// General functions.
bool LookupHQRegion(int holeNumber, RegionTable ®ionTable,
int &start, int &end, int &score) {
int regionLowIndex, regionHighIndex;
regionLowIndex = regionHighIndex = 0;
regionTable.LookupRegionsByHoleNumber(holeNumber,
regionLowIndex, regionHighIndex);
bool readHasGoodRegion = true;
int regionIndex = regionLowIndex;
while (regionIndex < regionHighIndex and
regionTable.GetType(regionIndex) != HQRegion) {
regionIndex++;
}
if (regionIndex == regionHighIndex) {
start = end = score = 0;
return false;
}
else {
start = regionTable.GetStart(regionIndex);
end = regionTable.GetEnd(regionIndex);
score = regionTable.GetScore(regionIndex);
return true;
}
}
//
// Collect region indices for either all region types, or just a few specific region types.
//
//
int CollectRegionIndices(SMRTSequence &read, RegionTable ®ionTable,
std::vector<int> ®ionIndices, RegionType *regionTypes,
int numRegionTypes) {
int regionLow, regionHigh;
int prevNumRegionIndices = regionIndices.size();
if (FindRegionIndices(read, ®ionTable, regionLow, regionHigh)) {
int i;
for (i = regionLow; i < regionHigh; i++) {
if (regionTypes == NULL) {
regionIndices.push_back(i);
}
else {
int t;
for (t = 0; t < numRegionTypes; t++) {
if (regionTable.GetType(i) == regionTypes[t]) {
regionIndices.push_back(i);
break;
}
}
}
}
}
return regionIndices.size() - prevNumRegionIndices;
}
// General functions.
bool LookupHQRegion(int holeNumber, RegionTable ®ionTable, int &start, int &end, int &score)
{
if (regionTable.HasHoleNumber(holeNumber)) {
RegionAnnotations zmwRegions = regionTable[holeNumber];
if (zmwRegions.HasHQRegion()) {
start = zmwRegions.HQStart();
end = zmwRegions.HQEnd();
score = zmwRegions.HQScore();
return true;
}
}
start = end = score = 0;
return false;
}
TEST(SubreadsTest, EndToEnd_Multiple)
{
// setup
const string movieName = "m140905_042212_sidney_c100564852550000001823085912221377_s1_X0";
vector<string> baxFilenames;
baxFilenames.push_back(tests::Data_Dir + "/" + movieName + ".1.bax.h5");
const string generatedBam = movieName + ".subreads.bam";
const string scrapBam = movieName + ".scraps.bam";
// run conversion
const int result = RunBax2Bam(baxFilenames, "--subread");
EXPECT_EQ(0, result);
// open BAX reader on original data
HDFBasReader baxReader;
baxReader.IncludeField("Basecall");
baxReader.IncludeField("DeletionQV");
baxReader.IncludeField("DeletionTag");
baxReader.IncludeField("InsertionQV");
baxReader.IncludeField("PreBaseFrames");
baxReader.IncludeField("MergeQV");
baxReader.IncludeField("SubstitutionQV");
baxReader.IncludeField("HQRegionSNR");
// not using SubTag or PulseWidth here
string baxBasecallerVersion;
string baxBindingKit;
string baxSequencingKit;
const int initOk = baxReader.Initialize(baxFilenames.front());
EXPECT_EQ(1, initOk);
if (initOk == 1) {
if (baxReader.scanDataReader.fileHasScanData && baxReader.scanDataReader.initializedRunInfoGroup) {
if (baxReader.scanDataReader.runInfoGroup.ContainsAttribute("BindingKit")) {
HDFAtom<std::string> bkAtom;
if (bkAtom.Initialize(baxReader.scanDataReader.runInfoGroup, "BindingKit")) {
bkAtom.Read(baxBindingKit);
bkAtom.dataspace.close();
}
}
if (baxReader.scanDataReader.runInfoGroup.ContainsAttribute("SequencingKit")) {
HDFAtom<std::string> skAtom;
if (skAtom.Initialize(baxReader.scanDataReader.runInfoGroup, "SequencingKit")) {
skAtom.Read(baxSequencingKit);
skAtom.dataspace.close();
}
}
}
baxReader.GetChangeListID(baxBasecallerVersion);
}
// read region table info
boost::scoped_ptr<HDFRegionTableReader> regionTableReader(new HDFRegionTableReader);
RegionTable regionTable;
std::string fn = baxFilenames.front();
EXPECT_TRUE(regionTableReader->Initialize(fn) != 0);
regionTable.Reset();
regionTableReader->ReadTable(regionTable);
regionTableReader->Close();
EXPECT_NO_THROW(
{
// open BAM file
BamFile bamFile(generatedBam);
// check BAM header information
const BamHeader& header = bamFile.Header();
EXPECT_EQ(string("1.5"), header.Version());
EXPECT_EQ(string("unknown"), header.SortOrder());
EXPECT_EQ(string("3.0.1"), header.PacBioBamVersion());
EXPECT_TRUE(header.Sequences().empty());
EXPECT_TRUE(header.Comments().empty());
ASSERT_FALSE(header.Programs().empty());
const vector<string> readGroupIds = header.ReadGroupIds();
ASSERT_FALSE(readGroupIds.empty());
const ReadGroupInfo& rg = header.ReadGroup(readGroupIds.front());
string rawId = movieName + "//SUBREAD";
string md5Id;
MakeMD5(rawId, md5Id, 8);
EXPECT_EQ(md5Id, rg.Id());
EXPECT_EQ(string("PACBIO"), rg.Platform());
EXPECT_EQ(movieName, rg.MovieName());
EXPECT_TRUE(rg.SequencingCenter().empty());
EXPECT_TRUE(rg.Date().empty());
EXPECT_TRUE(rg.FlowOrder().empty());
EXPECT_TRUE(rg.KeySequence().empty());
EXPECT_TRUE(rg.Library().empty());
EXPECT_TRUE(rg.Programs().empty());
EXPECT_TRUE(rg.PredictedInsertSize().empty());
EXPECT_TRUE(rg.Sample().empty());
EXPECT_EQ("SUBREAD", rg.ReadType());
EXPECT_EQ(baxBasecallerVersion, rg.BasecallerVersion());
EXPECT_EQ(baxBindingKit, rg.BindingKit());
EXPECT_EQ(baxSequencingKit, rg.SequencingKit());
EXPECT_EQ(75, std::stod(rg.FrameRateHz()));
EXPECT_EQ("dq", rg.BaseFeatureTag(BaseFeature::DELETION_QV));
EXPECT_EQ("dt", rg.BaseFeatureTag(BaseFeature::DELETION_TAG));
EXPECT_EQ("iq", rg.BaseFeatureTag(BaseFeature::INSERTION_QV));
EXPECT_EQ("ip", rg.BaseFeatureTag(BaseFeature::IPD));
EXPECT_EQ("mq", rg.BaseFeatureTag(BaseFeature::MERGE_QV));
