void Bam2FastQ::closeFiles()
{
// NULL out any duplicate file pointers
// so files are only closed once.
if(myFirstFile == myUnpairedFile)
{
myFirstFile = NULL;
}
if(mySecondFile == myUnpairedFile)
{
mySecondFile = NULL;
}
if(mySecondFile == myFirstFile)
{
mySecondFile = NULL;
}
if(myUnpairedFile != NULL)
{
ifclose(myUnpairedFile);
myUnpairedFile = NULL;
}
if(myFirstFile != NULL)
{
ifclose(myFirstFile);
myFirstFile = NULL;
}
if(mySecondFile != NULL)
{
ifclose(mySecondFile);
mySecondFile = NULL;
}
}
void Bam2FastQ::closeFiles()
{
// NULL out any duplicate file pointers
// so files are only closed once.
if(myFirstFile == myUnpairedFile)
{
myFirstFile = NULL;
}
if(mySecondFile == myUnpairedFile)
{
mySecondFile = NULL;
}
if(mySecondFile == myFirstFile)
{
mySecondFile = NULL;
}
if(myUnpairedFile != NULL)
{
ifclose(myUnpairedFile);
myUnpairedFile = NULL;
}
if(myFirstFile != NULL)
{
ifclose(myFirstFile);
myFirstFile = NULL;
}
if(mySecondFile != NULL)
{
ifclose(mySecondFile);
mySecondFile = NULL;
}
if(myFqList != NULL)
{
ifclose(myFqList);
myFqList = NULL;
}
// Loop through the fastq map and close those files.
for (OutFastqMap::iterator it=myOutFastqs.begin();
it!=myOutFastqs.end(); ++it)
{
ifclose(it->second);
it->second = NULL;
}
myOutFastqs.clear();
}
// Close a FastQFile.
FastQStatus::Status FastQFile::closeFile()
{
int closeStatus = 0; // Success.
// If a file has been opened, close it.
if(myFile != NULL)
{
// Close the file.
closeStatus = ifclose(myFile);
myFile = NULL;
}
if(closeStatus == 0)
{
// Success - either there wasn't a file to close or it was closed
// successfully.
return(FastQStatus::FASTQ_SUCCESS);
}
else
{
std::string errorMessage = "Failed to close file: ";
errorMessage += myFileName.c_str();
logMessage(errorMessage.c_str());
return(FastQStatus::FASTQ_CLOSE_ERROR);
}
}
void testReadLine()
{
IFILE filePtr = ifopen("testFiles/testFile.txt", "rb");
assert(filePtr != NULL);
String line = "";
line.ReadLine(filePtr);
assert(line == " Hello, I am a testFile. ");
line.Trim();
assert(line == "Hello, I am a testFile.");
// Does not compile in current version, but compiles in old verison.
// This can be added back in to ensure that it will catch the difference
// in return value for ReadLine (now: int; used to be: string&)
// testMethod(line.ReadLine(filePtr));
line.ReadLine(filePtr);
assert(temp1 == 0);
testMethod(line);
assert(temp1 == 1);
// line.ReadLine(filePtr).Trim();
line.ReadLine(filePtr);
line.Trim();
assert(line == "ThirdLine.");
ifclose(filePtr);
}
// Reset variables for each file.
void GlfFile::resetFile()
{
// Close the file.
if (myFilePtr != NULL)
{
// If we already have an open file, close it.
// First check to see if an end record needs to be written, which
// is the case if the state is RECORD.
if(myNextSection == RECORD)
{
if(!writeRecord(myEndMarker))
{
// Failed to write the end marker record.
myStatus.setStatus(GlfStatus::FAIL_IO,
"Failed to write end of chromosome/section marker.");
throw(GlfException(myStatus));
}
}
ifclose(myFilePtr);
myFilePtr = NULL;
}
myIsOpenForRead = false;
myIsOpenForWrite = false;
myRecordCount = 0;
myStatus = GlfStatus::SUCCESS;
myNextSection = HEADER;
}
void StringHash::ReadLinesFromFile(const char * filename)
{
IFILE f = ifopen(filename, "rb");
if (f == NULL) return;
ReadLinesFromFile(f);
ifclose(f);
}
BedFile::~BedFile() {
if ( iBimFile != NULL ) {
ifclose(iBimFile);
}
iBimFile = NULL;
if ( iFamFile != NULL ) {
ifclose(iFamFile);
}
iFamFile = NULL;
if ( iFile != NULL ) {
ifclose(iFile);
}
if ( pBedBuffer != NULL ) {
delete[] pBedBuffer;
}
}
void Imputation::PrintInfoFile(HaplotypeSet &rHap,HaplotypeSet &tHap, ImputationStatistics &stats)
{
cout<<endl<<" Writing summary (.info) files ... "<<endl;
IFILE info = ifopen(outFile + ".info", "wb");
ifprintf(info, "SNP\tREF(0)\tALT(1)\tALT_Frq\tMAF\tAvgCall\tRsq\tGenotyped\tLooRsq\tEmpR\tEmpRsq\tDose0\tDose1\n");
int i=0;
for (int index =0; index < rHap.RefTypedTotalCount; index++)
{
if(rHap.RefTypedIndex[index]==-1)
{
if(i>=rHap.PrintStartIndex && i <= rHap.PrintEndIndex)
{
ifprintf(info, "%s\t%s\t%s\t%.5f\t%.5f\t%.5f\t%.5f\t",
RsId? rHap.VariantList[i].rsid.c_str(): rHap.VariantList[i].name.c_str(),
rHap.VariantList[i].refAlleleString.c_str(),
rHap.VariantList[i].altAlleleString.c_str(),
stats.AlleleFrequency(i),
stats.AlleleFrequency(i) > 0.5 ? 1.0 - stats.AlleleFrequency(i) : stats.AlleleFrequency(i),
stats.AverageCallScore(i),
stats.Rsq(i));
if (!tHap.missing[i])
{
ifprintf(info, "Genotyped\t%.3f\t%.3f\t%.5f\t%.5f\t%.5f\n",
stats.LooRsq(i), stats.EmpiricalR(i), stats.EmpiricalRsq(i),
stats.LooMajorDose(i), stats.LooMinorDose(i));
}
else
ifprintf(info, "Imputed\t-\t-\t-\t-\t-\n");
}
i++;
}
else
{
variant ThisTypedVariant =tHap.TypedOnlyVariantList[rHap.RefTypedIndex[index]];
ifprintf(info, "%s\t%s\t%s\t%.5f\t%.5f\t-\t-\tTyped_Only\t-\t-\t-\t-\t-\n",
RsId? ThisTypedVariant.rsid.c_str(): ThisTypedVariant.name.c_str(),
ThisTypedVariant.refAlleleString.c_str(),
ThisTypedVariant.altAlleleString.c_str(),
tHap.AlleleFreq[rHap.RefTypedIndex[index]],
tHap.AlleleFreq[rHap.RefTypedIndex[index]] > 0.5 ?
1.0 - tHap.AlleleFreq[rHap.RefTypedIndex[index]] : tHap.AlleleFreq[rHap.RefTypedIndex[index]]);
}
}
ifclose(info);
cout<<endl<<" Summary information written to : "<<outFile<<".info"<<endl;
}
void Imputation::FlushPartialVcf(HaplotypeSet &rHap,HaplotypeSet &tHap,HaplotypeSet &PartialDosage, string &filename,int &Index)
{
string tempFileIndex(outFile),tempFileIndex1(outFile);
IFILE vcfdosepartial = ifopen(filename.c_str(), "wb", InputFile::BGZF);
for(int hapId=0;hapId<(int)PartialDosage.individualName.size();hapId++)
{
ifprintf(vcfdosepartial,"\t%s",PartialDosage.individualName[hapId].c_str());
}
ifprintf(vcfdosepartial,"\n");
int i=0;
for (int index =0; index < rHap.RefTypedTotalCount; index++)
{
if(rHap.RefTypedIndex[index]==-1)
{
if(i>=rHap.PrintStartIndex && i <= rHap.PrintEndIndex)
{
bool majorIsReference=false;
if(!rHap.major[i])
majorIsReference=true;
if(!tHap.AllMaleTarget)
PartialDosage.PrintDosageForVcfOutputForID(vcfdosepartial,i, majorIsReference,rHap.VariantList[i].refAllele);
else
PartialDosage.PrintDosageForVcfOutputForIDMaleSamples(vcfdosepartial,i, majorIsReference,rHap.VariantList[i].refAllele);
ifprintf(vcfdosepartial,"\n");
}
i++;
}
else
{
if(!tHap.AllMaleTarget)
PartialDosage.PrintDosageGWASOnlyForVcfOutputForID
(tHap,vcfdosepartial,rHap.RefTypedIndex[index]);
else
PartialDosage.PrintDosageGWASOnlyForVcfOutputForIDMaleSamples
(tHap,vcfdosepartial,rHap.RefTypedIndex[index]);
ifprintf(vcfdosepartial,"\n");
}
}
ifclose(vcfdosepartial);
}
VcfFile::~VcfFile()
{
// Close the file.
if (myFilePtr != NULL)
{
// If we already have an open file, close it.
ifclose(myFilePtr);
myFilePtr = NULL;
}
}
void GCContent::LoadRegions(String & regionsFile, GenomeSequence &genome, bool invertRegion)
{
if(regionsFile.Length()==0) return;
if(genome.sequenceLength()==0) error("No reference genome loaded!\n");
IFILE fhRegions;
fhRegions = ifopen(regionsFile.c_str(),"r");
if(fhRegions==NULL)
error("Open regions file %s failed!\n", regionsFile.c_str());
regionIndicator.resize(genome.sequenceLength());
StringArray tokens;
String buffer;
int len;
fprintf(stderr, "Loading region list...");
while (!ifeof(fhRegions)){
buffer.ReadLine(fhRegions);
if (buffer.IsEmpty() || buffer[0] == '#') continue;
tokens.AddTokens(buffer, WHITESPACE);
if(tokens.Length() < 3) continue;
genomeIndex_t startGenomeIndex = 0;
int chromosomeIndex = tokens[1].AsInteger();
// use chromosome name (token[0]) and position (token[1]) to query genome index.
startGenomeIndex = genome.getGenomePosition(tokens[0].c_str(), chromosomeIndex);
if(startGenomeIndex >= regionIndicator.size() ) {
//fprintf(stderr, "WARNING: region list section %s position %u is not found in the reference and skipped...\n", tokens[0].c_str(), chromosomeIndex);
continue;
}
len = tokens[2].AsInteger() - tokens[1].AsInteger() + 1;
for(uint32_t i=startGenomeIndex; i<startGenomeIndex+len; i++)
regionIndicator[i] = true;
tokens.Clear();
buffer.Clear();
}
if (invertRegion) {
fprintf(stderr, " invert region...");
for (uint32_t i = 0; i < regionIndicator.size(); i++) {
regionIndicator[i] = !regionIndicator[i];
}
}
ifclose(fhRegions);
fprintf(stderr, "DONE!\n");
}
void MarkovParameters::WriteErrorRates(StringArray & markerNames, const char * filename)
{
IFILE output = ifopen(filename, "wb");
if (output == NULL) return;
ifprintf(output, "MarkerName\tErrorRate\n");
for (int i = 0; i < markers; i++)
ifprintf(output, "%s\t%.5g\n", (const char *) markerNames[i], E[i]);
ifclose(output);
}
void glfHandler::Rewind()
{
if (isOpen())
{
ifrewind(handle);
if (!ReadHeader())
ifclose(handle);
endOfSection = true;
}
}
bool StringAlias::ReadFromFile(const char * filename)
{
IFILE input = ifopen(filename, "rt");
if (input == NULL)
return false;
ReadFromFile(input);
ifclose(input);
return true;
}
void MarkovParameters::WriteCrossoverRates(StringArray & markerNames, const char * filename)
{
IFILE output = ifopen(filename, "wb");
if (output == NULL) return;
ifprintf(output, "Interval\tSwitchRate\n");
for (int i = 0; i < markers - 1; i++)
ifprintf(output, "%s-%s\t%.5g\n",
(const char *) markerNames[i],
(const char *) markerNames[i+1], R[i]);
ifclose(output);
}
void VcfFile::reset()
{
// Reset the child class.
resetFile();
// Close the file.
if (myFilePtr != NULL)
{
// If we already have an open file, close it.
ifclose(myFilePtr);
myFilePtr = NULL;
}
myNumRecords = 0;
}
EXTERN_C void sread_atf(HDRTYPE* hdr) {
if (VERBOSE_LEVEL>6) fprintf(stdout,"SREAD ATF [%i,%i]\n",(unsigned)hdr->NRec, (unsigned)hdr->SPR);
if (hdr->AS.rawdata != NULL) return;
//if (hdr->NRec * hdr->SPR > 0)
hdr->AS.rawdata = malloc(hdr->NRec * hdr->SPR * hdr->AS.bpb);
ifseek(hdr, hdr->HeadLen, SEEK_SET);
size_t ll;
char *line = NULL;
if (VERBOSE_LEVEL>6) fprintf(stdout,"SREAD ATF\n");
size_t ln = 0;
while (~ifeof(hdr)) {
if (line!=NULL) { free(line); line=NULL; } // allocate line buffer as needed
ssize_t nc = getline(&line, &ll, hdr->FILE.FID);
if (nc < 0) break;
if (VERBOSE_LEVEL>8) fprintf(stdout,"SREAD ATF 2 %i\t<%s>\n",(unsigned)ln,line );
if ((hdr->NRec * hdr->SPR) <= (ln+1) ) {
hdr->NRec = max(1024, ln*2);
hdr->AS.rawdata = realloc(hdr->AS.rawdata, hdr->NRec * hdr->SPR * hdr->AS.bpb);
}
if (VERBOSE_LEVEL>8) fprintf(stdout,"SREAD ATF 4 %i\t<%s>\n",(unsigned)ln,line );
char *str = strtok(line,"\t");
typeof(hdr->NS) k;
for (k = 0; k < hdr->NS; k++) {
*(double*)(hdr->AS.rawdata + ln*hdr->AS.bpb + hdr->CHANNEL[k].bi) = strtod(str, &str);
// extract next value
//str = strtok(NULL,"\t");
}
if (VERBOSE_LEVEL>8) fprintf(stdout,"SREAD ATF 6 %i\t<%s>\n",(unsigned)ln,line );
ln++;
}
free(line);
ifclose(hdr);
hdr->NRec = ln;
hdr->AS.first = 0;
hdr->AS.length = hdr->NRec;
}
// Reset variables for each file.
