last_RefID < 0) last_RefID = al.RefID;
if (last_Position < 0) last_Position = al.Position;
if (al.RefID > last_RefID) {
// We've moved to the next reference sequence
// Clean up reads with mates expected here that haven't been seen
if (debug_ref_mate) {
cerr << "MISSED " << ref_mates.size() << " ref_mates on this reference "
<< last_RefID << " " << refs[last_RefID].RefName << endl;
}
for (stringMapI rmI = ref_mates.begin(); rmI != ref_mates.end(); ++rmI) {
++n_reads_skipped_ref_mate;
read1Map.erase(read1Map.find(rmI->first));
ref_mates.erase(ref_mates.find(rmI->first));
}
last_RefID = al.RefID;
last_Position = al.Position;
} else if (! isCoordinateSorted(al.RefID, al.Position, last_RefID, last_Position)) {
cerr << filename << " is not sorted, " << al.Name << " out of position" << endl;
return EXIT_FAILURE;
}
if (! al.IsMapped()) { ++n_reads_skipped_unmapped; continue; }
if (! al.IsMateMapped()) { ++n_reads_skipped_mate_unmapped; continue; }
alignmentMapI mI = read1Map.find(al.Name);
if (mI == read1Map.end()) {
// the read name has not been seen before
if (al.MateRefID < al.RefID
|| (al.MateRefID == al.RefID && al.MatePosition < al.Position)) {
// we should have seen its mate earlier, so skip it
++n_reads_skipped_wont_see_mate;
continue;
}
// If the mate likely to also be a link pair candidate, add the read
int32_t mate_tail_est = readTailS(al.IsMateMapped(), al.IsMateReverseStrand(),
al.MatePosition, refs[al.MateRefID].RefLength, max_read_length);
if (mate_tail_est <= mate_tail_est_crit) {
// the mate tail estimate suggests it might be a link pair candidate
read1Map[al.Name] = al; // add the read to the map
} else {
// the mate tail estimate appears too long for the mate to be a candidate
++n_reads_skipped_mate_tail_est;
continue;
}
if (read1Map.size() > max_reads_in_map) max_reads_in_map = read1Map.size();
if (al.MateRefID == al.RefID && al.MatePosition >= al.Position) {
// the mate is expected later on this contig
ref_mates[al.Name] = al.MateRefID;
}
} else {
// get the mate's alignment, and process the pair
const BamAlignment& al_mate = mI->second;
if (processReadPair(al, al_mate, refs, link_pair_total_tail,
link_pair_crit_tail, link_pair_diff_chrom)) {
++count;
// write to the new BAM file, if the string is not empty
if (! output_bam_filename.empty()) {
writer.SaveAlignment(al_mate); // the first one seen
writer.SaveAlignment(al); // the second one seen
}
}
read1Map.erase(mI);
if (al.MateRefID == al.RefID) {
stringMapI rmI = ref_mates.find(al.Name);
if (rmI == ref_mates.end()) {
cerr << "expected a ref_mate, couldn't find its name: " << al.Name << endl;
return EXIT_FAILURE;
}
ref_mates.erase(rmI);
}
}
}
cerr << "===============================" << endl;
cerr << read1Map.size() << " alignments left in read1Map" << endl;
cerr << max_reads_in_map << " maximum number of reads in read1Map" << endl;
cerr << count << " pairs processed" << endl;
cerr << "===============================" << endl;
cerr << n_reads << " total reads" << endl;
cerr << n_singleton_reads << " singleton reads" << endl;
cerr << n_reads_skipped_unmapped << " reads skipped because unmapped" << endl;
cerr << n_reads_skipped_mate_unmapped << " reads skipped because mate unmapped" << endl;
cerr << n_reads_skipped_wont_see_mate << " reads skipped because mate won't be seen" << endl;
cerr << n_reads_skipped_mate_tail_est << " reads skipped because mate tail appears too long" << endl;
cerr << n_reads_skipped_ref_mate << " reads skipped because mate not on reference" << endl;
reader.Close();
if (! output_bam_filename.empty()) {
writer.Close();
}
return EXIT_SUCCESS;
}
void BedIntersect::IntersectBam(string bamFile) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
_bedB = new BedFile(_bedBFile);
_bedB->loadBedFileIntoMap();
// create a dummy BED A file for printing purposes if not
// using BAM output.
if (_bamOutput == false) {
_bedA = new BedFile(_bedAFile);
_bedA->bedType = 12;
}
// open the BAM file
BamReader reader;
BamWriter writer;
reader.Open(bamFile);
// get header & reference information
string bamHeader = reader.GetHeaderText();
RefVector refs = reader.GetReferenceData();
// open a BAM output to stdout if we are writing BAM
if (_bamOutput == true) {
// set compression mode
BamWriter::CompressionMode compressionMode = BamWriter::Compressed;
if ( _isUncompressedBam ) compressionMode = BamWriter::Uncompressed;
writer.SetCompressionMode(compressionMode);
// open our BAM writer
writer.Open("stdout", bamHeader, refs);
}
vector<BED> hits;
// reserve some space
hits.reserve(100);
BamAlignment bam;
// get each set of alignments for each pair.
while (reader.GetNextAlignment(bam)) {
// save an unaligned read if -v
if (!bam.IsMapped()) {
if (_noHit == true)
writer.SaveAlignment(bam);
continue;
}
// break alignment into discrete blocks,
bedVector bed_blocks;
string chrom = refs.at(bam.RefID).RefName;
GetBamBlocks(bam, chrom, bed_blocks, false, true);
// create a basic BED entry from the BAM alignment
BED bed;
MakeBedFromBam(bam, chrom, bed_blocks, bed);
bool overlapsFound = false;
if ((_bamOutput == true) && (_obeySplits == false))
{
overlapsFound = _bedB->anyHits(bed.chrom, bed.start, bed.end,
bed.strand, _sameStrand, _diffStrand,
_overlapFraction, _reciprocal);
}
else if ( ((_bamOutput == true) && (_obeySplits == true)) ||
((_bamOutput == false) && (_obeySplits == true)) )
{
// find the hits that overlap with the full span of the blocked BED
_bedB->allHits(bed.chrom, bed.start, bed.end, bed.strand,
hits, _sameStrand, _diffStrand,
_overlapFraction, _reciprocal);
// find the overlaps between the block in A and B
overlapsFound = FindBlockedOverlaps(bed, bed_blocks, hits, _bamOutput);
}
else if ((_bamOutput == false) && (_obeySplits == false))
{
FindOverlaps(bed, hits);
}
// save the BAM alignment if overlap reqs. were met
if (_bamOutput == true) {
if ((overlapsFound == true) && (_noHit == false))
writer.SaveAlignment(bam);
else if ((overlapsFound == false) && (_noHit == true))
writer.SaveAlignment(bam);
}
hits.clear();
}
// close the relevant BAM files.
reader.Close();
if (_bamOutput == true) {
writer.Close();
}
}
int main (int argc, char** argv)
{
// Print Commandline
string ss(argv[0]); // convert Char to String
string commandline = "##Print Command line " + ss;
int c;
FastaReference* reference = NULL;
int minbaseQ = 10; //default
int windowlen = 40; //by default
string regionstr;
string RegionFile;
string bamfile;
bool STdin = false;
bool has_region = false;
bool has_regionFile = false;
bool has_bamfile = false;
bool has_ref = false;
int ploidy = 2;
bool SetLowComplexityRegionSWGapExt = false;
bool SetLowComplexityRegion = false;
if (argc < 2)
{
printSummary(argv);
exit(1);
}
while (true)
{
static struct option long_options[] =
{
{"help", no_argument, 0, 'h'},
{"ploidy", required_argument, 0, 'p'},
{"window-size", required_argument, 0, 'w'},
{"reference", required_argument, 0, 'f'},
{"min-base-quality", required_argument, 0,'q'},
{"Region", required_argument, 0, 'R'},
{"STdin", no_argument, 0, 's'},
{"bam", required_argument, 0, 'b'},
{"Repeat-Extgap", no_argument, 0, 'E'},
{"LowCompex", no_argument, 0, 'l'},
{0, 0, 0, 0}
};
int option_index = 0;
c = getopt_long (argc, argv, "hslEf:q:w:s:r:R:p:b:", long_options, &option_index);
/* Detect the end of the options. */
if (c == -1)
break;
switch (c)
{
case 'f':
reference = new FastaReference(optarg); // will exit on open failure
commandline = commandline + " -f " + optarg;
has_ref = true;
break;
case 'b':
has_bamfile = true;
bamfile = optarg;
commandline = commandline + " -b " + optarg;
break;
case 'r':
regionstr = optarg;
has_region = true;
commandline = commandline + " -r " + optarg;
break;
case 'R':
RegionFile = optarg;
has_regionFile = true;
commandline = commandline + " -R " + optarg;
break;
case 's':
STdin = true;
commandline = commandline + " -s ";
break;
case 'q':
minbaseQ = atoi(optarg);
commandline = commandline + " -q " + optarg;
break;
case 'w':
windowlen = atoi(optarg);
commandline = commandline + " -w " + optarg;
break;
case 'p':
ploidy = atoi(optarg);
commandline = commandline + " -p " + optarg;
break;
case 'E':
SetLowComplexityRegionSWGapExt = true;
commandline = commandline + " -E ";
break;
case 'l':
SetLowComplexityRegion = true;
commandline = commandline + " -l ";
break;
case 'h':
printSummary(argv);
commandline = commandline + " -h ";
exit(0);
break;
case '?':
printSummary(argv);
exit(1);
break;
default:
abort();
break;
}
}
//// Open Error log files
ofstream cerrlog("bonsaiReport.txt");
streambuf *cerrsave = std::cerr.rdbuf();
// Redirect stream buffers
if (cerrlog.is_open())
cerr.rdbuf(cerrlog.rdbuf());
cerr << commandline << endl;
//Check for Reference Fasta sequence
if (!has_ref)
{
cerr << "no FASTA reference provided, cannot realign" << endl;
exit(1);
}
////Check for reader
BamReader reader;
if (STdin == true)
{
if (!reader.Open("stdin"))
{
cerr << "could not open stdin bam for reading" << endl;
cerr << reader.GetErrorString() << endl;
reader.Close();
printSummary(argv);
}
}
else
{
if (has_bamfile == true)
{
if (!reader.Open(bamfile))
{
cerr << "ERROR: could not open bam files from stdin ... Aborting" << endl;
cerr << reader.GetErrorString() << endl;
reader.Close();
printSummary(argv);
}
if ( !reader.LocateIndex() )
reader.CreateIndex();
}
else
{
cerr << "--bam flag is set but no bamfile is provided... Aborting" << endl;
reader.Close();
printSummary(argv);
}
}
//// Check Region Tags
if ( (has_regionFile == true) && (has_region == true) )
{
cerr << "ERROR: You provide both region and has provide a Set Region List... Aborting" << endl;
exit(1);
}
//// store the names of all the reference sequences in the BAM file
vector<RefData> referencedata = reader.GetReferenceData();
//// Store Region LIST
vector<BamRegion> regionlist;
if (has_region == true)
{
BamRegion region;
ParseRegionString(regionstr, reader, region);
regionlist.push_back(region);
}
else if (has_regionFile == true)
{