EXPECT_EQ("sq", rg.BaseFeatureTag(BaseFeature::SUBSTITUTION_QV));
EXPECT_FALSE(rg.HasBaseFeature(BaseFeature::SUBSTITUTION_TAG));
EXPECT_EQ(FrameCodec::V1, rg.IpdCodec());
// compare 1st record from each file
SMRTSequence baxRecord;
UInt holeNumber = 0;
vector<float> hqSnr;
size_t intervalIdx = 0;
vector<SubreadInterval> subreadIntervals;
size_t numTested = 0;
EntireFileQuery entireFile(bamFile);
for (BamRecord& bamRecord : entireFile) {
if (intervalIdx >= subreadIntervals.size())
{
while (baxReader.GetNext(baxRecord))
{
holeNumber = baxRecord.zmwData.holeNumber;
ComputeSubreadIntervals(&subreadIntervals, regionTable, holeNumber);
/* this is for debugging subread interval problems
int hqStart = 0;
int hqEnd = 0;
int hqScore = 0;
LookupHQRegion(holeNumber,
regionTable,
hqStart,
hqEnd,
hqScore);
vector<ReadInterval> subreadIntervals_;
CollectSubreadIntervals(baxRecord, ®ionTable, subreadIntervals_);
for (int i = subreadIntervals_.size() - 1; i >= 0; --i)
{
auto& in = subreadIntervals_[i];
int inStart = max(hqStart, in.start);
int inEnd = min(hqEnd, in.end);
if (inEnd <= inStart)
subreadIntervals_.erase(subreadIntervals_.begin() + i);
}
cerr << "hqRegion: " << hqStart << ", " << hqEnd << endl;
cerr << "subreadRegions:" << endl;
for (const auto& in : subreadIntervals_)
cerr << " l, r: " << in.start << ", " << in.end << endl;
cerr << "adapterDerived:" << endl;
for (const auto& in : subreadIntervals)
cerr << " l, r: " << in.Start << ", " << in.End << endl;
cerr << endl;
// */
if (subreadIntervals.empty())
continue;
intervalIdx = 0;
hqSnr.clear();
hqSnr.push_back(baxRecord.HQRegionSnr('A'));
hqSnr.push_back(baxRecord.HQRegionSnr('C'));
hqSnr.push_back(baxRecord.HQRegionSnr('G'));
hqSnr.push_back(baxRecord.HQRegionSnr('T'));
EXPECT_GT(hqSnr[0], 0);
EXPECT_GT(hqSnr[1], 0);
EXPECT_GT(hqSnr[2], 0);
EXPECT_GT(hqSnr[3], 0);
goto compare;
}
goto cleanup;
}
compare:
const BamRecordImpl& bamRecordImpl = bamRecord.Impl();
EXPECT_EQ(4680,bamRecordImpl.Bin());
EXPECT_EQ(0, bamRecordImpl.InsertSize());
EXPECT_EQ(255, bamRecordImpl.MapQuality());
EXPECT_EQ(-1, bamRecordImpl.MatePosition());
EXPECT_EQ(-1, bamRecordImpl.MateReferenceId());
EXPECT_EQ(-1, bamRecordImpl.Position());
EXPECT_EQ(-1, bamRecordImpl.ReferenceId());
EXPECT_FALSE(bamRecordImpl.IsMapped());
const int subreadStart = subreadIntervals[intervalIdx].Start;
const int subreadEnd = subreadIntervals[intervalIdx].End;
const string expectedName = movieName + "/" +
to_string(holeNumber) + "/" +
to_string(subreadStart) + "_" +
to_string(subreadEnd);
EXPECT_EQ(expectedName, bamRecordImpl.Name());
using PacBio::BAM::QualityValue;
using PacBio::BAM::QualityValues;
const DNALength length = subreadEnd - subreadStart;
string expectedSequence;
expectedSequence.assign((const char*)baxRecord.seq + subreadStart, length);
const string bamSequence = bamRecord.Sequence();
const QualityValues bamQualities = bamRecord.Qualities();
EXPECT_EQ(expectedSequence, bamSequence);
EXPECT_TRUE(bamQualities.empty());
const QualityValues bamDeletionQVs = bamRecord.DeletionQV();
const QualityValues bamInsertionQVs = bamRecord.InsertionQV();
const QualityValues bamMergeQVs = bamRecord.MergeQV();
const QualityValues bamSubstitutionQVs = bamRecord.SubstitutionQV();
for (size_t i = 0; i < length; ++i) {
const size_t pos = subreadStart + i;
EXPECT_EQ((QualityValue)baxRecord.GetDeletionQV(pos), bamDeletionQVs.at(i));
EXPECT_EQ((QualityValue)baxRecord.GetInsertionQV(pos), bamInsertionQVs.at(i));
EXPECT_EQ((QualityValue)baxRecord.GetMergeQV(pos), bamMergeQVs.at(i));
EXPECT_EQ((QualityValue)baxRecord.GetSubstitutionQV(pos), bamSubstitutionQVs.at(i));
}
if (baxRecord.deletionTag)
{
string expectedDeletionTags;
expectedDeletionTags.assign((char*)baxRecord.deletionTag + subreadStart,
(char*)baxRecord.deletionTag + subreadStart + length);
const string& bamDeletionTags = bamRecord.DeletionTag();
EXPECT_EQ(expectedDeletionTags, bamDeletionTags);
}
if (baxRecord.substitutionTag)
{
string expectedSubstitutionTags;
expectedSubstitutionTags.assign((char*)baxRecord.substitutionTag + subreadStart,
(char*)baxRecord.substitutionTag + subreadStart + length);
const string& bamSubstitutionTags = bamRecord.SubstitutionTag();
EXPECT_EQ(expectedSubstitutionTags, bamSubstitutionTags);
}
// TODO: IPDs
const LocalContextFlags ctxFlags = subreadIntervals[intervalIdx].LocalContextFlags;
EXPECT_EQ(md5Id, bamRecord.ReadGroupId());
EXPECT_EQ(movieName, bamRecord.MovieName());
EXPECT_EQ(1, bamRecord.NumPasses());
EXPECT_EQ(holeNumber, bamRecord.HoleNumber());
EXPECT_EQ(subreadStart, bamRecord.QueryStart());
EXPECT_EQ(subreadEnd, bamRecord.QueryEnd());
EXPECT_EQ(hqSnr, bamRecord.SignalToNoise());
EXPECT_EQ(ctxFlags, bamRecord.LocalContextFlags());
numTested++;
intervalIdx++;
}
cleanup:
EXPECT_GT(numTested, 1);
// cleanup
baxReader.Close();
RemoveFile(generatedBam);
RemoveFile(scrapBam);
}); // EXPECT_NO_THROW
int main(int argc, char* argv[]) {
string program = "pls2fasta";
string versionString = VERSION;
AppendPerforceChangelist(PERFORCE_VERSION_STRING, versionString);
string plsFileName, fastaOutName;
vector<string> plsFileNames;
bool trimByRegion, maskByRegion;
trimByRegion = false;
maskByRegion = false;
int argi = 3;
RegionTable regionTable;
string regionsFOFNName = "";
vector<string> regionFileNames;
bool splitSubreads = true;