void SamFile::resetFile()
{
// Close the file.
if (myFilePtr != NULL)
{
// If we already have an open file, close it.
ifclose(myFilePtr);
myFilePtr = NULL;
}
if(myInterfacePtr != NULL)
{
delete myInterfacePtr;
myInterfacePtr = NULL;
}
myIsOpenForRead = false;
myIsOpenForWrite = false;
myHasHeader = false;
mySortedType = UNSORTED;
myPrevReadName.Clear();
myPrevCoord = -1;
myPrevRefID = 0;
myRecordCount = 0;
myStatus = SamStatus::SUCCESS;
// Reset indexed bam values.
myIsBamOpenForRead = false;
myRefID = BamIndex::REF_ID_ALL;
myStartPos = -1;
myEndPos = -1;
myNewSection = false;
myOverlapSection = true;
myCurrentChunkEnd = 0;
myChunksToRead.clear();
if(myBamIndex != NULL)
{
delete myBamIndex;
myBamIndex = NULL;
}
// If statistics are being generated, reset them.
if(myStatistics != NULL)
{
myStatistics->reset();
}
myRefName.clear();
}
void put_cde_slots(IFILE *file, int ifaxq) /*;put_cde_slots*/
{
long dpos;
dpos = iftell(file); /* get current position */
putnum(file, "n-code_slots", tup_size(CODE_SLOTS));
putnum(file, "n-data-slots", tup_size(DATA_SLOTS));
putnum(file, "n-exception-slots", tup_size(EXCEPTION_SLOTS));
put_slot(file, CODE_SLOTS);
put_slot(file, DATA_SLOTS);
put_slot(file, EXCEPTION_SLOTS);
/* now replace word at start of file with long giving offset to
*start of information just written.
*/
file->fh_slots = dpos;
ifclose(file);
}
void VcfFile::reset() {
if ( iFile != NULL )
ifclose(iFile);
iFile = NULL;
for(int i=0; i < (int) vpVcfInds.size(); ++i) {
delete vpVcfInds[i];
}
vpVcfInds.clear();
for(int i=0; i < (int) vpVcfMarkers.size(); ++i) {
delete vpVcfMarkers[i];
}
vpVcfMarkers.clear();
asMetaKeys.Clear();
asMetaValues.Clear();
nNumLines = 0;
}
bool glfHandler::Open(const String & filename)
{
isStub = false;
handle = ifopen(filename, "rb");
if (handle == NULL)
{
isStub = true;
return false;
}
if (!ReadHeader())
ifclose(handle);
endOfSection = true;
return handle != NULL;
}
bool FilterStat::writeMergedVcf(const char* outFile) {
IFILE oFile = ifopen(outFile,"wb");
if ( oFile == NULL ) {
Logger::gLogger->error("Cannot open output file %s",outFile);
}
VcfFile vcf;
vcf.setSiteOnly(false);
vcf.setParseValues(true);
vcf.openForRead(sAnchorVcf.c_str(),1);
vcf.printVCFHeader(oFile);
VcfMarker* pMarker;
String STC, STR;
for( int i=0; vcf.iterateMarker(); ++i ) {
pMarker = vcf.getLastMarker();
int c[FILTER_STAT_COUNTS];
for(int j=0; j < FILTER_STAT_COUNTS; ++j) {
c[j] = vCounts[FILTER_STAT_COUNTS*i+j];
}
STC.printf("%d,%d,%d,%d,%d,%d",c[0],c[1],c[2],c[3],c[4],c[5]);
if ( ( c[0]+c[1] > 4 ) && ( c[1]+c[3] > 4 ) && ( c[0]+c[2] > 4 ) && ( c[1]+c[3] > 4 ) ) {
STR.printf("%.2lf",((c[0]+.5)*(c[3]+.5)-(c[1]+.5)*(c[2]+.5))/sqrt((c[0]+c[1]+1.)*(c[2]+c[3]+1.)*(c[0]+c[2]+1.)*(c[1]+c[3]+1.)));
}
else {
STR = "0";
}
pMarker->asInfoKeys.Add("STC");
pMarker->asInfoKeys.Add("STR");
pMarker->asInfoValues.Add(STC);
pMarker->asInfoValues.Add(STR);
pMarker->printVCFMarker(oFile,false);
}
ifclose(oFile);
return true;
}
// Read & parse the specified index file.
SamStatus::Status BamIndex::readIndex(const char* filename)
{
// Reset the index from anything that may previously be set.
resetIndex();
IFILE indexFile = ifopen(filename, "rb");
// Failed to open the index file.
if(indexFile == NULL)
{
return(SamStatus::FAIL_IO);
}
// generate the bam index structure.
// Read the magic string.
char magic[4];
if(ifread(indexFile, magic, 4) != 4)
{
// Failed to read the magic
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// If this is not an index file, set num references to 0.
if (magic[0] != 'B' || magic[1] != 'A' || magic[2] != 'I' || magic[3] != 1)
{
// Not a BAM Index file.
ifclose(indexFile);
return(SamStatus::FAIL_PARSE);
}
// It is a bam index file.
// Read the number of reference sequences.
if(ifread(indexFile, &n_ref, 4) != 4)
{
// Failed to read.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// Size the references.
myRefs.resize(n_ref);
for(int refIndex = 0; refIndex < n_ref; refIndex++)
{
// Read each reference.
Reference* ref = &(myRefs[refIndex]);
// Read the number of bins.
if(ifread(indexFile, &(ref->n_bin), 4) != 4)
{
// Failed to read the number of bins.
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_PARSE);
}
// If there are no bins, then there are no
// mapped/unmapped reads.
if(ref->n_bin == 0)
{
ref->n_mapped = 0;
ref->n_unmapped = 0;
}
// Resize the bins so they can be indexed by bin number.
ref->bins.resize(ref->n_bin + 1);
// Read each bin.
for(int binIndex = 0; binIndex < ref->n_bin; binIndex++)
{
uint32_t binNumber;
// Read in the bin number.
if(ifread(indexFile, &(binNumber), 4) != 4)
{
// Failed to read the bin number.
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// Add the bin to the reference and get the
// pointer back so the values can be set in it.
Bin* binPtr = &(ref->bins[binIndex]);
binPtr->bin = binNumber;
// Read in the number of chunks.
if(ifread(indexFile, &(binPtr->n_chunk), 4) != 4)
{
// Failed to read number of chunks.
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// Read in the chunks.
// Allocate space for the chunks.
uint32_t sizeOfChunkList = binPtr->n_chunk * sizeof(Chunk);
binPtr->chunks = (Chunk*)malloc(sizeOfChunkList);
if(ifread(indexFile, binPtr->chunks, sizeOfChunkList) != sizeOfChunkList)
{
// Failed to read the chunks.
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// Determine the min/max for this bin if it is not the max bin.
if(binPtr->bin != MAX_NUM_BINS)
{
for(int i = 0; i < binPtr->n_chunk; i++)
{
if(binPtr->chunks[i].chunk_beg < ref->minChunkOffset)
{
ref->minChunkOffset = binPtr->chunks[i].chunk_beg;
}
if(binPtr->chunks[i].chunk_end > ref->maxChunkOffset)
{
ref->maxChunkOffset = binPtr->chunks[i].chunk_end;
}
if(binPtr->chunks[i].chunk_end > maxOverallOffset)
{
maxOverallOffset = binPtr->chunks[i].chunk_end;
}
}
}
else
{
// Mapped/unmapped are the last chunk of the
// MAX BIN
ref->n_mapped = binPtr->chunks[binPtr->n_chunk - 1].chunk_beg;
ref->n_unmapped = binPtr->chunks[binPtr->n_chunk - 1].chunk_end;
}
}
// Read the number of intervals.
if(ifread(indexFile, &(ref->n_intv), 4) != 4)
{
// Failed to read, set to 0.
ref->n_intv = 0;
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
// Allocate space for the intervals and read them.
uint32_t linearIndexSize = ref->n_intv * sizeof(uint64_t);
ref->ioffsets = (uint64_t*)malloc(linearIndexSize);
if(ifread(indexFile, ref->ioffsets, linearIndexSize) != linearIndexSize)
{
// Failed to read the linear index.
// Return failure.
ifclose(indexFile);
return(SamStatus::FAIL_IO);
}
}
int32_t numUnmapped = 0;
if(ifread(indexFile, &numUnmapped, sizeof(int32_t)) == sizeof(int32_t))
{
myUnMappedNumReads = numUnmapped;
}
// Successfully read the bam index file.
ifclose(indexFile);
return(SamStatus::SUCCESS);
}
int main(int argc, char ** argv)
{
setbuf(stdout, NULL);
time_t start = time(NULL);
printf("MiniMac - Imputation into phased haplotypes\n"
"(c) 2011 Goncalo Abecasis\n");
#ifdef __VERSION__
printf("VERSION 5.0\n");
#else
printf("UNDOCUMENTED RELEASE\n");
#endif
int rounds = 5, states = 200, cpus = 0;
bool em = false, gzip = false, phased = false;
String referenceHaplotypes, referenceSnps;
String haplotypes, snps;
String prefix("minimac");
String firstMarker, lastMarker;
String recombinationRates, errorRates;
BEGIN_LONG_PARAMETERS(longParameters)
LONG_PARAMETER_GROUP("Reference Haplotypes")
LONG_STRINGPARAMETER("refHaps", &referenceHaplotypes)
LONG_STRINGPARAMETER("refSnps", &referenceSnps)
LONG_PARAMETER_GROUP("Target Haplotypes")
LONG_STRINGPARAMETER("haps", &haplotypes)
LONG_STRINGPARAMETER("snps", &snps)
LONG_PARAMETER_GROUP("Starting Parameters")
LONG_STRINGPARAMETER("rec", &recombinationRates)
LONG_STRINGPARAMETER("erate", &errorRates)
LONG_PARAMETER_GROUP("Parameter Fitting")
LONG_INTPARAMETER("rounds", &rounds)
LONG_INTPARAMETER("states", &states)
LONG_PARAMETER("em", &em)
LONG_PARAMETER_GROUP("Output Files")
LONG_STRINGPARAMETER("prefix", &prefix)
LONG_PARAMETER("phased", &phased)
LONG_PARAMETER("gzip", &gzip)
// LONG_PARAMETER_GROUP("Clipping Window")
// LONG_STRINGPARAMETER("start", &firstMarker)
// LONG_STRINGPARAMETER("stop", &lastMarker)
#ifdef _OPENMP
LONG_PARAMETER_GROUP("Multi-Threading")
LONG_INTPARAMETER("cpus", &cpus)
#endif
END_LONG_PARAMETERS();
ParameterList pl;
pl.Add(new LongParameters("Command Line Options", longParameters));
pl.Read(argc, argv);
pl.Status();
#ifdef _OPENMP
if (cpus > 0)
omp_set_num_threads(cpus);
#endif
// Read marker list
printf("Reading Reference Marker List ...\n");
StringArray refMarkerList;
refMarkerList.Read(referenceSnps);
// Index markers
StringIntHash referenceHash;
for (int i = 0; i < refMarkerList.Length(); i++)
referenceHash.Add(refMarkerList[i].Trim(), i);
printf(" %d Markers in Reference Haplotypes...\n\n", refMarkerList.Length());
// Load reference haplotypes
printf("Loading reference haplotypes ...\n");
HaplotypeSet reference;
reference.markerCount = refMarkerList.Length();
reference.LoadHaplotypes(referenceHaplotypes);
printf(" %d Reference Haplotypes Loaded ...\n\n", reference.count);
// Read framework marker list
printf("Reading Framework Marker List ...\n");
StringArray markerList;
markerList.Read(snps);
ClipReference(reference, refMarkerList, referenceHash, markerList,
firstMarker, lastMarker);
// Crossref Marker Names to Reference Panel Positions
IntArray markerIndex;
markerIndex.Dimension(markerList.Length());
int matches = 0;
for (int i = 0; i < markerList.Length(); i++)
{
markerIndex[i] = referenceHash.Integer(markerList[i].Trim());
if (markerIndex[i] >= 0) matches++;
}
printf(" %d Markers in Framework Haplotypes Overlap Reference ...\n", matches);
if (matches == 0)
error("No markers overlap between target and reference\n"
"Please check correct reference is being used and markers are named consistently");
printf(" %d Other Markers in Framework Haplotypes Discarded ...\n\n", markerList.Length() - matches);
// Check for flips in reference vs. target haplotypes
int flips = 0;
int previous = -1;
for (int i = 0; i < markerIndex.Length(); i++)
if (markerIndex[i] >= 0)
if (markerIndex[i] < previous)
{
if (flips++ < 10)
printf(" -> Marker %s precedes %s in reference, but follows it in target\n",
(const char *) refMarkerList[previous],
(const char *) markerList[i]);
previous = markerIndex[i];
}
if (flips > 10)
printf(" -> %d Additional Marker Order Changes Not Listed\n", flips - 10);
if (flips)
printf(" %d Marker Pairs Change Order in Target vs Framework Haplotypes\n", flips);
// Load target haplotypes
printf("Loading target haplotypes ...\n");
HaplotypeSet target;
target.markerCount = markerList.Length();
target.LoadHaplotypes(haplotypes, true);
reference.CalculateFrequencies();
target.CalculateFrequencies();
target.CompareFrequencies(reference, markerIndex, markerList);
printf(" %d Target Haplotypes Loaded ...\n\n", target.count);
int startIndex = firstMarker.IsEmpty() ? 0 : referenceHash.Integer(firstMarker);
