void VcfHelper::printArrayJoin(IFILE oFile, const StringArray& arr, const char* sep, const char* empty, int start, int end) {
for(int i=start; i < end; ++i) {
if ( i > start ) {
ifprintf(oFile,"%s",sep);
}
ifprintf(oFile,"%s",arr[i].c_str());
}
}
void VcfHelper::printArrayDoubleJoin(IFILE oFile, const StringArray& arr1, const StringArray& arr2, const char* sep1, const char* sep2, const char* empty, int start, int end) {
for(int i=start; i < end; ++i) {
if ( i > start ) {
ifprintf(oFile,"%s",sep1);
}
ifprintf(oFile,"%s%s%s",arr1[i].c_str(),sep2,arr2[i].c_str());
}
}
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);
}
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 VcfHelper::printArrayJoin(IFILE oFile, const StringArray& arr, const char* sep, const char* empty) {
int len = arr.Length();
if ( len == 0 ) {
ifprintf(oFile,"%s",empty);
}
else if ( len == 1 ) {
ifprintf(oFile,"%s",arr[0].c_str());
}
else {
printArrayJoin(oFile,arr,sep,empty,0,len);
}
}
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 VcfFile::printVCFHeader(IFILE oFile) {
for(int i=0; i < getMetaCount(); ++i) {
ifprintf(oFile,"##%s=%s\n",getMetaKey(i).c_str(), getMetaValue(i, "<na>").c_str());
}
ifprintf(oFile,"#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO");
if ( ( getSampleCount() > 0 ) && ( !bSiteOnly ) ) {
ifprintf(oFile,"\tFORMAT");
for(int i=0; i < getSampleCount(); ++i) {
ifprintf(oFile,"\t%s",vpVcfInds[i]->sIndID.c_str());
}
}
ifprintf(oFile,"\n");
}
void VcfFile::printVCFHeaderSubset(IFILE oFile, std::vector<int>& subsetIndices) {
//fprintf(stderr,"foo\n");
for(int i=0; i < getMetaCount(); ++i) {
ifprintf(oFile,"##%s=%s\n",getMetaKey(i).c_str(), getMetaValue(i, "<na>").c_str());
}
ifprintf(oFile,"#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO");
ifprintf(oFile,"\tFORMAT");
for(int j=0; j < (int)subsetIndices.size(); ++j) {
int i = subsetIndices[j];
ifprintf(oFile,"\t%s",vpVcfInds[i]->sIndID.c_str());
}
ifprintf(oFile,"\n");
}
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 Bam2FastQ::writeFastQ(SamRecord& samRec, IFILE filePtr,
const char* readNameExt)
{
static int16_t flag;
static std::string sequence;
static String quality;
if(filePtr == NULL)
{
return;
}
flag = samRec.getFlag();
const char* readName = samRec.getReadName();
sequence = samRec.getSequence();
quality = samRec.getQuality();
if(SamFlag::isReverse(flag) && myReverseComp)
{
// It is reverse, so reverse compliment the sequence
BaseUtilities::reverseComplement(sequence);
// Reverse the quality.
quality.Reverse();
}
else
{
// Ensure it is all capitalized.
int seqLen = sequence.size();
for (int i = 0; i < seqLen; i++)
{
sequence[i] = (char)toupper(sequence[i]);
}
}
if(myRNPlus)
{
ifprintf(filePtr, "@%s%s\n%s\n+%s%s\n%s\n", readName, readNameExt,
sequence.c_str(), readName, readNameExt, quality.c_str());
}
else
{
ifprintf(filePtr, "@%s%s\n%s\n+\n%s\n", readName, readNameExt,
sequence.c_str(), quality.c_str());
}
// Release the record.
