private static void PostProcessAndWriteResults(SegmentationInput segmentationInput, string outPartitionedFile,
PloidyInfo referencePloidy, GenomeSegmentationResults segmentationResults)
{
var segments = segmentationInput.PostProcessSegments(segmentationResults, referencePloidy);
segmentationInput.WriteCanvasPartitionResults(outPartitionedFile, segments);
}
// dataType: "logratio" (aCGH, ROMA, etc.) or "binary" (LOH)
public Segmentation(string inputBinPath, string forbiddenBedPath, string dataType = "logratio",
int maxInterBinDistInSegment = 1000000)
{
this.InputBinPath = inputBinPath;
this.DataType = dataType;
this.SegmentationResults = null;
this.ForbiddenIntervalBedPath = forbiddenBedPath;
this.MaxInterBinDistInSegment = maxInterBinDistInSegment;
// Read the input file:
this.ReadBEDInput();
}
public void WriteCanvasPartitionResults(string outPath, GenomeSegmentationResults segmentationResults)
{
Dictionary <string, bool> starts = new Dictionary <string, bool>();
Dictionary <string, bool> stops = new Dictionary <string, bool>();
foreach (string chr in segmentationResults.SegmentByChr.Keys)
{
for (int segmentIndex = 0; segmentIndex < segmentationResults.SegmentByChr[chr].Length; segmentIndex++)
{
Segmentation.Segment segment = segmentationResults.SegmentByChr[chr][segmentIndex];
starts[chr + ":" + segment.start] = true;
stops[chr + ":" + segment.end] = true;
}
}
Dictionary <string, List <SampleGenomicBin> > excludedIntervals = new Dictionary <string, List <SampleGenomicBin> >();
if (!string.IsNullOrEmpty(ForbiddenIntervalBedPath))
{
excludedIntervals = CanvasCommon.Utilities.LoadBedFile(ForbiddenIntervalBedPath);
}
using (GzipWriter writer = new GzipWriter(outPath))
{
int segmentNum = -1;
foreach (string chr in StartByChr.Keys)
{
List <SampleGenomicBin> excludeIntervals = null;
if (excludedIntervals.ContainsKey(chr))
{
excludeIntervals = excludedIntervals[chr];
}
int excludeIndex = 0; // Points to the first interval which *doesn't* end before our current position
uint previousBinEnd = 0;
for (int pos = 0; pos < StartByChr[chr].Length; pos++)
{
uint start = StartByChr[chr][pos];
uint end = EndByChr[chr][pos];
string key = chr + ":" + start;
bool newSegment = IsNewSegment(starts, key, excludeIntervals, previousBinEnd, end, start, ref excludeIndex);
if (newSegment)
{
segmentNum++;
}
writer.WriteLine(string.Format($"{chr}\t{start}\t{end}\t{ScoreByChr[chr][pos]}\t{segmentNum}"));
previousBinEnd = end;
}
}
}
}
public Dictionary <string, List <SegmentWithBins> > PostProcessSegments(
GenomeSegmentationResults segmentationResults,
PloidyInfo referencePloidy)
{
var excludedIntervals = new Dictionary <string, List <SampleGenomicBin> >();
if (!string.IsNullOrEmpty(ForbiddenIntervalBedPath))
{
excludedIntervals = CanvasCommon.Utilities.LoadBedFile(ForbiddenIntervalBedPath);
}
return(_processor.PostProcessSegments(segmentationResults, referencePloidy, excludedIntervals,
CoverageInfo));
}
/// <summary>
/// Wavelets: unbalanced HAAR wavelets segmentation
/// </summary>
/// <param name="threshold">wavelets coefficient threshold</param>
private void Wavelets(bool isGermline, double thresholdLower = 5, double thresholdUpper = 80, int minSize = 10, int verbose = 1)
{
Dictionary <string, int[]> inaByChr = new Dictionary <string, int[]>();
Dictionary <string, double[]> finiteScoresByChr = new Dictionary <string, double[]>();
List <ThreadStart> tasks = new List <ThreadStart>();
foreach (KeyValuePair <string, double[]> scoreByChrKVP in ScoreByChr)
{
tasks.Add(new ThreadStart(() =>
{
string chr = scoreByChrKVP.Key;
int[] ina;
Helper.GetFiniteIndices(scoreByChrKVP.Value, out ina); // not NaN, -Inf, Inf
double[] scores;
if (ina.Length == scoreByChrKVP.Value.Length)
{
scores = scoreByChrKVP.Value;
}
else
{
Helper.ExtractValues <double>(scoreByChrKVP.Value, ina, out scores);
}
lock (finiteScoresByChr)
{
finiteScoresByChr[chr] = scores;
inaByChr[chr] = ina;
}
}));
}
Isas.Shared.Utilities.DoWorkParallelThreads(tasks);
// Quick sanity-check: If we don't have any segments, then return a dummy result.