int stopIndex = lastMarker.IsEmpty() ? reference.markerCount - 1 : referenceHash.Integer(lastMarker);
if (startIndex < 0 || stopIndex < 0)
error("Clipping requested, but no position available for one of the endpoints");
printf("Setting up Markov Model...\n\n");
// Setup Markov Model
MarkovParameters mp;
mp.Allocate(reference.markerCount);
if (rounds > 0)
printf("Initializing Model Parameters (using %s and up to %d haplotypes)\n",
em ? "E-M" : "MCMC", states);
// Simple initial estimates of error and recombination rate
for (int i = 0; i < reference.markerCount; i++)
mp.E[i] = 0.01;
for (int i = 0; i < reference.markerCount - 1; i++)
mp.R[i] = 0.001;
if (mp.ReadErrorRates(errorRates))
printf(" Updated error rates using data in %s ...\n", (const char *) errorRates);
if (mp.ReadCrossoverRates(recombinationRates))
printf(" Updated recombination rates using %s ...\n", (const char *) recombinationRates);
// Parameter estimation loop
for (int round = 0; round < rounds; round++)
{
printf(" Round %d of Parameter Refinement ...\n", round + 1);
int iterations = states < reference.count ? states : reference.count;
MarkovModel original;
original.CopyParameters(mp);
#pragma omp parallel for
for (int i = 0; i < iterations; i++)
{
MarkovModel mm;
mm.Allocate(reference.markerCount, reference.count - 1);
mm.CopyParameters(original);
// Reference leave one out (loo) panel
char ** reference_loo = new char * [reference.count - 1];
for (int in = 0, out = 0; in < reference.count; in++)
if (in != i)
reference_loo[out++] = reference.haplotypes[in];
mm.WalkLeft(reference.haplotypes[i], reference_loo, reference.freq);
if (em)
mm.CountExpected(reference.haplotypes[i], reference_loo, reference.freq);
else
{
#pragma omp critical
{ mm.ProfileModel(reference.haplotypes[i], reference_loo, reference.freq); }
}
delete [] reference_loo;
#pragma omp critical
mp += mm;
}
if (round >= rounds / 2)
{
int iterations = states < target.count ? states : target.count;
#pragma omp parallel for
for (int i = 0; i < iterations; i++)
{
MarkovModel mm;
mm.Allocate(reference.markerCount, reference.count);
mm.CopyParameters(original);
// Padded version of target haplotype, including missing sites
char * padded = new char [reference.markerCount];
for (int k = 0; k < reference.markerCount; k++)
padded[k] = 0;
// Copy current haplotype into padded vector
for (int j = 0; j < target.markerCount; j++)
if (markerIndex[j] >= 0)
padded[markerIndex[j]] = target.haplotypes[i][j];
mm.WalkLeft(padded, reference.haplotypes, reference.freq);
if (em)
mm.CountExpected(padded, reference.haplotypes, reference.freq);
else
{
#pragma omp critical
{ mm.ProfileModel(padded, reference.haplotypes, reference.freq); }
}
delete [] padded;
#pragma omp critical
mp += mm;
}
}
mp.UpdateModel();
double crossovers = 0;
for (int i = 0; i < reference.markerCount - 1; i++)
crossovers += mp.R[i];
double errors = 0;
for (int i = 0; i < reference.markerCount; i++)
{
double heterozygosity = 1.0 - square(reference.freq[1][i])
- square(reference.freq[2][i])
- square(reference.freq[3][i])
- square(reference.freq[4][i]);
errors += mp.E[i] * heterozygosity;
}
errors /= reference.markerCount + 1e-30;
printf(" %.0f mosaic crossovers expected per haplotype\n", crossovers);
printf(" %.1f%% of crossovers are due to reference flips\n", mp.empiricalFlipRate * 100.);
printf(" %.3g errors in mosaic expected per marker\n", errors);
}
if (rounds > 0)
{
printf(" Saving estimated parameters for future use ...\n");
mp.WriteParameters(refMarkerList, prefix, gzip);
}
printf("\n");
// List the major allele at each location
reference.ListMajorAlleles();
printf("Generating Draft .info File ...\n\n");
// Output some basic information
IFILE info = ifopen(prefix + ".info.draft", "wt");
ifprintf(info, "SNP\tAl1\tAl2\tFreq1\tGenotyped\n");
for (int i = 0, j = 0; i <= stopIndex; i++)
if (i >= startIndex)
ifprintf(info, "%s\t%s\t%s\t%.4f\t%s\n",
(const char *) refMarkerList[i],
reference.MajorAlleleLabel(i), reference.MinorAlleleLabel(i),
reference.freq[reference.major[i]][i],
j < markerIndex.Length() && i == markerIndex[j] ? (j++, "Genotyped") : "-");
else
if (j < markerIndex.Length() && i == markerIndex[j])
j++;
ifclose(info);
printf("Imputing Genotypes ...\n");
IFILE dosages = ifopen(prefix + ".dose" + (gzip ? ".gz" : ""), "wt");
IFILE hapdose, haps;
if (phased)
{
hapdose = ifopen(prefix + ".hapDose" + (gzip ? ".gz" : ""), "wt");
haps = ifopen(prefix + ".haps" + (gzip ? ".gz" : ""), "wt");
}
ImputationStatistics stats(reference.markerCount);
// Impute each haplotype
#pragma omp parallel for
for (int i = 0; i < target.count; i++)
{
if (i != 0 && target.labels[i] == target.labels[i-1])
continue;
MarkovModel mm;
mm.Allocate(reference.markerCount, reference.count);
mm.ClearImputedDose();
mm.CopyParameters(mp);
// Padded version of target haplotype, including missing sites
char * padded = new char [reference.markerCount];
for (int j = 0; j < reference.markerCount; j++)
padded[j] = 0;
int k = i;
do {
printf(" Processing Haplotype %d of %d ...\n", k + 1, target.count);
// Copy current haplotype into padded vector
for (int j = 0; j < target.markerCount; j++)
if (markerIndex[j] >= 0)
padded[markerIndex[j]] = target.haplotypes[k][j];
mm.WalkLeft(padded, reference.haplotypes, reference.freq);
mm.Impute(reference.major, padded, reference.haplotypes, reference.freq);
#pragma omp critical
{ stats.Update(mm.imputedHap, mm.leaveOneOut, padded, reference.major); }
#pragma omp critical
if (phased)
{
ifprintf(hapdose, "%s\tHAPLO%d", (const char *) target.labels[i], k - i + 1);
ifprintf(haps, "%s\tHAPLO%d", (const char *) target.labels[i], k - i + 1);
for (int j = startIndex; j <= stopIndex; j++)
{
ifprintf(hapdose, "\t%.3f", mm.imputedHap[j]);
ifprintf(haps, "%s%c", j % 8 == 0 ? " " : "", mm.imputedAlleles[j]);
}
ifprintf(hapdose, "\n");
ifprintf(haps, "\n");
}
k++;
} while (k < target.count && target.labels[k] == target.labels[i]);
printf(" Outputting Individual %s ...\n", (const char *) target.labels[i]);
#pragma omp critical
{
ifprintf(dosages, "%s\tDOSE", (const char *) target.labels[i]);
for (int j = startIndex; j <= stopIndex; j++)
ifprintf(dosages, "\t%.3f", mm.imputedDose[j]);
ifprintf(dosages, "\n");
}
delete [] padded;
}
ifclose(dosages);
if (phased)
{
ifclose(hapdose);
ifclose(haps);
}
// Output some basic information
info = ifopen(prefix + ".info" + (gzip ? ".gz" : ""), "wt");
ifprintf(info, "SNP\tAl1\tAl2\tFreq1\tMAF\tAvgCall\tRsq\tGenotyped\tLooRsq\tEmpR\tEmpRsq\tDose1\tDose2\n");
// Padded version of target haplotype, including missing sites
char * padded = new char [reference.markerCount];
for (int k = 0; k < reference.markerCount; k++)
padded[k] = 0;
// Mark genotyped SNPs in padded vector
for (int j = 0; j < target.markerCount; j++)
if (markerIndex[j] >= 0)
padded[markerIndex[j]] = 1;
for (int i = startIndex; i <= stopIndex; i++)
{
ifprintf(info, "%s\t%s\t%s\t%.5f\t%.5f\t%.5f\t%.5f\t",
(const char *) refMarkerList[i],
reference.MajorAlleleLabel(i),
reference.MinorAlleleLabel(i),
stats.AlleleFrequency(i),
stats.AlleleFrequency(i) > 0.5 ? 1.0 - stats.AlleleFrequency(i) : stats.AlleleFrequency(i),
stats.AverageCallScore(i),
stats.Rsq(i));
if (padded[i])
ifprintf(info, "Genotyped\t%.5f\t%.5f\t%.5f\t%.5f\t%.5f\n",
stats.LooRsq(i), stats.EmpiricalR(i), stats.EmpiricalRsq(i),
stats.LooMajorDose(i), stats.LooMinorDose(i));
else
ifprintf(info, "-\t-\t-\t-\t-\t-\n");
}
ifclose(info);
delete [] padded;
time_t stop = time(NULL);
int seconds = stop - start;
printf("\nRun completed in %d hours, %d mins, %d seconds on %s\n\n",
seconds / 3600, (seconds % 3600) / 60, seconds % 60,
ctime(&stop));
}
void Imputation::MergeFinalVcfAllVariants(HaplotypeSet &rHap,HaplotypeSet &tHap,ImputationStatistics &stats,int MaxIndex)
{
cout<<" ------------------------------------------------------------------------------"<<endl;
cout<<" FINAL VCF MERGE "<<endl;
cout<<" ------------------------------------------------------------------------------"<<endl;
printf("\n Merging partial VCF files to final output VCF File : %s ",(outFile + ".dose.vcf" + (gzip ? ".gz" : "")).c_str() );
cout<<endl<<endl;
IFILE vcfdosepartial = ifopen(outFile + ".dose.vcf" + (gzip ? ".gz" : ""), "a", gzip ?InputFile::BGZF:InputFile::UNCOMPRESSED);
vector<IFILE> vcfdosepartialList(MaxIndex);
for(int i=1;i<=MaxIndex;i++)
{
string tempFileIndex(outFile);
stringstream strs;
strs<<(i);
tempFileIndex+=(".dose.vcf.part." +
(string)(strs.str())+(gzip ? ".gz" : ""));
vcfdosepartialList[i-1] = ifopen(tempFileIndex.c_str(), "r");
}
string line;
for(int i=1;i<=MaxIndex;i++)
{
line.clear();
vcfdosepartialList[i-1]->readLine(line);
ifprintf(vcfdosepartial,"%s",line.c_str());
}
int i=0;
for (int index =0; index < rHap.RefTypedTotalCount; index++)
{
//abort();
if(index%10000==0)
{
printf(" Merging marker %d of %d [%.1f%%] to VCF File ...", index + 1, rHap.RefTypedTotalCount,100*(double)(index + 1)/(int)rHap.RefTypedTotalCount);
cout<<endl;
}
if(rHap.RefTypedIndex[index]==-1)
{
if(i>=rHap.PrintStartIndex && i <= rHap.PrintEndIndex)
{
ifprintf(vcfdosepartial,"\n%s\t%d\t%s\t%s\t%s\t.\tPASS\tMAF=%.5f;R2=%.5f",
rHap.VariantList[i].chr.c_str(),rHap.VariantList[i].bp,
RsId?rHap.VariantList[i].rsid.c_str():rHap.VariantList[i].name.c_str(),rHap.VariantList[i].refAlleleString.c_str(),
rHap.VariantList[i].altAlleleString.c_str(),stats.AlleleFrequency(i) > 0.5 ? 1.0 - stats.AlleleFrequency(i) : stats.AlleleFrequency(i),stats.Rsq(i));
if(!tHap.missing[i])
ifprintf(vcfdosepartial,";ER2=%.5f",stats.EmpiricalRsq(i));
ifprintf(vcfdosepartial,"\t%s",GT?(DS?(GP?"GT:DS:GP":"GT:DS"):(GP?"GT:GP":"GT")):(DS?(GP?"DS:GP":"DS"):(GP?"GP":"")));
for(int j=1;j<=MaxIndex;j++)
{
string tempFileIndex(outFile);
stringstream strs;
strs<<(j);
tempFileIndex+=(".dose.vcf.part."
+ (string)(strs.str())
+(gzip ? ".gz" : ""));
line.clear();
vcfdosepartialList[j-1]->readLine(line);
ifprintf(vcfdosepartial,"%s",line.c_str());
}
}
i++;
}
else
{
variant ThisTypedVariant =tHap.TypedOnlyVariantList[rHap.RefTypedIndex[index]];
ifprintf(vcfdosepartial,"\n%s\t%d\t%s\t%s\t%s\t.\tPASS\t",
ThisTypedVariant.chr.c_str(),
ThisTypedVariant.bp,
RsId? ThisTypedVariant.rsid.c_str():ThisTypedVariant.name.c_str(),
ThisTypedVariant.refAlleleString.c_str(),
ThisTypedVariant.altAlleleString.c_str());
ifprintf(vcfdosepartial,"GENOTYPED_ONLY;AN=%d;MAF=%.5f",
tHap.TotalSample[rHap.RefTypedIndex[index]],
tHap.AlleleFreq[rHap.RefTypedIndex[index]]);
//cout<<rHap.RefTypedIndex[index]<<" " <<tHap.TotalSample[rHap.RefTypedIndex[index]]<<" " << tHap.AlleleFreq[rHap.RefTypedIndex[index]]/(double)tHap.TotalSample[rHap.RefTypedIndex[index]]<< endl;
ifprintf(vcfdosepartial,"\t%s",GT?(DS?(GP?"GT:DS:GP":"GT:DS"):(GP?"GT:GP":"GT")):(DS?(GP?"DS:GP":"DS"):(GP?"GP":"")));
for(int j=1;j<=MaxIndex;j++)
{
string tempFileIndex(outFile);
stringstream strs;
strs<<(j);
tempFileIndex+=(".dose.vcf.part."