ifstream RG(RegionFile.c_str(), ios_base::in);
string line;
while(getline(RG,line))
{
BamRegion region;
ParseRegionString(line, reader, region);
regionlist.push_back(region);
}
RG.close();
}
else if ( (has_regionFile == false) && (has_region == false) )
{
for (int i= 0; i < (int)referencedata.size(); i++)
{
string regionstr = referencedata.at(i).RefName;
BamRegion region;
ParseRegionString(regionstr, reader, region);
if (!reader.SetRegion(region)) // Bam region will get [0,101) = 0 to 100 => [closed, half-opened)
{
cerr << "ERROR: set region " << regionstr << " failed. Check that REGION describes a valid range... Aborting" << endl;
reader.Close();
exit(1);
}
else
regionlist.push_back(region);
}
}
////
BamWriter writer;
if (!writer.Open("stdout", reader.GetHeaderText(), reader.GetReferenceData()))
{
cerr << "could not open stdout for writing" << endl;
exit(1);
}
//// Smallest start position and Largest end position for Req Seq
vector<RefData>::iterator refdataIter = referencedata.begin();
vector<BamRegion>::iterator regionListIter = regionlist.begin();
// CLASS
RealignFunctionsClass RealignFunction;
map<int, string> RefIDRedName;
vector<SalRealignInfo> AlGroups;
multimap<int, BamAlignment> SortRealignedAlignmentsMultimap;
int refid = 0;
BamAlignment alignment;
bool IsNextAlignment = reader.GetNextAlignment(alignment);
//cerr << " " << alignment.Name << " Chr " << alignment.RefID << " Startpos: " << alignment.Position << " Endpos: " << alignment.GetEndPosition() << " Length: " << alignment.Length << endl;
int windowrealigned = 0;
int TotalWindowDetected = 0;
int TotalReadsAligned = 0;
int TotalWindow = 0;
int TotalReads = 0;
while (refdataIter != referencedata.end() )
{
string refname = refdataIter->RefName;
RefIDRedName[refid] = refname;
int reflength = refdataIter->RefLength;
int winstartpos, winendpos;
int AllowableBasesInWindow = 1;
bool nextChrName = false;
cerr << "##HeaderINFO: RefID = " << refdataIter->RefName << "\t" << "RefLen = " << reflength << endl;
while (nextChrName == false )
{
vector<int> minmaxRefSeqPos;
bool IsPassDetectorNoRealignment = false;
minmaxRefSeqPos.push_back(-1);
minmaxRefSeqPos.push_back(0);
//cerr << " region: " << (*regionListIter).LeftRefID << " : " << (*regionListIter).LeftPosition << " .. " << (*regionListIter).RightPosition << endl;
if ((refid == (int)referencedata.size() - 1) && ((*regionListIter).LeftRefID == refid) && ((has_region==true) || (has_regionFile==true)) )
{
////
if ( (has_region == true) || (has_regionFile == true) )
{
winstartpos = (*regionListIter).LeftPosition;
winendpos = winstartpos + windowlen - 1;
reflength = (*regionListIter).RightPosition;
if (reflength < winendpos)
reflength = winendpos;
// Get Next Alignment First
if ( (refid == alignment.RefID) && (winstartpos == (*regionListIter).LeftPosition) && (IsNextAlignment == false) )
IsNextAlignment = reader.GetNextAlignment(alignment);
}
else if (has_region == false)
{
winstartpos = 0;
winendpos = winstartpos + windowlen - 1;
// Get Next Alignment First
if ( (refid == alignment.RefID) && (winstartpos == 0) && (IsNextAlignment == false) )
IsNextAlignment = reader.GetNextAlignment(alignment);
}
//cerr << " winstart: " << winstartpos << " ; winend: " << winendpos;
//cerr << " " << alignment.Name << " Chr " << alignment.RefID << " Startpos: " << alignment.Position << " Endpos: " << alignment.GetEndPosition() << " Length: " << alignment.Length << endl;
////
while ((winstartpos < reflength))
{
//// Check window end position
if (winendpos > reflength)
winendpos = reflength;
// Reinitialized
unsigned int NewReadMappedcount = 0;
//// Save and Erase alignments that are outside of window (Deque?)
if (!AlGroups.empty())
{
minmaxRefSeqPos.at(0) = -1;
minmaxRefSeqPos.at(1) = 0;
//cerr << "#Start: Keep alignments with start position exceed the right end of the window/Region " << endl;
vector<SalRealignInfo>::iterator Iter = AlGroups.begin();
while (Iter != AlGroups.end())
{
// Erase alignment s
if ((*Iter).al.GetEndPosition() < winstartpos)
{
//cerr << " ToWrite: " << (*Iter).second.size() << " ; " << (*Iter).al.Name << " ; " << (*Iter).al.Position << " < " << winstartpos << " : " << (*Iter).al.GetEndPosition() << endl;
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > ((*Iter).al.Position, (*Iter).al));
AlGroups.erase(Iter);
//cerr << " ToWrite: DONE " << endl;
}
else
{
string referenceSequence = reference->getSubSequence(RefIDRedName[(*Iter).al.RefID], (*Iter).al.Position, 2*(*Iter).al.Length);
if ((*Iter).HasRealign == true )
{
(*Iter).currentReadPosition = 0;
(*Iter).currentGenomeSeqPosition = 0;
(*Iter).currentAlPosition = (*Iter).al.Position;
(*Iter).cigarindex = 0;
}
(*Iter).CigarSoftclippingLength = 0;
SalRealignInfo talr = (*Iter);
//cerr << " ToKEEP: " << (*Iter).al.Name << " ; " << (*Iter).al.Position << " < " << winstartpos << " : " << (*Iter).al.GetEndPosition() << endl;
RealignFunction.ParseAlignmentsAndExtractVariantsByBaseQualv(AlGroups, talr, Iter, (*Iter).al, referenceSequence, minmaxRefSeqPos, winstartpos, winendpos, (float) minbaseQ, false);
++Iter; //Increment iterator
}
}
}
// Write Sorted Alignments that are outside of window
//cerr << "SortRealignedAlignmentsMultimap: " << SortRealignedAlignmentsMultimap.size() << " minmaxRefSeqPos.at(0)= " << minmaxRefSeqPos.at(0) << endl;
if (!SortRealignedAlignmentsMultimap.empty()) // && (winWrite < winstartpos ) )
{
//cerr << "#Start: Write alignments and delete alignments with start position exceed the right end of the window/Region " << endl;
multimap<int, BamAlignment>::iterator sraIter = SortRealignedAlignmentsMultimap.begin();
while (sraIter != SortRealignedAlignmentsMultimap.end())
{
//cerr << " (*sraIter).first= " << (*sraIter).first << " minmaxRefSeqPos.at(0)= " << minmaxRefSeqPos.at(0) << " winstartpos - ((windowlen - 1)*0.9)= " << winstartpos - ((windowlen - 1)*0.9) << endl;
if (((float) (*sraIter).first < floor((float) (winstartpos - ((windowlen - 1)*0.9)))) && ((minmaxRefSeqPos.at(0) > 0) && ((*sraIter).first < minmaxRefSeqPos.at(0)))) {
//writer.SaveAlignment((*sraIter).second); // Why sometimes, it doesn't work ?????
if (!writer.SaveAlignment((*sraIter).second))
cerr << writer.GetErrorString() << endl;
SortRealignedAlignmentsMultimap.erase(sraIter++);
} else {
++sraIter;
}
}
//cerr << "#Done: Write alignments and delete alignments with start position exceed the right end of the window/Region " << endl;
}
//cerr << " winstart: " << winstartpos << " ; winend: " << winendpos;
//cerr << " " << alignment.Name << " Chr " << alignment.RefID << " Startpos: " << alignment.Position << " Endpos: " << alignment.GetEndPosition() << " Length: " << alignment.Length << endl;
//cerr << ": " << alignment.RefID << " :" << RefIDRedName[alignment.RefID] << " : " << RefIDRedName[alignment.RefID] << endl;
//cerr << "Start: Gather aligmenets that lie (fully or partially) within the window frame and group INDELs if there are ... " << endl;
// Gather Reads within a window frame
while ((IsNextAlignment) && (refid == alignment.RefID)) // Neeed more conditions
{
if (SetLowComplexityRegion == true)
{
string sequenceInWindow = reference->getSubSequence(RefIDRedName[alignment.RefID], winstartpos, (winendpos-winstartpos+1) );
if (IsWindowInRepeatRegion(sequenceInWindow) == true)
{
if ((IsWithinWindow(alignment, winstartpos, winendpos, AllowableBasesInWindow)) == 0)
{
TotalReads++;
if (alignment.IsMapped())
{
string referenceSequence = reference->getSubSequence(RefIDRedName[alignment.RefID], alignment.Position, 2*alignment.Length);
vector<SalRealignInfo>::iterator tIter;
SalRealignInfo alr;
alr.al = alignment;
alr.currentReadPosition = 0;
alr.currentGenomeSeqPosition = 0;
alr.currentAlPosition = alignment.Position;
alr.cigarindex = 0;
alr.HasRealign = false;
alr.CigarSoftclippingLength = 0;
string str = "ZZZZZZZZZZZZZZZZZ";
if (alignment.Name.find(str) != string::npos) {
stringstream cigar;
for (vector<CigarOp>::const_iterator cigarIter = alignment.CigarData.begin(); cigarIter != alignment.CigarData.end(); ++cigarIter)
cigar << cigarIter->Length << cigarIter->Type;
string cigarstr = cigar.str();
cerr << " TRACKING: " << alignment.RefID << " " << alignment.Name << " pos: " << alignment.Position << " cigar: " << cigarstr << endl;
}
RealignFunction.ParseAlignmentsAndExtractVariantsByBaseQualv(AlGroups, alr, tIter, alignment, referenceSequence, minmaxRefSeqPos, winstartpos, winendpos, (float) minbaseQ, true);
NewReadMappedcount++;
}
else
{
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > (alignment.Position, alignment));
cerr << "UNmapped : " << alignment.Name << endl;
}
}
else if ((IsWithinWindow(alignment, winstartpos, winendpos, AllowableBasesInWindow)) == 1)
{
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > (alignment.Position, alignment));
}
else
break;
} else {
if ((IsWithinWindow(alignment, winstartpos, winendpos, AllowableBasesInWindow)) < 2)
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > (alignment.Position, alignment));
else
break;
}
}
else // (SetLowComplexityRegion == false)
{
if ((IsWithinWindow(alignment, winstartpos, winendpos, AllowableBasesInWindow)) == 0)
{
TotalReads++;
if (alignment.IsMapped())
{
string referenceSequence = reference->getSubSequence(RefIDRedName[alignment.RefID], alignment.Position, 2 * alignment.Length);
vector<SalRealignInfo>::iterator tIter;
SalRealignInfo alr;
alr.al = alignment;
alr.currentReadPosition = 0;
alr.currentGenomeSeqPosition = 0;
alr.currentAlPosition = alignment.Position;
alr.cigarindex = 0;
alr.HasRealign = false;
alr.CigarSoftclippingLength = 0;
string str = "ZZZZZZZZZZZZZZZZZ";
if (alignment.Name.find(str) != string::npos)
{
stringstream cigar;
for (vector<CigarOp>::const_iterator cigarIter = alignment.CigarData.begin(); cigarIter != alignment.CigarData.end(); ++cigarIter)
cigar << cigarIter->Length << cigarIter->Type;
string cigarstr = cigar.str();
