// Dump the specified Bam Index file.
int DumpIndex::execute(int argc, char **argv)
{
// Extract command line arguments.
static const int UNSPECIFIED_INT = -1;
String indexFile = "";
int refID = UNSPECIFIED_INT;
bool summary = false;
bool params = false;
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_STRINGPARAMETER("bamIndex", &indexFile)
LONG_INTPARAMETER("refID", &refID)
LONG_PARAMETER("summary", &summary)
LONG_PARAMETER("params", ¶ms)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters",
longParameterList));
inputParameters.Read(argc-1, &(argv[1]));
// Check to see if the index file was specified, if not, report an error.
if(indexFile == "")
{
usage();
inputParameters.Status();
// mandatory argument was not specified.
std::cerr << "Missing mandatory argument: --bamIndex" << std::endl;
return(-1);
}
if(params)
{
inputParameters.Status();
}
// Read the index.
BamIndex bamIndex;
SamStatus status;
status = bamIndex.readIndex(indexFile);
if(status != SamStatus::SUCCESS)
{
// Failed to read the index, return.
fprintf(stderr, "%s\n", status.getStatusMessage());
return(status.getStatus());
}
// Print the index file.
bamIndex.printIndex(refID, summary);
return(status.getStatus());
}
int main(int argc, char ** argv)
{
printf("glfMultiples -- SNP calls based on .glf or .glz files\n");
printf("(c) 2008-2011 Goncalo Abecasis, Sebastian Zoellner, Yun Li\n\n");
String pedfile;
String positionfile;
String callfile;
String glfAliases;
String glfPrefix;
String glfSuffix;
ParameterList pl;
double posterior = 0.50;
int mapQuality = 0;
int minTotalDepth = 1;
int maxTotalDepth = INT_MAX;
bool verbose = false;
bool mapQualityStrict = false;
bool hardFilter = false;
bool smartFilter = false;
bool softFilter = true;
bool robustPrior = true;
bool uniformPrior = false;
String xLabel("X"), yLabel("Y"), mitoLabel("MT");
int xStart = 2699520, xStop = 154931044;
BEGIN_LONG_PARAMETERS(longParameters)
LONG_PARAMETER_GROUP("Pedigree File")
LONG_STRINGPARAMETER("ped", &pedfile)
LONG_PARAMETER_GROUP("Map Quality Filter")
LONG_INTPARAMETER("minMapQuality", &mapQuality)
LONG_PARAMETER("strict", &mapQualityStrict)
LONG_PARAMETER_GROUP("Total Depth Filter")
LONG_INTPARAMETER("minDepth", &minTotalDepth)
LONG_INTPARAMETER("maxDepth", &maxTotalDepth)
LONG_PARAMETER_GROUP("Position Filter")
LONG_STRINGPARAMETER("positionFile", &positionfile)
LONG_PARAMETER_GROUP("Chromosome Labels")
LONG_STRINGPARAMETER("xChr", &xLabel)
LONG_STRINGPARAMETER("yChr", &yLabel)
LONG_STRINGPARAMETER("mito", &mitoLabel)
LONG_INTPARAMETER("xStart", &xStart)
LONG_INTPARAMETER("xStop", &xStop)
LONG_PARAMETER_GROUP("Filtering Options")
EXCLUSIVE_PARAMETER("hardFilter", &hardFilter)
EXCLUSIVE_PARAMETER("smartFilter", &smartFilter)
EXCLUSIVE_PARAMETER("softFilter", &softFilter)
LONG_PARAMETER_GROUP("Prior Options")
EXCLUSIVE_PARAMETER("uniformPrior", &uniformPrior)
EXCLUSIVE_PARAMETER("robustPrior", &robustPrior)
LONG_PARAMETER_GROUP("Output")
LONG_PARAMETER("verbose", &verbose)
LONG_PARAMETER_GROUP("Sample Names")
LONG_STRINGPARAMETER("glfAliases", &glfAliases)
LONG_PARAMETER_GROUP("Prefixes and Suffixes")
LONG_STRINGPARAMETER("glfPrefix",&glfPrefix)
LONG_STRINGPARAMETER("glfSuffix",&glfSuffix)
END_LONG_PARAMETERS();
pl.Add(new StringParameter('b', "Base Call File", callfile));
pl.Add(new DoubleParameter('p', "Posterior Threshold", posterior));
pl.Add(new LongParameters("Additional Options", longParameters));
int argstart = pl.ReadWithTrailer(argc, argv) + 1;
pl.Status();
if (posterior < 0.50)
error("Posterior threshold for genotype calls (-p option) must be > 0.50.");
time_t t;
time(&t);
printf("Analysis started on %s\n", ctime(&t));
fflush(stdout);
int n = argc - argstart;
argv += argstart;
Pedigree ped;
if (!pedfile.IsEmpty())
{
ped.pd.AddStringColumn("glfFile");
ped.Load(pedfile);
n = ped.count;
}
else
if (n == 0)
error("No pedigree file present and no glf files listed at the end of command line\n");
// Prior for finding difference from the reference at a particular site
//BgzfFileType::setRequireEofBlock(false);
double prior = 0.0;
for (int i = 1; i <= 2 * n; i++)
prior += 1.0 / i;
prior *= 0.001;
glfHandler * glf = new glfHandler[n];
bool firstGlf = n;
if (ped.count)
{
bool warn = false;
for (int i = n - 1; i > 0; i++)
{
if (!glf[i].Open(ped[i].strings[0]))
{
printf("Failed to open genotype likelihood file [%s] for individual %s:%s\n",
(const char *) ped[i].strings[0],
(const char *) ped[i].famid,
(const char *) ped[i].pid);
glf[i].OpenStub();
firstGlf = i;
}
if (warn)
printf("\n");
if (firstGlf == n)
error("No genotype likelihood files could be opened");
}
}
else
{
for (int i = firstGlf = 0; i < n; i++)
{
String glfName = glfPrefix + String(argv[i]) + glfSuffix;
if (!glf[i].Open(glfName))
error("Failed to open genotype likelihood file [%s]\n", glfName.c_str());
}
}
StringAlias aliases;
aliases.ReadFromFile(glfAliases);
printf("Calling genotypes for files ...\n");
for (int i = 0; i < n; i++)
printf("%s\n", ped.count ? (const char *) ped[i].strings[0] : argv[i]);
printf("\n");
baseCalls = fopen(callfile, "wt");
if (baseCalls != NULL)
{
fprintf(baseCalls, "##fileformat=VCFv4.0\n");
ReportDate(baseCalls);
fprintf(baseCalls, "##source=glfMultiples\n");
fprintf(baseCalls, "##minDepth=%d\n", minTotalDepth);
fprintf(baseCalls, "##maxDepth=%d\n", maxTotalDepth);
fprintf(baseCalls, "##minMapQuality=%d\n", mapQuality);
fprintf(baseCalls, "##minPosterior=%.4f\n", posterior);
fprintf(baseCalls, "##INFO=<ID=DP,Number=1,Type=Integer,Description=\"Total Depth\">\n");
fprintf(baseCalls, "##INFO=<ID=MQ,Number=1,Type=Integer,Description=\"Root Mean Squared Mapping Quality\">\n");
fprintf(baseCalls, "##INFO=<ID=NS,Number=1,Type=Integer,Description=\"Number of samples with coverage\">\n");
fprintf(baseCalls, "##INFO=<ID=AN,Number=1,Type=Integer,Description=\"Total number of alleles (with coverage)\">\n");