+ (string)(strs.str())
+(gzip ? ".gz" : ""));
line.clear();
vcfdosepartialList[j-1]->readLine(line);
ifprintf(vcfdosepartial,"%s",line.c_str());
}
// ifprintf(vcfdosepartial,"\n");
}
}
for(int i=1;i<=MaxIndex;i++)
{
ifclose(vcfdosepartialList[i-1]);
string tempFileIndex(outFile);
stringstream strs;
strs<<(i);
tempFileIndex+=(".dose.vcf.part." +
(string)(strs.str())+
(gzip ? ".gz" : ""));
remove(tempFileIndex.c_str());
}
ifclose(vcfdosepartial);
printf("\n Merging Finished ..." );
cout<<endl <<endl;
}
int Stats::execute(int argc, char **argv)
{
// Extract command line arguments.
String inFile = "";
String indexFile = "";
bool basic = false;
bool noeof = false;
bool params = false;
bool qual = false;
bool phred = false;
int maxNumReads = -1;
bool unmapped = false;
String pBaseQC = "";
String cBaseQC = "";
String regionList = "";
int excludeFlags = 0;
int requiredFlags = 0;
bool withinRegion = false;
int minMapQual = 0;
String dbsnp = "";
PosList *dbsnpListPtr = NULL;
bool baseSum = false;
int bufferSize = PileupHelper::DEFAULT_WINDOW_SIZE;
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_PARAMETER_GROUP("Required Parameters")
LONG_STRINGPARAMETER("in", &inFile)
LONG_PARAMETER_GROUP("Types of Statistics")
LONG_PARAMETER("basic", &basic)
LONG_PARAMETER("qual", &qual)
LONG_PARAMETER("phred", &phred)
LONG_STRINGPARAMETER("pBaseQC", &pBaseQC)
LONG_STRINGPARAMETER("cBaseQC", &cBaseQC)
LONG_PARAMETER_GROUP("Optional Parameters")
LONG_INTPARAMETER("maxNumReads", &maxNumReads)
LONG_PARAMETER("unmapped", &unmapped)
LONG_STRINGPARAMETER("bamIndex", &indexFile)
LONG_STRINGPARAMETER("regionList", ®ionList)
LONG_INTPARAMETER("excludeFlags", &excludeFlags)
LONG_INTPARAMETER("requiredFlags", &requiredFlags)
LONG_PARAMETER("noeof", &noeof)
LONG_PARAMETER("params", ¶ms)
LONG_PARAMETER_GROUP("Optional phred/qual Only Parameters")
LONG_PARAMETER("withinRegion", &withinRegion)
LONG_PARAMETER_GROUP("Optional BaseQC Only Parameters")
LONG_PARAMETER("baseSum", &baseSum)
LONG_INTPARAMETER("bufferSize", &bufferSize)
LONG_INTPARAMETER("minMapQual", &minMapQual)
LONG_STRINGPARAMETER("dbsnp", &dbsnp)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters",
longParameterList));
inputParameters.Read(argc-1, &(argv[1]));
// If no eof block is required for a bgzf file, set the bgzf file type to
// not look for it.
if(noeof)
{
// Set that the eof block is not required.
BgzfFileType::setRequireEofBlock(false);
}
// Check to see if the in file was specified, if not, report an error.
if(inFile == "")
{
usage();
inputParameters.Status();
// In file was not specified but it is mandatory.
std::cerr << "--in is a mandatory argument for stats, "
<< "but was not specified" << std::endl;
return(-1);
}
// Use the index file if unmapped or regionList is not empty.
bool useIndex = (unmapped|| (!regionList.IsEmpty()));
// IndexFile is required, so check to see if it has been set.
if(useIndex && (indexFile == ""))
{
// In file was not specified, so set it to the in file
// + ".bai"
indexFile = inFile + ".bai";
}
////////////////////////////////////////
// Setup in case pileup is used.
Pileup<PileupElementBaseQCStats> pileup(bufferSize);
// Initialize start/end positions.
myStartPos = 0;
myEndPos = -1;
// Open the output qc file if applicable.
IFILE baseQCPtr = NULL;
if(!pBaseQC.IsEmpty() && !cBaseQC.IsEmpty())
{
usage();
inputParameters.Status();
// Cannot specify both types of baseQC.
std::cerr << "Cannot specify both --pBaseQC & --cBaseQC." << std::endl;
return(-1);
}
else if(!pBaseQC.IsEmpty())
{
baseQCPtr = ifopen(pBaseQC, "w");
PileupElementBaseQCStats::setPercentStats(true);
}
else if(!cBaseQC.IsEmpty())
{
baseQCPtr = ifopen(cBaseQC, "w");
PileupElementBaseQCStats::setPercentStats(false);
}
if(baseQCPtr != NULL)
{
PileupElementBaseQCStats::setOutputFile(baseQCPtr);
PileupElementBaseQCStats::printHeader();
}
if((baseQCPtr != NULL) || baseSum)
{
PileupElementBaseQCStats::setMapQualFilter(minMapQual);
PileupElementBaseQCStats::setBaseSum(baseSum);
}
if(params)
{
inputParameters.Status();
}
// Open the file for reading.
SamFile samIn;
if(!samIn.OpenForRead(inFile))
{
fprintf(stderr, "%s\n", samIn.GetStatusMessage());
return(samIn.GetStatus());
}
samIn.SetReadFlags(requiredFlags, excludeFlags);
// Set whether or not basic statistics should be generated.
samIn.GenerateStatistics(basic);
// Read the sam header.
SamFileHeader samHeader;
if(!samIn.ReadHeader(samHeader))
{
fprintf(stderr, "%s\n", samIn.GetStatusMessage());
return(samIn.GetStatus());
}
// Open the bam index file for reading if we are
// doing unmapped reads (also set the read section).
if(useIndex)
{
samIn.ReadBamIndex(indexFile);
if(unmapped)
{
samIn.SetReadSection(-1);
}
if(!regionList.IsEmpty())
{
myRegionList = ifopen(regionList, "r");
}
}
//////////////////////////
// Read dbsnp if specified and doing baseQC
if(((baseQCPtr != NULL) || baseSum) && (!dbsnp.IsEmpty()))
{
// Read the dbsnp file.
IFILE fdbSnp;
fdbSnp = ifopen(dbsnp,"r");
// Determine how many entries.
const SamReferenceInfo& refInfo = samHeader.getReferenceInfo();
int maxRefLen = 0;
for(int i = 0; i < refInfo.getNumEntries(); i++)
{
int refLen = refInfo.getReferenceLength(i);
if(refLen >= maxRefLen)
{
maxRefLen = refLen + 1;
}
}
dbsnpListPtr = new PosList(refInfo.getNumEntries(),maxRefLen);
if(fdbSnp==NULL)
{
std::cerr << "Open dbSNP file " << dbsnp.c_str() << " failed!\n";
}
else if(dbsnpListPtr == NULL)
{
std::cerr << "Failed to init the memory allocation for the dbsnpList.\n";
}
else
{
// Read the dbsnp file.
StringArray tokens;
String buffer;
int position = 0;
int refID = 0;
// Loop til the end of the file.
while (!ifeof(fdbSnp))
{
// Read the next line.
buffer.ReadLine(fdbSnp);
// If it does not have at least 2 columns,
// continue to the next line.
if (buffer.IsEmpty() || buffer[0] == '#') continue;
tokens.AddTokens(buffer);
if(tokens.Length() < 2) continue;
if(!tokens[1].AsInteger(position))
{
std::cerr << "Improperly formatted region line, start position "
<< "(2nd column) is not an integer: "
<< tokens[1]
<< "; Skipping to the next line.\n";
continue;
}
// Look up the reference name.
refID = samHeader.getReferenceID(tokens[0]);
if(refID != SamReferenceInfo::NO_REF_ID)
{
// Reference id was found, so add it to the dbsnp
dbsnpListPtr->addPosition(refID, position);
}
tokens.Clear();
buffer.Clear();
}
}
ifclose(fdbSnp);
}
// Read the sam records.
SamRecord samRecord;
int numReads = 0;
//////////////////////
// Setup in case doing a quality count.
// Quality histogram.
const int MAX_QUAL = 126;
const int START_QUAL = 33;
uint64_t qualCount[MAX_QUAL+1];
for(int i = 0; i <= MAX_QUAL; i++)
{
qualCount[i] = 0;
}
const int START_PHRED = 0;
const int PHRED_DIFF = START_QUAL - START_PHRED;
const int MAX_PHRED = MAX_QUAL - PHRED_DIFF;
uint64_t phredCount[MAX_PHRED+1];
for(int i = 0; i <= MAX_PHRED; i++)
{
phredCount[i] = 0;
}
int refPos = 0;
Cigar* cigarPtr = NULL;
char cigarChar = '?';
// Exclude clips from the qual/phred counts if unmapped reads are excluded.
bool qualExcludeClips = excludeFlags & SamFlag::UNMAPPED;
//////////////////////////////////
// When not reading by sections, getNextSection returns true
// the first time, then false the next time.
while(getNextSection(samIn))
{
// Keep reading records from the file until SamFile::ReadRecord
// indicates to stop (returns false).
while(((maxNumReads < 0) || (numReads < maxNumReads)) && samIn.ReadRecord(samHeader, samRecord))
{
// Another record was read, so increment the number of reads.
++numReads;
// See if the quality histogram should be genereated.
if(qual || phred)
{
// Get the quality.
const char* qual = samRecord.getQuality();
// Check for no quality ('*').
if((qual[0] == '*') && (qual[1] == 0))
{
// This record does not have a quality string, so no
// quality processing is necessary.
}
else
{
int index = 0;
cigarPtr = samRecord.getCigarInfo();
cigarChar = '?';
refPos = samRecord.get0BasedPosition();
if(!qualExcludeClips && (cigarPtr != NULL))
{
// Offset the reference position by any soft clips
// by subtracting the queryIndex of this start position.
// refPos is now the start position of the clips.
refPos -= cigarPtr->getQueryIndex(0);
}
while(qual[index] != 0)
{
// Skip this quality if it is clipped and we are skipping clips.
if(cigarPtr != NULL)
{
cigarChar = cigarPtr->getCigarCharOpFromQueryIndex(index);
}
if(qualExcludeClips && Cigar::isClip(cigarChar))
{
// Skip a clipped quality.
++index;
// Increment the position.
continue;
}
if(withinRegion && (myEndPos != -1) && (refPos >= myEndPos))
{
// We have hit the end of the region, stop processing this
// quality string.
break;
}
if(withinRegion && (refPos < myStartPos))
{
// This position is not in the target.
++index;
// Update the position if this is found in the reference or a clip.
if(Cigar::foundInReference(cigarChar) || Cigar::isClip(cigarChar))
{
++refPos;
}
continue;
}
// Check for valid quality.
if((qual[index] < START_QUAL) || (qual[index] > MAX_QUAL))
{
if(qual)
{
std::cerr << "Invalid Quality found: " << qual[index]
<< ". Must be between "
<< START_QUAL << " and " << MAX_QUAL << ".\n";
}
if(phred)
{
std::cerr << "Invalid Phred Quality found: " << qual[index] - PHRED_DIFF
<< ". Must be between "
<< START_QUAL << " and " << MAX_QUAL << ".\n";
}
// Skip an invalid quality.
++index;
// Update the position if this is found in the reference or a clip.
if(Cigar::foundInReference(cigarChar) || Cigar::isClip(cigarChar))
{
++refPos;
}
continue;
}
// Increment the count for this quality.
++(qualCount[(int)(qual[index])]);
++(phredCount[(int)(qual[index]) - PHRED_DIFF]);
// Update the position if this is found in the reference or a clip.
if(Cigar::foundInReference(cigarChar) || Cigar::isClip(cigarChar))
{
++refPos;
}
++index;
}
}
}
// Check the next thing to do for the read.
if((baseQCPtr != NULL) || baseSum)
{
// Pileup the bases for this read.
pileup.processAlignmentRegion(samRecord, myStartPos, myEndPos, dbsnpListPtr);
}
}
// Done with a section, move on to the next one.