cerr << " TRACKING: " << alignment.RefID << " " << alignment.Name << " pos: " << alignment.Position << " cigar: " << cigarstr << endl;
}
RealignFunction.ParseAlignmentsAndExtractVariantsByBaseQualv(AlGroups, alr, tIter, alignment, referenceSequence, minmaxRefSeqPos, winstartpos, winendpos, (float) minbaseQ, true);
//cerr << " winstart: " << winstartpos << " ; winend: " << winendpos;
//cerr << " INDEL: " << alignment.RefID << " " << alignment.Name << " pos: " << alignment.Position << " Length: " << alignment.Length << " CIGARstr: " << cigarstr << endl;
NewReadMappedcount++;
}
else
{
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > (alignment.Position, alignment));
cerr << "UNmapped : " << alignment.Name << endl;
}
}
else if ((IsWithinWindow(alignment, winstartpos, winendpos, AllowableBasesInWindow)) == 1) {
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > (alignment.Position, alignment));
}
else
break;
}
////Get next alignment
IsNextAlignment = reader.GetNextAlignment(alignment);
}
//cerr << "Done: Gather aligmenets that lie (fully or partially) within the window frame and group INDELs if there are ... " << endl;
//// Detector Corner
bool ToRealign = MeetIndelDetectorThresholdv(AlGroups);
cerr << "MeetIndelDetectorThresholdv(AlGroups).size()= " << AlGroups.size() << endl;
// **************
if (ToRealign)
{
//cerr << " ToRealign: " << refdataIter->RefName << "\t" << reflength << "\t" << winstartpos << "\t" << winendpos << "\t" << AlGroups.size() << endl;
//cerr << " minmaxRefSeqPos.at(1)= " << minmaxRefSeqPos.at(1) << " minmaxRefSeqPos.at(0)= " << minmaxRefSeqPos.at(0) << endl;
////// Perform Realign routines
int TotalAlR = 0; // Total number of alignments to be realigned
int NumAlR = 0; // Now many alignments are aligned
TotalWindowDetected++;
cerr << "#Start: Meet Threshold, Realigning ... " << endl;
if (minmaxRefSeqPos.at(1) < winendpos)
minmaxRefSeqPos.at(1) = winendpos;
if (minmaxRefSeqPos.at(0) > winstartpos)
minmaxRefSeqPos.at(0) = winstartpos;
bool IsToRealign = RealignFunction.PruningByNaiveSelectionProcedureAndConstructHaplotypes2(winstartpos, winendpos, refid, refname, minmaxRefSeqPos, reference);
if (IsToRealign == true)
{
RealignFunction.SelectHaplotypeCandidates_SmithWatermanBSv(AlGroups, minmaxRefSeqPos, SetLowComplexityRegionSWGapExt);
minmaxRefSeqPos.at(0) = -1;
minmaxRefSeqPos.at(1) = 0;
int nextwinstartpos = winendpos + 1;
int nextwinendpos = winstartpos + windowlen - 1;
if (nextwinendpos > reflength)
nextwinendpos = reflength;
//cerr << " Before Realign : " << SortRealignedAlignmentsMultimap.size() << endl;
RealignFunction.AdjustCigarsWRTChosenMultipleHaplotypesAndPrepareAlignmentsTobeWrittenOut(AlGroups, SortRealignedAlignmentsMultimap, reference, RefIDRedName, minmaxRefSeqPos, nextwinstartpos, nextwinendpos, minbaseQ, TotalAlR, NumAlR, ploidy);
IsPassDetectorNoRealignment = false; // Set flag to false to deactivate write functions
//cerr << " After Realign : " << SortRealignedAlignmentsMultimap.size() << endl;
TotalReadsAligned += NumAlR;
if (NumAlR > 0) // Realignment done
windowrealigned++;
} else
cerr << "#Done: Meet Threshold, Realigning ... " << endl;
}
if (NewReadMappedcount > 0)
TotalWindow++;
RealignFunction.Clear();
//// Move the window frame
winstartpos = winendpos + 1;
winendpos = winstartpos + windowlen - 1;
}
//// Save and Erase remaining alignments that are outside of window (Deque?)
if ((!AlGroups.empty())) {
cerr << "#Start: Write Remaining alignments and delete all alignments" << endl;
for (vector<SalRealignInfo>::iterator Iter = AlGroups.begin(); Iter != AlGroups.end(); ++Iter) {
//cerr << " Remain alignment start: " << (*Iter).al.Name << " " << Iter->al.Position << " < " << winstartpos << " " << winendpos << endl;
SortRealignedAlignmentsMultimap.insert(pair<int, BamAlignment > ((*Iter).al.Position, (*Iter).al));
}
cerr << "#Done: Write Remaining alignments and delete all alignments" << endl;
}
AlGroups.clear();
// Write Sorted remaining Alignments that are outside of window
if (!SortRealignedAlignmentsMultimap.empty())
{
for (multimap<int, BamAlignment>::iterator sraIter = SortRealignedAlignmentsMultimap.begin(); sraIter != SortRealignedAlignmentsMultimap.end(); ++sraIter)
{
//writer.SaveAlignment((*sraIter).second);
if (!writer.SaveAlignment((*sraIter).second))
cerr << writer.GetErrorString() << endl;
}
SortRealignedAlignmentsMultimap.clear();
}
}
++regionListIter;
if ((*regionListIter).LeftRefID > refid)
nextChrName = true;
}
//// If End of the chromosome position
//// increament iterator
++refdataIter;
++refid;
}
reader.Close();
writer.Close();
cerr << "##-Completed- " << endl;
cerr << " Total Reads processed = " << TotalReads << endl;
cerr << " Total Reads Aligned = " << TotalReadsAligned << endl;
cerr << " Total Window processed = " << TotalWindow << endl;
cerr << " Total Window Detected = " << TotalWindowDetected << endl;
cerr << " Total Windows Aligned = " << windowrealigned << endl;
// Restore cerr's stream buffer before terminating
if (cerrlog.is_open())
cerr.rdbuf(cerrsave);
commandline.clear();
return 0;
}
// generates mutiple sorted temp BAM files from single unsorted BAM file
bool SortTool::SortToolPrivate::GenerateSortedRuns(void) {
// open input BAM file
BamReader reader;
if ( !reader.Open(m_settings->InputBamFilename) ) {
cerr << "bamtools sort ERROR: could not open " << m_settings->InputBamFilename
<< " for reading... Aborting." << endl;
return false;
}
// get basic data that will be shared by all temp/output files
SamHeader header = reader.GetHeader();
header.SortOrder = ( m_settings->IsSortingByName
? Constants::SAM_HD_SORTORDER_QUERYNAME
: Constants::SAM_HD_SORTORDER_COORDINATE );
m_headerText = header.ToString();
m_references = reader.GetReferenceData();
// set up alignments buffer
BamAlignment al;
vector<BamAlignment> buffer;
buffer.reserve( (size_t)(m_settings->MaxBufferCount*1.1) );
bool bufferFull = false;
// if sorting by name, we need to generate full char data
// so can't use GetNextAlignmentCore()
if ( m_settings->IsSortingByName ) {
// iterate through file
while ( reader.GetNextAlignment(al)) {
// check buffer's usage
bufferFull = ( buffer.size() >= m_settings->MaxBufferCount );
// store alignments until buffer is "full"
if ( !bufferFull )
buffer.push_back(al);
// if buffer is "full"
else {
// push any unmapped reads into buffer,
// don't want to split these into a separate temp file
if ( !al.IsMapped() )
buffer.push_back(al);
// "al" is mapped, so create a sorted temp file with current buffer contents
// then push "al" into fresh buffer
else {
CreateSortedTempFile(buffer);
buffer.push_back(al);
}
}
}
}
// sorting by position, can take advantage of GNACore() speedup
else {
// iterate through file
while ( reader.GetNextAlignmentCore(al) ) {
// check buffer's usage
bufferFull = ( buffer.size() >= m_settings->MaxBufferCount );
// store alignments until buffer is "full"
if ( !bufferFull )
buffer.push_back(al);
// if buffer is "full"
else {
// push any unmapped reads into buffer,
// don't want to split these into a separate temp file
if ( !al.IsMapped() )
buffer.push_back(al);
// "al" is mapped, so create a sorted temp file with current buffer contents
// then push "al" into fresh buffer
else {
CreateSortedTempFile(buffer);
buffer.push_back(al);
}
}
}
}
// handle any leftover buffer contents
if ( !buffer.empty() )
CreateSortedTempFile(buffer);
// close reader & return success
reader.Close();
return true;
}
//{{{bool sort_inter_chrom_bam(string in_file_name,
bool sort_inter_chrom_bam(string in_file_name,
string out_file_name)
{
// open input BAM file
BamReader reader;
if ( !reader.Open(in_file_name) ) {
cerr << "sort ERROR: could not open " <<
in_file_name << " for reading... Aborting." << endl;
return false;
}
SamHeader header = reader.GetHeader();
if ( !header.HasVersion() )
header.Version = Constants::SAM_CURRENT_VERSION;
string header_text = header.ToString();
RefVector ref = reader.GetReferenceData();
// set up alignments buffer
BamAlignment al;
vector<BamAlignment> buffer;
buffer.reserve( (size_t)(SORT_DEFAULT_MAX_BUFFER_COUNT*1.1) );
bool bufferFull = false;
int buff_count = 0;
// iterate through file
while ( reader.GetNextAlignment(al)) {
// check buffer's usage
bufferFull = ( buffer.size() >= SORT_DEFAULT_MAX_BUFFER_COUNT );
// store alignments until buffer is "full"
if ( !bufferFull )
buffer.push_back(al);
// if buffer is "full"
else {
// so create a sorted temp file with current buffer contents
// then push "al" into fresh buffer
create_sorted_temp_file(buffer,
out_file_name,
buff_count,
header_text,
ref);
++buff_count;
buffer.push_back(al);
}
}
// handle any leftover buffer contents
if ( !buffer.empty() ) {
create_sorted_temp_file(buffer,
out_file_name,
buff_count,
header_text,
ref);
++buff_count;
}
reader.Close();
return merge_sorted_files(out_file_name, buff_count, header_text, ref);
/*
for (int i = 0; i < buff_count; ++i) {
stringstream temp_name;
temp_name << out_file_name << i;
}
*/
}
int main (int argc, char *argv[]) {
bool produceUnCompressedBAM=false;
bool verbose=false;
bool ancientDNA=false;
bool keepOrig=false;
string adapter_F=options_adapter_F_BAM;
string adapter_S=options_adapter_S_BAM;
string adapter_chimera=options_adapter_chimera_BAM;
string key="";
bool allowMissing=false;
int trimCutoff=1;
bool allowAligned=false;
bool printLog=false;
string logFileName;
BamReader reader;
BamWriter writer;
string bamFile;
string bamFileOUT="";
string key1;
string key2;
bool useDist=false;
double location=-1.0;
double scale =-1.0;
bool fastqFormat=false;
string fastqfile1 = "";
string fastqfile2 = "";
string fastqoutfile = "";
bool singleEndModeFQ=true;
const string usage=string(string(argv[0])+
" [options] BAMfile"+"\n"+
"\nThis program takes an unaligned BAM where mates are consecutive\nor fastq files and trims and merges reads\n"+
"\n\tYou can specify a unaligned bam file or one or two fastq :\n"+
"\t\t"+"-fq1" +"\t\t"+"First fastq"+"\n"+