fprintf(baseCalls, "##INFO=<ID=AC,Number=.,Type=Integer,Description=\"Alternative allele count (with coverage)\">\n");
fprintf(baseCalls, "##INFO=<ID=AF,Number=.,Type=Float,Description=\"Alternate allele frequency\">\n");
fprintf(baseCalls, "##INFO=<ID=AB,Number=1,Type=Float,Description=\"Estimated allele balance between the alleles\">\n");
if ( mapQuality > 0 ) {
fprintf(baseCalls, "##FILTER=<ID=mq%d,Description=\"Mapping Quality less than %d\">\n",mapQuality,mapQuality);
}
if ( minTotalDepth > 1 ) {
fprintf(baseCalls, "##FILTER=<ID=dp%d,Description=\"Total Read Depth less than %d\">\n",minTotalDepth,minTotalDepth);
}
if ( minTotalDepth < INT_MAX ) {
fprintf(baseCalls, "##FILTER=<ID=DP%d,Description=\"Total Read Depth greater than %d\">\n",maxTotalDepth,maxTotalDepth);
}
fprintf(baseCalls, "##FORMAT=<ID=GT,Number=1,Type=String,Description=\"Most Likely Genotype\">\n");
fprintf(baseCalls, "##FORMAT=<ID=GQ,Number=1,Type=Integer,Description=\"Genotype Call Quality\">\n");
fprintf(baseCalls, "##FORMAT=<ID=DP,Number=1,Type=Integer,Description=\"Read Depth\">\n");
fprintf(baseCalls, "##FORMAT=<ID=PL,Number=3,Type=Integer,Description=\"Genotype Likelihoods for Genotypes 0/0,0/1,1/1\">\n");
fprintf(baseCalls, "##FORMAT=<ID=PL3,Number=6,Type=Integer,Description=\"Genotype Likelihoods for Genotypes 0/0,0/1,1/1,0/2,1/2,2/2\">\n");
fprintf(baseCalls, "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT");
for (int i = 0; i < n; i++)
fprintf(baseCalls, "\t%s", ped.count ?
(const char *) (ped[i].famid + ":" + ped[i].pid) :
(const char *) aliases.GetAlias(argv[i]));
fprintf(baseCalls, "\n");
}
StringArray buffer, tokens;
StringHash positions;
buffer.Read(positionfile);
for (int i = 0; i < buffer.Length(); i++)
{
tokens.ReplaceTokens(buffer[i], " \t:");
if (tokens.Length() != 2) continue;
positions.Add(tokens[0] + ":" + (int(tokens[1].AsInteger() - 1)));
}
int chromosomeType = 0;
while (glf[firstGlf].NextSection())
{
for (int i = firstGlf + 1; i < n; i++)
{
if (glf[i].isStub) continue;
glf[i].NextSection();
if (glf[firstGlf].maxPosition != glf[i].maxPosition || glf[firstGlf].label != glf[i].label)
{
error("Genotype files '%s' and '%s' are not compatible ...\n"
" File '%s' has section %s with %d entries ...\n"
" File '%s' section %s with %d entries ...\n",
ped.count ? (const char *) ped[firstGlf].strings[0] : argv[firstGlf],
ped.count ? (const char *) ped[i].strings[0] : argv[i],
ped.count ? (const char *) ped[firstGlf].strings[0] : argv[firstGlf],
(const char *) glf[firstGlf].label, glf[firstGlf].maxPosition,
ped.count ? (const char *) ped[i].strings[0] : argv[i],
(const char *) glf[i].label, glf[i].maxPosition);
}
}
chromosomeType = CT_AUTOSOME;
if (ped.count)
{
if (glf[firstGlf].label == xLabel) chromosomeType = CT_CHRX;
if (glf[firstGlf].label == yLabel) chromosomeType = CT_CHRY;
if (glf[firstGlf].label == mitoLabel) chromosomeType = CT_MITO;
}
printf("Processing section %s with %d entries\n",
(const char *) glf[firstGlf].label, glf[firstGlf].maxPosition);
int refBase = 0;
int position = 0;
int mapQualityFilter = 0;
int depthFilter = 0;
int homozygousReference = 0;
int transitions = 0;
int transversions = 0;
int otherPolymorphisms = 0;
int sinkFilter = 0;
int smartFilterHits = 0;
int baseCounts[5] = {0, 0, 0, 0, 0};
String filter;
while (true)
{
if (position > 0)
{
// Check whether we have reached the end of the current chromosome
bool done = true;
for (int i = 0; i < n; i++)
if (glf[i].data.recordType != 0)
done = false;
if (done) break;
}
// Advance to the next position where needed
for (int i = 0; i < n; i++)
if (glf[i].position == position)
glf[i].NextBaseEntry();
// Figure out the current analysis position
refBase = glf[0].data.refBase;
position = glf[0].position;
for (int i = 1; i < n; i++)
if (position > glf[i].position)
{
position = glf[i].position;
refBase = glf[i].data.refBase;
}
// Avoid alignments that extend past the end of the chromosome
if (position >= glf[firstGlf].maxPosition)
break;
baseCounts[(int)refBase]++;
// These lines can be uncommented for debugging purposes
// for (int i = 0; i < n; i++)
// printf("GLF %d : position %d, refBase %d\n", i, position, refBase);
// printf("Position: %d, refBase: %d\n", position, refBase);
if (positions.Entries())
{
filter = glf[firstGlf].label + ":" + position;
if (positions.Find(filter) < 0) continue;
}
if (refBase == 0) continue;
// Corrected calculation of root-mean-square Map Quality score
// and check if we have at least one sample with good quality data
int currentDepth = 0, totalDepth = 0, numCovered = 0;
double currentQuality = 0.0, averageMapQuality = 0.0;
bool passMapQualityFilter = false;
for (int i = 0; i < n; i++)
{
currentDepth = glf[i].GetDepth(position);
if (currentDepth != 0)
{
totalDepth += currentDepth;
numCovered++; // not currently used -- will be "NS"
currentQuality = glf[i].GetMapQuality(position);
averageMapQuality +=
currentDepth * currentQuality * currentQuality;
if (currentQuality >= mapQuality)
passMapQualityFilter = true;
}
}
averageMapQuality = sqrt(averageMapQuality / totalDepth);
filter.Clear();
if (!passMapQualityFilter)
{
if (filter.Length() == 0) mapQualityFilter++;
if (hardFilter) continue;
filter.catprintf("%smq%d", filter.Length() ? ";" : "", mapQuality);
}
if (totalDepth < minTotalDepth)
{
if (filter.Length() == 0) depthFilter++;
if (hardFilter) continue;
filter.catprintf("%sdp%d", filter.Length() ? ";" : "", minTotalDepth);
}
if (totalDepth > maxTotalDepth)
{
if (filter.Length() == 0) depthFilter++;
if (hardFilter) continue;
filter.catprintf("%sDP%d", filter.Length() ? ";" : "", maxTotalDepth);
}
// Create convenient aliases for each base
unsigned char transition = (((refBase - 1) ^ 2) + 1);
unsigned char transvers1 = (((refBase - 1) ^ 3) + 1);
unsigned char transvers2 = (((refBase - 1) ^ 1) + 1);
int homRef = glf[0].GenotypeIndex(refBase, refBase);
// Calculate likelihood assuming every is homozygous for the reference