int minSubreadLength = 0;
bool addSimulatedData = false;
bool printSimulatedCoordinate = false;
bool printSimulatedSequenceIndex = false;
bool printFastq = false;
bool printCcs = false;
int lineLength = 50;
int minReadScore = 0;
vector<int> holeNumbers;
CommandLineParser clp;
bool printOnlyBest = false;
clp.SetProgramName(program);
clp.SetVersion(versionString);
clp.RegisterStringOption("in.pls.h5", &plsFileName, "Input pls.h5/bax.h5/fofn file.", true);
clp.RegisterStringOption("out.fasta", &fastaOutName, "Output fasta/fastq file.", true);
clp.RegisterPreviousFlagsAsHidden();
clp.RegisterFlagOption("trimByRegion", &trimByRegion, "Trim away low quality regions.");
clp.RegisterFlagOption("maskByRegion", &maskByRegion, "Mask low quality regions with 'N'.");
clp.RegisterStringOption("regionTable", ®ionsFOFNName, "Optional HDF file with a /PulseData/Regions dataset.");
clp.RegisterIntOption("minSubreadLength", &minSubreadLength, "Do not write subreads less than the specified length.", CommandLineParser::PositiveInteger);
clp.RegisterFlagOption("noSplitSubreads", &splitSubreads, "Do not split reads on adapter sequences.");
clp.RegisterIntListOption("holeNumber", &holeNumbers, "Only print this hole number (or list of numbers).");
clp.RegisterFlagOption("fastq", &printFastq, "Print in FASTQ format with quality.");
clp.RegisterFlagOption("ccs", &printCcs, "Print de novo CCS sequences");
clp.RegisterIntOption("lineLength", &lineLength, "Specify fasta/fastq line length", CommandLineParser::PositiveInteger);
clp.RegisterIntOption("minReadScore", &minReadScore, "Minimum read score to print a read. The score is "
"a number between 0 and 1000 and represents the expected accuracy percentage * 10. "
"A typical value would be between 750 and 800. This does not apply to ccs reads.", CommandLineParser::NonNegativeInteger);
clp.RegisterFlagOption("best", &printOnlyBest, "If a CCS sequence exists, print this. Otherwise, print the longest"
"subread. This does not support fastq.");
string description = ("Converts pls.h5/bax.h5/fofn files to fasta or fastq files. Although fasta files are provided"
" with every run, they are not trimmed nor split into subreads. This program takes "
"additional annotation information, such as the subread coordinates and high quality regions "
"and uses them to create fasta sequences that are substrings of all bases called. Most of the time "
"you will want to trim low quality reads, so you should specify -trimByRegion.");
clp.SetProgramSummary(description);
clp.ParseCommandLine(argc, argv);
cerr << "[INFO] " << GetTimestamp() << " [" << program << "] started." << endl;
if (trimByRegion and maskByRegion) {
cout << "ERROR! You cannot both trim and mask regions. Use one or the other." << endl;
exit(1);
}
if (printFastq) {
// Setting lineLength to 0 flags to print on one line.
lineLength = 0;
}
if (FileOfFileNames::IsFOFN(plsFileName)) {
FileOfFileNames::FOFNToList(plsFileName, plsFileNames);
}
else {
plsFileNames.push_back(plsFileName);
}
if (regionsFOFNName == "") {
regionFileNames = plsFileNames;
}
else {
if (FileOfFileNames::IsFOFN(regionsFOFNName)) {
FileOfFileNames::FOFNToList(regionsFOFNName, regionFileNames);
}
else {
regionFileNames.push_back(regionsFOFNName);
}
}
ofstream fastaOut;
CrucialOpen(fastaOutName, fastaOut);
int plsFileIndex;
HDFRegionTableReader hdfRegionReader;
sort(holeNumbers.begin(), holeNumbers.end());
for (plsFileIndex = 0; plsFileIndex < plsFileNames.size(); plsFileIndex++) {
if (trimByRegion or maskByRegion or splitSubreads) {
hdfRegionReader.Initialize(regionFileNames[plsFileIndex]);
hdfRegionReader.ReadTable(regionTable);
regionTable.SortTableByHoleNumber();
}
ReaderAgglomerate reader;
HDFBasReader ccsReader;
if (printOnlyBest) {
ccsReader.SetReadBasesFromCCS();
ccsReader.Initialize(plsFileNames[plsFileIndex]);
}
if (printCcs == false) {
reader.IgnoreCCS();
}
else {
reader.hdfBasReader.SetReadBasesFromCCS();
}
if (addSimulatedData) {
reader.hdfBasReader.IncludeField("SimulatedCoordinate");
reader.hdfBasReader.IncludeField("SimulatedSequenceIndex");
}
if (reader.SetReadFileName(plsFileNames[plsFileIndex]) == 0) {
cout << "ERROR, could not determine file type."
<< plsFileNames[plsFileIndex] << endl;
exit(1);
}
if (reader.Initialize() == 0) {
cout << "ERROR, could not initialize file "
<< plsFileNames[plsFileIndex] << endl;
exit(1);
}
DNALength simulatedCoordinate;
DNALength simulatedSequenceIndex;
reader.SkipReadQuality();
SMRTSequence seq;
vector<ReadInterval> subreadIntervals;;
SMRTSequence ccsSeq;
while (reader.GetNext(seq)) {
if (printOnlyBest) {
ccsReader.GetNext(ccsSeq);
}
if (holeNumbers.size() != 0 and
binary_search(holeNumbers.begin(), holeNumbers.end(), seq.zmwData.holeNumber) == false) {
continue;
}
if (seq.length == 0) {
continue;
}
if (addSimulatedData) {
reader.hdfBasReader.simulatedCoordinateArray.Read(reader.hdfBasReader.curRead-1, reader.hdfBasReader.curRead, &simulatedCoordinate);
reader.hdfBasReader.simulatedSequenceIndexArray.Read(reader.hdfBasReader.curRead-1, reader.hdfBasReader.curRead, &simulatedSequenceIndex);
}
if (printCcs == true) {
if (printFastq == false) {
seq.PrintSeq(fastaOut);
}
else {
seq.PrintFastq(fastaOut, lineLength);
}
continue;
}
//
// Determine the high quality boundaries of the read. This is
// the full read is no hq regions exist, or it is stated to
// ignore regions.