myPool.releaseRecord(&samRec);
}
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 VerifyBamID::printPerMarkerInfo(const char* filename, int indIdx) {
IFILE oFile = ifopen(filename,"wb");
int nMarkers = (int)(pGenotypes->chroms.size());
char base, a1, a2;
ifprintf(oFile,"#CHROM\tPOS\tA1\tA2\tAF\tGENO\t#REF\t#ALT\t#OTHERS\tBASES\tQUALS\tMAPQS\n");
for(int i=0; i < nMarkers; ++i) {
int counts[3] = {0,0,0};
std::vector<char> bases;
std::vector<char> quals;
std::vector<char> mqs;
ifprintf(oFile,"%s\t%d\t%c\t%c\t%.4lf\t",pGenotypes->chroms[i].c_str(),pGenotypes->positions[i],pGenotypes->refBases[i],pGenotypes->altBases[i],pGenotypes->alleleFrequencies[i]);
int geno = pGenotypes->getGenotype(indIdx,i);
switch(geno) {
case 0: // MISSING
ifprintf(oFile,"./.");
break;
case 1: // HOMREF;
ifprintf(oFile,"0/0");
break;
case 2: // HET;
ifprintf(oFile,"0/1");
break;
case 3: // HOMALT;
ifprintf(oFile,"1/1");
break;
default:
Logger::gLogger->error("Unrecognized genotype %d at ind %d, marker %d",indIdx,i);
}
a1 = pGenotypes->refBases[i];
a2 = pGenotypes->altBases[i];
for(int j=(int)pPile->nBegins[i]; j < (int)pPile->nEnds[i]; ++j) {
// obtain b (base), (error), and readgroup info
base = pPile->cBases[j];
if ( base == a1 ) {
++counts[0];
}
else if ( base == a2 ) {
++counts[1];
}
else {
++counts[2];
}
bases.push_back(base);
quals.push_back(pPile->cQuals[j]);
mqs.push_back(((uint8_t)(pPile->cMapQs[j]) > 90) ? '~' : static_cast<char>(pPile->cMapQs[j]+33));
}
ifprintf(oFile,"\t%d\t%d\t%d\t%.3lf\t",counts[0],counts[1],counts[2],(counts[0]+counts[1] == 0) ? 0.5 : (double)counts[0]/(double)(counts[0]+counts[1]));
ifprintf(oFile,"\t");
for(int j=0; j < (int)bases.size(); ++j)
ifprintf(oFile,"%c",bases[j]);
ifprintf(oFile,"\t");
for(int j=0; j < (int)quals.size(); ++j)
ifprintf(oFile,"%c",quals[j]);
ifprintf(oFile,"\t");
for(int j=0; j < (int)mqs.size(); ++j)
ifprintf(oFile,"%c",mqs[j]);
ifprintf(oFile,"\n");
}
}
void Imputation::PrintHaplotypeData(HaplotypeSet &rHap,HaplotypeSet &tHap,
IFILE hapdose, IFILE haps,
vector<float> &ThisimputedHap,vector<bool> ThisimputedAlleles,
int ThisHapId, int ThisSampleId)
{
char labels[]= {0, 'A', 'C', 'G', 'T', 'D', 'I', 'R'};
printf(" Outputting HAPLO%d of Individual %s for Haplotype File...",
tHap.AllMaleTarget?1:(ThisHapId%2+1) ,tHap.individualName[ThisSampleId].c_str());
cout<<endl;
ifprintf(hapdose, "%s\tHAPLO%d", tHap.individualName[ThisSampleId].c_str(), tHap.AllMaleTarget?1:(ThisHapId%2+1) );
ifprintf(haps, "%s\tHAPLO%d\t", tHap.individualName[ThisSampleId].c_str(), tHap.AllMaleTarget?1:(ThisHapId%2+1) );
int i=0;
for (int index =0; index < rHap.RefTypedTotalCount; index++)
{
if(rHap.RefTypedIndex[index]==-1)
{
if(i>=rHap.PrintStartIndex && i <= rHap.PrintEndIndex)
{
ifprintf(hapdose, "\t%.5f", ThisimputedHap[i]);
ifprintf(haps, "%c", labels[(int) (ThisimputedAlleles[i]?
