/// <summary>
/// Perform de-novo CNV calling in two steps:
/// 1. Filter REF variants and common CNVs, this step relies only on total CN calls with associated shortcomings
/// 2. Assign de-novo quality based on joint likelihood across pedigree using marginalisation operations
/// </summary>
/// <param name="canvasSegments"></param>
/// <param name="samplesInfo"></param>
/// <param name="parentIDs"></param>
/// <param name="offspringIDs"></param>
/// <param name="copyNumbersLikelihoods"></param>
private void SetDenovoQualityScores(ISampleMap <CanvasSegment> canvasSegments, ISampleMap <SampleMetrics> samplesInfo, List <SampleId> parentIDs, List <SampleId> offspringIDs,
JointLikelihoods copyNumbersLikelihoods)
{
foreach (var probandId in offspringIDs)
{
// targeted proband is REF
if (IsReferenceVariant(canvasSegments, samplesInfo, probandId))
{
continue;
}
// common variant
if (CanvasPedigreeCaller.IsSharedCnv(canvasSegments, samplesInfo, parentIDs, probandId, _callerParameters.MaximumCopyNumber))
{
continue;
}
// other offsprings are ALT
if (!offspringIDs.Except(probandId.ToEnumerable()).All(id => IsReferenceVariant(canvasSegments, samplesInfo, id)))
{
continue;
}
// not all q-scores are above the threshold
if (parentIDs.Concat(probandId).Any(id => !IsPassVariant(canvasSegments, id)))
{
continue;
}
double deNovoQualityScore = CanvasPedigreeCaller.GetConditionalDeNovoQualityScore(canvasSegments, copyNumbersLikelihoods, samplesInfo, parentIDs, probandId);
// adjustment so that denovo quality score threshold is 20 (rather than 10) to match Manta
deNovoQualityScore *= 2;
if (Double.IsInfinity(deNovoQualityScore) | deNovoQualityScore > _callerParameters.MaxQscore)
{
deNovoQualityScore = _callerParameters.MaxQscore;
}
canvasSegments[probandId].DqScore = deNovoQualityScore;
}
}
/// <summary>
/// Identify variant with the highest likelihood at a given setPosition and assign relevant scores
/// </summary>
public void CallVariant(ISampleMap <CanvasSegment> canvasSegments, ISampleMap <SampleMetrics> samplesInfo,
ISampleMap <ICopyNumberModel> copyNumberModel, PedigreeInfo pedigreeInfo)
{
var singleSampleLikelihoods = _copyNumberLikelihoodCalculator.GetCopyNumbersLikelihoods(canvasSegments, samplesInfo, copyNumberModel, _callerParameters.NumberOfTrimmedBins);
(var pedigreeCopyNumbers, var pedigreeLikelihoods) = GetPedigreeCopyNumbers(pedigreeInfo, singleSampleLikelihoods);
var nonPedigreeCopyNumbers = CanvasPedigreeCaller.GetNonPedigreeCopyNumbers(canvasSegments, pedigreeInfo, singleSampleLikelihoods);
var mergedCopyNumbers = pedigreeCopyNumbers.Concat(nonPedigreeCopyNumbers).OrderBy(canvasSegments.SampleIds);
EstimateQScores(canvasSegments, samplesInfo, pedigreeInfo, singleSampleLikelihoods, pedigreeLikelihoods, mergedCopyNumbers);
// TODO: this will be integrated with GetCopyNumbers* on a model level as a part of https://jira.illumina.com/browse/CANV-404
if (CanvasPedigreeCaller.UseAlleleCountsInformation(canvasSegments, _callerParameters.MinAlleleCountsThreshold, _callerParameters.MinAlleleNumberInSegment) &&
pedigreeInfo.HasFullPedigree())
{
AssignMccWithPedigreeInfo(canvasSegments, copyNumberModel, pedigreeInfo);
}
if (CanvasPedigreeCaller.UseAlleleCountsInformation(canvasSegments, _callerParameters.MinAlleleCountsThreshold, _callerParameters.MinAlleleNumberInSegment) &&
pedigreeInfo.HasOther())
{
AssignMccNoPedigreeInfo(canvasSegments.Where(segment => pedigreeInfo.OtherIds.Contains(segment.SampleId)).ToSampleMap(), copyNumberModel, _genotypes);
}
}
public void CallVariant(ISampleMap <CanvasSegment> canvasSegments, ISampleMap <SampleMetrics> samplesInfo,
ISampleMap <ICopyNumberModel> copyNumberModel, PedigreeInfo pedigreeInfo)
{
var coverageLikelihoods = _copyNumberLikelihoodCalculator.GetCopyNumbersLikelihoods(canvasSegments, samplesInfo, copyNumberModel);
// if number and properties of SNPs in the segment are above threshold, calculate likelihood from SNPs and merge with
// coverage likelihood to form merged likelihoods
int nBalleles = canvasSegments.Values.First().Balleles.Size();
// If allele information is available (i.e. segment has enough SNPs) merge coverage and allele likelihood obtained by GetGenotypeLogLikelihoods
// into singleSampleLikelihoods using JoinLikelihoods function.