//
DNALength hqReadStart, hqReadEnd;
int hqRegionScore;
if (GetReadTrimCoordinates(seq, seq.zmwData, regionTable, hqReadStart, hqReadEnd, hqRegionScore) == false or
(trimByRegion == false and maskByRegion == false)) {
hqReadStart = 0;
hqReadEnd = seq.length;
}
//
// Mask off the low quality portions of the reads.
//
if (maskByRegion) {
if (hqReadStart > 0) {
fill(&seq.seq[0], &seq.seq[hqReadStart], 'N');
}
if (hqReadEnd != seq.length) {
fill(&seq.seq[hqReadEnd], &seq.seq[seq.length], 'N');
}
}
//
// Now possibly print the full read with masking. This could be handled by making a
//
if (splitSubreads == false) {
ReadInterval wholeRead(0, seq.length);
// The set of subread intervals is just the entire read.
subreadIntervals.clear();
subreadIntervals.push_back(wholeRead);
}
else {
//
// Print subread coordinates no matter whether or not reads have subreads.
//
subreadIntervals.clear(); // clear old, new intervals are appended.
CollectSubreadIntervals(seq, ®ionTable, subreadIntervals);
}
//
// Output all subreads as separate sequences.
//
int intvIndex;
SMRTSequence bestSubreadSequence;
int bestSubreadScore = -1;
int bestSubreadIndex = 0;
int bestSubreadStart = 0, bestSubreadEnd = 0;
SMRTSequence bestSubread;
for (intvIndex = 0; intvIndex < subreadIntervals.size(); intvIndex++) {
SMRTSequence subreadSequence, subreadSequenceRC;
subreadSequence.subreadStart = subreadIntervals[intvIndex].start;
subreadSequence.subreadEnd = subreadIntervals[intvIndex].end;
//
// When trimming by region, only output the parts of the
// subread that overlap the hq region.
//
if (trimByRegion == true) {
subreadSequence.subreadStart = max((DNALength) subreadIntervals[intvIndex].start, hqReadStart);
subreadSequence.subreadEnd = min((DNALength) subreadIntervals[intvIndex].end, hqReadEnd);
}
if (subreadSequence.subreadStart >= subreadSequence.subreadEnd or
subreadSequence.subreadEnd - subreadSequence.subreadStart <= minSubreadLength) {
//
// There is no high qualty portion of this subread. Skip it.
//
continue;
}
if (hqRegionScore < minReadScore) {
continue;
}
//
// Print the subread, adding the coordinates as part of the title.
//
subreadSequence.ReferenceSubstring(seq, subreadSequence.subreadStart,
subreadSequence.subreadEnd - subreadSequence.subreadStart);
stringstream titleStream;
titleStream << seq.title;
if (splitSubreads) {
//
// Add the subread coordinates if splitting on subread.
//
titleStream << "/"
<< subreadSequence.subreadStart
<< "_" << subreadSequence.subreadEnd;
}
//
// If running on simulated data, add where the values were simulated from.
//
if (addSimulatedData) {
titleStream << ((FASTASequence*)&seq)->title << "/chrIndex_"
<< simulatedSequenceIndex << "/position_"<< simulatedCoordinate;
((FASTASequence*)&seq)->CopyTitle(titleStream.str());
}
subreadSequence.CopyTitle(titleStream.str());
//
// Eventually replace with WriterAgglomerate.
//
if (printOnlyBest == false) {
if (subreadSequence.length > 0) {
if (printFastq == false) {
((FASTASequence*)&subreadSequence)->PrintSeq(fastaOut);
}
else {
subreadSequence.PrintFastq(fastaOut, lineLength);
}
}
delete[] subreadSequence.title;
}
else {
int subreadWeightedScore = subreadSequence.length * hqRegionScore;
if (subreadWeightedScore > bestSubreadScore) {
bestSubreadIndex = intvIndex;
bestSubread = subreadSequence;
bestSubreadScore = subreadWeightedScore;
}
}
}
if (printOnlyBest) {
if (ccsSeq.length > 0) {
if (printFastq == false) {
ccsSeq.PrintSeq(fastaOut);
}
else {
ccsSeq.PrintFastq(fastaOut, ccsSeq.length);
}
}
else {
if (bestSubreadScore >= 0) {
if (printFastq == false) {
bestSubread.PrintSeq(fastaOut);
}
else {
bestSubread.PrintFastq(fastaOut, bestSubread.length);
}
bestSubread.Free();
}
}
ccsSeq.Free();
}
seq.Free();
}
reader.Close();
hdfRegionReader.Close();
}
cerr << "[INFO] " << GetTimestamp() << " [" << program << "] ended." << endl;
}
bool SubreadConverter::ConvertFile(HDFBasReader* reader,
PacBio::BAM::BamWriter* writer,
PacBio::BAM::BamWriter* scrapsWriter)
{
assert(reader);
// initialize with default values (shared across all unmapped subreads)
BamRecordImpl bamRecord;
// read region table info
std::unique_ptr<HDFRegionTableReader> const regionTableReader(new HDFRegionTableReader);
RegionTable regionTable;
string fn = filenameForReader_[reader];
assert(!fn.empty());
if (regionTableReader->Initialize(fn) == 0) {
AddErrorMessage("could not read region table on "+fn);
return false;
}
regionTable.Reset();
regionTableReader->ReadTable(regionTable);
regionTableReader->Close();
// initialize read scores
InitReadScores(reader);
// fetch records from HDF5 file
SMRTSequence smrtRecord;
while (reader->GetNext(smrtRecord)) {
// compute subread & adapter intervals
SubreadInterval hqInterval;
deque<SubreadInterval> subreadIntervals;
deque<SubreadInterval> adapterIntervals;
try {
hqInterval = ComputeSubreadIntervals(&subreadIntervals,
&adapterIntervals,
regionTable,
smrtRecord.zmwData.holeNumber,
smrtRecord.length);
} catch (runtime_error& e) {
AddErrorMessage(string(e.what()));
smrtRecord.Free();
return false;
}
// sequencing ZMW
if (IsSequencingZmw(smrtRecord))
{
// write subreads to main BAM file
for (const SubreadInterval& interval : subreadIntervals)
{
// skip invalid or 0-sized intervals
if (interval.End <= interval.Start)
continue;
if (!WriteSubreadRecord(smrtRecord,
interval.Start,
interval.End,
ReadGroupId(),
static_cast<uint8_t>(interval.LocalContextFlags),
writer))
{
smrtRecord.Free();
return false;
}
}
// if scraps BAM file present
if (scrapsWriter)
{
// write 5-end LQ sequence
if (hqInterval.Start > 0)
{
if (!WriteLowQualityRecord(smrtRecord,
0,
hqInterval.Start,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write adapters
for (const SubreadInterval& interval : adapterIntervals) {
// skip invalid or 0-sized adapters