rHap.VariantList[i].altAllele
:rHap.VariantList[i].refAllele)]);
}
i++;
}
else
{
int MarkerIndex=rHap.RefTypedIndex[index];
bool a1;
a1=tHap.GWASOnlyMissingSampleUnscaffolded[ThisHapId][MarkerIndex];
double outAllele1=0.0;
if(a1)
{
outAllele1=tHap.AlleleFreq[MarkerIndex];
a1=round(outAllele1)==1?true:false;
}
else
{
a1=tHap.GWASOnlyhaplotypesUnscaffolded[ThisHapId][MarkerIndex];
if(a1)
outAllele1=1.0;
}
// if(!tHap.major[MarkerIndex])
// outAllele1=1-outAllele1;
ifprintf(haps, "%c", labels[(int) (a1?
tHap.TypedOnlyVariantList[MarkerIndex].altAllele
:tHap.TypedOnlyVariantList[MarkerIndex].refAllele)]);
ifprintf(hapdose, "\t%.5f",outAllele1);
}
}
ifprintf(hapdose, "\n");
ifprintf(haps, "\n");
}
void VcfFile::printBEDHeader(IFILE oBedFile, IFILE oFamFile) {
for(int i=0; i < getSampleCount(); ++i) {
if ( vpVcfInds[i]->sFamID.Length() == 0 ) {
ifprintf(oFamFile,"%s",vpVcfInds[i]->sIndID.c_str());
}
else {
ifprintf(oFamFile,"%s",vpVcfInds[i]->sFamID.c_str());
}
ifprintf(oFamFile,"\t%s",vpVcfInds[i]->sIndID.c_str());
if ( vpVcfInds[i]->sFatID.Length() == 0 ) {
ifprintf(oFamFile,"\t0");
}
else {
ifprintf(oFamFile,"%s",vpVcfInds[i]->sFatID.c_str());
}
if ( vpVcfInds[i]->sMotID.Length() == 0 ) {
ifprintf(oFamFile,"\t0");
}
else {
ifprintf(oFamFile,"\t%s",vpVcfInds[i]->sMotID.c_str());
}
switch( vpVcfInds[i]->gender ) {
case VcfInd::UNKNOWN:
ifprintf(oFamFile,"\t0");
break;
case VcfInd::MALE:
ifprintf(oFamFile,"\t1");
break;
case VcfInd::FEMALE:
ifprintf(oFamFile,"\t2");
break;
default:
throw VcfFileException("Unrecognized value for gender");
break;
}
ifprintf(oFamFile,"\t-9\n");
}
char magicNumbers[3] = {0x6c,0x1b,0x01};
oBedFile->ifwrite(magicNumbers, 3);
}
void Imputation::PrintDosageData(HaplotypeSet &rHap,HaplotypeSet &tHap,
IFILE dosages, vector<float> &ThisDosage,
int ThisSampleId)
{
printf(" Outputting Individual %s for Dosage file...", tHap.individualName[ThisSampleId].c_str());
cout<<endl;
ifprintf(dosages, "%s\tDOSE",tHap.individualName[ThisSampleId].c_str());
int i=0;
for (int index =0; index < rHap.RefTypedTotalCount; index++)
{
if(rHap.RefTypedIndex[index]==-1)
{
if(i>=rHap.PrintStartIndex && i <= rHap.PrintEndIndex)
{
ifprintf(dosages, "\t%.3f", ThisDosage[i]);
}
i++;
}
else
{
int MarkerIndex=rHap.RefTypedIndex[index];
bool a1,a2;
double outAllele1=0.0,outAllele2=0.0;
if(tHap.AllMaleTarget)
{
a1=tHap.GWASOnlyMissingSampleUnscaffolded[ThisSampleId][MarkerIndex];
if(a1)
{
outAllele1=tHap.AlleleFreq[MarkerIndex];
a1=round(outAllele1)==1?true:false;
}
else
{
a1=tHap.GWASOnlyhaplotypesUnscaffolded[ThisSampleId][MarkerIndex];
if(a1)
outAllele1=1.0;
}
// if(!tHap.major[MarkerIndex])
// outAllele1=1-outAllele1;
ifprintf(dosages, "\t%.3f", outAllele1);
}
else
{
a1=tHap.GWASOnlyMissingSampleUnscaffolded[2*ThisSampleId][MarkerIndex];
a2=tHap.GWASOnlyMissingSampleUnscaffolded[2*ThisSampleId+1][MarkerIndex];
if(a1 || a2)
{
outAllele1=tHap.AlleleFreq[MarkerIndex];
outAllele2=outAllele1;
a1=round(outAllele1)==1?true:false;
a2=a1;
}
else
{
a1=tHap.GWASOnlyhaplotypesUnscaffolded[2*ThisSampleId][MarkerIndex];
a2=tHap.GWASOnlyhaplotypesUnscaffolded[2*ThisSampleId+1][MarkerIndex];
if(a1)
outAllele1=1.0;
if(a2)
outAllele2=1.0;
}
// if(!tHap.major[MarkerIndex])
// {
// outAllele1=1-outAllele1;
// outAllele2=1-outAllele2;
// }
ifprintf(dosages, "\t%.3f", outAllele1+outAllele2);