double lRef = log(1.0 - prior);
for (int i = 0; i < n; i++)
lRef += log(glf[i].GetLikelihoods(position)[homRef]);
// Calculate maximum likelihood for a variant
if (smartFilter)
{
double anyVariant = log(prior) + FilteringLikelihood(glf, n, position, refBase);
if (exp(lRef - anyVariant) > (1.0 - posterior)/posterior)
{
smartFilterHits++;
continue;
}
}
//fprintf(stderr,"position = %d\n",position);
double pTs = uniformPrior ? 1./3. : 2./3.;
double pTv = uniformPrior ? 1./3. : 1./6.;
// Calculate likelihoods for the most likelily SNP configurations
double refTransition = log(prior * pTs) + PolymorphismLikelihood(glf, n, position, refBase, transition);
double refTransvers1 = log(prior * pTv) + PolymorphismLikelihood(glf, n, position, refBase, transvers1);
double refTransvers2 = log(prior * pTv) + PolymorphismLikelihood(glf, n, position, refBase, transvers2);
// Calculate likelihoods for less likely SNP configurations
double transitiontv1 = log(prior * 0.001) + PolymorphismLikelihood(glf, n, position, transition, transvers1);
double transitiontv2 = log(prior * 0.001) + PolymorphismLikelihood(glf, n, position, transition, transvers2);
double transvers1tv2 = log(prior * 0.001) + PolymorphismLikelihood(glf, n, position, transvers1, transvers2);
// Calculate the likelihood for unusual configurations where everyone is heterozygous ...
double sink = n > 10 ? log(prior * 1e-8) + SinkLikelihood(glf, n, position) : -1e100;
double lmax = max(
max(max(lRef, refTransition),max(refTransvers1, refTransvers2)),
max(max(transitiontv1, transitiontv2), max(transvers1tv2, sink)));
double sum = exp(lRef - lmax) + exp(refTransition -lmax) +
exp(refTransvers1 - lmax) + exp(refTransvers2 - lmax) +
exp(transitiontv1 - lmax) + exp(transitiontv2 - lmax) +
exp(transvers1tv2 - lmax) + exp(sink - lmax);
if (sum == 0.0) continue;
if (exp(lRef - lmax)/sum > 1.0 - prior)
{
if (filter.Length() == 0) homozygousReference++;
if (positions.Entries())
ReportSNP(glf, n, position, refBase, refBase, refBase, filter, totalDepth, averageMapQuality, lRef / sum);
continue;
}
double quality = 1.0 - exp(lRef - lmax) / sum;
if (verbose)
{
DumpDetails(glf, n, position, refBase);
printf("%.3f %.3f %.3f %.3f %.3f %.3f %.3f\n",
lRef, refTransition, refTransvers1, refTransvers2,
transitiontv1, transitiontv2, transvers1tv2);
}
if (exp(refTransition - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, refBase, transition,
filter, totalDepth, averageMapQuality, quality /* refTransition/sum */);
if (filter.Length() == 0) transitions++;
}
else if (exp(refTransvers1 - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, refBase, transvers1,
filter, totalDepth, averageMapQuality, quality /* refTransvers1/sum */);
if (filter.Length() == 0) transversions++;
}
else if (exp(refTransvers2 - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, refBase, transvers2,
filter, totalDepth, averageMapQuality, quality /* refTransvers2/sum */);
if (filter.Length() == 0) transversions++;
}
else if (exp(transitiontv1 - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, transition, transvers1,
filter, totalDepth, averageMapQuality, quality /* transitiontv1/sum */);
if (filter.Length() == 0) otherPolymorphisms++;
}
else if (exp(transitiontv2 - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, transition, transvers2,
filter, totalDepth, averageMapQuality, quality /* transitiontv2/sum */);
if (filter.Length() == 0) otherPolymorphisms++;
}
else if (exp(transvers1tv2 - lmax)/sum > posterior)
{
ReportSNP(glf, n, position, refBase, transvers1, transvers2,
filter, totalDepth, averageMapQuality, quality /* transvers1tv2/sum */);
if (filter.Length() == 0) otherPolymorphisms++;
}
else if (exp(sink - lmax)/sum > posterior)
sinkFilter++;
}
int actualBases = glf[firstGlf].maxPosition - baseCounts[0];
printf(" Missing bases = %9d (%.3f%%)\n",
baseCounts[0], baseCounts[0] * 100. / glf[firstGlf].maxPosition);
printf(" Reference bases = %9d (%.3f%%)\n",
glf[firstGlf].maxPosition - baseCounts[0], (glf[firstGlf].maxPosition - baseCounts[0]) * 100. / glf[firstGlf].maxPosition);
printf(" A/T bases = %9d (%.3f%%, %d A, %d T)\n",
baseCounts[1] + baseCounts[4],
(baseCounts[1] + baseCounts[4]) * 100. / actualBases,
baseCounts[1], baseCounts[4]);
printf(" G/C bases = %9d (%.3f%%, %d G, %d C)\n",
baseCounts[3] + baseCounts[2],
(baseCounts[3] + baseCounts[2]) * 100. / actualBases,
baseCounts[3], baseCounts[2]);
printf(" Depth Filter = %9d bases (%.3f%%)\n",
depthFilter, depthFilter * 100. / actualBases);
printf(" Map Quality Filter = %9d bases (%.3f%%)\n",
mapQualityFilter, mapQualityFilter * 100. / actualBases);
printf(" Non-Polymorphic = %9d bases (%.3f%%)\n",
homozygousReference, homozygousReference * 100. / actualBases);
printf(" Transitions = %9d bases (%.3f%%)\n",
transitions, transitions * 100. / actualBases);
printf(" Transversions = %9d bases (%.3f%%)\n",
transversions, transversions * 100. / actualBases);
printf(" Other Polymorphisms = %9d bases (%.3f%%)\n",
otherPolymorphisms, otherPolymorphisms * 100. / actualBases);
if (n > 10)
printf(" Homology Sink = %9d bases (%.3f%%)\n",
sinkFilter, sinkFilter * 100. / actualBases);
if (smartFilter)
printf(" Smart Filter = %9d bases (%.3f%%)\n",
smartFilterHits, smartFilterHits * 100. / actualBases);
int noCalls = actualBases - homozygousReference - transitions - transversions - otherPolymorphisms - sinkFilter;
printf(" No call = %9d bases (%.3f%%)\n",
noCalls, noCalls * 100. / actualBases);
fflush(stdout);
}
if (baseCalls != NULL)
fclose(baseCalls);
time(&t);
printf("\nAnalysis completed on %s\n", ctime(&t));
fflush(stdout);
}
int ReadReference::execute(int argc, char **argv)
{
static const int UNSPECIFIED_INT = -1;
String refFile = "";
String refName = "";
int start = UNSPECIFIED_INT;
int numBases = UNSPECIFIED_INT;
int end = UNSPECIFIED_INT;
bool params = false;
// Read in the parameters.