if (interval.End <= interval.Start)
continue;
if (!WriteAdapterRecord(smrtRecord,
interval.Start,
interval.End,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write 3'-end LQ sequence
if (hqInterval.End < smrtRecord.length)
{
if (!WriteLowQualityRecord(smrtRecord,
hqInterval.End,
smrtRecord.length,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
}
} // sequencing ZMW
// non-sequencing ZMW
else
{
assert(!IsSequencingZmw(smrtRecord));
// only write these if scraps BAM present & we are in 'internal mode'
if (settings_.isInternal && scrapsWriter)
{
// write 5-end LQ sequence to scraps BAM
if (hqInterval.Start > 0)
{
if (!WriteLowQualityRecord(smrtRecord,
0,
hqInterval.Start,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write subreads & adapters to scraps BAM, sorted by query start
while (!subreadIntervals.empty() && !adapterIntervals.empty()) {
const SubreadInterval& subread = subreadIntervals.front();
const SubreadInterval& adapter = adapterIntervals.front();
assert(subread.Start != adapter.Start);
if (subread.Start < adapter.Start)
{
if (!WriteFilteredRecord(smrtRecord,
subread.Start,
subread.End,
ScrapsReadGroupId(),
static_cast<uint8_t>(subread.LocalContextFlags),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
subreadIntervals.pop_front();
}
else
{
if (!WriteAdapterRecord(smrtRecord,
adapter.Start,
adapter.End,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
adapterIntervals.pop_front();
}
}
// flush any traling subread intervals
while (!subreadIntervals.empty())
{
assert(adapterIntervals.empty());
const SubreadInterval& subread = subreadIntervals.front();
if (!WriteFilteredRecord(smrtRecord,
subread.Start,
subread.End,
ScrapsReadGroupId(),
static_cast<uint8_t>(subread.LocalContextFlags),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
subreadIntervals.pop_front();
}
// flush any remaining adapter intervals
while (!adapterIntervals.empty())
{
assert(subreadIntervals.empty());
const SubreadInterval& adapter = adapterIntervals.front();
if (!WriteAdapterRecord(smrtRecord,
adapter.Start,
adapter.End,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
adapterIntervals.pop_front();
}
// write 3'-end LQ sequence to scraps BAM
if (hqInterval.End < smrtRecord.length)
{
if (!WriteLowQualityRecord(smrtRecord,
hqInterval.End,
smrtRecord.length,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
}
} // non-sequencing ZMW
smrtRecord.Free();
}
// if we get here, all OK
return true;
}
bool HqRegionConverter::ConvertFile(HDFBasReader* reader,
PacBio::BAM::BamWriter* writer,
PacBio::BAM::BamWriter* scrapsWriter)
{
assert(reader);
// read region table info
std::unique_ptr<HDFRegionTableReader> const regionTableReader(new HDFRegionTableReader);
RegionTable regionTable;
std::string fn = filenameForReader_[reader];
assert(!fn.empty());
if (regionTableReader->Initialize(fn) == 0) {
AddErrorMessage("could not read region table on "+fn);
return false;
}
regionTable.Reset();
regionTableReader->ReadTable(regionTable);
regionTableReader->Close();
// initialize read scores
InitReadScores(reader);
// fetch records from HDF5 file
SMRTSequence smrtRecord;
int hqStart, hqEnd, score;
while (reader->GetNext(smrtRecord)) {
// attempt get high quality region
if (!LookupHQRegion(smrtRecord.zmwData.holeNumber,
regionTable,
hqStart,
hqEnd,
score))
{
stringstream s;
s << "could not find HQ region for hole number: " << smrtRecord.zmwData.holeNumber;
AddErrorMessage(s.str());
smrtRecord.Free();
return false;
}
// Catch and repair 1-off errors in the HQ region
hqEnd = (hqEnd == static_cast<int>(smrtRecord.length)-1) ? smrtRecord.length
: hqEnd;
// sequencing ZMW
if (IsSequencingZmw(smrtRecord))
{
// write HQRegion to main BAM file
if (hqStart < hqEnd)
{
if (!WriteRecord(smrtRecord,
hqStart,
hqEnd,
ReadGroupId(),
writer))
{
smrtRecord.Free();
return false;
}
}
// if scraps BAM file present
if (scrapsWriter)
{
// write 5'-end LQ sequence
if (hqStart > 0)
{
if (!WriteLowQualityRecord(smrtRecord,
0,
hqStart,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write 3'-end LQ sequence
if (static_cast<size_t>(hqEnd) < smrtRecord.length)
{
if (!WriteLowQualityRecord(smrtRecord,
hqEnd,
smrtRecord.length,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
}
}
// non-sequencing ZMW
else
{
assert(!IsSequencingZmw(smrtRecord));
// only write these if scraps BAM present & we are in 'internal mode'
if (settings_.isInternal && scrapsWriter)
{
// write 5'-end LQ sequence
if (hqStart > 0)
{
if (!WriteLowQualityRecord(smrtRecord,
0,
hqStart,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write HQRegion to scraps BAM file
if (hqStart < hqEnd)
{
if (!WriteFilteredRecord(smrtRecord,
hqStart,
hqEnd,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
// write 3'-end LQ sequence
if (static_cast<size_t>(hqEnd) < smrtRecord.length)
{
if (!WriteLowQualityRecord(smrtRecord,
hqEnd,
smrtRecord.length,
ScrapsReadGroupId(),
scrapsWriter))
{
smrtRecord.Free();
return false;
}
}
}
}
smrtRecord.Free();
}
// if we get here, all OK
return true;
}
int main(int argc, char* argv[]) {
string refGenomeFileName = "";
string lengthModelFileName = "";
string outputModelFileName = "";
DNALength numBasesPerFile = 0;
string sourceReadsFileName = "";
string titleTableFileName = "";
int numBasH5Files = 1;
string basH5BaseFileName = "simulated";
string movieName = "m101211_092754_00114_cSIM_s1_p0";
bool doRandGenInit = true;
bool usePosMap = false;
bool printPercentRepeat = false;
string posMapFileName = "";
vector<string> movieNames;
bool useLengthModel = false;
bool useFixedLength = false;
ofstream posMapFile;
int scaledLength = 0;
int fixedLength = 0;
int nBasFiles = 1;
bool useLengthsModel = true;
bool printHelp = false;
// Look to see if the refAsReads flag is specified anywhere before
// parsing the command line.