}
}
}
ifprintf(dosages,"\n");
}
void VcfMarker::printVCFMarker(IFILE oFile, bool siteOnly) {
String line;
ifprintf(oFile,"%s",sChrom.c_str());
ifprintf(oFile,"\t%d",nPos);
ifprintf(oFile,"\t%s",sID.c_str());
ifprintf(oFile,"\t%s",sRef.c_str());
if ( asAlts.Length() == 1 ) {
ifprintf(oFile,"\t%s",asAlts[0].c_str());
}
else {
ifprintf(oFile,"\t");
VcfHelper::printArrayJoin(oFile, asAlts, ",", ".");
}
if ( fQual < 0 ) {
ifprintf(oFile,"\t.");
}
else {
ifprintf(oFile,"\t%.0f",fQual);
}
if ( asFilters.Length() == 1 ) {
ifprintf(oFile,"\t%s",asFilters[0].c_str());
}
else {
ifprintf(oFile,"\t");
VcfHelper::printArrayJoin(oFile, asFilters, ";", "PASS");
}
ifprintf(oFile,"\t");
VcfHelper::printArrayDoubleJoin(oFile, asInfoKeys, asInfoValues, ";", "=", ".");
if ( !siteOnly ) {
if ( asSampleValues.Length() > 0 ) {
ifprintf(oFile,"\t");
VcfHelper::printArrayJoin(oFile, asFormatKeys, ":", ".");
for(int i=0; i < getSampleSize(); ++i) {
ifprintf(oFile,"\t");
VcfHelper::printArrayJoin(oFile, asSampleValues, ":", ".", i*asFormatKeys.Length(), (i+1)*asFormatKeys.Length());
}
}
else if ( vnSampleGenotypes.size() > 0 ) {
ifprintf(oFile,"\tGT",line.c_str());
for(int i=0; i < (int)vnSampleGenotypes.size(); ++i) {
if ( vnSampleGenotypes[i] == 0xffff ) {
ifprintf(oFile,"\t./.");
}
else {
ifprintf(oFile,"\t%d/%d",((vnSampleGenotypes[i] & 0xff00) >> 8),(vnSampleGenotypes[i] & 0xff));
}
}
}
}
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;
}
bool VcfRecord::write(IFILE filePtr, bool siteOnly)
{
if(filePtr == NULL)
{
myStatus.setStatus(StatGenStatus::FAIL_ORDER,
"Error writing VCF record before opening the file.");
return(false);
}
int numWritten = 0;
int numExpected = 0;
if(myChrom.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", myChrom.c_str());
numExpected += myChrom.length() + 1;
}
if(false) //my1BasedPos.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
std::string strPos = std::to_string((long long int)my1BasedPosNum);
numWritten += ifprintf(filePtr, "%s\t", strPos.c_str());
numExpected += strPos.length() + 1;
}
if(myID.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", myID.c_str());
numExpected += myID.length() + 1;
}
if(myRef.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", myRef.c_str());
numExpected += myRef.length() + 1;
}
if(myAlt.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", myAlt.c_str());
numExpected += myAlt.length() + 1;
}
if(myQual.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", myQual.c_str());
numExpected += myQual.length() + 1;
}
const std::string& filterString = myFilter.getString();
if(filterString.length() == 0)
{
numWritten += ifprintf(filePtr, ".\t");
numExpected += 2;
}
else
{
numWritten += ifprintf(filePtr, "%s\t", filterString.c_str());
numExpected += filterString.length() + 1;
}
// Write the info.
bool writeSuccess = myInfo.write(filePtr);
// Only write the format & genotype if we are not just writing siteOnly
// data and there is at least one sample
if((!siteOnly) && (myGenotype.getNumSamples() != 0))
{
writeSuccess &= myGenotype.write(filePtr);
}
// Write the new line.
numWritten += ifprintf(filePtr, "\n");
numExpected += 1;
return((numWritten == numExpected) && writeSuccess);
}
int Bam2FastQ::execute(int argc, char **argv)
{
// Extract command line arguments.