"\t\t"+"-fq2" +"\t\t"+"Second fastq file (for paired-end)"+"\n"+
"\t\t"+"-fqo" +"\t\t"+"Output fastq prefix"+"\n\n"+
//"\t"+"-p , --PIPE"+"\n\t\t"+"Read BAM from and write it to PIPE"+"\n"+
"\t"+"-o , --outfile" +"\t\t"+"Output (BAM format)."+"\n"+
"\t"+"-u " +"\t\t"+"Produce uncompressed bam (good for pipe)"+"\n"+
// "\t"+" , --outprefix" +"\n\t\t"+"Prefix for output files (default '"+outprefix+"')."+"\n"+
//"\t"+" , --SAM" +"\n\t\t"+"Output SAM not BAM."+"\n"+
"\t"+"--aligned" +"\t\t"+"Allow reads to be aligned (default "+boolStringify(allowAligned)+")"+"\n"+
"\t"+"-v , --verbose" +"\t\t"+"Turn all messages on (default "+boolStringify(verbose)+")"+"\n"+
"\t"+"--log [log file]" +"\t"+"Print a tally of merged reads to this log file (default only to stderr)"+"\n"+
"\n\t"+"Paired End merging/Single Read trimming options"+"\n"+
"\t\t"+"You can specify either:"+"\n"+
"\t\t\t"+"--ancientdna"+"\t\t\t"+"ancient DNA (default "+boolStringify(ancientDNA)+")"+"\n"+
"\t\t"+" "+"\t\t\t\t"+"this allows for partial overlap"+"\n"+
"\n\t\t"+"or if you know your size length distribution:"+"\n"+
"\t\t\t"+"--loc"+"\t\t\t\t"+"Location for lognormal dist. (default none)"+"\n"+
"\t\t\t"+"--scale"+"\t\t\t\t"+"Scale for lognormal dist. (default none)"+"\n"+
// "\t\t\t\t\t\t\tGood for merging ancient DNA reads into a single sequence\n\n"
"\n\t\t"+"--keepOrig"+"\t\t\t\t"+"Write original reads if they are trimmed or merged (default "+boolStringify(keepOrig)+")"+"\n"+
"\t\t\t\t\t\t\tSuch reads will be marked as PCR duplicates\n\n"
"\t\t"+"-f , --adapterFirstRead" +"\t\t\t"+"Adapter that is observed after the forward read (def. Multiplex: "+options_adapter_F_BAM.substr(0,30)+")"+"\n"+
"\t\t"+"-s , --adapterSecondRead" +"\t\t"+"Adapter that is observed after the reverse read (def. Multiplex: "+options_adapter_S_BAM.substr(0,30)+")"+"\n"+
"\t\t"+"-c , --FirstReadChimeraFilter" +"\t\t"+"If the forward read looks like this sequence, the cluster is filtered out.\n\t\t\t\t\t\t\tProvide several sequences separated by comma (def. Multiplex: "+options_adapter_chimera_BAM.substr(0,30)+")"+"\n"+
"\t\t"+"-k , --key"+"\t\t\t\t"+"Key sequence with which each sequence starts. Comma separate for forward and reverse reads. (default '"+key+"')"+"\n"+
"\t\t"+"-i , --allowMissing"+"\t\t\t"+"Allow one base in one key to be missing or wrong. (default "+boolStringify(allowMissing)+")"+"\n"+
"\t\t"+"-t , --trimCutoff"+"\t\t\t"+"Lowest number of adapter bases to be observed for single Read trimming (default "+stringify(trimCutoff)+")");
if( (argc== 1) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cout<<"Usage:"<<endl;
cout<<""<<endl;
cout<<usage<<endl;
return 1;
}
for(int i=1;i<(argc-1);i++){ //all but the last arg
if(strcmp(argv[i],"-fq1") == 0 ){
fastqfile1=string(argv[i+1]);
fastqFormat=true;
i++;
continue;
}
if(strcmp(argv[i],"-fq2") == 0 ){
fastqfile2=string(argv[i+1]);
fastqFormat=true;
singleEndModeFQ=false;
i++;
continue;
}
if(strcmp(argv[i],"-fqo") == 0 ){
fastqoutfile=string(argv[i+1]);
fastqFormat=true;
i++;
continue;
}
if(strcmp(argv[i],"--log") == 0 ){
logFileName =string(argv[i+1]);
printLog=true;
i++;
continue;
}
if(strcmp(argv[i],"-p") == 0 || strcmp(argv[i],"--PIPE") == 0 ){
cerr<<"This version no longer works with pipe, exiting"<<endl;
return 1;
}
if(strcmp(argv[i],"-u") == 0 ){
produceUnCompressedBAM=true;
continue;
}
if(strcmp(argv[i],"--aligned") == 0 ){
allowAligned=true;
continue;
}
if(strcmp(argv[i],"-o") == 0 || strcmp(argv[i],"--outfile") == 0 ){
bamFileOUT =string(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-v") == 0 || strcmp(argv[i],"--verbose") == 0 ){
verbose=true;
continue;
}
if(strcmp(argv[i],"--ancientdna") == 0 ){
ancientDNA=true;
continue;
}
if(strcmp(argv[i],"--keepOrig") == 0 ){
keepOrig=true;
continue;
}
if(strcmp(argv[i],"--loc") == 0 ){
location =destringify<double>(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"--scale") == 0 ){
scale =destringify<double>(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-f") == 0 || strcmp(argv[i],"--adapterFirstRead") == 0 ){
adapter_F =string(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-s") == 0 || strcmp(argv[i],"--adapterSecondRead") == 0 ){
adapter_S =string(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-c") == 0 || strcmp(argv[i],"--FirstReadChimeraFilter") == 0 ){
adapter_chimera =string(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-k") == 0 || strcmp(argv[i],"--keys") == 0 ){
key =string(argv[i+1]);
i++;
continue;
}
if(strcmp(argv[i],"-i") == 0 || strcmp(argv[i],"--allowMissing") == 0 ){
allowMissing=true;
continue;
}
if(strcmp(argv[i],"-t") == 0 || strcmp(argv[i],"--trimCutoff") == 0 ){
trimCutoff=atoi(argv[i+1]);
i++;
continue;
}
cerr<<"Unknown option "<<argv[i] <<" exiting"<<endl;
return 1;
}
bamFile=argv[argc-1];
if( (location != -1.0 && scale == -1.0) ||
(location == -1.0 && scale != -1.0) ){
cerr<<"Cannot specify --location without specifying --scale"<<endl;
return 1;
}
if( (location != -1.0 && scale != -1.0) ){
useDist=true;
if(ancientDNA){
cerr<<"Cannot specify --location/--scale and --ancientDNA"<<endl;
return 1;
}
}
MergeTrimReads mtr (adapter_F,adapter_S,adapter_chimera,
key1,key2,
trimCutoff,allowMissing,ancientDNA,location,scale,useDist);
fqwriters onereadgroup;
if(fastqFormat){
if( bamFileOUT != "" || produceUnCompressedBAM || allowAligned){
cerr<<"ERROR : Cannot specify options like -o, -u or --allowAligned for fastq"<<endl;
return 1;
}
if(fastqfile1 == ""){
cerr<<"ERROR : Must specify as least the first file for fastq"<<endl;
return 1;
}
FastQParser * fqp1;
FastQParser * fqp2;
if(singleEndModeFQ){
fqp1 = new FastQParser (fastqfile1);
string outdirs = fastqoutfile+".fq.gz";
string outdirsf = fastqoutfile+".fail.fq.gz";
onereadgroup.single.open(outdirs.c_str(), ios::out);
onereadgroup.singlef.open(outdirsf.c_str(), ios::out);
if(!onereadgroup.single.good()){ cerr<<"Cannot write to file "<<outdirs<<endl; return 1; }
if(!onereadgroup.singlef.good()){ cerr<<"Cannot write to file "<<outdirsf<<endl; return 1; }
}else{
fqp1 = new FastQParser (fastqfile1);
fqp2 = new FastQParser (fastqfile2);
string outdirs = fastqoutfile+".fq.gz";
string outdir1 = fastqoutfile+"_r1.fq.gz";
string outdir2 = fastqoutfile+"_r2.fq.gz";
string outdirsf = fastqoutfile+".fail.fq.gz";
string outdir1f = fastqoutfile+"_r1.fail.fq.gz";
string outdir2f = fastqoutfile+"_r2.fail.fq.gz";
onereadgroup.single.open(outdirs.c_str(), ios::out);
onereadgroup.pairr1.open(outdir1.c_str(), ios::out);
onereadgroup.pairr2.open(outdir2.c_str(), ios::out);
onereadgroup.singlef.open(outdirsf.c_str(), ios::out);
onereadgroup.pairr1f.open(outdir1f.c_str(), ios::out);
onereadgroup.pairr2f.open(outdir2f.c_str(), ios::out);
if(!onereadgroup.single.good()){ cerr<<"Cannot write to file "<<outdirs<<endl; return 1; }
if(!onereadgroup.pairr1.good()){ cerr<<"Cannot write to file "<<outdir1<<endl; return 1; }
if(!onereadgroup.pairr2.good()){ cerr<<"Cannot write to file "<<outdir2<<endl; return 1; }
if(!onereadgroup.singlef.good()){ cerr<<"Cannot write to file "<<outdirsf<<endl; return 1; }
if(!onereadgroup.pairr1f.good()){ cerr<<"Cannot write to file "<<outdir1f<<endl; return 1; }
if(!onereadgroup.pairr2f.good()){ cerr<<"Cannot write to file "<<outdir2f<<endl; return 1; }
}
unsigned int totalSeqs=0;
while(fqp1->hasData()){
FastQObj * fo1=fqp1->getData();
vector<string> def1=allTokens( *(fo1->getID()), ' ' );
string def1s=def1[0];
FastQObj * fo2;
string def2s;
string ext2s;
if(!singleEndModeFQ){
if(!fqp2->hasData()){
cerr << "ERROR: Discrepency between fastq files at record " << *(fo1->getID()) <<endl;
return 1;
}
fo2=fqp2->getData();
vector<string> def2=allTokens( *(fo2->getID()), ' ' );
def2s=def2[0];
if(strEndsWith(def1s,"/1")){
def1s=def1s.substr(0,def1s.size()-2);
}
if(strEndsWith(def2s,"/2")){
def2s=def2s.substr(0,def2s.size()-2);
}
if(strBeginsWith(def1s,"@")){
def1s=def1s.substr(1,def1s.size()-1);
}
if(strBeginsWith(def2s,"@")){
def2s=def2s.substr(1,def2s.size()-1);
}
if(def1s != def2s){
cerr << "ERROR: Discrepency between fastq files, different names " << *(fo1->getID()) <<" and "<< *(fo2->getID()) <<endl;
return 1;
}
merged result= mtr.process_PE(*(fo1->getSeq()),*(fo1->getQual()),
*(fo2->getSeq()),*(fo2->getQual()));
mtr.incrementCountall();
if(result.code != ' '){ //keys or chimeras
if(result.code == 'K'){
mtr.incrementCountfkey();
}else{
if(result.code == 'D'){
mtr.incrementCountchimera();
}else{
cerr << "leehom: Wrong return code =\""<<result.code<<"\""<<endl;
exit(1);
}
}
onereadgroup.pairr2f<<"@"<<def2s<<"/2" <<endl <<*(fo2->getSeq())<<endl<<"+"<<endl <<*(fo2->getQual())<<endl;
onereadgroup.pairr1f<<"@"<<def1s<<"/1" <<endl <<*(fo1->getSeq())<<endl<<"+"<<endl <<*(fo1->getQual())<<endl;
continue;
}else{
if(result.sequence != ""){ //new sequence
onereadgroup.single<<"@"<<def1s<<"" <<endl << result.sequence<<endl<<"+"<<endl <<result.quality<<endl;
if( result.sequence.length() > max(fo1->getSeq()->length(),fo2->getSeq()->length()) ){
mtr.incrementCountmergedoverlap();
}else{
mtr.incrementCountmerged();
}
}else{ //keep as is
mtr.incrementCountnothing();
onereadgroup.pairr2<<"@"<<def2s<<"/2" <<endl <<*(fo2->getSeq())<<endl<<"+"<<endl <<*(fo2->getQual())<<endl;
onereadgroup.pairr1<<"@"<<def1s<<"/1" <<endl <<*(fo1->getSeq())<<endl<<"+"<<endl <<*(fo1->getQual())<<endl;
}
}
}else{
merged result=mtr.process_SR(*(fo1->getSeq()),*(fo1->getQual()));
mtr.incrementCountall();
if(result.code != ' '){ //either chimera or missing key
if(result.code == 'K'){
mtr.incrementCountfkey();
}else{
if(result.code == 'D'){
mtr.incrementCountchimera();
}else{
cerr << "leehom: Wrong return code =\""<<result.code<<"\""<<endl;
exit(1);
}
}
onereadgroup.singlef<<""<<*(fo1->getID())<<"" <<endl << *(fo1->getSeq())<<endl<<"+"<<endl <<*(fo1->getQual())<<endl;
continue;
}
if(result.sequence != ""){ //new sequence
mtr.incrementCounttrimmed();
onereadgroup.single<<""<<*(fo1->getID())<<"" <<endl << result.sequence<<endl<<"+"<<endl <<result.quality<<endl;