CommandLineParser clp;
string commandLine;
string helpString;
SetHelp(helpString);
vector<string> fns;
clp.RegisterStringOption("genome", &refGenomeFileName, "");
clp.RegisterIntOption("numBasesPerFile", (int*)&numBasesPerFile, "",
CommandLineParser::PositiveInteger);
clp.RegisterStringOption("sourceReads", &sourceReadsFileName, "");
clp.RegisterStringOption("lengthModel", &lengthModelFileName, "");
clp.RegisterIntOption("fixedLength", &fixedLength, "",
CommandLineParser::PositiveInteger);
clp.RegisterFlagOption("lengthModel", &useLengthModel, "");
clp.RegisterStringOption("movieName", &movieName, "");
clp.RegisterStringOption("titleTable", &titleTableFileName, "");
clp.RegisterStringOption("baseFileName", &basH5BaseFileName, "");
clp.RegisterIntOption("nFiles", &nBasFiles, "",
CommandLineParser::PositiveInteger);
clp.RegisterIntOption("meanLength", &scaledLength, "",
CommandLineParser::PositiveInteger);
clp.RegisterStringOption("posMap", &posMapFileName, "");
clp.RegisterFlagOption("printPercentRepeat", &printPercentRepeat, "");
clp.RegisterFlagOption("h", &printHelp, "");
clp.SetHelp(helpString);
clp.ParseCommandLine(argc, argv, fns);
clp.CommandLineToString(argc, argv, commandLine);
clp.SetProgramName("alchemy");
outputModelFileName = fns[0];
if (argc <= 1 or printHelp or outputModelFileName == "") {
cout << helpString << endl;
exit(0);
}
if (usePosMap) {
CrucialOpen(posMapFileName, posMapFile, std::ios::out);
}
if (sourceReadsFileName == "" and fixedLength == 0) {
useLengthModel = true;
}
if (useLengthModel and fixedLength != 0) {
cout << "ERROR! You must either use a length model or a fixed length." << endl;
exit(1);
}
if (sourceReadsFileName == "" and numBasesPerFile == 0) {
cout << "ERROR! You must specify either a set of read to use as " << endl
<< "original reads for simulation or the total number of bases " << endl
<< "to simulate in each bas.h5 file." << endl;
exit(1);
}
if (sourceReadsFileName == "" and refGenomeFileName == "") {
cout << "ERROR! You must specify a genome to sample reads from or a set of read "<<endl
<< "to use as original reads for simulation." << endl;
exit(1);
}
if (fixedLength != 0 and refGenomeFileName == "") {
cout << "ERROR! You must specify a genome file if using a fixed length." << endl;
exit(1);
}
if ((fixedLength != 0 or scaledLength != 0) and sourceReadsFileName != "") {
cout << "ERROR! You cannot specify a fixed length nor mean length with a source " << endl
<< "reads file. The read lengths are taken from the source reads or the length model." << endl;
exit(1);
}
LengthHistogram lengthHistogram;
OutputSampleListSet outputModel(0);
TitleTable titleTable;
if (doRandGenInit) {
InitializeRandomGeneratorWithTime();
}
//
// Read models.
//
if (titleTableFileName != "") {
titleTable.Read(titleTableFileName);
}
outputModel.Read(outputModelFileName);
if (useLengthModel) {
lengthHistogram.BuildFromAlignmentLengths(outputModel.lengths);
}
vector<int> alignmentLengths;
int meanAlignmentLength;
if (scaledLength != 0 and useLengthModel) {
//
// Scale the histogram so that the average length is 'scaledLength'.
//
// 1. Integrate histogram
long totalLength = 0;
long totalSamples = 0;
int hi;
for (hi = 0; hi < lengthHistogram.lengthHistogram.cdf.size()-1; hi++) {
int ni;
ni = lengthHistogram.lengthHistogram.cdf[hi+1] - lengthHistogram.lengthHistogram.cdf[hi];
totalLength += ni * lengthHistogram.lengthHistogram.data[hi];
}
totalSamples = lengthHistogram.lengthHistogram.cdf[lengthHistogram.lengthHistogram.cdf.size()-1];
float meanSampleLength = totalLength / (1.0*totalSamples);
float fractionIncrease = scaledLength / meanSampleLength;
for (hi = 0; hi < lengthHistogram.lengthHistogram.cdf.size(); hi++) {
lengthHistogram.lengthHistogram.data[hi] *= fractionIncrease;
}
}
FASTAReader inReader, seqReader;
vector<FASTASequence> reference;
DNALength refLength = 0;
int i;
if (refGenomeFileName != "") {
inReader.Init(refGenomeFileName);
inReader.ReadAllSequences(reference);
for (i = 0; i < reference.size(); i++) {
refLength += reference[i].length;
}
}
if (sourceReadsFileName != "") {
seqReader.Init(sourceReadsFileName);
}
ofstream readsFile;
//
// Create and simulate bas.h5 files.
//
int baseFileIndex;
bool readsRemain = true;
for (baseFileIndex = 0; ((sourceReadsFileName == "" and baseFileIndex < nBasFiles) // case 1 is reads are generated by file
or (sourceReadsFileName != "" and readsRemain)); // case 2 is reads are generated by an input file.
baseFileIndex++) {
//
// Prep the base file for writing.
//
stringstream fileNameStrm, movieNameStrm;
//string movieName = "m000000_000000_00000_cSIMULATED_s";
movieNameStrm << movieName << baseFileIndex << "_p0";
string fullMovieName = movieNameStrm.str();
fileNameStrm << fullMovieName << ".bas.h5";
HDFBasWriter basWriter;
HDFRegionTableWriter regionWriter;
//
// This is mainly used to create the atributes.
//
RegionTable regionTable;
regionTable.CreateDefaultAttributes();
basWriter.SetPlatform(Springfield);
//
// Use a fixed set of fields for now.