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_STRINGPARAMETER("refFile", &refFile)
LONG_STRINGPARAMETER("refName", &refName)
LONG_INTPARAMETER("start", &start)
LONG_INTPARAMETER("end", &end)
LONG_INTPARAMETER("numBases", &numBases)
LONG_PARAMETER("params", ¶ms)
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((refName == "") || (start == UNSPECIFIED_INT) ||
((end == UNSPECIFIED_INT) && (numBases == UNSPECIFIED_INT)))
{
usage();
inputParameters.Status();
std::cerr << "Missing Required Parameter\n\n";
return(-1);
}
if((end != UNSPECIFIED_INT) && (numBases != UNSPECIFIED_INT))
{
usage();
inputParameters.Status();
std::cerr << "Only --end or --numBases can be specified\n\n";
return(-1);
}
else if(numBases != UNSPECIFIED_INT)
{
end = start + numBases;
}
if(params)
{
inputParameters.Status();
}
// Open the reference.
GenomeSequence reference(refFile);
uint32_t refStart =
reference.getGenomePosition(refName.c_str());
if(refStart == INVALID_GENOME_INDEX)
{
std::cerr << "Reference Name: " << refName.c_str()
<< " not found in the reference file\n";
return(-1);
}
std::string refString;
reference.getString(refString, refStart + start, end - start);
std::cout << refString << std::endl;
return(0);
}
int ClipOverlap::execute(int argc, char **argv)
{
// Extract command line arguments.
String inFile = "";
String outFile = "";
String storeOrig = "";
bool readName = false;
bool noRNValidate = false;
bool stats = false;
int poolSize = DEFAULT_POOL_SIZE;
bool unmapped = false;
bool noeof = false;
bool params = false;
String excludeFlags = "0xF0C";
// TODO, cleanup legacy parameters
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_PARAMETER_GROUP("Required Parameters")
LONG_STRINGPARAMETER("in", &inFile)
LONG_STRINGPARAMETER("out", &outFile)
LONG_PARAMETER_GROUP("Optional Parameters")
LONG_STRINGPARAMETER("storeOrig", &storeOrig)
LONG_PARAMETER("readName", &readName)
LONG_PARAMETER ("noRNValidate", &noRNValidate)
LONG_PARAMETER ("stats", &stats)
LONG_PARAMETER ("overlapsOnly", &myOverlapsOnly)
LONG_STRINGPARAMETER ("excludeFlags", &excludeFlags)
LONG_PARAMETER("unmapped", &unmapped)
LONG_PARAMETER("noeof", &noeof)
LONG_PARAMETER("params", ¶ms)
LONG_PARAMETER_GROUP("Coordinate Processing Optional Parameters")
LONG_INTPARAMETER("poolSize", &poolSize)
LONG_PARAMETER("poolSkipOverlap", &myPoolSkipOverlap)
LONG_PHONEHOME(VERSION)
BEGIN_LEGACY_PARAMETERS()
LONG_PARAMETER ("clipsOnly", &myOverlapsOnly)
LONG_PARAMETER("poolSkipClip", &myPoolSkipOverlap)
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);
}
// Check to see if the in file was specified, if not, report an error.
if(inFile == "")
{
printUsage(std::cerr);
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);
}
// Check to see if the out file was specified, if not, report an error.
if(outFile == "")
{
printUsage(std::cerr);
inputParameters.Status();
// Out file was not specified but it is mandatory.
std::cerr << "--out is a mandatory argument, "
<< "but was not specified" << std::endl;
return(-1);
}
if((storeOrig.Length() != 0) && (storeOrig.Length() != 2))
{
printUsage(std::cerr);
inputParameters.Status();
std::cerr << "--storeOrig tag name must be 2 characters.\n";
return(-1);
}
myOverlapHandler = new OverlapClipLowerBaseQual();
if(myOverlapHandler == NULL)
{
printUsage(std::cerr);
inputParameters.Status();
std::cerr << "Failed to allocate the overlap handler\n";
return(-1);
}
if(unmapped)
{
myOverlapHandler->markAsUnmapped();
}
// Setup the overlap handler.
myOverlapHandler->keepStats(stats);
if(storeOrig.Length() != 0)
{
myOverlapHandler->storeOrigCigar(storeOrig);
}
myIntExcludeFlags = excludeFlags.AsInteger();
if(params)
{
inputParameters.Status();
}
// For each step process the file.
// Open the files & read/write the sam header.