String inFile = "";
bool readName = false;
String refFile = "";
String firstOut = "";
String secondOut = "";
String unpairedOut = "";
bool interleave = false;
bool noeof = false;
bool gzip = false;
bool params = false;
myOutBase = "";
myNumMateFailures = 0;
myNumPairs = 0;
myNumUnpaired = 0;
mySplitRG = false;
myQField = "";
myNumQualTagErrors = 0;
myReverseComp = true;
myRNPlus = false;
myFirstRNExt = DEFAULT_FIRST_EXT;
mySecondRNExt = DEFAULT_SECOND_EXT;
myCompression = InputFile::DEFAULT;
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_PARAMETER_GROUP("Required Parameters")
LONG_STRINGPARAMETER("in", &inFile)
LONG_PARAMETER_GROUP("Optional Parameters")
LONG_PARAMETER("readName", &readName)
LONG_PARAMETER("splitRG", &mySplitRG)
LONG_STRINGPARAMETER("qualField", &myQField)
LONG_PARAMETER("merge", &interleave)
LONG_STRINGPARAMETER("refFile", &refFile)
LONG_STRINGPARAMETER("firstRNExt", &myFirstRNExt)
LONG_STRINGPARAMETER("secondRNExt", &mySecondRNExt)
LONG_PARAMETER("rnPlus", &myRNPlus)
LONG_PARAMETER("noReverseComp", &myReverseComp)
LONG_PARAMETER("gzip", &gzip)
LONG_PARAMETER("noeof", &noeof)
LONG_PARAMETER("params", ¶ms)
LONG_PARAMETER_GROUP("Optional OutputFile Names")
LONG_STRINGPARAMETER("outBase", &myOutBase)
LONG_STRINGPARAMETER("firstOut", &firstOut)
LONG_STRINGPARAMETER("secondOut", &secondOut)
LONG_STRINGPARAMETER("unpairedOut", &unpairedOut)
LONG_PHONEHOME(VERSION)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters",
longParameterList));
// parameters start at index 2 rather than 1.
inputParameters.Read(argc, argv, 2);
// 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);
}
if(gzip)
{
myCompression = InputFile::GZIP;
}
// 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, "
<< "but was not specified" << std::endl;
return(-1);
}
// Cannot specify both interleaved & secondOut since secondOut would be N/A.
if(interleave && !secondOut.IsEmpty())
{
usage();
inputParameters.Status();
std::cerr << "ERROR: Cannot specify --merge & --secondOut.\n";
return(-1);
}
// Cannot specify both interleaved & secondOut since secondOut would be N/A.
if(interleave && !secondOut.IsEmpty())
{
usage();
inputParameters.Status();
std::cerr << "ERROR: Cannot specify --merge & --secondOut.\n";
return(-1);
}
// Cannot specify both splitRG & firstOut/secondOut/unpairedOut
// since it needs a different file for each RG.
if(mySplitRG && (!firstOut.IsEmpty() ||
!secondOut.IsEmpty() || !unpairedOut.IsEmpty()))
{
usage();
inputParameters.Status();
std::cerr << "ERROR: Cannot specify --splitRG & --firstOut/--secondOut/--unpairedOut.\n";
std::cerr << "Use --outBase instead.\n";
return(-1);
}
// Cannot specify splitRG & output to stdout.
if(mySplitRG && (myOutBase[0] == '-'))
{
usage();
inputParameters.Status();
std::cerr << "ERROR: Cannot specify --splitRG & write to stdout.\n";
return(-1);
}
// Check to see if the out file was specified, if not, generate it from
// the input filename.