}else{
mtr.incrementCountnothing();
onereadgroup.single<<""<<*(fo1->getID())<<"" <<endl << *(fo1->getSeq())<<endl<<"+"<<endl <<*(fo1->getQual())<<endl;
}
}
totalSeqs++;
}
delete fqp1;
if(!singleEndModeFQ){
delete fqp2;
}
if(singleEndModeFQ){
onereadgroup.single.close();
onereadgroup.singlef.close();
}else{
onereadgroup.single.close();
onereadgroup.pairr1.close();
onereadgroup.pairr2.close();
onereadgroup.singlef.close();
onereadgroup.pairr1f.close();
onereadgroup.pairr2f.close();
}
//fastq
}else{
//else BAM
// initMerge();
// set_adapter_sequences(adapter_F,
// adapter_S,
// adapter_chimera);
// set_options(trimCutoff,allowMissing,mergeoverlap);
if(key != ""){
size_t found=key.find(",");
if (found == string::npos){ //single end reads
key1=key;
key2="";
} else{ //paired-end
key1=key.substr(0,found);
key2=key.substr(found+1,key.length()-found+1);
}
}
if( bamFileOUT == "" ){
cerr<<"The output must be a be specified, exiting"<<endl;
return 1;
}
if ( !reader.Open(bamFile) ) {
cerr << "Could not open input BAM file "<<bamFile << endl;
return 1;
}
SamHeader header = reader.GetHeader();
string pID = "mergeTrimReadsBAM";
string pName = "mergeTrimReadsBAM";
string pCommandLine = "";
for(int i=0;i<(argc);i++){
pCommandLine += (string(argv[i])+" ");
}
putProgramInHeader(&header,pID,pName,pCommandLine,returnGitHubVersion(string(argv[0]),".."));
const RefVector references = reader.GetReferenceData();
//we will not call bgzip with full compression, good for piping into another program to
//lessen the load on the CPU
if(produceUnCompressedBAM)
writer.SetCompressionMode(BamWriter::Uncompressed);
if ( !writer.Open(bamFileOUT,header,references) ) {
cerr << "Could not open output BAM file "<<bamFileOUT << endl;
return 1;
}
SamHeader sh=reader.GetHeader();
//Up to the user to be sure that a sequence is followed by his mate
// if(!sh.HasSortOrder() ||
// sh.SortOrder != "queryname"){
// cerr << "Bamfile must be sorted by queryname" << endl;
// return 1;
// }
BamAlignment al;
BamAlignment al2;
bool al2Null=true;
while ( reader.GetNextAlignment(al) ) {
if(al.IsMapped() || al.HasTag("NM") || al.HasTag("MD") ){
if(!allowAligned){
cerr << "Reads should not be aligned" << endl;
return 1;
}else{
//should we remove tags ?
}
}
if(al.IsPaired() &&
al2Null ){
al2=al;
al2Null=false;
continue;
}else{
if(al.IsPaired() &&
!al2Null){
bool result = mtr.processPair(al,al2);
if( result ){//was merged
BamAlignment orig;
BamAlignment orig2;
if(keepOrig){
orig2 = al2;
orig = al;
}
writer.SaveAlignment(al);
if(keepOrig){
orig.SetIsDuplicate(true);
orig2.SetIsDuplicate(true);
writer.SaveAlignment(orig2);
writer.SaveAlignment(orig);
}
//the second record is empty
}else{
//keep the sequences as pairs
writer.SaveAlignment(al2);
writer.SaveAlignment(al);
}
//
// SINGLE END
//
}else{
BamAlignment orig;
if(keepOrig){
orig =al;
}
mtr.processSingle(al);
if(keepOrig){
//write duplicate
if(orig.QueryBases.length() != al.QueryBases.length()){
orig.SetIsDuplicate(true);
writer.SaveAlignment(orig);
}
}
writer.SaveAlignment(al);
} //end single end
al2Null=true;
}//second pair
} //while al
reader.Close();
writer.Close();
} //else BAM
cerr <<mtr.reportSingleLine()<<endl;
if(printLog){
ofstream fileLog;
fileLog.open(logFileName.c_str());
if (fileLog.is_open()){
fileLog <<mtr.reportMultipleLines() <<endl;
}else{
cerr << "Unable to print to file "<<logFileName<<endl;
}
fileLog.close();
}
return 0;
}
int main (int argc, char *argv[]) {
// bool mapped =false;
// bool unmapped=false;
int bpToDecrease5=1;
int bpToDecrease3=2;
const string usage=string(string(argv[0])+" [options] input.bam out.bam"+"\n\n"+
"\tThis program takes a BAM file as input and produces\n"+
"\tanother where the putative deaminated bases have\n"+
"\ta base quality score of "+intStringify(baseQualForDeam)+"\n"+
"\tgiven an "+intStringify(offset)+" offset \n"+
"\n"+
"\tOptions:\n"+
"\t\t"+"-n5" +"\t\t\t"+"Decrease the nth bases surrounding the 5' ends (Default:"+stringify(bpToDecrease5)+") "+"\n"+
"\t\t"+"-n3" +"\t\t\t"+"Decrease the nth bases surrounding the 3' ends (Default:"+stringify(bpToDecrease3)+") "+"\n"
);
// "\t"+"-u , --unmapped" +"\n\t\t"+"For an unmapped bam file"+"\n"+
// "\t"+"-m , --mapped" +"\n\t\t"+"For an mapped bam file"+"\n");
if( (argc== 1) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cout<<"Usage:"<<endl;
cout<<usage<<endl;
cout<<""<<endl;
return 1;
}
for(int i=1;i<(argc-2);i++){ //all but the last arg
if( string(argv[i]) == "-n5" ){
bpToDecrease5 = destringify<int>(argv[i+1]);
i++;
continue;
}
if( string(argv[i]) == "-n3" ){
bpToDecrease3 = destringify<int>(argv[i+1]);
i++;
continue;
}
cerr<<"Unknown option "<<argv[i] <<" exiting"<<endl;
return 1;
}
if(argc < 3){
cerr<<"Error: Must specify the input and output BAM files";
return 1;
}
string inbamFile =argv[argc-2];
string outbamFile=argv[argc-1];
if(inbamFile == outbamFile){
cerr<<"Input and output files are the same"<<endl;
return 1;
}
// if(!mapped && !unmapped){
// cerr << "Please specify whether you reads are mapped or unmapped" << endl;
// return 1;
// }
// if(mapped && unmapped){
// cerr << "Please specify either mapped or unmapped but not both" << endl;
// return 1;
// }
BamReader reader;
if ( !reader.Open(inbamFile) ) {
cerr << "Could not open input BAM files." << endl;
return 1;
}
vector<RefData> testRefData=reader.GetReferenceData();
SamHeader header = reader.GetHeader();
string pID = "decrQualDeaminated";
string pName = "decrQualDeaminated";
string pCommandLine = "";
for(int i=0;i<(argc);i++){
pCommandLine += (string(argv[i])+" ");
}
putProgramInHeader(&header,pID,pName,pCommandLine);
const RefVector references = reader.GetReferenceData();
BamWriter writer;
if ( !writer.Open(outbamFile, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
BamAlignment al;
// BamAlignment al2;
// bool al2Null=true;
while ( reader.GetNextAlignment(al) ) {
if(al.IsPaired() ){
if(al.IsFirstMate() ){ //5' end, need to check first base only
if(al.IsReverseStrand()){ //
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads :" <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//5' of first mate reversed
indexToCheck=al.QueryBases.length()-1;
for(int i=0;i<bpToDecrease5;i++){
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=max(indexToCheck-1,0);
}
}else{
int indexToCheck;
//5' of first mate
indexToCheck=0;
for(int i=0;i<bpToDecrease5;i++){ //first base
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=min(indexToCheck+1,int(al.Qualities.size()));
}
}
}else{ //3' end, need to check last two bases only
if( al.IsSecondMate() ){
if(al.IsReverseStrand()){ //
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads :" <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//3' of second mate reversed
indexToCheck=al.QueryBases.length()-1;
for(int i=0;i<bpToDecrease3;i++){
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=max(indexToCheck-1,0);
}
}else{
int indexToCheck;
//3' of second mate forward
indexToCheck=0;
for(int i=0;i<bpToDecrease3;i++){ //first base
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=min(indexToCheck+1,int(al.Qualities.size()));
}
}
}else{
cerr << "Wrong state" << endl;
return 1;
}
}
}//end of paired end
else{//we consider single reads to have been sequenced from 5' to 3'
if(al.IsReverseStrand()){ //need to consider
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads :" <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//5' of single read reversed
indexToCheck=al.QueryBases.length()-1;
for(int i=0;i<bpToDecrease5;i++){
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=max(indexToCheck-1,0);
}
//3' of single read reversed
indexToCheck=0;
for(int i=0;i<bpToDecrease3;i++){ //first base
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=min(indexToCheck+1,int(al.Qualities.size()));
}
}else{
int indexToCheck;
//5' of single read
indexToCheck=0;
for(int i=0;i<bpToDecrease5;i++){ //first base
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=min(indexToCheck+1,int(al.Qualities.size()));
}
//3' of single read
indexToCheck=al.QueryBases.length()-1;
for(int i=0;i<bpToDecrease3;i++){
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
}
indexToCheck=max(indexToCheck-1,0);
}
}
}//end of single end
writer.SaveAlignment(al);
}// while ( reader.GetNextAlignment(al) ) {
reader.Close();
writer.Close();
cerr<<"Program terminated gracefully"<<endl;
return 0;
}
// Same as ParseRegionString() above, but accepts a BamMultiReader
bool ParseRegionString(const string& regionString,
const BamReader& reader,
BamRegion& region)
{
// -------------------------------
// parse region string
// check first for empty string
if ( regionString.empty() )
return false;
//cerr << "ParseRegionString Input: " << regionString << endl;
// non-empty string, look for a colom
size_t foundFirstColon = regionString.find(':');
// store chrom strings, and numeric positions
string startChrom;
string stopChrom;
int startPos;
int stopPos;
// no colon found
// going to use entire contents of requested chromosome
// just store entire region string as startChrom name
// use BamReader methods to check if its valid for current BAM file
if ( foundFirstColon == string::npos ) {
startChrom = regionString;
startPos = 0;
stopChrom = regionString;
stopPos = -1;
}
// colon found, so we at least have some sort of startPos requested
else {
// store start chrom from beginning to first colon
startChrom = regionString.substr(0,foundFirstColon);
// look for ".." after the colon
size_t foundRangeDots = regionString.find("..", foundFirstColon+1);
// no dots found
// so we have a startPos but no range
// store contents before colon as startChrom, after as startPos
if ( foundRangeDots == string::npos )
{
startPos = atoi( regionString.substr(foundFirstColon+1).c_str() );
stopChrom = startChrom;
stopPos = -1;
}
// ".." found, so we have some sort of range selected
else {
// store startPos between first colon and range dots ".."