// New section, so flush the pileup.
pileup.flushPileup();
}
// Flush the rest of the pileup.
if((baseQCPtr != NULL) || baseSum)
{
// Pileup the bases.
pileup.processAlignmentRegion(samRecord, myStartPos, myEndPos, dbsnpListPtr);
PileupElementBaseQCStats::printSummary();
ifclose(baseQCPtr);
}
std::cerr << "Number of records read = " <<
samIn.GetCurrentRecordCount() << std::endl;
if(basic)
{
std::cerr << std::endl;
samIn.PrintStatistics();
}
// Print the quality stats.
if(qual)
{
std::cerr << std::endl;
std::cerr << "Quality\tCount\n";
for(int i = START_QUAL; i <= MAX_QUAL; i++)
{
std::cerr << i << "\t" << qualCount[i] << std::endl;
}
}
// Print the phred quality stats.
if(phred)
{
std::cerr << std::endl;
std::cerr << "Phred\tCount\n";
for(int i = START_PHRED; i <= MAX_PHRED; i++)
{
std::cerr << i << "\t" << phredCount[i] << std::endl;
}
}
SamStatus::Status status = samIn.GetStatus();
if(status == SamStatus::NO_MORE_RECS)
{
// A status of NO_MORE_RECS means that all reads were successful.
status = SamStatus::SUCCESS;
}
return(status);
}
void Imputation::performImputation(HaplotypeSet &tHap,HaplotypeSet &rHap, String Golden)
{
vector<int> optStructure=rHap.optEndPoints;
int time_prev = time(0),time_load,vcfSampleIndex=0;;
includeGwas=true;
MarkovParameters* MP=createEstimates(rHap,tHap,rHap.optEndPoints,1-includeGwas);
cout<<" ------------------------------------------------------------------------------"<<endl;
cout<<" MAIN IMPUTATION "<<endl;
cout<<" ------------------------------------------------------------------------------"<<endl;
ImputationStatistics stats(rHap.numMarkers );
IFILE dosages=NULL, hapdose=NULL, haps=NULL,vcfdosepartial=NULL;
HaplotypeSet DosageForVcfPartial;
DosageForVcfPartial.unphasedOutput=unphasedOutput;
DosageForVcfPartial.TypedOnly=tHap.TypedOnly;
DosageForVcfPartial.GWASOnlycounter=tHap.GWASOnlycounter;
if(tHap.TypedOnly)
{
printf("\n Calculating Allele Frequency for Typed-Only variants ... ");
cout<<endl;
tHap.CalculateGWASOnlyFreq();
}
cout << "\n Starting Imputation ...";
printf("\n\n Setting up Markov Model for Imputation ...");
cout<<endl<<endl;
if (phased && !unphasedOutput)
{
hapdose = ifopen(outFile + ".hapDose" + (gzip ? ".gz" : ""), "wb", gzip ?InputFile::BGZF:InputFile::UNCOMPRESSED);
haps = ifopen(outFile + ".hapLabel" + (gzip ? ".gz" : ""), "wb", gzip ?InputFile::BGZF:InputFile::UNCOMPRESSED);
}
int maxVcfSample=200,NumVcfWritten=0,NumVcfCreated=0,NovcfParts=1;
if((maxVcfSample)>=tHap.numSamples)
maxVcfSample=tHap.numSamples;
if(vcfOutput)
{
vcfdosepartial = ifopen(outFile + ".dose.vcf" + (gzip ? ".gz" : ""), "wb", gzip ?InputFile::BGZF:InputFile::UNCOMPRESSED);
ifprintf(vcfdosepartial,"##fileformat=VCFv4.1\n");
time_t t = time(0);
struct tm * now = localtime( & t );
ifprintf(vcfdosepartial,"##filedate=%d.%d.%d\n",(now->tm_year + 1900),(now->tm_mon + 1) ,now->tm_mday);
ifprintf(vcfdosepartial,"##source=Minimac3\n");
if(GT)
ifprintf(vcfdosepartial,"##FORMAT=<ID=GT,Number=1,Type=String,Description=\"Genotype\">\n");
if(tHap.AllMaleTarget)
{
if(DS)
ifprintf(vcfdosepartial,"##FORMAT=<ID=DS,Number=1,Type=Float,Description=\"Estimated Alternate Allele Dosage (For Male Chr: X) : [P(Alt Allele)]\">\n");
if(GP)
ifprintf(vcfdosepartial,"##FORMAT=<ID=GP,Number=2,Type=Float,Description=\"Estimated Posterior Probabilities for Genotypes 0 and 1 (For Male Chr: X) \">\n");
}
else
{
if(DS)
ifprintf(vcfdosepartial,"##FORMAT=<ID=DS,Number=1,Type=Float,Description=\"Estimated Alternate Allele Dosage : [P(0/1)+2*P(1/1)]\">\n");
if(GP)
ifprintf(vcfdosepartial,"##FORMAT=<ID=GP,Number=3,Type=Float,Description=\"Estimated Posterior Probabilities for Genotypes 0/0, 0/1 and 1/1 \">\n");
}
ifprintf(vcfdosepartial,"##INFO=<ID=MAF,Number=1,Type=Float,Description=\"Estimated Alternate Allele Frequency\">\n");
ifprintf(vcfdosepartial,"##INFO=<ID=R2,Number=1,Type=Float,Description=\"Estimated Imputation Accuracy\">\n");
ifprintf(vcfdosepartial,"##INFO=<ID=ER2,Number=1,Type=Float,Description=\"Empirical (Leave-One-Out) R-square (available only for genotyped variants)\">\n");
ifprintf(vcfdosepartial,"#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT");
ifclose(vcfdosepartial);
if(!tHap.AllMaleTarget)
DosageForVcfPartial.InitializePartialDosageForVcfOutput((2*maxVcfSample),rHap.numMarkers,format);
else
DosageForVcfPartial.InitializePartialDosageForVcfOutputMaleSamples(maxVcfSample<MaxSample?maxVcfSample:MaxSample,rHap.numMarkers,format);
}
if(doseOutput)
dosages = ifopen(outFile + ".dose" + (gzip ? ".gz" : ""), "wb",(gzip ? InputFile::BGZF:InputFile::UNCOMPRESSED) );
#pragma omp parallel for
for(int hapId=0;hapId<MaxSample;hapId++)
{
if (hapId %2==1)
{
if(rHap.finChromosome!="X")
continue;
else if(!tHap.AllMaleTarget)
continue;
}
vector<float> foldedProb,recomProb,noRecomProb, rightProb,probAlleleNoStandardize(8,0.0),tempDoseHap1;
vector<bool> tempHap(rHap.numMarkers),tempMissHap(rHap.numMarkers);
vector<bool> tempDoseAlleleHap1;
MarkovModel MM(tHap,rHap,tHap.missing,rHap.major);
MM.CopyParameters(MP);
int hapIdIndiv=hapId;
do{
MM.initializeMatrices(tHap,rHap,optStructure,rHap.ReducedStructureInfo);
printf(" Processing Haplotype %d of %d ...", hapIdIndiv + 1, MaxSample);
cout<<endl;
MM.ThisHapId=hapIdIndiv;
for(int group=1;group<(int)optStructure.size();group++)
{
MM.foldProbabilities(foldedProb,group-1,rHap.ReducedStructureInfo[group-1],0,refCount);
MM.leftNoRecoProb[group-1][0]=foldedProb;
if(group==1 && !tHap.missing[0])
if(!tHap.getMissingScaffoldedHaplotype(hapIdIndiv,0))
{
Condition(rHap,0,foldedProb,MM.leftNoRecoProb[group-1][0],MM.Error[0],
tHap.getScaffoldedHaplotype(hapIdIndiv,0)? rHap.AlleleFreq[0] : 1-rHap.AlleleFreq[0],
tHap.getScaffoldedHaplotype(hapIdIndiv,0),MM.backgroundError,
foldedProb.size(),rHap.ReducedStructureInfo[0]);
}
MM.WalkLeft(tHap,hapIdIndiv,MM.leftProb[group-1],MM.leftNoRecoProb[group-1],
foldedProb,optStructure[group-1],optStructure[group],
rHap.ReducedStructureInfo[group-1],rHap.AlleleFreq);
splitFoldedProb(recomProb,MM.leftProb[group-1][optStructure[group]-optStructure[group-1]],MM.leftNoRecoProb[group-1][optStructure[group]-optStructure[group-1]]);
MM.unfoldProbabilities(group-1,recomProb,MM.leftNoRecoProb[group-1][optStructure[group]-optStructure[group-1]],foldedProb,0,rHap.ReducedStructureInfo,refCount);
}
for(int group=optStructure.size()-1;group>0;group--)
{
MM.foldProbabilities(foldedProb,group-1,rHap.ReducedStructureInfo[group-1],1,refCount);
rightProb=foldedProb;
noRecomProb=foldedProb;
MM.Impute(tHap,foldedProb,hapIdIndiv,MM.leftProb[group-1],MM.leftNoRecoProb[group-1],rightProb,noRecomProb,MM.junctionLeftProb[group-1],
MM.junctionRightProb[group],optStructure[group-1], optStructure[group],rHap.ReducedStructureInfo[group-1],1,rHap.AlleleFreq);
splitFoldedProb(recomProb,rightProb,noRecomProb);
MM.unfoldProbabilities(group-1,recomProb,noRecomProb,foldedProb,1,rHap.ReducedStructureInfo,refCount);
}
for(int jjj=0;jjj<rHap.numMarkers;jjj++)
{
tempHap[jjj]=tHap.getScaffoldedHaplotype(hapIdIndiv,jjj);
tempMissHap[jjj]=tHap.getMissingScaffoldedHaplotype(hapIdIndiv,jjj);
}
if(vcfOutput)
{
if(hapIdIndiv%2==0)
{
tempDoseHap1= MM.imputedHap;
tempDoseAlleleHap1= MM.imputedAlleleNumber;
}
}
#pragma omp critical
{
stats.Update(MM.imputedHap, MM.leaveOneOut,tempHap,tempMissHap,rHap.major);
}
#pragma omp critical
if (phased && !unphasedOutput)
{
PrintHaplotypeData(rHap, tHap, hapdose, haps,
MM.imputedHap, MM.imputedAlleleNumber,
hapIdIndiv, tHap.AllMaleTarget?hapId:hapId/2);
}
if(tHap.AllMaleTarget)
break;
hapIdIndiv++;
}while(hapIdIndiv<MaxSample && hapIdIndiv%2==1);
#pragma omp critical
if(doseOutput)
{
PrintDosageData(rHap, tHap, dosages, MM.imputedDose, tHap.AllMaleTarget?hapId:hapId/2);
}
#pragma omp critical
if(vcfOutput)
{
printf(" Saving Individual %s for VCF File...\n", tHap.individualName[tHap.AllMaleTarget?hapId:hapId/2].c_str());
if(!tHap.AllMaleTarget)
DosageForVcfPartial.SaveDosageForVcfOutputSampleWise(NumVcfCreated-NumVcfWritten,
tHap.individualName[tHap.AllMaleTarget?hapId:hapId/2],
tempDoseHap1,MM.imputedHap,
tempDoseAlleleHap1,MM.imputedAlleleNumber);
else
DosageForVcfPartial.SaveDosageForVcfOutputSampleWiseChrX(NumVcfCreated-NumVcfWritten,
tHap.individualName[tHap.AllMaleTarget?hapId:hapId/2],
MM.imputedHap,
MM.imputedAlleleNumber);
if(DosageForVcfPartial.TypedOnly)
{
DosageForVcfPartial.SaveIndexForGWASOnlyForVcfOutput(NumVcfCreated-NumVcfWritten,
tHap.AllMaleTarget?hapId:hapId/2);
}
NumVcfCreated++;
vcfSampleIndex++;
if(NumVcfCreated%maxVcfSample==0 || NumVcfCreated==(tHap.AllMaleTarget?MaxSample:MaxSample/2))
{
string PartialVcfFileName(outFile),tempFileIndex1(outFile);
stringstream strs;
strs<<(NovcfParts);
PartialVcfFileName+=(".dose.vcf.part." +
(string)(strs.str())
+(gzip ? ".gz" : ""));
if(!tHap.AllMaleTarget)
printf("\n --->>> Saving samples %d-%d in VCF file : %s ...\n\n",
(NumVcfWritten)+1,(MaxSample/2<(NumVcfWritten+maxVcfSample)?MaxSample/2:(NumVcfWritten+maxVcfSample)),
PartialVcfFileName.c_str());
else
printf("\n --->>> Saving samples %d-%d in VCF file : %s ...\n\n",
(NumVcfWritten)+1,(MaxSample<(NumVcfWritten+maxVcfSample)?MaxSample:(NumVcfWritten+maxVcfSample)),
PartialVcfFileName.c_str());
//if(NovcfParts==2)
// abort();
FlushPartialVcf(rHap,tHap,DosageForVcfPartial,PartialVcfFileName,NovcfParts);
if(NumVcfCreated<(tHap.AllMaleTarget?MaxSample:MaxSample/2))
{
NovcfParts++;
NumVcfWritten+=maxVcfSample;
//int gg=maxVcfSample<(((tHap.AllMaleTarget?MaxSample:MaxSample/2))-NumVcfWritten)?