var singleSampleLikelihoods = CanvasPedigreeCaller.UseAlleleCountsInformation(canvasSegments,
_callerParameters.MinAlleleCountsThreshold, _callerParameters.MinAlleleNumberInSegment)
? JoinLikelihoods(GetGenotypeLogLikelihoods(canvasSegments, copyNumberModel, _PhasedGenotypes), coverageLikelihoods, nBalleles)
: ConvertToLogLikelihood(coverageLikelihoods);
// estimate joint likelihood across pedigree samples from singleSampleLikelihoods using either only coverage or coverage + allele counts
(var pedigreeCopyNumbers, var pedigreeLikelihoods) = GetPedigreeCopyNumbers(pedigreeInfo, singleSampleLikelihoods);
var nonPedigreeCopyNumbers = CanvasPedigreeCaller.GetNonPedigreeCopyNumbers(canvasSegments, pedigreeInfo, singleSampleLikelihoods);
var mergedCopyNumbers = nonPedigreeCopyNumbers.Concat(pedigreeCopyNumbers).OrderBy(canvasSegments.SampleIds);
AssignCNandScores(canvasSegments, samplesInfo, pedigreeInfo, singleSampleLikelihoods,
pedigreeLikelihoods, mergedCopyNumbers);
}
static int Main(string[] args)
{
CanvasCommon.Utilities.LogCommandLine(args);
string outDir = null;
var segmentFiles = new List <string>();
var variantFrequencyFiles = new List <string>();
string ploidyBedPath = null;
string pedigreeFile = null;
string referenceFolder = null;
var sampleNames = new List <string>();
bool needHelp = false;
int? qScoreThreshold = null;
int? dqScoreThreshold = null;
string parameterconfigPath = Path.Combine(Utilities.GetAssemblyFolder(typeof(Program)), "PedigreeCallerParameters.json");
var caller = new CanvasPedigreeCaller();
var p = new OptionSet()
{
{ "i|infile=", "file containing bins, their counts, and assigned segments (obtained from CanvasPartition.exe)", v => segmentFiles.Add(v) },
{ "v|varfile=", "file containing variant frequencies (obtained from CanvasSNV.exe)", v => variantFrequencyFiles.Add(v) },
{ "o|outdir=", "name of output directory", v => outDir = v },
{ "r|reference=", "reference genome folder that contains GenomeSize.xml", v => referenceFolder = v },
{ "n|sampleName=", "sample name for output VCF header (optional)", v => sampleNames.Add(v) },
{ "f|pedigree=", "relationship withoin pedigree (parents/proband)", v => pedigreeFile = v },
{ "p|ploidyBed=", "bed file specifying reference ploidy (e.g. for sex chromosomes) (optional)", v => ploidyBedPath = v },
{ "h|help", "show this message and exit", v => needHelp = v != null },
{ "q|qscore=", $"quality filter threshold (default {caller.QualityFilterThreshold})", v => qScoreThreshold = int.Parse(v) },
{ "d|dqscore=", $"de novo quality filter threshold (default {caller.DeNovoQualityFilterThreshold})", v => dqScoreThreshold = int.Parse(v) },
{ "c|config=", $"parameter configuration path (default {parameterconfigPath})", v => parameterconfigPath = v }
};
var extraArgs = p.Parse(args);
if (extraArgs.Count > 0)
{
Console.WriteLine("* Error: I don't understand the argument '{0}'", extraArgs[0]);
needHelp = true;
}
if (needHelp)
{
ShowHelp(p);
return(0);
}
if (!segmentFiles.Any() || !variantFrequencyFiles.Any() || string.IsNullOrEmpty(referenceFolder) || string.IsNullOrEmpty(outDir))
{
ShowHelp(p);
return(0);
}
foreach (string segmentFile in segmentFiles)
{
if (File.Exists(segmentFile))
{
continue;
}
Console.WriteLine($"CanvasPedigreeCaller.exe: File {segmentFile} does not exist! Exiting.");
return(1);
}
foreach (string variantFrequencyFile in variantFrequencyFiles)
{
if (File.Exists(variantFrequencyFile))
{
continue;
}
Console.WriteLine($"CanvasPedigreeCaller.exe: File {variantFrequencyFile} does not exist! Exiting.");
return(1);
}
if (!File.Exists(Path.Combine(referenceFolder, "GenomeSize.xml")))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {Path.Combine(referenceFolder, "GenomeSize.xml")} does not exist! Exiting.");
return(1);
}
if (!File.Exists(parameterconfigPath))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {parameterconfigPath} does not exist! Exiting.");
return(1);
}
var parameterconfigFile = new FileLocation(parameterconfigPath);
caller.CallerParameters = Deserialize <PedigreeCallerParameters>(parameterconfigFile);
