/// <summary>
/// Writes copy-number data (cnd) file.
/// </summary>
/// <param name="fragmentCountFile"></param>
/// <param name="referenceCountFile"></param>
/// <param name="ratios"></param>
/// <param name="outputFile"></param>
public static void WriteCndFile(IFileLocation fragmentCountFile, IFileLocation referenceCountFile,
IEnumerable <SampleGenomicBin> ratios, IFileLocation outputFile)
{
IEnumerable <SampleGenomicBin> fragmentCounts = CanvasIO.IterateThroughTextFile(fragmentCountFile.FullName);
IEnumerable <SampleGenomicBin> referenceCounts = CanvasIO.IterateThroughTextFile(referenceCountFile.FullName);
using (var eFragment = fragmentCounts.GetEnumerator())
using (var eReference = referenceCounts.GetEnumerator())
using (var eRatio = ratios.GetEnumerator())
using (FileStream stream = new FileStream(outputFile.FullName, FileMode.Create, FileAccess.Write))
using (StreamWriter writer = new StreamWriter(stream))
{
writer.WriteLine(CSVWriter.GetLine("Fragment Count", "Reference Count", "Chromosome",
"Start", "End", "Unsmoothed Log Ratio"));
while (eFragment.MoveNext() && eReference.MoveNext() && eRatio.MoveNext())
{
// Some bins could have been skipped when calculating the ratios
while (!eFragment.Current.IsSameBin(eRatio.Current))
{
if (!eFragment.MoveNext()) // Ran out of fragment bins
{
throw new Illumina.Common.IlluminaException("Fragment bins and ratio bins are not in the same order.");
}
}
while (!eReference.Current.IsSameBin(eRatio.Current))
{
if (!eReference.MoveNext()) // Ran out of reference bins
{
throw new Illumina.Common.IlluminaException("Reference bins and ratio bins are not in the same order.");
}
}
if (!eFragment.Current.IsSameBin(eReference.Current) ||
!eFragment.Current.IsSameBin(eRatio.Current))
{
throw new Illumina.Common.IlluminaException("Bins are not in the same order.");
}
writer.WriteLine(CSVWriter.GetLine(eFragment.Current.Count.ToString(),
eReference.Current.Count.ToString(), eFragment.Current.GenomicBin.Chromosome,
eFragment.Current.Start.ToString(), eFragment.Current.Stop.ToString(),
eRatio.Current.Count.ToString()));
}
}
}
public IEnumerable <SampleGenomicBin> Run(IFileLocation sampleBedFile, IFileLocation referenceBedFile)
{
if (!sampleBedFile.Exists)
{
throw new FileNotFoundException(sampleBedFile.FullName + " does not exist.");
}
if (!referenceBedFile.Exists)
{
throw new FileNotFoundException(referenceBedFile.FullName + " does not exist.");
}
var sampleBins = CanvasIO.IterateThroughTextFile(sampleBedFile.FullName);
var referenceBins = CanvasIO.IterateThroughTextFile(referenceBedFile.FullName);
using (var eSampleBins = sampleBins.GetEnumerator())
using (var eReferenceBins = referenceBins.GetEnumerator())
{
while (eSampleBins.MoveNext() && eReferenceBins.MoveNext())
{
var sampleBin = eSampleBins.Current;
var referenceBin = eReferenceBins.Current;
// Bins with extreme reference counts introduce large variance into the ratios.
// It would be better to just drop these bins so we don't introduce too much noise into segmentation and CNV calling.
if (referenceBin.Count < _minReferenceCount)
{
continue;
} // skip the bin
if (referenceBin.Count > _maxReferenceCount)
{
continue;
} // skip the bin
double sampleCount = eSampleBins.Current.Count;
double ratio = sampleBin.Count / referenceBin.Count;
yield return(new SampleGenomicBin(sampleBin.GenomicBin.Chromosome, sampleBin.Start, sampleBin.Stop, sampleBin.GenomicBin.GC, (float)ratio));
}
}
}
public IEnumerable <SampleGenomicBin> Run(IFileLocation sampleBedFile, IFileLocation referenceBedFile)
{
if (!sampleBedFile.Exists)
{
throw new FileNotFoundException(sampleBedFile.FullName + " does not exist.");
}
if (!referenceBedFile.Exists)
{
throw new FileNotFoundException(referenceBedFile.FullName + " does not exist.");
}
var sampleBins = CanvasIO.IterateThroughTextFile(sampleBedFile.FullName);
var referenceBins = CanvasIO.IterateThroughTextFile(referenceBedFile.FullName);
double sampleMedian = (new BinCounts(sampleBins, manifest: _manifest)).OnTargetMedianBinCount;
double referenceMedian = (new BinCounts(referenceBins, manifest: _manifest)).OnTargetMedianBinCount;
double librarySizeFactor = (sampleMedian > 0 && referenceMedian > 0) ? referenceMedian / sampleMedian : 1;
using (var eSampleBins = sampleBins.GetEnumerator())
using (var eReferenceBins = referenceBins.GetEnumerator())
{
while (eSampleBins.MoveNext() && eReferenceBins.MoveNext())
{
var sampleBin = eSampleBins.Current;
var referenceBin = eReferenceBins.Current;
// The weighted average count of a bin could be less than 1.
// Using these small counts for coverage normalization creates large ratios.
// It would be better to just drop these bins so we don't introduce too much noise into segmentation and CNV calling.
if (referenceBin.Count < 1)
{
continue;
} // skip the bin
double ratio = sampleBin.Count / referenceBin.Count * librarySizeFactor;
yield return(new SampleGenomicBin(sampleBin.GenomicBin.Chromosome, sampleBin.Start, sampleBin.Stop, sampleBin.GenomicBin.GC, (float)ratio));
}
}
}
private static void LoadBinCounts(string binnedPath, NexteraManifest manifest, out List <double> binCounts,
out List <int> onTargetIndices)
{
LoadBinCounts(CanvasIO.IterateThroughTextFile(binnedPath), manifest, out binCounts, out onTargetIndices);
}