//2*maxVcfSample:2*(((tHap.AllMaleTarget?MaxSample:MaxSample/2))-NumVcfWritten);
//
//
//abort();
if(!tHap.AllMaleTarget)
DosageForVcfPartial.InitializePartialDosageForVcfOutput(maxVcfSample<(MaxSample/2-NumVcfWritten)?2*maxVcfSample:2*(MaxSample/2-NumVcfWritten),rHap.numMarkers,format);
else
DosageForVcfPartial.InitializePartialDosageForVcfOutputMaleSamples(maxVcfSample<(MaxSample-NumVcfWritten)?maxVcfSample:(MaxSample-NumVcfWritten),rHap.numMarkers,format);
}
}
}
}
cout<<endl<<" Imputation Finished ... "<<endl;
if (phased && !unphasedOutput)
{
ifclose(hapdose);
ifclose(haps);
cout<<endl<<" Haplotype Dosage information written to : "<<
outFile + ".hapDose" + (gzip ? ".gz" : "")<<endl;
cout<<endl<<" Haplotype Allele information written to : "<<
outFile + ".hapLabel" + (gzip ? ".gz" : "")<<endl;
}
if(doseOutput)
{
ifclose(dosages);
cout<<endl<<" Dosage information written to : "<<
outFile + ".dose" + (gzip ? ".gz" : "")<<endl;
}
PrintInfoFile(rHap,tHap,stats);
time_load = time(0) - time_prev;
cout << "\n Time taken for imputation = " << time_load << " seconds."<<endl<<endl;
if(vcfOutput)
MergeFinalVcfAllVariants(rHap,tHap,stats,NovcfParts);
}
void glfHandler::Close()
{
if (isOpen())
ifclose(handle);
}
// main function of verifyBamID
int execute(int argc, char** argv) {
printf("verifyBamID %s -- verify identity and purity of sequence data\n"
"(c) 2010-2014 Hyun Min Kang, Goo Jun, and Goncalo Abecasis\n\n", VERSION);
VerifyBamIDArgs args;
ParameterList pl;
BEGIN_LONG_PARAMETERS(longParameters)
LONG_PARAMETER_GROUP("Input Files")
LONG_STRINGPARAMETER("vcf",&args.sVcfFile)
LONG_STRINGPARAMETER("bam",&args.sBamFile)
LONG_STRINGPARAMETER("subset",&args.sSubsetInds)
LONG_STRINGPARAMETER("smID",&args.sSMID)
LONG_PARAMETER_GROUP("VCF analysis options")
LONG_DOUBLEPARAMETER("genoError",&args.genoError)
LONG_DOUBLEPARAMETER("minAF",&args.minAF)
LONG_DOUBLEPARAMETER("minCallRate",&args.minCallRate)
LONG_PARAMETER_GROUP("Individuals to compare with chip data")
EXCLUSIVE_PARAMETER("site",&args.bSiteOnly)
EXCLUSIVE_PARAMETER("self",&args.bSelfOnly)
EXCLUSIVE_PARAMETER("best",&args.bFindBest)
LONG_PARAMETER_GROUP("Chip-free optimization options")
EXCLUSIVE_PARAMETER("free-none",&args.bFreeNone)
EXCLUSIVE_PARAMETER("free-mix",&args.bFreeMixOnly)
EXCLUSIVE_PARAMETER("free-refBias",&args.bFreeRefBiasOnly)
EXCLUSIVE_PARAMETER("free-full",&args.bFreeFull)
LONG_PARAMETER_GROUP("With-chip optimization options")
EXCLUSIVE_PARAMETER("chip-none",&args.bChipNone)
EXCLUSIVE_PARAMETER("chip-mix",&args.bChipMixOnly)
EXCLUSIVE_PARAMETER("chip-refBias",&args.bChipRefBiasOnly)
EXCLUSIVE_PARAMETER("chip-full",&args.bChipFull)
LONG_PARAMETER_GROUP("BAM analysis options")
LONG_PARAMETER("ignoreRG",&args.bIgnoreRG)
LONG_PARAMETER("ignoreOverlapPair",&args.bIgnoreOverlapPair)
LONG_PARAMETER("noEOF",&args.bNoEOF)
LONG_PARAMETER("precise",&args.bPrecise)
LONG_INTPARAMETER("minMapQ",&args.minMapQ)
LONG_INTPARAMETER("maxDepth",&args.maxDepth)
LONG_INTPARAMETER("minQ",&args.minQ)
LONG_INTPARAMETER("maxQ",&args.maxQ)
LONG_DOUBLEPARAMETER("grid",&args.grid)
LONG_PARAMETER_GROUP("Modeling Reference Bias")
LONG_DOUBLEPARAMETER("refRef",&args.pRefRef)
LONG_DOUBLEPARAMETER("refHet",&args.pRefHet)
LONG_DOUBLEPARAMETER("refAlt",&args.pRefAlt)
LONG_PARAMETER_GROUP("Output options")
LONG_STRINGPARAMETER("out",&args.sOutFile)
LONG_PARAMETER("verbose",&args.bVerbose)
LONG_PHONEHOME(VERSION)
END_LONG_PARAMETERS();
pl.Add(new LongParameters("Available Options",longParameters));
pl.Read(argc, argv);
pl.Status();
// check the validity of input files
if ( args.sVcfFile.IsEmpty() ) {
error("--vcf [vcf file] required");
}
if ( args.sBamFile.IsEmpty() ) {
error("--bam [bam file] is required");
}
if ( args.sOutFile.IsEmpty() ) {
error("--out [output prefix] is required");
}
Logger::gLogger = new Logger((args.sOutFile + ".log").c_str(), args.bVerbose);
if ( ! ( args.bSiteOnly || args.bSelfOnly || args.bFindBest ) ) {
warning("--self option was autotomatically turned on by default. Specify --best option if you wanted to check across all possible samples in the VCF");
args.bSelfOnly = true;
}
if ( ( args.maxDepth > 20 ) && ( !args.bPrecise ) ) {
warning("--precise option is not turned on at --maxDepth %d : may be prone to precision errors",args.maxDepth);
}
if ( ( args.bChipRefBiasOnly ) && ( !args.bSelfOnly ) ) {
error("--self must be set for --chip-refBias to work. Skipping..");
}
// check timestamp
time_t t;
time(&t);
Logger::gLogger->writeLog("Analysis started on %s",ctime(&t));
// load arguments
VerifyBamID vbid(&args);
// load input VCF and BAM files
Logger::gLogger->writeLog("Opening Input Files");
vbid.loadFiles(args.sBamFile.c_str(), args.sVcfFile.c_str());
// Check which genotype-free method is used
if ( args.bFreeNone ) { // if no genotype-free mode is tested. skip it
// do nothing for genotype-free estimation
Logger::gLogger->writeLog("Skipping chip-free estimation of sample mixture");
}
else if ( args.bFreeMixOnly ) { // only mixture is estimated.
// genotype-free method
Logger::gLogger->writeLog("Performing chip-free estimation of sample mixture at fixed reference bias parameters (%lf, %lf, %lf)",args.pRefRef,args.pRefHet,args.pRefAlt);
// scan across multiple readgroups
for(int rg=-1; rg < vbid.nRGs - (int)args.bIgnoreRG; ++rg) {
VerifyBamID::mixLLK mix(&vbid);
mix.OptimizeLLK(rg);
Logger::gLogger->writeLog("Optimal per-sample fMix = %lf, LLK0 = %lf, LLK1 = %lf\n",mix.fMix,mix.llk0,mix.llk1);
vbid.mixOut.llk0s[rg+1] = mix.llk0;
vbid.mixOut.llk1s[rg+1] = mix.llk1;
vbid.mixOut.fMixs[rg+1] = mix.fMix;
}
//vbid.mixRefHet = 0.5;
//vbid.mixRefAlt = 0.00;
}
else if ( args.bFreeRefBiasOnly ) {
Logger::gLogger->writeLog("Performing chip-free estimation of reference-bias without sample mixture");
for(int rg=-1; rg < vbid.nRGs - (int)args.bIgnoreRG; ++rg) {
VerifyBamID::refBiasMixLLKFunc myFunc(&vbid, rg);
AmoebaMinimizer myMinimizer;
Vector startingPoint(2);
startingPoint[0] = 0; // pRefHet = 0.5
startingPoint[1] = -4.595; // pRefAlt = 0.01
myMinimizer.func = &myFunc;
myMinimizer.Reset(2);
myMinimizer.point = startingPoint;
myMinimizer.Minimize(1e-6);
double pRefHet = VerifyBamID::invLogit(myMinimizer.point[0]);
double pRefAlt = VerifyBamID::invLogit(myMinimizer.point[1]);
Logger::gLogger->writeLog("Reference Bias Estimated as ( Pr[refBase|HET] = %lf, Pr[refBase|ALT] = %lf) with LLK = %lf at readGroup %d",pRefHet,pRefAlt,myMinimizer.fmin,rg);
//vbid.setRefBiasParams(1.0, pRefHet, pRefAlt);
vbid.mixOut.llk0s[rg+1] = myFunc.llk0;
vbid.mixOut.llk1s[rg+1] = myFunc.llk1;
vbid.mixOut.refHets[rg+1] = myFunc.pRefHet;
vbid.mixOut.refAlts[rg+1] = myFunc.pRefAlt;
}
}
else if ( args.bFreeFull ) {
Logger::gLogger->writeLog("Performing chip-free estimation of reference-bias and sample mixture together");
for(int rg = -1; rg < vbid.nRGs - args.bIgnoreRG; ++rg) {
VerifyBamID::fullMixLLKFunc myFunc(&vbid, rg);
AmoebaMinimizer myMinimizer;
Vector startingPoint(3);
startingPoint[0] = -3.91; // start with fMix = 0.01
startingPoint[1] = 0; // pRefHet = 0.5
startingPoint[2] = -4.595; // pRefAlt = 0.01
myMinimizer.func = &myFunc;
myMinimizer.Reset(3);
myMinimizer.point = startingPoint;
myMinimizer.Minimize(1e-6);
double fMix = VerifyBamID::invLogit(myMinimizer.point[0]);
if ( fMix > 0.5 )
fMix = 1.-fMix;
double pRefHet = VerifyBamID::invLogit(myMinimizer.point[1]);
double pRefAlt = VerifyBamID::invLogit(myMinimizer.point[2]);
Logger::gLogger->writeLog("Optimal per-sample fMix = %lf\n",fMix);
Logger::gLogger->writeLog("Reference Bias Estimated as ( Pr[refBase|HET] = %lf, Pr[refBase|ALT] = %lf) with LLK = %lf",pRefHet,pRefAlt,myMinimizer.fmin);
//vbid.setRefBiasParams(1.0, pRefHet, pRefAlt);
vbid.mixOut.llk0s[rg+1] = myFunc.llk0;
vbid.mixOut.llk1s[rg+1] = myFunc.llk1;
vbid.mixOut.fMixs[rg+1] = myFunc.fMix;
vbid.mixOut.refHets[rg+1] = myFunc.pRefHet;
vbid.mixOut.refAlts[rg+1] = myFunc.pRefAlt;
}
}
Logger::gLogger->writeLog("calculating depth distribution");
vbid.calculateDepthDistribution(args.maxDepth, vbid.mixOut);
Logger::gLogger->writeLog("finished calculating depth distribution");
std::vector<int> bestInds(vbid.nRGs+1,-1);
std::vector<int> selfInds(vbid.nRGs+1,-1);
if ( args.bChipNone ) {
// do nothing
Logger::gLogger->writeLog("Skipping with-chip estimation of sample mixture");
}
else if ( args.bChipMixOnly ) {
Logger::gLogger->writeLog("Performing with-chip estimation of sample mixture at fixed reference bias parameter (%lf, %lf, %lf)",args.pRefRef,args.pRefHet,args.pRefAlt);
for(int rg=-1; rg < (vbid.nRGs - (int)args.bIgnoreRG); ++rg) {
double maxIBD = -1;
VerifyBamID::ibdLLK ibd(&vbid);
for(int i=0; i < (int)vbid.pGenotypes->indids.size(); ++i) {
double fIBD = ibd.OptimizeLLK(i, rg);
Logger::gLogger->writeLog("Comparing with individual %s.. Optimal fIBD = %lf, LLK0 = %lf, LLK1 = %lf for readgroup %d",vbid.pGenotypes->indids[i].c_str(),fIBD, ibd.llk0, ibd.llk1, rg);
if ( maxIBD < fIBD ) {
bestInds[rg+1] = i;
vbid.bestOut.llk0s[rg+1] = ibd.llk0;
vbid.bestOut.llk1s[rg+1] = ibd.llk1;
vbid.bestOut.fMixs[rg+1] = 1-ibd.fIBD;
maxIBD = ibd.fIBD;
}
if ( ( (rg < 0) && (vbid.pPile->sBamSMID == vbid.pGenotypes->indids[i] ) ) || ( ( rg >= 0 ) && ( vbid.pPile->vsSMIDs[rg] == vbid.pGenotypes->indids[i]) ) ) {
selfInds[rg+1] = i;
vbid.selfOut.llk0s[rg+1] = ibd.llk0;
vbid.selfOut.llk1s[rg+1] = ibd.llk1;
vbid.selfOut.fMixs[rg+1] = 1-ibd.fIBD;
}
}
if ( bestInds[rg+1] >= 0 ) {
Logger::gLogger->writeLog("Best Matching Individual is %s with IBD = %lf",vbid.pGenotypes->indids[bestInds[rg+1]].c_str(),maxIBD);
vbid.calculateDepthByGenotype(bestInds[rg+1],rg,vbid.bestOut);
}
if ( selfInds[rg+1] >= 0 ) {
Logger::gLogger->writeLog("Self Individual is %s with IBD = %lf",vbid.pGenotypes->indids[selfInds[rg+1]].c_str(),vbid.selfOut.fMixs[rg+1]);
vbid.calculateDepthByGenotype(selfInds[rg+1],rg,vbid.selfOut);
}
}
}
else if ( args.bChipRefBiasOnly ) {
Logger::gLogger->writeLog("Performing with-chip estimation of reference-bias without sample mixture");
if ( args.bSelfOnly ) {
for(int rg=-1; rg < (vbid.nRGs - (int)args.bIgnoreRG); ++rg) {
VerifyBamID::refBiasIbdLLKFunc myFunc(&vbid, rg);
AmoebaMinimizer myMinimizer;
Vector startingPoint(2);
startingPoint[0] = 0; // pRefHet = 0.5
startingPoint[1] = -4.595; // pRefAlt = 0.01
myMinimizer.func = &myFunc;
myMinimizer.Reset(2);
myMinimizer.point = startingPoint;
myMinimizer.Minimize(1e-6);
double pRefHet = VerifyBamID::invLogit(myMinimizer.point[0]);
double pRefAlt = VerifyBamID::invLogit(myMinimizer.point[1]);
Logger::gLogger->writeLog("Reference Bias Estimated as ( Pr[refBase|HET] = %lf, Pr[refBase|ALT] = %lf) with LLK = %lf",pRefHet,pRefAlt,myMinimizer.fmin);
//vbid.setRefBiasParams(1.0, pRefHet, pRefAlt);
vbid.selfOut.llk0s[rg+1] = myFunc.llk0;
vbid.selfOut.llk1s[rg+1] = myFunc.llk1;