startPos = atoi( regionString.substr(foundFirstColon+1, foundRangeDots-foundFirstColon-1).c_str() );
// look for second colon
size_t foundSecondColon = regionString.find(':', foundRangeDots+1);
// no second colon found
// so we have a "standard" chrom:start..stop input format (on single chrom)
if ( foundSecondColon == string::npos ) {
stopChrom = startChrom;
stopPos = atoi( regionString.substr(foundRangeDots+2).c_str() );
}
// second colon found
// so we have a range requested across 2 chrom's
else {
stopChrom = regionString.substr(foundRangeDots+2, foundSecondColon-(foundRangeDots+2));
stopPos = atoi( regionString.substr(foundSecondColon+1).c_str() );
}
}
}
// -------------------------------
// validate reference IDs & genomic positions
const RefVector references = reader.GetReferenceData();
// if startRefID not found, return false
int startRefID = reader.GetReferenceID(startChrom);
if ( startRefID == -1 ) return false;
// startPos cannot be greater than or equal to reference length
const RefData& startReference = references.at(startRefID);
if ( startPos >= startReference.RefLength ) return false;
// if stopRefID not found, return false
int stopRefID = reader.GetReferenceID(stopChrom);
if ( stopRefID == -1 ) return false;
// stopPosition cannot be larger than reference length
const RefData& stopReference = references.at(stopRefID);
if ( stopPos > stopReference.RefLength ) return false;
// if no stopPosition specified, set to reference end
if ( stopPos == -1 ) stopPos = stopReference.RefLength;
// -------------------------------
// set up Region struct & return
region.LeftRefID = startRefID;
region.LeftPosition = startPos;
region.RightRefID = stopRefID;;
region.RightPosition = stopPos;
//cerr << "ParseRegionString " << region.LeftRefID << " " << region.LeftPosition << " " << region.RightPosition << endl;
return true;
}
int main (int argc, char *argv[]) {
bool mapped =false;
bool unmapped=false;
const string usage=string(string(argv[0])+" [options] input.bam out.bam"+"\n\n"+
"This program takes a BAM file as input and produces\n"+
"another where the putative deaminated bases have\n"+
"have been cut\n"+
"\n"+
"Options:\n");
// "\t"+"-u , --unmapped" +"\n\t\t"+"For an unmapped bam file"+"\n"+
// "\t"+"-m , --mapped" +"\n\t\t"+"For an mapped bam file"+"\n");
if( (argc== 1) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cout<<"Usage:"<<endl;
cout<<usage<<endl;
cout<<""<<endl;
return 1;
}
// for(int i=1;i<(argc-1);i++){ //all but the last arg
// if(strcmp(argv[i],"-m") == 0 || strcmp(argv[i],"--mapped") == 0 ){
// mapped=true;
// continue;
// }
// if(strcmp(argv[i],"-u") == 0 || strcmp(argv[i],"--unmapped") == 0 ){
// unmapped=true;
// continue;
// }
// cerr<<"Unknown option "<<argv[i] <<" exiting"<<endl;
// return 1;
// }
if(argc != 3){
cerr<<"Error: Must specify the input and output BAM files";
return 1;
}
string inbamFile =argv[argc-2];
string outbamFile=argv[argc-1];
// if(!mapped && !unmapped){
// cerr << "Please specify whether you reads are mapped or unmapped" << endl;
// return 1;
// }
// if(mapped && unmapped){
// cerr << "Please specify either mapped or unmapped but not both" << endl;
// return 1;
// }
BamReader reader;
if ( !reader.Open(inbamFile) ) {
cerr << "Could not open input BAM files." << endl;
return 1;
}
vector<RefData> testRefData=reader.GetReferenceData();
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
BamWriter writer;
if ( !writer.Open(outbamFile, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
BamAlignment al;
// BamAlignment al2;
// bool al2Null=true;
while ( reader.GetNextAlignment(al) ) {
if(al.IsPaired() ){
if(al.IsFirstMate() ){ //5' end, need to check first base only
if(al.IsReverseStrand()){ //
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads: " <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//last
indexToCheck=al.QueryBases.length()-1;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
}
}else{
int indexToCheck;
//first base
indexToCheck=0;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
}
}
}else{ //3' end, need to check last two bases only
if( al.IsSecondMate() ){
if(al.IsReverseStrand()){ //
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads: " <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//second to last
indexToCheck=al.QueryBases.length()-2;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
}else{
//last
indexToCheck=al.QueryBases.length()-1;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
}
}
}else{
int indexToCheck;
//second base
indexToCheck=1;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
}else{
//first base
indexToCheck=0;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
//al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
}
}
}
}else{
cerr << "Wrong state" << endl;
return 1;
}
}
}//end of paired end
else{//we consider single reads to have been sequenced from 5' to 3'
if(al.IsReverseStrand()){ //need to consider
if(!al.IsMapped()){
cerr << "Cannot have reverse complemented unmapped reads: " <<al.Name<< endl;
//return 1;
}
int indexToCheck;
//second base
indexToCheck=1;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"51 "<<al.QueryBases<<endl;
// cout<<"51 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
// cout<<"52 "<<al.QueryBases<<endl;
// cout<<"52 "<<al.Qualities<<endl;
}else{
//first base
indexToCheck=0;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"61 "<<al.QueryBases<<endl;
// cout<<"61 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
// cout<<"62 "<<al.QueryBases<<endl;
// cout<<"62 "<<al.Qualities<<endl;
}
}
//last
indexToCheck=al.QueryBases.length()-1;
if(toupper(al.QueryBases[indexToCheck]) == 'A'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"21 "<<al.QueryBases<<endl;
// cout<<"21 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
// cout<<"22 "<<al.QueryBases<<endl;
// cout<<"22 "<<al.Qualities<<endl;
}
}else{
int indexToCheck;
//first base
indexToCheck=0;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"11 "<<al.QueryBases<<endl;
// cout<<"11 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(indexToCheck+1);
al.Qualities = al.Qualities.substr(indexToCheck+1);
// cout<<"12 "<<al.QueryBases<<endl;
// cout<<"12 "<<al.Qualities<<endl;
}
//second to last
indexToCheck=al.QueryBases.length()-2;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"31 "<<al.QueryBases<<endl;
// cout<<"31 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
// cout<<"32 "<<al.QueryBases<<endl;
// cout<<"32 "<<al.Qualities<<endl;
}else{
//last
indexToCheck=al.QueryBases.length()-1;
if(toupper(al.QueryBases[indexToCheck]) == 'T'){
// al.Qualities[indexToCheck]=char(offset+baseQualForDeam);
// cout<<"41 "<<al.QueryBases<<endl;
// cout<<"41 "<<al.Qualities<<endl;
al.QueryBases = al.QueryBases.substr(0,indexToCheck);
al.Qualities = al.Qualities.substr(0, indexToCheck);
// cout<<"42 "<<al.QueryBases<<endl;
// cout<<"42 "<<al.Qualities<<endl;
}
}
}
}//end of single end
writer.SaveAlignment(al);
}// while ( reader.GetNextAlignment(al) ) {
reader.Close();
writer.Close();
return 0;
}
int main (int argc, char *argv[]) {
int minBaseQuality = 0;
string usage=string(""+string(argv[0])+" [in BAM file] [in VCF file] [chr name] [deam out BAM] [not deam out BAM]"+
"\nThis program divides aligned single end reads into potentially deaminated\n"+
"\nreads and the puts the rest into another bam file if the deaminated positions are not called as the alternative base in the VCF.\n"+
"\nTip: if you do not need one of them, use /dev/null as your output\n"+
"arguments:\n"+
"\t"+"--bq [base qual] : Minimum base quality to flag a deaminated site (Default: "+stringify(minBaseQuality)+")\n"+
"\n");
if(argc == 1 ||
argc < 4 ||
(argc == 2 && (string(argv[0]) == "-h" || string(argv[0]) == "--help") )
){
cerr << "Usage "<<usage<<endl;
return 1;
}
for(int i=1;i<(argc-2);i++){
if(string(argv[i]) == "--bq"){
minBaseQuality=destringify<int>(argv[i+1]);
i++;
continue;
}
}
string bamfiletopen = string( argv[ argc-5 ] );
string vcffiletopen = string( argv[ argc-4 ] );
string chrname = string( argv[ argc-3 ] );
string deambam = string( argv[ argc-2 ] );
string nondeambam = string( argv[ argc-1 ] );
//dummy reader, will need to reposition anyway
VCFreader vcfr (vcffiletopen,
vcffiletopen+".tbi",
chrname,
1,
1,
0);
BamReader reader;
if ( !reader.Open(bamfiletopen) ) {
cerr << "Could not open input BAM file"<< bamfiletopen << endl;
return 1;
}
// if ( !reader.LocateIndex() ) {
// cerr << "The index for the BAM file cannot be located" << endl;
// return 1;
// }
// if ( !reader.HasIndex() ) {
// cerr << "The BAM file has not been indexed." << endl;
// return 1;
// }
//positioning the bam file
int refid=reader.GetReferenceID(chrname);
if(refid < 0){
cerr << "Cannot retrieve the reference ID for "<< chrname << endl;
return 1;
}
//cout<<"redif "<<refid<<endl;
//setting the BAM reader at that position
reader.SetRegion(refid,
0,
refid,
-1);
vector<RefData> testRefData=reader.GetReferenceData();
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
BamWriter writerDeam;
if ( !writerDeam.Open(deambam, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
BamWriter writerNoDeam;
if ( !writerNoDeam.Open(nondeambam, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
unsigned int totalReads =0;
unsigned int deaminatedReads =0;
unsigned int ndeaminatedReads =0;
unsigned int skipped =0;
//iterating over the alignments for these regions
BamAlignment al;
int i;
while ( reader.GetNextAlignment(al) ) {
// cerr<<al.Name<<endl;
//skip unmapped
if(!al.IsMapped()){
skipped++;
continue;
}
//skip paired end !