SamStatus::Status runStatus = SamStatus::SUCCESS;
for(int i = 1; i <= myOverlapHandler->numSteps(); i++)
{
// Open the file for reading.
mySamHeader.resetHeader();
SamFile samIn(inFile, SamFile::READ, &mySamHeader);
SamFile* samOutPtr = NULL;
// Check if writing, if so, open the output file.
if(i == myOverlapHandler->numSteps())
{
samOutPtr = new SamFile(outFile, SamFile::WRITE, &mySamHeader);
}
if(readName)
{
if(!noRNValidate)
{
samIn.setSortedValidation(SamFile::QUERY_NAME);
}
runStatus = handleSortedByReadName(samIn, samOutPtr);
}
else
{
// Coordinate sorted, so work with the pools.
samIn.setSortedValidation(SamFile::COORDINATE);
myPool.setMaxAllocatedRecs(poolSize);
// Reset the number of failures
myNumMateFailures = 0;
myNumPoolFail = 0;
myNumPoolFailNoHandle = 0;
myNumPoolFailHandled = 0;
myNumOutOfOrder = 0;
// Run by coordinate
if(samOutPtr != NULL)
{
// Setup the output buffer for writing.
SamCoordOutput outputBuffer(myPool);
outputBuffer.setOutputFile(samOutPtr, &mySamHeader);
runStatus = handleSortedByCoord(samIn, &outputBuffer);
// Cleanup the output buffer.
if(!outputBuffer.flushAll())
{
std::cerr << "ERROR: Failed to flush the output buffer\n";
runStatus = SamStatus::FAIL_IO;
}
}
else
{
runStatus = handleSortedByCoord(samIn, NULL);
}
}
if(runStatus != SamStatus::SUCCESS)
{
break;
}
// Close the input file, it will be reopened if there are
// multiple steps.
samIn.Close();
if(samOutPtr != NULL)
{
samOutPtr->Close();
delete samOutPtr;
samOutPtr = NULL;
}
}
// Done processing.
// Print Stats
myOverlapHandler->printStats();
if(myNumMateFailures != 0)
{
std::cerr << "WARNING: did not find expected overlapping mates for "
<< myNumMateFailures << " records." << std::endl;
}
if(myNumPoolFail != 0)
{
// Had to skip clipping some records due to running out of
// memory and not being able to wait for the mate.
std::cerr << "WARNING: " << myNumPoolFail
<< " record pool failures\n";
if(myNumPoolFailNoHandle != 0)
{
std::cerr << "Due to hitting the max record poolSize, skipped handling "
<< myNumPoolFailNoHandle << " records." << std::endl;
}
if(myNumPoolFailHandled != 0)
{
std::cerr << "Due to hitting the max record poolSize, default handled "
<< myNumPoolFailHandled << " records." << std::endl;
}
if(myNumOutOfOrder != 0)
{
std::cerr << "WARNING: Resulting File out of Order by "
<< myNumOutOfOrder << " records.\n";
}
}
if(runStatus == SamStatus::SUCCESS)
{
if(myNumPoolFail == 0)
{
std::cerr << "Completed ClipOverlap Successfully.\n";
}
else
{
runStatus = SamStatus::NO_MORE_RECS;
std::cerr << "Completed ClipOverlap with WARNINGS.\n";
}
}
else
{
std::cerr << "Failed to complete ClipOverlap.\n";
}
return(runStatus);
}
int Validate::execute(int argc, char **argv)
{
// Extract command line arguments.
String inFile = "";
int maxErrors = -1;
int printableErrors = 100;
bool so_flag = false;
bool so_coord = false;
bool so_query = false;
bool noeof = false;
bool disableStatistics = false;
bool verbose = false;
bool params = false;
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_STRINGPARAMETER("in", &inFile)
LONG_PARAMETER("noeof", &noeof)
LONG_INTPARAMETER("maxErrors", &maxErrors)
LONG_PARAMETER("verbose", &verbose)
LONG_INTPARAMETER("printableErrors", &printableErrors)
LONG_PARAMETER("disableStatistics", &disableStatistics)
LONG_PARAMETER("params", ¶ms)
LONG_PARAMETER_GROUP("SortOrder")
EXCLUSIVE_PARAMETER("so_flag", &so_flag)
EXCLUSIVE_PARAMETER("so_coord", &so_coord)
EXCLUSIVE_PARAMETER("so_query", &so_query)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters",
longParameterList));
inputParameters.Read(argc-1, &(argv[1]));
// Determine the sort type for validation based on the parameters.
SamFile::SortedType sortType = SamFile::UNSORTED;
if(so_flag)
{
sortType = SamFile::FLAG;
}
else if(so_coord)
{
sortType = SamFile::COORDINATE;
}
else if(so_query)
{
sortType = SamFile::QUERY_NAME;
}
// 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 validate, "
<< "but was not specified" << std::endl;
return(-1);
}
if(params)
{
inputParameters.Status();
}
// Since we want to accumulate multiple errors, use RETURN rather
// than throwing exceptions.
SamFile samIn(ErrorHandler::RETURN);
// Open the file for reading.
if(!samIn.OpenForRead(inFile))
{
fprintf(stderr, "%s\n", samIn.GetStatusMessage());
return(samIn.GetStatus());
}
// Set the sorting validation type.
samIn.setSortedValidation(sortType);
// Set that statistics should be generated.
samIn.GenerateStatistics(!disableStatistics);
// Read the sam header.
SamFileHeader samHeader;
if(!samIn.ReadHeader(samHeader))
{
fprintf(stderr, "%s\n", samIn.GetStatusMessage());
return(samIn.GetStatus());
}
// Read the sam records.
SamRecord samRecord(ErrorHandler::RETURN);
// Track the status.
SamStatus::Status status = SamStatus::SUCCESS;
// Keep reading records until the end of the file is reached.
int numValidRecords = 0;
int numInvalidRecords = 0;
int numErrorRecords = 0;
int numRecords = 0;
int numReportedErrors = 0;
int totalErrorRecords = 0;
std::map<SamStatus::Status, uint64_t> errorStats;
std::map<SamValidationError::Type, uint64_t> invalidStats;
SamValidationErrors invalidSamErrors;
// Keep reading records from the file until SamFile::ReadRecord
// indicates to stop (returns false).
while( ( (maxErrors < 0) || (totalErrorRecords < maxErrors) ) &&
( (samIn.ReadRecord(samHeader, samRecord)) || (SamStatus::isContinuableStatus(samIn.GetStatus())) ) )
{
++numRecords;
if(samIn.GetStatus() == SamStatus::SUCCESS)
{
// Successfully set the record, so check to see if it is valid.
// Clear any errors in the list.
invalidSamErrors.clear();
if(!SamValidator::isValid(samHeader, samRecord, invalidSamErrors))
{
// The record is not valid.