vbid.selfOut.refHets[rg+1] = myFunc.pRefHet;
vbid.selfOut.refAlts[rg+1] = myFunc.pRefAlt;
vbid.calculateDepthByGenotype(0,rg,vbid.selfOut);
}
}
else {
Logger::gLogger->warning("--self must be set for --chip-refBias to work. Skipping..");
}
}
else if ( args.bChipFull ) {
Logger::gLogger->writeLog("Performing with-chip estimation of reference-bias and sample mixture together");
for(int rg=-1; rg < (vbid.nRGs - (int)args.bIgnoreRG); ++rg) {
double maxIBD = -1;
for(int i=0; i < (int)vbid.pGenotypes->indids.size(); ++i) {
VerifyBamID::fullIbdLLKFunc myFunc(&vbid,i,rg);
AmoebaMinimizer myMinimizer;
Vector startingPoint(3);
startingPoint[0] = 3.91; // start with fIBD = 0.99
startingPoint[1] = 0; // pRefHet = 0.5
startingPoint[2] = -4.595; // pRefAlt = 0.01
myMinimizer.func = &myFunc;
myFunc.indIdx = i;
myMinimizer.Reset(3);
myMinimizer.point = startingPoint;
myMinimizer.Minimize(1e-6);
double fIBD = VerifyBamID::invLogit(myMinimizer.point[0]);
double pRefHet = VerifyBamID::invLogit(myMinimizer.point[1]);
double pRefAlt = VerifyBamID::invLogit(myMinimizer.point[2]);
Logger::gLogger->writeLog("Comparing with individual %s.. Optimal fIBD = %lf, LLK0 = %lf, LLK1 = %lf for readgroup %d",vbid.pGenotypes->indids[i].c_str(), fIBD, myFunc.llk0, myFunc.llk1, rg);
//Logger::gLogger->writeLog("Optimal per-sample fIBD = %lf, ",fIBD);
Logger::gLogger->writeLog("Reference Bias Estimated as ( Pr[refBase|HET] = %lf, Pr[refBase|ALT] = %lf ) with LLK = %lf",pRefHet,pRefAlt,myMinimizer.fmin);
if ( maxIBD < fIBD ) {
bestInds[rg+1] = i;
maxIBD = fIBD;
vbid.bestOut.llk0s[rg+1] = myFunc.llk0;
vbid.bestOut.llk1s[rg+1] = myFunc.llk1;
vbid.bestOut.fMixs[rg+1] = 1.-myFunc.fIBD;
vbid.bestOut.refHets[rg+1] = myFunc.pRefHet;
vbid.bestOut.refAlts[rg+1] = myFunc.pRefAlt;
}
if ( ( (rg < 0) && (vbid.pPile->sBamSMID == vbid.pGenotypes->indids[i] ) ) || ( ( rg >= 0 ) && ( vbid.pPile->vsSMIDs[rg] == vbid.pGenotypes->indids[i]) ) ) {
selfInds[rg+1] = i;
vbid.selfOut.llk0s[rg+1] = myFunc.llk0;
vbid.selfOut.llk1s[rg+1] = myFunc.llk1;
vbid.selfOut.fMixs[rg+1] = 1.-myFunc.fIBD;
vbid.selfOut.refHets[rg+1] = myFunc.pRefHet;
vbid.selfOut.refAlts[rg+1] = myFunc.pRefAlt;
vbid.calculateDepthByGenotype(i, rg, vbid.selfOut);
}
}
//vbid.setRefBiasParams(1.0, pRefHet, pRefAlt);
if ( bestInds[rg+1] >= 0 ) {
Logger::gLogger->writeLog("Best Matching Individual is %s with IBD = %lf",vbid.pGenotypes->indids[bestInds[rg+1]].c_str(),maxIBD);
vbid.calculateDepthByGenotype(bestInds[rg+1], rg, vbid.bestOut);
}
if ( selfInds[rg+1] >= 0 ) {
Logger::gLogger->writeLog("Self Individual is %s with IBD = %lf",vbid.pGenotypes->indids[selfInds[rg+1]].c_str(),vbid.selfOut.fMixs[rg+1]);
vbid.calculateDepthByGenotype(selfInds[rg+1],rg,vbid.selfOut);
}
}
}
// PRINT OUTPUT FILE - ".selfSM"
// [SEQ_ID] : SAMPLE ID in the sequence file
// [CHIP_ID] : SAMPLE ID in the chip file (NA if not available)
// [#SNPS] : Number of markers evaluated
// [#READS] : Number of reads evaluated
// [AVG_DP] : Mean depth
// [FREEMIX] : Chip-free estimated alpha (% MIX in 0-1 scale), NA if unavailable
// [FREELK1] : Chip-free log-likelihood at estimated alpha
// [FREELK0] : Chip-free log-likelihood at 0% contamination
// [CHIPIBD] : With-chip estimated alpha (% MIX in 0-1 scale)
// [CHIPLK1] : With-chip log-likelihood at estimated alpha
// [CHIPLK0] : With-chip log-likelihood at 0% contamination
// [DPREF] : Depth at reference site in the chip
// [RDPHET] : Relative depth at HET site in the chip
// [RDPALT] : Relative depth at HOMALT site in the chip
// [FREE_RF] : Pr(Ref|Ref) site estimated without chip data
// [FREE_RH] : Pr(Ref|Het) site estimated without chip data
// [FREE_RA] : Pr(Ref|Alt) site estimated without chip data
// [CHIP_RF] : Pr(Ref|Ref) site estimated with chip data
// [CHIP_RH] : Pr(Ref|Het) site estimated with chip data
// [CHIP_RA] : Pr(Ref|Alt) site estimated with chip data
// [DPREF] : Depth at reference alleles
// [RDPHET] : Relative depth at heterozygous alleles
// [RDPALT] : Relative depth at hom-alt alleles
String selfSMFN = args.sOutFile + ".selfSM";
String bestSMFN = args.sOutFile + ".bestSM";
String selfRGFN = args.sOutFile + ".selfRG";
String bestRGFN = args.sOutFile + ".bestRG";
String dpSMFN = args.sOutFile + ".depthSM";
String dpRGFN = args.sOutFile + ".depthRG";
IFILE selfSMF = ifopen(selfSMFN,"wb");
IFILE bestSMF = (args.bFindBest ? ifopen(bestSMFN,"wb") : NULL);
IFILE selfRGF = (args.bIgnoreRG ? NULL : ifopen(selfRGFN,"wb"));
IFILE bestRGF = (args.bFindBest && !args.bIgnoreRG) ? ifopen(bestRGFN,"wb") : NULL;
IFILE dpSMF = ifopen(dpSMFN,"wb");
IFILE dpRGF = (args.bIgnoreRG ? NULL : ifopen(dpRGFN,"wb"));
if ( selfSMF == NULL ) {
Logger::gLogger->error("Cannot write to %s",selfSMF);
}
if ( args.bFindBest && ( bestSMF == NULL ) ) {
Logger::gLogger->error("Cannot write to %s",bestSMF);
}
if ( dpSMF == NULL ) {
Logger::gLogger->error("Cannot write to %s",dpSMF);
}
ifprintf(dpSMF,"#RG\tDEPTH\t#SNPs\t%%SNPs\t%%CUMUL\n");
int nCumMarkers = 0;
for(int i=args.maxDepth; i >= 0; --i) {
nCumMarkers += vbid.mixOut.depths[i];
ifprintf(dpSMF,"ALL\t%d\t%d\t%.5lf\t%.5lf\n",i, vbid.mixOut.depths[i],(double) vbid.mixOut.depths[i]/(double)vbid.nMarkers,(double)nCumMarkers/(double)vbid.nMarkers);
}
ifclose(dpSMF);
if ( dpRGF != NULL ) {
ifprintf(dpRGF,"#RG\tDEPTH\t#SNPs\t%%SNPs\t%%CUMUL\n");
for(int rg=0; rg < (vbid.nRGs - (int)args.bIgnoreRG); ++rg) {
const char* rgID = vbid.pPile->vsRGIDs[rg].c_str();
int nMarkers = 0;
for(int i=args.maxDepth; i >= 0; --i) {
nMarkers += vbid.mixOut.depths[(rg+1)*(args.maxDepth+1) + i];
}
nCumMarkers = 0;
for(int i=args.maxDepth; i >= 0; --i) {
int d = vbid.mixOut.depths[(rg+1)*(args.maxDepth+1) + i];
nCumMarkers += d;
ifprintf(dpRGF,"%s\t%d\t%d\t%.5lf\t%.5lf\n",rgID,i,d,(double)d/(double)vbid.nMarkers,(double)nCumMarkers/(double)nMarkers);
}
}
ifclose(dpRGF);
}
const char* headers[] = {"#SEQ_ID","RG","CHIP_ID","#SNPS","#READS","AVG_DP","FREEMIX","FREELK1","FREELK0","FREE_RH","FREE_RA","CHIPMIX","CHIPLK1","CHIPLK0","CHIP_RH","CHIP_RA","DPREF","RDPHET","RDPALT"};
int nheaders = sizeof(headers)/sizeof(headers[0]);
for(int i=0; i < nheaders; ++i) { ifprintf(selfSMF,"%s%s",i>0 ? "\t" : "",headers[i]); }
ifprintf(selfSMF,"\n");
ifprintf(selfSMF,"%s\tALL",vbid.pPile->sBamSMID.c_str());
ifprintf(selfSMF,"\t%s",selfInds[0] >= 0 ? vbid.pGenotypes->indids[selfInds[0]].c_str() : "NA");
ifprintf(selfSMF,"\t%d\t%d\t%.2lf",vbid.nMarkers,vbid.mixOut.numReads[0],(double)vbid.mixOut.numReads[0]/(double)vbid.nMarkers);
if ( args.bFreeNone ) { ifprintf(selfSMF,"\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bFreeMixOnly ) { ifprintf(selfSMF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA",vbid.mixOut.fMixs[0],vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0]); }
else if ( args.bFreeRefBiasOnly ) { ifprintf(selfSMF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0],vbid.mixOut.refHets[0],vbid.mixOut.refAlts[0]); }
else if ( args.bFreeFull ) { ifprintf(selfSMF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.fMixs[0],vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0],vbid.mixOut.refHets[0],vbid.mixOut.refAlts[0]); }
else { error("Invalid option in handling bFree"); }
if ( args.bChipNone || bestInds[0] < 0 ) { ifprintf(selfSMF,"\tNA\tNA\tNA\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bChipMixOnly ) { ifprintf(selfSMF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA\t%.3lf\t%.4lf\t%.4lf",vbid.selfOut.fMixs[0],vbid.selfOut.llk1s[0],vbid.selfOut.llk0s[0],(double)vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[1], (double)vbid.selfOut.numReads[2]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[2], (double)vbid.selfOut.numReads[3]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[3]); }
else if ( args.bChipMixOnly ) { ifprintf(selfSMF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf",vbid.selfOut.llk1s[0], vbid.selfOut.llk0s[0], vbid.selfOut.refHets[0], vbid.selfOut.refAlts[0], (double)vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[1], (double)vbid.selfOut.numReads[2]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[2], (double)vbid.selfOut.numReads[3]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[3]); }
else if ( args.bChipFull ) { ifprintf(selfSMF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf", vbid.selfOut.fMixs[0], vbid.selfOut.llk1s[0], vbid.selfOut.llk0s[0], vbid.selfOut.refHets[0], vbid.selfOut.refAlts[0], (double)vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[1], (double)vbid.selfOut.numReads[2]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[2], (double)vbid.selfOut.numReads[3]*vbid.selfOut.numGenos[1]/vbid.selfOut.numReads[1]/vbid.selfOut.numGenos[3]); }
else { error("Invalid option in handling bChip"); }
ifprintf(selfSMF,"\n");
ifclose(selfSMF);
if ( bestSMF != NULL ) {
for(int i=0; i < nheaders; ++i) { ifprintf(bestSMF,"%s%s",i>0 ? "\t" : "",headers[i]); }
ifprintf(bestSMF,"\n");
ifprintf(bestSMF,"%s\tALL",vbid.pPile->sBamSMID.c_str());
ifprintf(bestSMF,"\t%s",bestInds[0] >= 0 ? vbid.pGenotypes->indids[bestInds[0]].c_str() : "NA");
ifprintf(bestSMF,"\t%d\t%d\t%.2lf",vbid.nMarkers,vbid.mixOut.numReads[0],(double)vbid.mixOut.numReads[0]/(double)vbid.nMarkers);
if ( args.bFreeNone ) { ifprintf(bestSMF,"\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bFreeMixOnly ) { ifprintf(bestSMF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA",vbid.mixOut.fMixs[0],vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0]); }
else if ( args.bFreeRefBiasOnly ) { ifprintf(bestSMF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0],vbid.mixOut.refHets[0],vbid.mixOut.refAlts[0]); }
else if ( args.bFreeFull ) { ifprintf(bestSMF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.fMixs[0],vbid.mixOut.llk1s[0],vbid.mixOut.llk0s[0],vbid.mixOut.refHets[0],vbid.mixOut.refAlts[0]); }
else { error("Invalid option in handling bFree"); }
if ( args.bChipNone || bestInds[0] < 0 ) { ifprintf(bestSMF,"\tNA\tNA\tNA\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bChipMixOnly ) { ifprintf(bestSMF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA\t%.3lf\t%.4lf\t%.4lf",vbid.bestOut.fMixs[0],vbid.bestOut.llk1s[0],vbid.bestOut.llk0s[0],(double)vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[1], (double)vbid.bestOut.numReads[2]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[2], (double)vbid.bestOut.numReads[3]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[3]); }
else if ( args.bChipMixOnly ) { ifprintf(bestSMF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf",vbid.bestOut.llk1s[0], vbid.bestOut.llk0s[0], vbid.bestOut.refHets[0], vbid.bestOut.refAlts[0], (double)vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[1], (double)vbid.bestOut.numReads[2]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[2], (double)vbid.bestOut.numReads[3]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[3]); }