if(al.IsPaired() ){
continue;
// cerr<<"Paired end not yet coded"<<endl;
// return 1;
}
string reconstructedReference = reconstructRef(&al);
char refeBase;
char readBase;
bool isDeaminated;
if(al.Qualities.size() != reconstructedReference.size()){
cerr<<"Quality line is not the same size as the reconstructed reference"<<endl;
return 1;
}
isDeaminated=false;
if(al.IsReverseStrand()){
//first base next to 3'
i = 0 ;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
vcfr.repositionIterator(chrname,al.Position+1,al.Position+1);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
// cout<<*toprint<<endl;
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<"Problem1 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
//if the VCF has a at least one G but no A
if( toprint->hasAtLeastOneG() &&
!toprint->hasAtLeastOneA() ){
isDeaminated=true;
}
}
}
//second base next to 3'
i = 1;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'G' &&
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
vcfr.repositionIterator(chrname,al.Position+2,al.Position+2);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
// cout<<*toprint<<endl;
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<"Problem2 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
//if the VCF has at least one G but no A
// if(toprint->hasAtLeastOneG() &&
// toprint->getAlt().find("A") == string::npos){
if( toprint->hasAtLeastOneG() &&
!toprint->hasAtLeastOneA() ){
isDeaminated=true;
}
}
}
//last base next to 5'
i = (al.QueryBases.length()-1) ;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'G' &&
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,0);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
vcfr.repositionIterator(chrname,positionJump,positionJump);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<lengthMatches<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<positionJump<<endl;
cerr<<"Problem3 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
//if the VCF has at least one G but no A
if( toprint->hasAtLeastOneG() &&
!toprint->hasAtLeastOneA() ){
isDeaminated=true;
}
}
}
}else{
//first base next to 5'
i = 0;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'C'
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
vcfr.repositionIterator(chrname,al.Position+1,al.Position+1);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
//cout<<*toprint<<endl;
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<"Problem4 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
//if the VCF has at least one C but no T
if( toprint->hasAtLeastOneC() &&
!toprint->hasAtLeastOneT() ){
isDeaminated=true;
}
}
//cout<<al.Position+
}
//second last base next to 3'
i = (al.QueryBases.length()-2);
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'C' &&
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,1);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
vcfr.repositionIterator(chrname,positionJump,positionJump);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<lengthMatches<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<positionJump<<endl;
cerr<<"Problem5 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
if( toprint->hasAtLeastOneC() &&
!toprint->hasAtLeastOneT() ){
isDeaminated=true;
}
}
}
//last base next to 3'
i = (al.QueryBases.length()-1);
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//&& refeBase == 'C'
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,0);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
vcfr.repositionIterator(chrname,positionJump,positionJump);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
//skip deletions in the alt
if(toprint->getRef().length() != 1 )
continue;
if(toprint->getRef()[0] != refeBase){
cerr<<reconstructedReference<<endl;
cerr<<al.Position<<endl;
cerr<<lengthMatches<<endl;
cerr<<numberOfDeletions(&al)<<endl;
cerr<<positionJump<<endl;
cerr<<"Problem6 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
exit(1);
}
if( toprint->hasAtLeastOneC() &&
!toprint->hasAtLeastOneT() ){
isDeaminated=true;
}
}
}
}
totalReads++;
if(isDeaminated){
deaminatedReads++;
writerDeam.SaveAlignment(al);
}else{
ndeaminatedReads++;
writerNoDeam.SaveAlignment(al);
}
}//end for each read
reader.Close();
writerDeam.Close();
writerNoDeam.Close();
cerr<<"Program finished sucessfully, out of "<<totalReads<<" mapped reads (skipped: "<<skipped<<" reads) we flagged "<<deaminatedReads<<" as deaminated and "<<ndeaminatedReads<<" as not deaminated"<<endl;
return 0;
}
int main (int argc, char *argv[]) {
int minBaseQuality = 0;
string usage=string(""+string(argv[0])+" [in BAM file] [in VCF file] [chr name] [deam out BAM] [not deam out BAM]"+
"\nThis program divides aligned single end reads into potentially deaminated\n"+
"\nreads and the puts the rest into another bam file if the deaminated positions are not called as the alternative base in the VCF.\n"+
"\nThis is like filterDeaminatedVCF but it loads the VCF before then labels the reads instead of doing it on the fly\n"+
"\nwhich is good if you have many reads in the bam file.\n"+
"\nTip: if you do not need one of them, use /dev/null as your output\n"+
"\narguments:\n"+
"\t"+"--bq [base qual] : Minimum base quality to flag a deaminated site (Default: "+stringify(minBaseQuality)+")\n"+
"\t"+"--1000g [vcf file] : VCF file from 1000g to get the putative A and T positions in modern humans (Default: "+vcf1000g+")\n"+
"\n");
if(argc == 1 ||
argc < 4 ||
(argc == 2 && (string(argv[0]) == "-h" || string(argv[0]) == "--help") )
){
cerr << "Usage "<<usage<<endl;
return 1;
}
for(int i=1;i<(argc-2);i++){
if(string(argv[i]) == "--bq"){
minBaseQuality=destringify<int>(argv[i+1]);
i++;
continue;
}
if(string(argv[i]) == "--1000g"){
vcf1000g=string(argv[i+1]);
i++;
continue;
}
}
unsigned int maxSizeChromosome=250000000;//larger than chr1 hg19
bool * hasCnoT;
bool * hasGnoA;
bool * thousandGenomesHasA;
bool * thousandGenomesHasT;
cerr<<"Trying to allocating memory"<<endl;
try{
hasCnoT = new bool[ maxSizeChromosome ];
hasGnoA = new bool[ maxSizeChromosome ];
thousandGenomesHasA = new bool[ maxSizeChromosome ];
thousandGenomesHasT = new bool[ maxSizeChromosome ];
}catch(bad_alloc& exc){
cerr<<"ERROR: allocating memory failed"<<endl;
return 1;
}
cerr<<"Success in allocating memory"<<endl;
for(unsigned int i = 0;i<maxSizeChromosome;i++){
hasCnoT[i]=false;
hasGnoA[i]=false;
thousandGenomesHasA[i]=false;
thousandGenomesHasT[i]=false;
}
string bamfiletopen = string( argv[ argc-5 ] );
string vcffiletopen = string( argv[ argc-4 ] );
string chrname = string( argv[ argc-3 ] );
string deambam = string( argv[ argc-2 ] );
string nondeambam = string( argv[ argc-1 ] );
cerr<<"Reading consensus VCF "<<vcffiletopen<<" ... "<<endl;
VCFreader vcfr (vcffiletopen,
// vcffiletopen+".tbi",
// chrname,
// 1,
// maxSizeChromosome,
0);
while(vcfr.hasData()){
SimpleVCF * toprint=vcfr.getData();
if(toprint->getRef().length() != 1 )
continue;
//if the VCF has a at least one G but no A
if( toprint->hasAtLeastOneG() &&
!toprint->hasAtLeastOneA() ){
hasGnoA[ toprint->getPosition() ] =true;
}
if( toprint->hasAtLeastOneC() &&
!toprint->hasAtLeastOneT() ){
hasCnoT[ toprint->getPosition() ] =true;
}
}
cerr<<"done reading VCF"<<endl;
cerr<<"Reading 1000g VCF :"<<vcf1000g<<" ..."<<endl;
string line1000g;
ifstream myFile1000g;
myFile1000g.open(vcf1000g.c_str(), ios::in);
if (myFile1000g.is_open()){
while ( getline (myFile1000g,line1000g)){
vector<string> fields=allTokens(line1000g,'\t');
//0 chr
//1 pos
//2 id
//3 ref
//4 alt
//check if same chr
if(fields[0] != chrname){
cerr <<"Error, wrong chromosome in 1000g file for line= "<<line1000g<<endl;
return 1;
}
//skip indels
if(fields[3].size() != 1 ||
fields[4].size() != 1 )
continue;
char ref=toupper(fields[3][0]);
char alt=toupper(fields[4][0]);
unsigned int pos=destringify<unsigned int>( fields[1] );
thousandGenomesHasA[ pos ] = ( (ref=='A') || (alt=='A') );
thousandGenomesHasT[ pos ] = ( (ref=='T') || (alt=='T') );
}
myFile1000g.close();
}else{
cerr <<"Unable to open file "<<vcf1000g<<endl;
return 1;
}
cerr<<"done reading 1000g VCF"<<endl;
BamReader reader;
if ( !reader.Open(bamfiletopen) ) {
cerr << "Could not open input BAM file"<< bamfiletopen << endl;
return 1;
}
//positioning the bam file
int refid=reader.GetReferenceID(chrname);
if(refid < 0){
cerr << "Cannot retrieve the reference ID for "<< chrname << endl;
return 1;
}
//cout<<"redif "<<refid<<endl;
//setting the BAM reader at that position
reader.SetRegion(refid,
0,
refid,
-1);
vector<RefData> testRefData=reader.GetReferenceData();
const SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
BamWriter writerDeam;
if ( !writerDeam.Open(deambam, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
BamWriter writerNoDeam;
if ( !writerNoDeam.Open(nondeambam, header, references) ) {
cerr << "Could not open output BAM file" << endl;
return 1;
}
unsigned int totalReads =0;
unsigned int deaminatedReads =0;
unsigned int ndeaminatedReads =0;
unsigned int skipped =0;
//iterating over the alignments for these regions
BamAlignment al;
int i;
while ( reader.GetNextAlignment(al) ) {
// cerr<<al.Name<<endl;
//skip unmapped
if(!al.IsMapped()){
skipped++;
continue;
}
//skip paired end !