++numInvalidRecords;
++totalErrorRecords;
if(verbose && (numReportedErrors < printableErrors))
{
std::cerr << "Record " << numRecords << std::endl
<< invalidSamErrors << std::endl;
++numReportedErrors;
}
// Update the statistics for all validation errors found in this record.
invalidSamErrors.resetErrorIter();
const SamValidationError* errorPtr = invalidSamErrors.getNextError();
while(errorPtr != NULL)
{
++invalidStats[errorPtr->getType()];
errorPtr = invalidSamErrors.getNextError();
}
// If the status is not yet set, set it.
if(status == SamStatus::SUCCESS)
{
status = SamStatus::INVALID;
}
}
else
{
// Valid record, so increment the counter.
++numValidRecords;
}
}
else
{
// Error reading the record.
++numErrorRecords;
++totalErrorRecords;
if(verbose && (numReportedErrors < printableErrors))
{
// report error.
std::cerr << "Record " << numRecords << std::endl
<< samIn.GetStatusMessage() << std::endl
<< std::endl;
++numReportedErrors;
}
// Increment the statistics
++errorStats[samIn.GetStatus()];
// If the status is not yet set, set it.
if(status == SamStatus::SUCCESS)
{
status = samIn.GetStatus();
}
}
}
if( (samIn.GetStatus() != SamStatus::NO_MORE_RECS) &&
(totalErrorRecords < maxErrors) )
{
// The last read call had a failure, so report it.
// If the number of errors is >= ,maxErrors we don't
// want to print any more failures.
++numErrorRecords;
++totalErrorRecords;
if(numReportedErrors < printableErrors)
{
std::cerr << "Record " << numRecords << ": ";
std::cerr << std::endl << samIn.GetStatusMessage() << std::endl;
}
// Increment the statistics
++errorStats[samIn.GetStatus()];
if(status == SamStatus::SUCCESS)
{
status = samIn.GetStatus();
}
}
if(totalErrorRecords == maxErrors)
{
if(maxErrors == 0)
{
std::cerr << "WARNING file was not read at all due to maxErrors setting, but returning Success.\n";
}
else
{
// Print a note that the entire file was not read.
std::cerr << "File was not completely read due to the number of errors.\n";
std::cerr << "Statistics only reflect the part of the file that was read.\n";
}
}
fprintf(stderr, "\nNumber of records read = %d\n", numRecords);
fprintf(stderr, "Number of valid records = %d\n", numValidRecords);
std::cerr << std::endl;
if(numRecords != numValidRecords)
{
std::cerr << "Error Counts:\n";
// Loop through the non-validation errors.
std::map<SamStatus::Status, uint64_t>::iterator statusIter;
for(statusIter = errorStats.begin(); statusIter != errorStats.end(); statusIter++)
{
std::cerr << "\t" << SamStatus::getStatusString(statusIter->first) << ": "
<< statusIter->second << std::endl;
}
std::map<SamValidationError::Type, uint64_t>::iterator invalidIter;
for(invalidIter = invalidStats.begin(); invalidIter != invalidStats.end(); invalidIter++)
{
std::cerr << "\t" << SamValidationError::getTypeString(invalidIter->first) << ": "
<< invalidIter->second << std::endl;
}
std::cerr << std::endl;
}
samIn.PrintStatistics();
fprintf(stderr, "Returning: %d (%s)\n", status, SamStatus::getStatusString(status));
return(status);
}
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));
}
// 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 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);
}
int main(int argc, char ** argv)
{
ParameterList inputParameters;
String filename;
int minReadLength = 10;
int printableErrors = 20;
int maxErrors = -1;
String testParam;
BaseAsciiMap::SPACE_TYPE myBaseType = BaseAsciiMap::UNKNOWN;
// Read the parameters from the command line.
bool baseSpace = false;
bool colorSpace = false;
bool autoDetect = false;
bool ignoreErrors = false;
bool baseComposition = false;
bool avgQual = false;
bool quiet = false;
bool noeof = false;
bool params = false;
bool disableSeqIDCheck = false;
bool interleaved = false;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_STRINGPARAMETER("file", &filename)
LONG_PARAMETER("baseComposition", &baseComposition)
LONG_PARAMETER("avgQual", &avgQual)
LONG_PARAMETER("disableSeqIDCheck", &disableSeqIDCheck)
LONG_PARAMETER("interleaved", &interleaved)
LONG_PARAMETER("noeof", &noeof)
LONG_PARAMETER("quiet", &quiet)
LONG_PARAMETER("params", ¶ms)
LONG_INTPARAMETER("minReadLen", &minReadLength)
LONG_INTPARAMETER("maxErrors", &maxErrors)
LONG_PARAMETER_GROUP("Space Type")
EXCLUSIVE_PARAMETER("baseSpace", &baseSpace)
EXCLUSIVE_PARAMETER("colorSpace", &colorSpace)
EXCLUSIVE_PARAMETER("auto", &autoDetect)
LONG_PARAMETER_GROUP("Errors")
EXCLUSIVE_PARAMETER("ignoreErrors", &ignoreErrors)
LONG_SMARTINTPARAMETER("printableErrors", &printableErrors)
BEGIN_LEGACY_PARAMETERS()
LONG_PARAMETER("printBaseComp", &baseComposition)
LONG_PARAMETER("disableAllMessages", &quiet)
LONG_INTPARAMETER("quitAfterErrorNum", &maxErrors)
LONG_PARAMETER_GROUP("Space Type")
EXCLUSIVE_PARAMETER("baseSpace", &baseSpace)
EXCLUSIVE_PARAMETER("colorSpace", &colorSpace)
EXCLUSIVE_PARAMETER("autoDetect", &autoDetect)
LONG_PARAMETER_GROUP("Errors")
EXCLUSIVE_PARAMETER("ignoreAllErrors", &ignoreErrors)
LONG_SMARTINTPARAMETER("maxReportedErrors", &printableErrors)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters", longParameterList));
inputParameters.Read(argc, argv);
if(ignoreErrors)
{
// Ignore all errors, so set printableErrors to 0.
printableErrors = 0;
}
// Set the base type based on the passed in parameters.
if(baseSpace)
{
// Base Space
myBaseType = BaseAsciiMap::BASE_SPACE;
}
else if(colorSpace)
{
myBaseType = BaseAsciiMap::COLOR_SPACE;
}
else
{
myBaseType = BaseAsciiMap::UNKNOWN;
// Set autoDetect
autoDetect = true;
}
// 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);
}
// DO not print status if set to quiet.
if((!quiet) && params)
{
inputParameters.Status();
}
if(filename == "")
{
if(quiet)
{
return(-1);
}
// No filename was specified so print a usage description.