//
// These are all pulled from the outputModel.
basWriter.IncludeField("Basecall");
basWriter.IncludeField("QualityValue");
basWriter.IncludeField("SubstitutionQV");
basWriter.IncludeField("SubstitutionTag");
basWriter.IncludeField("InsertionQV");
basWriter.IncludeField("DeletionQV");
basWriter.IncludeField("DeletionTag");
basWriter.IncludeField("WidthInFrames");
basWriter.IncludeField("PreBaseFrames");
basWriter.IncludeField("PulseIndex");
vector<unsigned char> qualityValue, substitutionQV, substitutionTag, insertionQV, deletionQV, deletionTag;
vector<HalfWord> widthInFrames, preBaseFrames, pulseIndex;
// Just go from 0 .. hole Number
basWriter.IncludeField("HoleNumber");
// Fixed to 0.
basWriter.IncludeField("HoleXY");
if (usePosMap == false) {
basWriter.IncludeField("SimulatedSequenceIndex");
basWriter.IncludeField("SimulatedCoordinate");
}
basWriter.SetChangeListID("1.3.0.50.104380");
DNALength numSimulatedBases = 0;
FASTASequence sampleSeq;
//sampleSeq.length = readLength;
int maxRetry = 10000000;
int retryNumber = 0;
int numReads = 0;
int readLength = 0;
while (numBasesPerFile == 0 or numSimulatedBases < numBasesPerFile) {
DNALength seqIndex, seqPos;
if (useLengthModel or fixedLength) {
if (useLengthModel) {
lengthHistogram.GetRandomLength(readLength);
}
else {
readLength = fixedLength;
}
}
if (refGenomeFileName != "") {
FindRandomPos(reference, seqIndex, seqPos, readLength + (outputModel.keyLength - 1));
sampleSeq.seq = &reference[seqIndex].seq[seqPos];
sampleSeq.length = readLength + (outputModel.keyLength - 1);
assert(reference[seqIndex].length >= sampleSeq.length);
}
else if (sourceReadsFileName != "") {
if (seqReader.GetNext(sampleSeq) == false) {
readsRemain = false;
break;
}
if (sampleSeq.length < outputModel.keyLength) {
continue;
}
//
// Now attempt to parse the position from the fasta title.
//
if (useLengthModel) {
int tryNumber = 0;
readLength = 0;
int maxNTries = 1000;
int tryBuffer[5] = {-1,-1,-1,-1,-1};
while (tryNumber < maxNTries and readLength < outputModel.keyLength) {
lengthHistogram.GetRandomLength(readLength);
readLength = sampleSeq.length = min(sampleSeq.length, (unsigned int) readLength);
tryBuffer[tryNumber%5] = readLength;
tryNumber++;
}
if (tryNumber >= maxNTries) {
cout << "ERROR. Could not generate a read length greater than the " << outputModel.keyLength << " requried " <<endl
<< "minimum number of bases using the length model specified in the alchemy." <<endl
<< "model. Something is either wrong with the model or the context length is too large." <<endl;
cout << "The last few tries were: " << tryBuffer[0] << " " << tryBuffer[1] << " " << tryBuffer[2] << " " << tryBuffer[3] << " " << tryBuffer[4] << endl;
exit(1);
}
}
readLength = sampleSeq.length;
vector<string> tokens;
Tokenize(sampleSeq.title, "|", tokens);
if (tokens.size() == 4) {
seqPos = atoi(tokens[2].c_str());
if (titleTableFileName == "") {
seqIndex = 0;
}
else {
int index;
titleTable.Lookup(tokens[1], index);
seqIndex = index;
}
}
else {
seqPos = 0;
}
}
//
// If this is the first read printed to the base file, initialize it.
//
if (numSimulatedBases == 0) {
basWriter.Initialize(fileNameStrm.str(), movieNameStrm.str(), Springfield);
regionWriter.Initialize(basWriter.pulseDataGroup);
}
numSimulatedBases += readLength;
int p;
// create the sample sequence
int contextLength = outputModel.keyLength;
int contextMiddle = contextLength / 2;
string outputString;
int nDel = 0;
int nIns = 0;
//
// Simulate to beyond the sample length.
//
qualityValue.clear();
substitutionQV.clear();
substitutionTag.clear();
insertionQV.clear();
deletionQV.clear();
deletionTag.clear();
pulseIndex.clear();
widthInFrames.clear();
preBaseFrames.clear();
assert(sampleSeq.length > contextMiddle + 1);
for (p = contextMiddle;
p < sampleSeq.length - contextMiddle - 1; p++) {
string refContext;
refContext.assign((const char*) &sampleSeq.seq[p-contextMiddle], contextLength);
string outputContext;
int contextWasFound;
OutputSample sample;
int i;
for (i = 0; i < refContext.size(); i++) {
refContext[i] = toupper(refContext[i]);
}
outputModel.SampleRandomSample(refContext, sample);
if (sample.type == OutputSample::Deletion ) {
//
// There was a deletion. Advance in reference, then output
// the base after the deletion.
//
p++;
++nDel;
}
int cp;
//
// Add the sampled context, possibly multiple characters because of an insertion.
//
for (i = 0; i < sample.nucleotides.size(); i++) {
outputString.push_back(sample.nucleotides[i]);
qualityValue.push_back(sample.qualities[i].qv[0]);
deletionQV.push_back(sample.qualities[i].qv[1]);
insertionQV.push_back(sample.qualities[i].qv[2]);
substitutionQV.push_back(sample.qualities[i].qv[3]);
deletionTag.push_back(sample.qualities[i].tags[0]);
substitutionTag.push_back(sample.qualities[i].tags[1]);
pulseIndex.push_back(sample.qualities[i].frameValues[0]);
preBaseFrames.push_back(sample.qualities[i].frameValues[1]);
widthInFrames.push_back(sample.qualities[i].frameValues[2]);
}
nIns += sample.qualities.size() - 1;
}
if (outputString.find('N') != outputString.npos or
outputString.find('n') != outputString.npos) {
cout << "WARNING! The sampled string " << endl << outputString << endl
<< "should not contain N's, but it seems to. This is being ignored "<<endl
<< "for now so that simulation may continue, but this shouldn't happen"<<endl
<< "and is really a bug." << endl;
numSimulatedBases -= readLength;
continue;
}
//
// Ok, done creating the read, now time to create some quality values!!!!!