if (pedigreeFile.IsNullOrEmpty())
{
Console.WriteLine($"CanvasPedigreeCaller.exe: pedigreeFile option is not used! Calling CNV variants without family information.");
return(caller.CallVariants(variantFrequencyFiles, segmentFiles, outDir, ploidyBedPath, referenceFolder, sampleNames));
}
if (qScoreThreshold.HasValue & qScoreThreshold > 0 & qScoreThreshold < 60)
{
caller.QualityFilterThreshold = qScoreThreshold.Value;
Console.WriteLine($"CanvasPedigreeCaller.exe: Using user-supplied quality score threshold {qScoreThreshold}.");
}
if (dqScoreThreshold.HasValue & dqScoreThreshold > 0 & dqScoreThreshold < 60)
{
caller.DeNovoQualityFilterThreshold = dqScoreThreshold.Value;
Console.WriteLine($"CanvasPedigreeCaller.exe: Using user-supplied de novo quality score threshold {qScoreThreshold}.");
}
if (!File.Exists(pedigreeFile))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {pedigreeFile} does not exist! Exiting.");
return(1);
}
return(caller.CallVariantsInPedigree(variantFrequencyFiles, segmentFiles, outDir, ploidyBedPath, referenceFolder, sampleNames, pedigreeFile));
}
private static int Run(string[] args)
{
Utilities.LogCommandLine(args);
string outDir = null;
var segmentFiles = new List <string>();
var variantFrequencyFiles = new List <string>();
var sampleTypesString = new List <string>();
string ploidyBedPath = null;
string referenceFolder = null;
var sampleNames = new List <string>();
bool needHelp = false;
int? qScoreThresholdOption = null;
int? dqScoreThresholdOption = null;
string commonCnvsBedPath = null;
string parameterconfigPath = Path.Combine(Isas.Framework.Utilities.Utilities.GetAssemblyFolder(typeof(Program)), "PedigreeCallerParameters.json");
var p = new OptionSet()
{
{ "i|infile=", "file containing bins, their counts, and assigned segments (obtained from CanvasPartition.exe)", v => segmentFiles.Add(v) },
{ "v|varfile=", "file containing variant frequencies (obtained from CanvasSNV.exe)", v => variantFrequencyFiles.Add(v) },
{ "t|sampleType=", "sample types", v => sampleTypesString.Add(v) },
{ "o|outdir=", "name of output directory", v => outDir = v },
{ "r|reference=", "reference genome folder that contains GenomeSize.xml", v => referenceFolder = v },
{ "n|sampleName=", "sample name for output VCF header (optional)", v => sampleNames.Add(v) },
{ "p|ploidyBed=", "bed file specifying reference ploidy (e.g. for sex chromosomes) (optional)", v => ploidyBedPath = v },
{ "h|help", "show this message and exit", v => needHelp = v != null },
{ "q|qscore=", $"quality filter threshold (default {CanvasPedigreeCaller.DefaultQualityFilterThreshold})", v => qScoreThresholdOption = int.Parse(v) },
{ "commoncnvs=", "bed file with common CNVs (always include these intervals into segmentation results)", v => commonCnvsBedPath = v },
{ "d|dqscore=", $"de novo quality filter threshold (default {CanvasPedigreeCaller.DefaultDeNovoQualityFilterThreshold})", v => dqScoreThresholdOption = int.Parse(v) },
{ "c|config=", $"parameter configuration path (default {parameterconfigPath})", v => parameterconfigPath = v }
};
var extraArgs = p.Parse(args);
if (extraArgs.Count > 0)
{
Console.WriteLine("* Error: I don't understand the argument '{0}'", extraArgs[0]);
needHelp = true;
}
if (needHelp)
{
ShowHelp(p);
return(0);
}
if (!segmentFiles.Any() || !variantFrequencyFiles.Any() || string.IsNullOrEmpty(referenceFolder) || string.IsNullOrEmpty(outDir))
{
ShowHelp(p);
return(0);
}
foreach (string segmentFile in segmentFiles)
{
if (File.Exists(segmentFile))
{
continue;
}
Console.WriteLine($"CanvasPedigreeCaller.exe: File {segmentFile} does not exist! Exiting.");
return(1);
}
foreach (string variantFrequencyFile in variantFrequencyFiles)
{
if (File.Exists(variantFrequencyFile))
{
continue;
}
Console.WriteLine($"CanvasPedigreeCaller.exe: File {variantFrequencyFile} does not exist! Exiting.");
return(1);
}
var sampleTypesEnum = sampleTypesString.Select(GetSampleType).ToList();