if(al.IsPaired() ){
continue;
// cerr<<"Paired end not yet coded"<<endl;
// return 1;
}
string reconstructedReference = reconstructRef(&al);
char refeBase;
char readBase;
bool isDeaminated;
if(al.Qualities.size() != reconstructedReference.size()){
cerr<<"Quality line is not the same size as the reconstructed reference"<<endl;
return 1;
}
isDeaminated=false;
if(al.IsReverseStrand()){
//first base next to 3'
i = 0 ;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
if( hasGnoA[al.Position+1] &&
!thousandGenomesHasA[al.Position+1] )
isDeaminated=true;
// transformRef(&refeBase,&readBase);
// vcfr.repositionIterator(chrname,al.Position+1,al.Position+1);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// // cout<<*toprint<<endl;
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<"Problem1 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// //if the VCF has a at least one G but no A
// if( toprint->hasAtLeastOneG() &&
// !toprint->hasAtLeastOneA() ){
// isDeaminated=true;
// }
// }
}
//second base next to 3'
i = 1;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'G' &&
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
if( hasGnoA[al.Position+2] &&
!thousandGenomesHasA[al.Position+2] )
isDeaminated=true;
// transformRef(&refeBase,&readBase);
// vcfr.repositionIterator(chrname,al.Position+2,al.Position+2);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// // cout<<*toprint<<endl;
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<"Problem2 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// //if the VCF has at least one G but no A
// // if(toprint->hasAtLeastOneG() &&
// // toprint->getAlt().find("A") == string::npos){
// if( toprint->hasAtLeastOneG() &&
// !toprint->hasAtLeastOneA() ){
// isDeaminated=true;
// }
// }
}
//last base next to 5'
i = (al.QueryBases.length()-1) ;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'G' &&
if( readBase == 'A' && int(al.Qualities[i]-offset) >= minBaseQuality){ //isDeaminated=true; }
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,0);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
if(hasGnoA[positionJump] &&
!thousandGenomesHasA[positionJump] )
isDeaminated=true;
// vcfr.repositionIterator(chrname,positionJump,positionJump);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<lengthMatches<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<positionJump<<endl;
// cerr<<"Problem3 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// //if the VCF has at least one G but no A
// if( toprint->hasAtLeastOneG() &&
// !toprint->hasAtLeastOneA() ){
// isDeaminated=true;
// }
// }
}
}else{
//first base next to 5'
i = 0;
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'C'
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
// transformRef(&refeBase,&readBase);
if(hasCnoT[al.Position+1] &&
!thousandGenomesHasT[al.Position+1] )
isDeaminated=true;
// vcfr.repositionIterator(chrname,al.Position+1,al.Position+1);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// //cout<<*toprint<<endl;
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<"Problem4 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// //if the VCF has at least one C but no T
// if( toprint->hasAtLeastOneC() &&
// !toprint->hasAtLeastOneT() ){
// isDeaminated=true;
// }
// }
//cout<<al.Position+
}
//second last base next to 3'
i = (al.QueryBases.length()-2);
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//refeBase == 'C' &&
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
//transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,1);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
if(hasCnoT[positionJump] &&
!thousandGenomesHasT[positionJump] )
isDeaminated=true;
// vcfr.repositionIterator(chrname,positionJump,positionJump);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<lengthMatches<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<positionJump<<endl;
// cerr<<"Problem5 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// if( toprint->hasAtLeastOneC() &&
// !toprint->hasAtLeastOneT() ){
// isDeaminated=true;
// }
// }
}
//last base next to 3'
i = (al.QueryBases.length()-1);
refeBase=toupper(reconstructedReference[i]);
readBase=toupper( al.QueryBases[i]);
//&& refeBase == 'C'
if( readBase == 'T' && int(al.Qualities[i]-offset) >= minBaseQuality){
if( skipAlign(reconstructedReference,&al,&skipped) ){ continue; }
transformRef(&refeBase,&readBase);
int lengthMatches=countMatchesRecons(reconstructedReference,0);
int positionJump = al.Position+lengthMatches+numberOfDeletions(&al);
if(hasCnoT[positionJump] &&
!thousandGenomesHasT[positionJump] )
isDeaminated=true;
// vcfr.repositionIterator(chrname,positionJump,positionJump);
// while(vcfr.hasData()){
// SimpleVCF * toprint=vcfr.getData();
// //skip deletions in the alt
// if(toprint->getRef().length() != 1 )
// continue;
// if(toprint->getRef()[0] != refeBase){
// cerr<<reconstructedReference<<endl;
// cerr<<al.Position<<endl;
// cerr<<lengthMatches<<endl;
// cerr<<numberOfDeletions(&al)<<endl;
// cerr<<positionJump<<endl;
// cerr<<"Problem6 position "<<*toprint<<" does not have a "<<refeBase<<" as reference allele for read "<<al.Name<<endl;
// exit(1);
// }
// if( toprint->hasAtLeastOneC() &&
// !toprint->hasAtLeastOneT() ){
// isDeaminated=true;
// }
// }
}
}
totalReads++;
if(isDeaminated){
deaminatedReads++;
writerDeam.SaveAlignment(al);
}else{
ndeaminatedReads++;
writerNoDeam.SaveAlignment(al);
}
}//end for each read
reader.Close();
writerDeam.Close();
writerNoDeam.Close();
delete(hasCnoT);
delete(hasGnoA);
cerr<<"Program finished sucessfully, out of "<<totalReads<<" mapped reads (skipped: "<<skipped<<" reads) we flagged "<<deaminatedReads<<" as deaminated and "<<ndeaminatedReads<<" as not deaminated"<<endl;
return 0;
}
int main (int argc, char *argv[]) {
if( (argc!= 3) ||
(argc== 2 && string(argv[1]) == "-h") ||
(argc== 2 && string(argv[1]) == "-help") ||
(argc== 2 && string(argv[1]) == "--help") ){
cerr<<"Usage:splitByRG [in bam] [out prefix]"<<endl<<"this program creates one bam file per RG in the with the outprefix\nFor example splitByRG in.bam out will create\nout.rg1.bam\nout.rg2.bam\n"<<endl;
return 1;
}
string bamfiletopen = string(argv[1]);
// if(!strEndsWith(bamfiletopen,".bam")){
// }
string bamDirOutPrefix = string(argv[2]);
map<string,BamWriter *> rg2BamWriter;
// if(!isDirectory(bamDirOut)){
// cerr<<"ERROR: the out directory does not exist"<<endl;
// return 1;
// }
BamReader reader;
if ( !reader.Open(bamfiletopen) ) {
cerr << "Could not open input BAM files." << endl;
return 1;
}
SamHeader header = reader.GetHeader();
const RefVector references = reader.GetReferenceData();
vector<RefData> refData=reader.GetReferenceData();
string pID = "splitByRG";
string pName = "splitByRG";
string pCommandLine = "";
for(int i=0;i<(argc);i++){
pCommandLine += (string(argv[i])+" ");
}
putProgramInHeader(&header,pID,pName,pCommandLine,returnGitHubVersion(string(argv[0]),".."));
SamReadGroupDictionary srgd=header.ReadGroups;
for(SamReadGroupConstIterator srgci=srgd.ConstBegin();
srgci<srgd.ConstEnd();
srgci++){
//cout<<*srgci<<endl;
const SamReadGroup rg = (*srgci);
//cout<<rg.ID<<endl;
rg2BamWriter[rg.ID] = new BamWriter();
rg2BamWriter[rg.ID]->Open(bamDirOutPrefix+"."+rg.ID+".bam",header,references);
}
BamAlignment al;
unsigned int total=0;
while ( reader.GetNextAlignment(al) ) {
// al.SetIsFailedQC(false);
// writer.SaveAlignment(al);
// if(al.IsMapped () ){
// if(rg2BamWriter.find(refData[al.RefID].RefName) == rg2BamWriter.end()){ //new
// rg2BamWriter[refData[al.RefID].RefName] = new BamWriter();
// if ( !rg2BamWriter[refData[al.RefID].RefName]->Open(bamDirOutPrefix+"."+refData[al.RefID].RefName+".bam",header,references) ) {
// cerr << "Could not open output BAM file "<< bamDirOutPrefix<<"."<<refData[al.RefID].RefName<<".bam" << endl;
// return 1;
// }
// }else{
// rg2BamWriter[refData[al.RefID].RefName]->SaveAlignment(al);
// }
// }else{
// unmapped.SaveAlignment(al);
// }
if(al.HasTag("RG")){
string rgTag;
al.GetTag("RG",rgTag);
//cout<<rgTag<<endl;
if(rg2BamWriter.find(rgTag) == rg2BamWriter.end()){ //new
cerr<<"Found new RG "<<rgTag<<endl;
rg2BamWriter[rgTag] = new BamWriter();
if ( !rg2BamWriter[rgTag]->Open(bamDirOutPrefix+"."+rgTag+".bam",header,references) ) {
cerr << "Could not open output BAM file "<< bamDirOutPrefix<<"."<<rgTag<<".bam" << endl;
return 1;
}
rg2BamWriter[rgTag]->SaveAlignment(al);
}else{
rg2BamWriter[rgTag]->SaveAlignment(al);
}
}else{
string rgTag="unknown";
//cout<<rgTag<<endl;
if(rg2BamWriter.find(rgTag) == rg2BamWriter.end()){ //new
cerr<<"Found new RG "<<rgTag<<endl;
rg2BamWriter[rgTag] = new BamWriter();
if ( !rg2BamWriter[rgTag]->Open(bamDirOutPrefix+"."+rgTag+".bam",header,references) ) {
cerr << "Could not open output BAM file "<< bamDirOutPrefix<<"."<<rgTag<<".bam" << endl;
return 1;
}
rg2BamWriter[rgTag]->SaveAlignment(al);
}else{
rg2BamWriter[rgTag]->SaveAlignment(al);
}
// cerr << "Cannot get RG tag for " << al.Name<<endl;
// return 1;
}
total++;
} //while al
reader.Close();
// writer.Close();
// unmapped.Close();
map<string,BamWriter *>::iterator rg2BamWriterIt;
for (rg2BamWriterIt =rg2BamWriter.begin();
rg2BamWriterIt!=rg2BamWriter.end();
rg2BamWriterIt++){
rg2BamWriterIt->second->Close();
}
cerr<<"Wrote succesfully "<<total<<" reads"<<endl;
return 0;
}
// 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.
bool BamMultiReaderPrivate::ValidateReaders() const
{
m_errorString.clear();
// skip if 0 or 1 readers opened
if (m_readers.empty() || (m_readers.size() == 1)) return true;
// retrieve first reader
const MergeItem& firstItem = m_readers.front();
const BamReader* firstReader = firstItem.Reader;
if (firstReader == 0) return false;
// retrieve first reader's header data
const SamHeader& firstReaderHeader = firstReader->GetHeader();
const std::string& firstReaderSortOrder = firstReaderHeader.SortOrder;
// retrieve first reader's reference data
const RefVector& firstReaderRefData = firstReader->GetReferenceData();
const int firstReaderRefCount = firstReader->GetReferenceCount();
const int firstReaderRefSize = firstReaderRefData.size();
// iterate over all readers
std::vector<MergeItem>::const_iterator readerIter = m_readers.begin();
std::vector<MergeItem>::const_iterator readerEnd = m_readers.end();
for (; readerIter != readerEnd; ++readerIter) {
const MergeItem& item = (*readerIter);
BamReader* reader = item.Reader;
if (reader == 0) continue;
// get current reader's header data
const SamHeader& currentReaderHeader = reader->GetHeader();
const std::string& currentReaderSortOrder = currentReaderHeader.SortOrder;
// check compatible sort order
if (currentReaderSortOrder != firstReaderSortOrder) {
const std::string message =
std::string("mismatched sort order in ") + reader->GetFilename() + ", expected " +
firstReaderSortOrder + ", but found " + currentReaderSortOrder;
SetErrorString("BamMultiReader::ValidateReaders", message);
return false;
}
// get current reader's reference data
const RefVector currentReaderRefData = reader->GetReferenceData();
const int currentReaderRefCount = reader->GetReferenceCount();
const int currentReaderRefSize = currentReaderRefData.size();
// init reference data iterators
RefVector::const_iterator firstRefIter = firstReaderRefData.begin();
RefVector::const_iterator firstRefEnd = firstReaderRefData.end();
RefVector::const_iterator currentRefIter = currentReaderRefData.begin();
// compare reference counts from BamReader ( & container size, in case of BR error)
if ((currentReaderRefCount != firstReaderRefCount) ||
(firstReaderRefSize != currentReaderRefSize)) {
std::stringstream s;
s << "mismatched reference count in " << reader->GetFilename() << ", expected "
<< firstReaderRefCount << ", but found " << currentReaderRefCount;
SetErrorString("BamMultiReader::ValidateReaders", s.str());
return false;
}
// 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 (firstRefIter != firstRefEnd) {
const RefData& firstRef = (*firstRefIter);
const RefData& currentRef = (*currentRefIter);
// compare reference name & length
if ((firstRef.RefName != currentRef.RefName) ||
(firstRef.RefLength != currentRef.RefLength)) {
std::stringstream s;
s << "mismatched references found in" << reader->GetFilename()
<< "expected: " << std::endl;
// print first reader's reference data
RefVector::const_iterator refIter = firstReaderRefData.begin();
RefVector::const_iterator refEnd = firstReaderRefData.end();
for (; refIter != refEnd; ++refIter) {
const RefData& entry = (*refIter);
std::stringstream s;
s << entry.RefName << ' ' << std::endl;
}
s << "but found: " << std::endl;
// print current reader's reference data
refIter = currentReaderRefData.begin();
refEnd = currentReaderRefData.end();
for (; refIter != refEnd; ++refIter) {
const RefData& entry = (*refIter);
s << entry.RefName << ' ' << entry.RefLength << std::endl;
}
SetErrorString("BamMultiReader::ValidateReaders", s.str());
return false;
}
// update iterators
++firstRefIter;
++currentRefIter;
}
}
// if we get here, everything checks out
return true;
}