//
SMRTSequence read;
read.length = outputString.size();
read.Allocate(read.length);
memcpy(read.seq, outputString.c_str(), read.length * sizeof(unsigned char));
assert(qualityValue.size() == read.length * sizeof(unsigned char));
memcpy(read.qual.data, &qualityValue[0], read.length * sizeof(unsigned char));
memcpy(read.deletionQV.data, &deletionQV[0], read.length * sizeof(unsigned char));
memcpy(read.insertionQV.data, &insertionQV[0], read.length * sizeof(unsigned char));
memcpy(read.substitutionQV.data, &substitutionQV[0], read.length * sizeof(unsigned char));
memcpy(read.deletionTag, &deletionTag[0], read.length * sizeof(unsigned char));
memcpy(read.substitutionTag, &substitutionTag[0], read.length * sizeof(unsigned char));
memcpy(read.pulseIndex, &pulseIndex[0], read.length * sizeof(int));
memcpy(read.preBaseFrames, &preBaseFrames[0], read.length * sizeof(HalfWord));
memcpy(read.widthInFrames, &widthInFrames[0], read.length * sizeof(HalfWord));
//
// The pulse index for now is just fake data.
//
int i;
for (i = 0; i < read.length; i++) {
read.pulseIndex[i] = 1;
}
read.xy[0] = seqIndex;
read.xy[1] = seqPos;
read.zmwData.holeNumber = numReads;
basWriter.Write(read);
// Record where this was simulated from.
if (usePosMap == false) {
basWriter.WriteSimulatedCoordinate(seqPos);
basWriter.WriteSimulatedSequenceIndex(seqIndex);
}
else {
posMapFile << fullMovieName << "/" << numReads << "/0_" << read.length << " " << seqIndex << " "<< seqPos;
if (printPercentRepeat) {
DNALength nRepeat = sampleSeq.GetRepeatContent();
posMapFile << " " << nRepeat*1.0/sampleSeq.length;
}
posMapFile << endl;
}
RegionAnnotation region;
region.row[0] = read.zmwData.holeNumber;
region.row[1] = 1;
region.row[2] = 0;
region.row[3] = read.length;
region.row[4] = 1000; // Should be enough.
regionWriter.Write(region);
region.row[1] = 2; // Rewrite for hq region encompassing everything.
regionWriter.Write(region);
if (sourceReadsFileName != "") {
sampleSeq.Free();
}
read.Free();
++numReads;
}
regionWriter.Finalize(regionTable.columnNames,
regionTable.regionTypes,
regionTable.regionDescriptions,
regionTable.regionSources);
basWriter.Close();
numReads = 0;
//
// The bas writer should automatically flush on closing.
//
}
if (usePosMap) {
posMapFile.close();
}
for (i = 0; i < reference.size(); i++) {
reference[i].Free();
}
}
int main(int argc, char* argv[]) {
string inputFileName, outputFileName;
if (argc < 2) {
PrintUsage();
exit(0);
}
vector<string> inputFileNames;
inputFileName = argv[1];
outputFileName = argv[2];
int argi = 3;
RegionTable regionTable;
string regionsFOFNName = "";
vector<string> regionFileNames;
bool splitSubreads = true;
bool useCCS = false;
int minSubreadLength = 1;
while (argi < argc) {
if (strcmp(argv[argi], "-regionTable") == 0) {
regionsFOFNName = argv[++argi];
}
else if (strcmp(argv[argi], "-noSplitSubreads") == 0) {
splitSubreads = false;
}
else if (strcmp(argv[argi], "-minSubreadLength") == 0) {
minSubreadLength = atoi(argv[++argi]);
}
else if (strcmp(argv[argi], "-useccsdenovo") == 0) {
useCCS = true;
}
else {
PrintUsage();
cout << "ERROR! Option " << argv[argi] << " is not supported." << endl;
}
argi++;
}
if (FileOfFileNames::IsFOFN(inputFileName)) {
FileOfFileNames::FOFNToList(inputFileName, inputFileNames);
}
else {
inputFileNames.push_back(inputFileName);
}
if (regionsFOFNName == "") {
regionFileNames = inputFileNames;
}
else {
if (FileOfFileNames::IsFOFN(regionsFOFNName)) {
FileOfFileNames::FOFNToList(regionsFOFNName, regionFileNames);
}
else {
regionFileNames.push_back(regionsFOFNName);
}
}
ofstream fastaOut;
CrucialOpen(outputFileName, fastaOut);
int plsFileIndex;
HDFRegionTableReader hdfRegionReader;
AfgBasWriter afgWriter;
afgWriter.Initialize(outputFileName);
for (plsFileIndex = 0; plsFileIndex < inputFileNames.size(); plsFileIndex++) {
if (splitSubreads) {
hdfRegionReader.Initialize(regionFileNames[plsFileIndex]);
hdfRegionReader.ReadTable(regionTable);
regionTable.SortTableByHoleNumber();
}
ReaderAgglomerate reader;
// reader.SkipReadQuality(); // should have been taken care of by *Filter modules
if (useCCS){
reader.UseCCS();
} else {
reader.IgnoreCCS();
}
reader.Initialize(inputFileNames[plsFileIndex]);
CCSSequence seq;
int seqIndex = 0;
int numRecords = 0;
vector<ReadInterval> subreadIntervals;
while (reader.GetNext(seq)){
++seqIndex;
if (splitSubreads == false) {
if (seq.length >= minSubreadLength) {
afgWriter.Write(seq);
}
seq.Free();
continue;
}
DNALength hqReadStart, hqReadEnd;
int score;
GetReadTrimCoordinates(seq, seq.zmwData, regionTable, hqReadStart, hqReadEnd, score);
subreadIntervals.clear(); // clear old, new intervals are appended.
CollectSubreadIntervals(seq,®ionTable, subreadIntervals);
if (seq.length == 0 and subreadIntervals.size() > 0) {
cout << "WARNING! A high quality interval region exists for a read of length 0." <<endl;
cout << " The offending ZMW number is " << seq.zmwData.holeNumber << endl;
seq.Free();
continue;
}
for (int intvIndex = 0; intvIndex < subreadIntervals.size(); intvIndex++) {
SMRTSequence subreadSequence;
int subreadStart = subreadIntervals[intvIndex].start > hqReadStart ?
subreadIntervals[intvIndex].start : hqReadStart;
int subreadEnd = subreadIntervals[intvIndex].end < hqReadEnd ?
subreadIntervals[intvIndex].end : hqReadEnd;
int subreadLength = subreadEnd - subreadStart;
if (subreadLength < minSubreadLength) continue;
subreadSequence.subreadStart = subreadStart;
subreadSequence.subreadEnd = subreadEnd;
subreadSequence.ReferenceSubstring(seq, subreadStart, subreadLength);
stringstream titleStream;
titleStream << seq.title << "/" << subreadIntervals[intvIndex].start
<< "_" << subreadIntervals[intvIndex].end;
subreadSequence.CopyTitle(titleStream.str());
afgWriter.Write(subreadSequence);
delete[] subreadSequence.title;
}
seq.Free();
}
reader.Close();
hdfRegionReader.Close();
}
}