if(myOutBase == "")
{
// Just remove the extension from the input filename.
int extStart = inFile.FastFindLastChar('.');
if(extStart <= 0)
{
myOutBase = inFile;
}
else
{
myOutBase = inFile.Left(extStart);
}
}
if(mySplitRG)
{
std::string fqList = myOutBase.c_str();
fqList += ".list";
myFqList = ifopen(fqList.c_str(), "w");
ifprintf(myFqList, "MERGE_NAME\tFASTQ1\tFASTQ2\tRG\n");
}
// Check to see if the first/second/single-ended were specified and
// if not, set them.
myFirstFileNameExt = "_1.fastq";
mySecondFileNameExt = "_2.fastq";
myUnpairedFileNameExt = ".fastq";
if(interleave)
{
myFirstFileNameExt = "_interleaved.fastq";
myFirstFileNameExt = "_interleaved.fastq";
}
getFileName(firstOut, myFirstFileNameExt);
getFileName(secondOut, mySecondFileNameExt);
getFileName(unpairedOut, myUnpairedFileNameExt);
if(params)
{
inputParameters.Status();
}
// Open the files for reading/writing.
// Open prior to opening the output files,
// so if there is an error, the outputs don't get created.
SamFile samIn;
samIn.OpenForRead(inFile, &mySamHeader);
// Skip non-primary reads.
samIn.SetReadFlags(0, 0x0100);
// Open the output files if not splitting RG
if(!mySplitRG)
{
myUnpairedFile = ifopen(unpairedOut, "w", myCompression);
// Only open the first file if it is different than an already opened file.
if(firstOut != unpairedOut)
{
myFirstFile = ifopen(firstOut, "w", myCompression);
}
else
{
myFirstFile = myUnpairedFile;
}
// If it is interleaved or the 2nd file is not a new name, set it appropriately.
if(interleave || secondOut == firstOut)
{
mySecondFile = myFirstFile;
}
else if(secondOut == unpairedOut)
{
mySecondFile = myUnpairedFile;
}
else
{
mySecondFile = ifopen(secondOut, "w", myCompression);
}
if(myUnpairedFile == NULL)
{
std::cerr << "Failed to open " << unpairedOut
<< " so can't convert bam2FastQ.\n";
return(-1);
}
if(myFirstFile == NULL)
{
std::cerr << "Failed to open " << firstOut
<< " so can't convert bam2FastQ.\n";
return(-1);
}
if(mySecondFile == NULL)
{
std::cerr << "Failed to open " << secondOut
<< " so can't convert bam2FastQ.\n";
return(-1);
}
}
if((readName) || (strcmp(mySamHeader.getSortOrder(), "queryname") == 0))
{
readName = true;
}
else
{
// defaulting to coordinate sorted.
samIn.setSortedValidation(SamFile::COORDINATE);
}
// Setup the '=' translation if the reference was specified.
if(!refFile.IsEmpty())
{
GenomeSequence* refPtr = new GenomeSequence(refFile);
samIn.SetReadSequenceTranslation(SamRecord::BASES);
samIn.SetReference(refPtr);
}
SamRecord* recordPtr;
int16_t samFlag;
SamStatus::Status returnStatus = SamStatus::SUCCESS;
while(returnStatus == SamStatus::SUCCESS)
{
recordPtr = myPool.getRecord();
if(recordPtr == NULL)
{
// Failed to allocate a new record.
throw(std::runtime_error("Failed to allocate a new SAM/BAM record"));
}
if(!samIn.ReadRecord(mySamHeader, *recordPtr))
{
// Failed to read a record.
returnStatus = samIn.GetStatus();
continue;
}
// Have a record. Check to see if it is a pair or unpaired read.