std::cout << "ERROR: No filename specified. See below for usage help.";
std::cout << std::endl << std::endl;
std::cout << " Required Parameters:" << std::endl;
std::cout << "\t--file : FastQ filename with path to be prorcessed.\n";
std::cout << std::endl;
std::cout << " Optional Parameters:" << std::endl;
std::cout << "\t--minReadLen : Minimum allowed read length (Defaults to 10).\n";
std::cout << "\t--maxErrors : Number of errors to allow before quitting\n";
std::cout << "\t reading/validating the file.\n";
std::cout << "\t -1 (default) indicates to not quit until\n";
std::cout << "\t the entire file is read.\n";
std::cout << "\t 0 indicates not to read/validate anything\n";
std::cout << "\t--printableErrors : Maximum number of errors to print before\n";
std::cout << "\t suppressing them (Defaults to 20).\n";
std::cout << "\t Different than maxErrors since \n";
std::cout << "\t printableErrors will continue reading and\n";
std::cout << "\t validating the file until the end, but\n";
std::cout << "\t just doesn't print the errors.\n";
std::cout << "\t--ignoreErrors : Ignore all errors (same as printableErrors = 0)\n";
std::cout << "\t overwrites the printableErrors option.\n";
std::cout << "\t--baseComposition : Print the Base Composition Statistics.\n";
std::cout << "\t--avgQual : Print the average phred quality per cycle & overall average quality.\n";
std::cout << "\t--disableSeqIDCheck : Disable the unique sequence identifier check.\n";
std::cout << "\t Use this option to save memory since the sequence id\n";
std::cout << "\t check uses a lot of memory.\n";
std::cout << "\t--noeof : Disable checking that the eof block is present in gzipped files\n.";
std::cout << "\t--interleaved : Validate consequtive reads have the same sequence identifier\n";
std::cout << "\t (only allowed difference is 1/2, but not required) and validate\n";
std::cout << "\t that otherwise reads have unique sequence identifiers.\n";
std::cout << "\t Cannot be used if '--disableSeqIDCheck' is specified.\n";
std::cout << "\t--params : Print the parameter settings.\n";
std::cout << "\t--quiet : Suppresses the display of errors and summary statistics.\n";
std::cout << "\t Does not affect the printing of Base Composition Statistics.\n";
std::cout << "\n Optional Space Options for Raw Sequence (Last one specified is used):\n";
std::cout << "\t--auto : Determine baseSpace/colorSpace from the Raw Sequence in the file (Default).\n";
std::cout << "\t--baseSpace : ACTGN only\n";
std::cout << "\t--colorSpace : 0123. only\n";
std::cout << std::endl;
std::cout << " Usage:" << std::endl;
std::cout << "\t./fastQValidator --file <fileName> [--minReadLen <minReadLen>] [--maxErrors <numErrors>] [--printableErrors <printableErrors>|--ignoreErrors] [--baseComposition] [--disableSeqIDCheck] [--interleaved] [--quiet] [--baseSpace|--colorSpace|--auto] [--params]\n\n";
std::cout << " Examples:" << std::endl;
std::cout << "\t../fastQValidator --file testFile.txt\n";
std::cout << "\t../fastQValidator --file testFile.txt --minReadLen 10 --baseSpace --printableErrors 100\n";
std::cout << "\t./fastQValidator --file test/testFile.txt --minReadLen 10 --colorSpace --ignoreErrors\n";
std::cout << std::endl;
return (-1);
}
FastQFile validator(minReadLength, printableErrors);
if(quiet)
{
validator.disableMessages();
}
if(disableSeqIDCheck)
{
validator.disableSeqIDCheck();
}
if(interleaved)
{
validator.interleaved();
}
if(interleaved && disableSeqIDCheck)
{
if(!quiet)
{
std::cout << "ERROR: --interleaved and --disableSeqIDCheck cannot both be specified.\n";
}
return(-1);
}
validator.setMaxErrors(maxErrors);
FastQStatus::Status status = validator.validateFastQFile(filename, baseComposition, myBaseType, avgQual);
if(!quiet)
{
std::cout << "Returning: " << status << " : " << FastQStatus::getStatusString(status)
<< std::endl;
}
return(status);
}
int WriteRegion::execute(int argc, char **argv)
{
// Extract command line arguments.
String inFile = "";
String outFile = "";
String indexFile = "";
String readName = "";
String bed = "";
myStart = UNSPECIFIED_INT;
myEnd = UNSPECIFIED_INT;
myPrevStart = UNSPECIFIED_INT;
myPrevEnd = UNSPECIFIED_INT;
myRefID = UNSET_REF;
myRefName.Clear();
myPrevRefName.Clear();
myBedRefID = SamReferenceInfo::NO_REF_ID;
bool lshift = false;
bool noeof = false;
bool params = false;
myWithinReg = false;
myWroteReg = false;
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_PARAMETER_GROUP("Required Parameters")
LONG_STRINGPARAMETER("in", &inFile)
LONG_STRINGPARAMETER("out", &outFile)
LONG_PARAMETER_GROUP("Optional Region Parameters")
LONG_STRINGPARAMETER("bamIndex", &indexFile)
LONG_STRINGPARAMETER("refName", &myRefName)
LONG_INTPARAMETER("refID", &myRefID)
LONG_INTPARAMETER("start", &myStart)
LONG_INTPARAMETER("end", &myEnd)
LONG_STRINGPARAMETER("bed", &bed)
LONG_PARAMETER("withinReg", &myWithinReg)
LONG_STRINGPARAMETER("readName", &readName)
LONG_PARAMETER_GROUP("Optional Other Parameters")
LONG_PARAMETER("lshift", &lshift)
LONG_PARAMETER("noeof", &noeof)
LONG_PARAMETER("params", ¶ms)
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();
// mandatory argument was not specified.
inputParameters.Status();
std::cerr << "Missing mandatory argument: --in" << std::endl;
return(-1);
}
if(outFile == "")
{
usage();
// mandatory argument was not specified.
inputParameters.Status();
std::cerr << "Missing mandatory argument: --out" << std::endl;
return(-1);
}
if(indexFile == "")
{
// In file was not specified, so set it to the in file
// + ".bai"
indexFile = inFile + ".bai";
}
if(myRefID != UNSET_REF && myRefName.Length() != 0)
{
std::cerr << "Can't specify both refID and refName" << std::endl;
inputParameters.Status();
return(-1);
}
if(myRefID != UNSET_REF && bed.Length() != 0)
{
std::cerr << "Can't specify both refID and bed" << std::endl;
inputParameters.Status();
return(-1);
}
if(myRefName.Length() != 0 && bed.Length() != 0)
{
std::cerr << "Can't specify both refName and bed" << std::endl;
inputParameters.Status();
return(-1);
}
if(!bed.IsEmpty())
{
myBedFile = ifopen(bed, "r");
}
if(params)
{
inputParameters.Status();
}
// Open the file for reading.
mySamIn.OpenForRead(inFile);
// Open the output file for writing.