if (!File.Exists(Path.Combine(referenceFolder, "GenomeSize.xml")))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {Path.Combine(referenceFolder, "GenomeSize.xml")} does not exist! Exiting.");
return(1);
}
if (!File.Exists(parameterconfigPath))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {parameterconfigPath} does not exist! Exiting.");
return(1);
}
if (commonCnvsBedPath != null)
{
if (!File.Exists(commonCnvsBedPath))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {commonCnvsBedPath} does not exist! Exiting.");
return(1);
}
}
var parameterconfigFile = new FileLocation(parameterconfigPath);
var callerParameters = Deserialize <PedigreeCallerParameters>(parameterconfigFile);
int qScoreThreshold = CanvasPedigreeCaller.DefaultQualityFilterThreshold;
if (qScoreThresholdOption != null)
{
qScoreThreshold = qScoreThresholdOption.Value;
Console.WriteLine($"CanvasPedigreeCaller.exe: Using user-supplied quality score threshold {qScoreThresholdOption}.");
}
if (qScoreThreshold < 0 || qScoreThreshold >= callerParameters.MaxQscore)
{
throw new IlluminaException($"Quality score threshold must be >= 0 and < {callerParameters.MaxQscore}");
}
int dqScoreThreshold = CanvasPedigreeCaller.DefaultDeNovoQualityFilterThreshold;
if (dqScoreThresholdOption != null)
{
dqScoreThreshold = dqScoreThresholdOption.Value;
Console.WriteLine($"CanvasPedigreeCaller.exe: Using user-supplied de novo quality score threshold {qScoreThresholdOption}.");
}
if (dqScoreThreshold < 0 || dqScoreThreshold >= callerParameters.MaxQscore)
{
throw new IlluminaException($"De novo quality score threshold must be >= 0 and < {callerParameters.MaxQscore}");
}
var logger = new Logger(new[] { Console.Out }, new[] { Console.Error });
var settings = IsasConfigurationSettings.GetConfigSettings();
var outputDirectory = new DirectoryLocation(outDir);
var workerDirectory = new DirectoryLocation(Isas.Framework.Utilities.Utilities.GetAssemblyFolder(typeof(CanvasPedigreeCaller)));
var commandManager = new CommandManager(new ExecutableProcessor(settings, logger, workerDirectory));
var result = -1;
var pedigreeCallerWorkDirectory = outputDirectory.GetDirectoryLocation("CanvasPedigreeCaller");
WorkDoerFactory.RunWithWorkDoer(logger, settings, pedigreeCallerWorkDirectory, workDoer =>
{
var copyNumberLikelihoodCalculator = new CopyNumberLikelihoodCalculator(callerParameters.MaximumCopyNumber);
var variantCaller = callerParameters.DefaultCaller == CallerType.VariantCaller ?
(IVariantCaller) new VariantCaller(copyNumberLikelihoodCalculator, callerParameters, qScoreThreshold) :
new HaplotypeVariantCaller(copyNumberLikelihoodCalculator, callerParameters, qScoreThreshold);
var copyNumberModelFactory = new HaplotypeCopyNumberModelFactory();
var referenceGenome = new ReferenceGenomeFactory().GetReferenceGenome(new DirectoryLocation(referenceFolder));
var genomeMetadata = referenceGenome.GenomeMetadata;
var coverageBigWigWriterFactory =
new CoverageVisualizationWriterFactory(logger, workDoer, commandManager, genomeMetadata);
var roundingBedGraphWriter = new RoundingBedGraphWriter(new BedGraphWriterFacade(), 4);
var coverageBigWigWriter = coverageBigWigWriterFactory.CreateBinCoverageBigWigWriter(roundingBedGraphWriter);
var segmentCoverageBedGraphWriter = coverageBigWigWriterFactory.CreateSegmentBedGraphWriter(roundingBedGraphWriter);
var tabixWrapper = TabixWrapperFactory.GetTabixWrapper(logger, workDoer, commandManager);
var bgzfBedGraphWriter = new BgzfBedGraphWriter(new BedGraphWriterFacade(), tabixWrapper);
var copyNumberBedGraphWriter = new CopyNumberBedGraphWriter(bgzfBedGraphWriter, new CopyNumberBedGraphCalculator());
var caller = new CanvasPedigreeCaller(logger, qScoreThreshold, dqScoreThreshold, callerParameters, copyNumberLikelihoodCalculator, variantCaller, coverageBigWigWriter, copyNumberModelFactory, copyNumberBedGraphWriter, segmentCoverageBedGraphWriter);
var outVcf = outputDirectory.GetFileLocation("CNV.vcf.gz");
result = caller.CallVariants(variantFrequencyFiles, segmentFiles, outVcf, ploidyBedPath, referenceFolder, sampleNames, commonCnvsBedPath, sampleTypesEnum);
});
return(result);
}