samFlag = recordPtr->getFlag();
if(SamFlag::isPaired(samFlag))
{
if(readName)
{
handlePairedRN(*recordPtr);
}
else
{
handlePairedCoord(*recordPtr);
}
}
else
{
++myNumUnpaired;
writeFastQ(*recordPtr, myUnpairedFile,
myUnpairedFileNameExt);
}
}
// Flush All
cleanUpMateMap(0, true);
if(returnStatus == SamStatus::NO_MORE_RECS)
{
returnStatus = SamStatus::SUCCESS;
}
samIn.Close();
closeFiles();
// Output the results
std::cerr << "\nFound " << myNumPairs << " read pairs.\n";
std::cerr << "Found " << myNumUnpaired << " unpaired reads.\n";
if(myNumMateFailures != 0)
{
std::cerr << "Failed to find mates for " << myNumMateFailures
<< " reads, so they were written as unpaired\n"
<< " (not included in either of the above counts).\n";
}
if(myNumQualTagErrors != 0)
{
std::cerr << myNumQualTagErrors << " records did not have tag "
<< myQField.c_str() << " or it was invalid, so the quality field was used for those records.\n";
}
return(returnStatus);
}
void Bam2FastQ::writeFastQ(SamRecord& samRec, IFILE filePtr,
const std::string& fileNameExt, const char* readNameExt)
{
static int16_t flag;
static std::string sequence;
static String quality;
static std::string rg;
static std::string rgFastqExt;
static std::string rgListStr;
static std::string fileName;
static std::string fq2;
if(mySplitRG)
{
rg = samRec.getString("RG").c_str();
rgFastqExt = rg + fileNameExt;
OutFastqMap::iterator it;
it = myOutFastqs.find(rgFastqExt);
if(it == myOutFastqs.end())
{
// New file.
fileName = myOutBase.c_str();
if(rg != "")
{
fileName += '.';
}
else
{
rg = ".";
}
fileName += rgFastqExt;
filePtr = ifopen(fileName.c_str(), "w", myCompression);
myOutFastqs[rgFastqExt] = filePtr;
if(fileNameExt != mySecondFileNameExt)
{
// first end.
const char* sm = mySamHeader.getRGTagValue("SM", rg.c_str());
if(strcmp(sm, "") == 0){sm = myOutBase.c_str();}
rgListStr.clear();
SamHeaderRG* rgPtr = mySamHeader.getRG(rg.c_str());
if((rgPtr == NULL) || (!rgPtr->appendString(rgListStr)))
{
// No RG info for this record.
rgListStr = ".\n";
}
fq2 = ".";
if(fileNameExt == myFirstFileNameExt)
{
fq2 = myOutBase.c_str();
if(rg != ".")
{
fq2 += '.';
fq2 += rg;
}
fq2 += mySecondFileNameExt;
}
ifprintf(myFqList, "%s\t%s\t%s\t%s",
sm, fileName.c_str(), fq2.c_str(),
rgListStr.c_str());
}
}
else
{
filePtr = it->second;
}
}
if(filePtr == NULL)
{
throw(std::runtime_error("Programming ERROR/EXITING: Bam2FastQ filePtr not set."));
return;
}
flag = samRec.getFlag();
const char* readName = samRec.getReadName();
sequence = samRec.getSequence();
if(myQField.IsEmpty())
{
// Read the quality from the quality field
quality = samRec.getQuality();
}
else
{
// Read Quality from the specified tag
const String* qTagPtr = samRec.getStringTag(myQField.c_str());
if((qTagPtr != NULL) && (qTagPtr->Length() == (int)sequence.length()))
{
// Use the tag value for quality
quality = qTagPtr->c_str();
}
else
{
// Tag was not found, so use the quality field.
++myNumQualTagErrors;
if(myNumQualTagErrors == 1)
{
std::cerr << "Bam2FastQ: " << myQField.c_str()
<< " tag was not found/invalid, so using the quality field in records without the tag\n";
}
quality = samRec.getQuality();
}
}
if(SamFlag::isReverse(flag) && myReverseComp)
{
// It is reverse, so reverse compliment the sequence
BaseUtilities::reverseComplement(sequence);
// Reverse the quality.
quality.Reverse();
}
else
{
// Ensure it is all capitalized.
int seqLen = sequence.size();
for (int i = 0; i < seqLen; i++)
{
sequence[i] = (char)toupper(sequence[i]);
}
}
if(myRNPlus)
{
ifprintf(filePtr, "@%s%s\n%s\n+%s%s\n%s\n", readName, readNameExt,
sequence.c_str(), readName, readNameExt, quality.c_str());
}
else
{
ifprintf(filePtr, "@%s%s\n%s\n+\n%s\n", readName, readNameExt,
sequence.c_str(), quality.c_str());
}
// Release the record.
myPool.releaseRecord(&samRec);
}
// 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;
}
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));
}
int GapInfo::processFile(const char* inputFileName, const char* outputFileName,
const char* refFile, bool detailed,
bool checkFirst, bool checkStrand)
{
// Open the file for reading.