SamFile samOut;
samOut.OpenForWrite(outFile);
// Open the bam index file for reading if a region was specified.
if((myRefName.Length() != 0) || (myRefID != UNSET_REF) || (myBedFile != NULL))
{
mySamIn.ReadBamIndex(indexFile);
}
// Read & write the sam header.
mySamIn.ReadHeader(mySamHeader);
samOut.WriteHeader(mySamHeader);
// Read the sam records.
SamRecord samRecord;
// Track the status.
int numSectionRecords = 0;
// Set returnStatus to success. It will be changed
// to the failure reason if any of the writes fail.
SamStatus::Status returnStatus = SamStatus::SUCCESS;
while(getNextSection())
{
// Keep reading records until they aren't anymore.
while(mySamIn.ReadRecord(mySamHeader, samRecord))
{
if(!readName.IsEmpty())
{
// Check for readname.
if(strcmp(samRecord.getReadName(), readName.c_str()) != 0)
{
// not a matching read name, so continue to the next record.
continue;
}
}
// Check to see if the read has already been processed.
if(myPrevEnd != UNSPECIFIED_INT)
{
// Because we already know that the bed was sorted,
// we know that the previous section started before
// this one, so if the previous end is greater than
// this record's end position we know that it
// was already written in the previous section.
// Note: can't be equal to the previous end since
// the end range was exclusive, while
// get0BasedAlignmentEnd is inclusive.
// myPrevEnd is reset by getNextSection when a new
// chromosome is hit.
if(samRecord.get0BasedAlignmentEnd() < myPrevEnd)
{
// This record was already written.
continue;
}
}
// Shift left if applicable.
if(lshift)
{
samRecord.shiftIndelsLeft();
}
// Successfully read a record from the file, so write it.
samOut.WriteRecord(mySamHeader, samRecord);
++numSectionRecords;
}
myWroteReg = true;
}
if(myBedFile != NULL)
{
ifclose(myBedFile);
}
std::cerr << "Wrote " << outFile << " with " << numSectionRecords
<< " records.\n";
return(returnStatus);
}
int VcfMac::execute(int argc, char **argv)
{
String inputVcf = "";
int minAC = -1;
String sampleSubset = "";
String filterList = "";
bool params = false;
IntervalTree<int> regions;
std::vector<int> intersection;
// Read in the parameters.
ParameterList inputParameters;
BEGIN_LONG_PARAMETERS(longParameterList)
LONG_PARAMETER_GROUP("Required Parameters")
LONG_STRINGPARAMETER("in", &inputVcf)
LONG_PARAMETER_GROUP("Optional Parameters")
LONG_STRINGPARAMETER("sampleSubset", &sampleSubset)
LONG_INTPARAMETER("minAC", &minAC)
LONG_STRINGPARAMETER("filterList", &filterList)
LONG_PARAMETER("params", ¶ms)
LONG_PHONEHOME(VERSION)
END_LONG_PARAMETERS();
inputParameters.Add(new LongParameters ("Input Parameters",
longParameterList));
inputParameters.Read(argc-1, &(argv[1]));
// Check that all files were specified.
if(inputVcf == "")
{
usage();
inputParameters.Status();
std::cerr << "Missing \"--in\", a required parameter.\n\n";
return(-1);
}
if(params)
{
inputParameters.Status();
}
// Open the two input files.
VcfFileReader inFile;
VcfHeader header;
VcfRecord record;
// Open the file
if(sampleSubset.IsEmpty())
{
inFile.open(inputVcf, header);
}
else
{
inFile.open(inputVcf, header, sampleSubset, NULL, NULL);
}
// Add the discard rule for minor allele count.
if(minAC >= 0)
{
inFile.addDiscardMinMinorAlleleCount(minAC, NULL);
}
if(!filterList.IsEmpty())
{
// Open the filter list.
IFILE regionFile = ifopen(filterList, "r");
String regionLine;
StringArray regionColumn;
int start;
int end;
int intervalVal = 1;
if(regionFile == NULL)
{
std::cerr << "Failed to open " << filterList
<< ", so keeping all positions\n";
filterList.Clear();
}
else
{
while( regionFile->isOpen() && !regionFile->ifeof())
{
// Read the next interval
regionLine.Clear();
regionLine.ReadLine(regionFile);
if(regionLine.IsEmpty())
{
// Nothing on this line, continue to the next.
continue;
}
regionColumn.ReplaceColumns(regionLine, ' ');
if(regionColumn.Length() != 2)
{
std::cerr << "Improperly formatted region line: "
<< regionLine << "; skipping to the next line.\n";
continue;
}
// Convert the columns to integers.
if(!regionColumn[0].AsInteger(start))
{
// The start position (1st column) is not an integer.
std::cerr << "Improperly formatted region line, start position "
<< "(1st column) is not an integer: "
<< regionColumn[0]
<< "; Skipping to the next line.\n";
continue;
}
if(!regionColumn[1].AsInteger(end))
{
// The start position (1st column) is not an integer.
std::cerr << "Improperly formatted region line, end position "
<< "(2nd column) is not an integer: "
<< regionColumn[1]
<< "; Skipping to the next line.\n";
continue;
}
// Add 1-based inclusive intervals.
regions.add(start,end, intervalVal);
}
}
}
int numReadRecords = 0;
while( inFile.readRecord(record))
{
if(!filterList.IsEmpty())
{
// Check if the region should be kept.
intersection.clear();
regions.get_intersecting_intervals(record.get1BasedPosition(), intersection);
if(intersection.empty())
{
// not in the interval, so continue to the next record.
continue;
}
}
++numReadRecords;
// Loop through the number of possible alternates.
unsigned int numAlts = record.getNumAlts();
int minAlleleCount = -1;
int curAlleleCount = 0;
int totalAlleleCount = 0;
for(unsigned int i = 0; i <= numAlts; i++)
{
curAlleleCount = record.getAlleleCount(i);
if((minAlleleCount == -1) ||
(curAlleleCount < minAlleleCount))
{
minAlleleCount = curAlleleCount;
}
totalAlleleCount += curAlleleCount;
}
if(totalAlleleCount != 0)
{
double maf = (double)minAlleleCount/totalAlleleCount;
std::cout << record.getIDStr()
<< "\t" << minAlleleCount
<< "\t" << maf << "\n";
}
}
inFile.close();
// std::cerr << "\n\t# Records: " << numReadRecords << "\n";
// return success.
return(0);
}