else if ( args.bChipFull ) { ifprintf(bestSMF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf", vbid.bestOut.fMixs[0], vbid.bestOut.llk1s[0], vbid.bestOut.llk0s[0], vbid.bestOut.refHets[0], vbid.bestOut.refAlts[0], (double)vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[1], (double)vbid.bestOut.numReads[2]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[2], (double)vbid.bestOut.numReads[3]*vbid.bestOut.numGenos[1]/vbid.bestOut.numReads[1]/vbid.bestOut.numGenos[3]); }
else { error("Invalid option in handling bChip"); }
ifprintf(bestSMF,"\n");
ifclose(bestSMF);
}
if ( selfRGF != NULL ) {
for(int i=0; i < nheaders; ++i) { ifprintf(selfRGF,"%s%s",i>0 ? "\t" : "",headers[i]); }
ifprintf(selfRGF,"\n");
for(int rg=0; rg < vbid.nRGs; ++rg) {
ifprintf(selfRGF,"%s\t%s",vbid.pPile->sBamSMID.c_str(),vbid.pPile->vsRGIDs[rg].c_str());
ifprintf(selfRGF,"\t%s",bestInds[rg] >= 0 ? vbid.pGenotypes->indids[bestInds[rg]].c_str() : "NA");
ifprintf(selfRGF,"\t%d\t%d\t%.2lf",vbid.nMarkers,vbid.mixOut.numReads[(rg+1)*4],(double)vbid.mixOut.numReads[(rg+1)*4]/(double)vbid.mixOut.numGenos[(rg+1)*4]);
if ( args.bFreeNone ) { ifprintf(selfRGF,"\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bFreeMixOnly ) { ifprintf(selfRGF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA",vbid.mixOut.fMixs[rg+1],vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1]); }
else if ( args.bFreeRefBiasOnly ) { ifprintf(selfRGF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1],vbid.mixOut.refHets[rg+1],vbid.mixOut.refAlts[rg+1]); }
else if ( args.bFreeFull ) { ifprintf(selfRGF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.fMixs[rg+1],vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1],vbid.mixOut.refHets[rg+1],vbid.mixOut.refAlts[rg+1]); }
else { error("Invalid option in handling bFree"); }
if ( args.bChipNone || bestInds[0] < 0 ) { ifprintf(selfRGF,"\tNA\tNA\tNA\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bChipMixOnly ) { ifprintf(selfRGF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA\t%.3lf\t%.4lf\t%.4lf",vbid.selfOut.fMixs[rg+1], vbid.selfOut.llk1s[rg+1], vbid.selfOut.llk0s[rg+1], (double)vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+1], (double)vbid.selfOut.numReads[(rg+1)*4+2]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+2], (double)vbid.selfOut.numReads[(rg+1)*4+3]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+3]); }
else if ( args.bChipMixOnly ) { ifprintf(selfRGF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf",vbid.selfOut.llk1s[rg+1], vbid.selfOut.llk0s[rg+1], vbid.selfOut.refHets[rg+1], vbid.selfOut.refAlts[rg+1], (double)vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+1], (double)vbid.selfOut.numReads[(rg+1)*4+2]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4]/vbid.selfOut.numGenos[(rg+1)*4+2], (double)vbid.selfOut.numReads[(rg+1)*4+3]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+3]); }
else if ( args.bChipFull ) { ifprintf(selfRGF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf", vbid.selfOut.fMixs[rg+1], vbid.selfOut.llk1s[rg+1], vbid.selfOut.llk0s[rg+1], vbid.selfOut.refHets[rg+1], vbid.selfOut.refAlts[rg+1], (double)vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+1], (double)vbid.selfOut.numReads[(rg+1)*4+2]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+2], (double)vbid.selfOut.numReads[(rg+1)*4+3]*vbid.selfOut.numGenos[(rg+1)*4+1]/vbid.selfOut.numReads[(rg+1)*4+1]/vbid.selfOut.numGenos[(rg+1)*4+3]); }
else { error("Invalid option in handling bChip"); }
ifprintf(selfRGF,"\n");
}
ifclose(selfRGF);
}
if ( bestRGF != NULL ) {
for(int i=0; i < nheaders; ++i) { ifprintf(bestRGF,"%s%s",i>0 ? "\t" : "",headers[i]); }
ifprintf(bestRGF,"\n");
for(int rg=0; rg < vbid.nRGs; ++rg) {
ifprintf(bestRGF,"%s\t%s",vbid.pPile->sBamSMID.c_str(),vbid.pPile->vsRGIDs[rg].c_str());
ifprintf(bestRGF,"\t%s",bestInds[rg] >= 0 ? vbid.pGenotypes->indids[bestInds[rg]].c_str() : "NA");
ifprintf(bestRGF,"\t%d\t%d\t%.2lf",vbid.nMarkers,vbid.mixOut.numReads[(rg+1)*4],(double)vbid.mixOut.numReads[(rg+1)*4]/(double)vbid.mixOut.numGenos[(rg+1)*4]);
if ( args.bFreeNone ) { ifprintf(bestRGF,"\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bFreeMixOnly ) { ifprintf(bestRGF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA",vbid.mixOut.fMixs[rg+1],vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1]); }
else if ( args.bFreeRefBiasOnly ) { ifprintf(bestRGF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1],vbid.mixOut.refHets[rg+1],vbid.mixOut.refAlts[rg+1]); }
else if ( args.bFreeFull ) { ifprintf(bestRGF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf",vbid.mixOut.fMixs[rg+1],vbid.mixOut.llk1s[rg+1],vbid.mixOut.llk0s[rg+1],vbid.mixOut.refHets[rg+1],vbid.mixOut.refAlts[rg+1]); }
else { error("Invalid option in handling bFree"); }
if ( args.bChipNone || bestInds[0] < 0 ) { ifprintf(bestRGF,"\tNA\tNA\tNA\tNA\tNA\tNA\tNA\tNA"); }
else if ( args.bChipMixOnly ) { ifprintf(bestRGF,"\t%.5lf\t%.2lf\t%.2lf\tNA\tNA\t%.3lf\t%.4lf\t%.4lf",vbid.bestOut.fMixs[rg+1], vbid.bestOut.llk1s[rg+1], vbid.bestOut.llk0s[rg+1], (double)vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+1], (double)vbid.bestOut.numReads[(rg+1)*4+2]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+2], (double)vbid.bestOut.numReads[(rg+1)*4+3]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+3]); }
else if ( args.bChipMixOnly ) { ifprintf(bestRGF,"\tNA\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf",vbid.bestOut.llk1s[rg+1], vbid.bestOut.llk0s[rg+1], vbid.bestOut.refHets[rg+1], vbid.bestOut.refAlts[rg+1], (double)vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+1], (double)vbid.bestOut.numReads[(rg+1)*4+2]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4]/vbid.bestOut.numGenos[(rg+1)*4+2], (double)vbid.bestOut.numReads[(rg+1)*4+3]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+3]); }
else if ( args.bChipFull ) { ifprintf(bestRGF,"\t%.5lf\t%.2lf\t%.2lf\t%.5lf\t%.5lf\t%.3lf\t%.4lf\t%.4lf", vbid.bestOut.fMixs[rg+1], vbid.bestOut.llk1s[rg+1], vbid.bestOut.llk0s[rg+1], vbid.bestOut.refHets[rg+1], vbid.bestOut.refAlts[rg+1], (double)vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+1], (double)vbid.bestOut.numReads[(rg+1)*4+2]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+2], (double)vbid.bestOut.numReads[(rg+1)*4+3]*vbid.bestOut.numGenos[(rg+1)*4+1]/vbid.bestOut.numReads[(rg+1)*4+1]/vbid.bestOut.numGenos[(rg+1)*4+3]); }
else { error("Invalid option in handling bChip"); }
ifprintf(bestRGF,"\n");
}
ifclose(bestRGF);
}
time(&t);
Logger::gLogger->writeLog("Analysis finished on %s",ctime(&t));
return 0;
}
void main(int argc, char **argv)
{
int c;
int lib_opt = FALSE;
int errflg = 0;
extern int optind;
extern char *optarg;
char *libname,*fname;
char *t_name;
#ifdef IBM_PC
fprintf(stderr, "NYU Ada/Ed Librarian Version 1.11.2\n");
fprintf(stderr, "Copyright (C) 1985-1992 by New York University.\n");
#endif
#ifndef IBM_PC
while((c = getopt(argc,argv,"l:"))!=EOF) {
#else
while((c = getopt(argc,argv,"L:l:"))!=EOF) {
if (isupper(c)) c = tolower(c);
#endif
switch(c) {
case 'l': /* specify library name */
lib_opt = TRUE;
libname = strjoin(optarg,"");
break;
case '?':
errflg++;
}
}
fname = (char *)0;
if (optind < argc) fname = argv[optind];
if (!lib_opt && fname == (char *)0) {
fname = getenv("ADALIB");
if (fname!= (char *)0) {
#ifdef IBM_PC
fprintf(stderr, "L");
#else
fprintf(stderr, "l");
#endif
fprintf(stderr,"ibrary defined by ADALIB: %s\n", fname);
}
}
if ((!lib_opt && fname == (char *)0) || errflg) {
fprintf(stderr, "Usage: adalib [-l library]\n");
exit(1);
}
if (!lib_opt) {
libname = emalloc(strlen(fname) + 1);
strcpy(libname, fname);
}
t_name = libset(libname);
LIBFILE = ifopen(LIBFILENAME, "", "r", 0);
load_library();
exit(0);
}
static void load_library() /*;load_library*/
{
/*
* retrieve information from LIBFILE
* Called only if lib_option and not newlib.
*/
int i, j, n, m, unumber, nodes, symbols;
int comp_status, unit_count, cur_level;
char *comp_date, *status_str, *uname, *aisname, *tmp_str;
char *main_string;
int is_main, empty_unit_slots, parent;
int ignore;
unit_count = getnum(LIBFILE, "lib-unit-count");
n = getnum(LIBFILE, "lib-n");
empty_unit_slots = getnum(LIBFILE, "lib-empty-slots");
tmp_str = getstr(LIBFILE, "lib-tmp-str");
for (i = 1; i <= unit_count; i++) {
uname = getstr(LIBFILE, "lib-unit-name");
unumber = getnum(LIBFILE, "lib-unit-number");
aisname = getstr(LIBFILE, "lib-ais-name");
comp_date = getstr(LIBFILE, "unit-date");
symbols = getnum(LIBFILE, "lib-symbols");
nodes = getnum(LIBFILE, "lib-nodes");
is_main = getnum(LIBFILE, "lib-is-main");
if (is_main) {
if (streq(unit_name_type(uname), "ma"))
main_string = "(Interface)";
else
main_string = " (Main) ";
}
else {
main_string = "";
}
comp_status = getnum(LIBFILE, "lib-status");
status_str = (comp_status) ? "active " : "obsolete";
printf("%8s %11s %-15s %s\n",
status_str, main_string, convert_date(comp_date),
formatted_name(uname));
}
printf("\n");
n = getnum(LIBFILE, "lib-n");
if (n) {
printf("stubs \n\n");
for (i = 1; i <= n; i++) {
uname = getstr(LIBFILE, "lib-unit-name");
aisname = getstr(LIBFILE, "lib-ais-name");
parent = getnum(LIBFILE, "lib-parent");
cur_level = getnum(LIBFILE, "lib-cur-level");
m = getnum(LIBFILE, "stub-file-size");
for (j = 1; j <= m; j++)
ignore = getnum(LIBFILE, "stub-file");
printf("%s\n", formatted_stub(uname));
}
printf("\n");
}
ifclose(LIBFILE);
return;
#ifdef TBSL
n = getnum(LIBFILE, "precedes-map-size");
printf("precedes map\n");
for (i = 1; i <= n; i += 2) {
dom = getnum(LIBFILE, "precedes-map-dom");
m = getnum(LIBFILE, "precedes-map-nelt");
printf(" %4d:", dom);
for (j = 1; j <= m; j++) {
range = getnum(LIBFILE, "precedes-map-ent");
printf(" %4d", range);
}
printf("\n");
}
n = getnum(LIBFILE, "compilation_table_size");
if (n) {
printf("\ncompilation table\n");
for (i = 1; i <= n; i++) {
unum = (int) getnum(LIBFILE, "compilation-table-ent");
printf(" %d\n", unum);
}
printf("\n");
}
/* late_instances */
n = getnum(LIBFILE, "late-instances-size");
if (n) {
printf("late instances\n");
for (i = 1; i <= n; i++) {
str = (char *) getstr(LIBFILE, "late-instances-str");
printf(" %s\n", str);
}
}
/* current code segment */
n = getnum(LIBFILE, "unit-size");
printf("\ncurrent code segments\n");
printf(" unit cs\n");
for (i = 1; i <= n; i++) {
cs = getnum(LIBFILE, "current-code-segment");
if (cs) printf(" %d: %d\n", i, cs);
}
/* local reference maps */
n = getnum(LIBFILE, "unit-size");
get_local_ref_maps(LIBFILE, n);
cde_pos = get_cde_slots(LIBFILE, axq);
/* could free axq_data_slots, etc., but keep for now */
/* read out LIB_STUB map (always empty for now) */
ifclose(LIBFILE);
return;
#endif
}