SamFile samIn;
samIn.OpenForRead(inputFileName);
// Read the sam header.
SamFileHeader samHeader;
samIn.ReadHeader(samHeader);
SamRecord samRecord;
GenomeSequence* refPtr = NULL;
if(strcmp(refFile, "") != 0)
{
refPtr = new GenomeSequence(refFile);
}
IFILE outFile = ifopen(outputFileName, "w");
// Map for summary.
std::map<int, int> gapInfoMap;
// Keep reading records until ReadRecord returns false.
while(samIn.ReadRecord(samHeader, samRecord))
{
uint16_t samFlags = samRecord.getFlag();
if((!SamFlag::isMapped(samFlags)) ||
(!SamFlag::isMateMapped(samFlags)) ||
(!SamFlag::isPaired(samFlags)) ||
(samFlags & SamFlag::SECONDARY_ALIGNMENT) ||
(SamFlag::isDuplicate(samFlags)) ||
(SamFlag::isQCFailure(samFlags)))
{
// unmapped, mate unmapped, not paired,
// not the primary alignment,
// duplicate, fails vendor quality check
continue;
}
// No gap info if the chromosome names are different or
// are unknown.
int32_t refID = samRecord.getReferenceID();
if((refID != samRecord.getMateReferenceID()) || (refID == -1))
{
continue;
}
int32_t readStart = samRecord.get0BasedPosition();
int32_t mateStart = samRecord.get0BasedMatePosition();
// If the mate starts first, then the pair was processed by
// the mate.
if(mateStart < readStart)
{
continue;
}
if((mateStart == readStart) && (SamFlag::isReverse(samFlags)))
{
// read and mate start at the same position, so
// only process the forward strand.
continue;
}
// Process this read pair.
int32_t readEnd = samRecord.get0BasedAlignmentEnd();
int32_t gapSize = mateStart - readEnd - 1;
if(detailed)
{
// Output the gap info.
ifprintf(outFile, "%s\t%d\t%d",
samRecord.getReferenceName(), readEnd+1, gapSize);
// Check if it is not the first or if it is not the forward strand.
if(checkFirst && !SamFlag::isFirstFragment(samFlags))
{
ifprintf(outFile, "\tNotFirst");
}
if(checkStrand && SamFlag::isReverse(samFlags))
{
ifprintf(outFile, "\tReverse");
}
ifprintf(outFile, "\n");
}
else
{
// Summary.
// Skip reads that are not the forward strand.
if(SamFlag::isReverse(samFlags))
{
// continue
continue;
}
// Forward.
// Check the reference for 'N's.
if(refPtr != NULL)
{
genomeIndex_t chromStartIndex =
refPtr->getGenomePosition(samRecord.getReferenceName());
if(chromStartIndex == INVALID_GENOME_INDEX)
{
// Invalid position, so continue to the next one.
continue;
}
bool skipRead = false;
for(int i = readEnd + 1; i < mateStart; i++)
{
if((*refPtr)[i] == 'N')
{
// 'N' in the reference, so continue to the next read.
skipRead = true;
break;
}
}
if(skipRead)
{
continue;
}
}
// Update the gapInfo.
gapInfoMap[gapSize]++;
}
}
if(!detailed)
{
// Output the summary.
ifprintf(outFile, "GapSize\tNumPairs\n");
for(std::map<int,int>::iterator iter = gapInfoMap.begin();
iter != gapInfoMap.end(); iter++)
{
ifprintf(outFile, "%d\t%d\n", (*iter).first, (*iter).second);
}
}
SamStatus::Status returnStatus = samIn.GetStatus();
if(returnStatus == SamStatus::NO_MORE_RECS)
{
return(SamStatus::SUCCESS);
}
return(returnStatus);
}