int n = 0;
foreach (var list in finiteScoresByChr.Values)
{
n += list.Length;
}
if (n == 0)
{
this.SegmentationResults = this.GetDummySegmentationResults();
return;
}
Dictionary <string, Segment[]> segmentByChr = new Dictionary <string, Segment[]>();
// when parallelizing we need an RNG for each chromosome to get deterministic results
Random seedGenerator = new MersenneTwister(0);
Dictionary <string, Random> perChromosomeRandom = new Dictionary <string, Random>();
foreach (string chr in this.ScoreByChr.Keys)
{
perChromosomeRandom[chr] = new MersenneTwister(seedGenerator.NextFullRangeInt32(), true);
}
tasks = new List <ThreadStart>();
foreach (string chr in ScoreByChr.Keys)
{
tasks.Add(new ThreadStart(() =>
{
int[] ina = inaByChr[chr];
List <int> breakpoints = new List <int>();
int sizeScoreByChr = this.ScoreByChr[chr].Length;
if (sizeScoreByChr > minSize)
{
WaveletSegmentation.HaarWavelets(this.ScoreByChr[chr].ToArray(), thresholdLower, thresholdUpper, breakpoints, isGermline);
}
List <int> startBreakpointsPos = new List <int>();
List <int> endBreakpointPos = new List <int>();
List <int> lengthSeg = new List <int>();
if (breakpoints.Count() >= 2 && sizeScoreByChr > 10)
{
startBreakpointsPos.Add(breakpoints[0]);
endBreakpointPos.Add(breakpoints[1] - 1);
lengthSeg.Add(breakpoints[1] - 1);
for (int i = 1; i < breakpoints.Count - 1; i++)
{
startBreakpointsPos.Add(breakpoints[i]);
endBreakpointPos.Add(breakpoints[i + 1] - 1);
lengthSeg.Add(breakpoints[i + 1] - 1 - breakpoints[i]);
}
startBreakpointsPos.Add(breakpoints[breakpoints.Count - 1]);
endBreakpointPos.Add(sizeScoreByChr - 1);
lengthSeg.Add(sizeScoreByChr - breakpoints[breakpoints.Count - 1] - 1);
}
else
{
startBreakpointsPos.Add(0);
endBreakpointPos.Add(sizeScoreByChr - 1);
lengthSeg.Add(sizeScoreByChr - 1);
}
// estimate segment means
double[] segmentMeans = new double[lengthSeg.Count()];
int ss = 0, ee = 0;
for (int i = 0; i < lengthSeg.Count(); i++)
{
ee += lengthSeg[i];
// Works even if weights == null
segmentMeans[i] = Helper.WeightedAverage(this.ScoreByChr[chr], null, iStart: ss, iEnd: ee);
ss = ee;
}
Segment[] segments = new Segment[startBreakpointsPos.Count];
for (int i = 0; i < startBreakpointsPos.Count; i++)
{
int start = startBreakpointsPos[i];
int end = endBreakpointPos[i];
segments[i] = new Segment();
segments[i].start = this.StartByChr[chr][start]; // Genomic start
segments[i].end = this.EndByChr[chr][end]; // Genomic end
segments[i].nMarkers = lengthSeg[i];
segments[i].mean = segmentMeans[i];
}
lock (segmentByChr)
{
segmentByChr[chr] = segments;
}
}));
}
Console.WriteLine("{0} Launching wavelet tasks", DateTime.Now);
Isas.Shared.Utilities.DoWorkParallelThreads(tasks);
Console.WriteLine("{0} Completed wavelet tasks", DateTime.Now);
this.SegmentationResults = new GenomeSegmentationResults(segmentByChr);
Console.WriteLine("{0} Segmentation results complete", DateTime.Now);
}
private GenomeSegmentationResults GetDummySegmentationResults()
{
GenomeSegmentationResults results = new GenomeSegmentationResults(new Dictionary <string, Segment[]>());
return(results);
}
/// <summary>
/// CBS: circular binary segmentation porting the R function segment in DNAcopy
/// </summary>
/// <param name="alpha">Now in this.Alpha</param>
/// <param name="nPerm"></param>
/// <param name="pMethod">"hybrid" or "perm"</param>
/// <param name="minWidth"></param>
/// <param name="kMax"></param>
/// <param name="nMin"></param>
/// <param name="eta"></param>
/// <param name="sbdry"></param>
/// <param name="trim"></param>
/// <param name="undoSplit">"none" or "prune" or "sdundo"; now in this.UndoMethod</param>
/// <param name="undoPrune"></param>
/// <param name="undoSD"></param>
/// <param name="verbose"></param>
private void CBS(uint nPerm = 10000, string pMethod = "hybrid", int minWidth = 2, int kMax = 25,
uint nMin = 200, double eta = 0.05, uint[] sbdry = null, double trim = 0.025,
double undoPrune = 0.05, double undoSD = 3, int verbose = 1)
{
if (minWidth < 2 || minWidth > 5)
{
Console.Error.WriteLine("Minimum segment width should be between 2 and 5");
Environment.Exit(1);
}
if (nMin < 4 * kMax)
{
Console.Error.WriteLine("nMin should be >= 4 * kMax");
Environment.Exit(1);
}
if (sbdry == null)
{
GetBoundary.ComputeBoundary(nPerm, this.Alpha, eta, out sbdry);
}
Dictionary <string, int[]> inaByChr = new Dictionary <string, int[]>();
Dictionary <string, double[]> finiteScoresByChr = new Dictionary <string, double[]>();
List <ThreadStart> tasks = new List <ThreadStart>();
foreach (KeyValuePair <string, double[]> scoreByChrKVP in ScoreByChr)
{
tasks.Add(new ThreadStart(() =>
{
string chr = scoreByChrKVP.Key;
int[] ina;
Helper.GetFiniteIndices(scoreByChrKVP.Value, out ina); // not NaN, -Inf, Inf
double[] scores;
if (ina.Length == scoreByChrKVP.Value.Length)
{
scores = scoreByChrKVP.Value;
}
else
{
Helper.ExtractValues <double>(scoreByChrKVP.Value, ina, out scores);
}
lock (finiteScoresByChr)
{
finiteScoresByChr[chr] = scores;
inaByChr[chr] = ina;
}
}));
}
//Parallel.ForEach(tasks, t => { t.Invoke(); });
Isas.Shared.Utilities.DoWorkParallelThreads(tasks);
// Quick sanity-check: If we don't have any segments, then return a dummy result.
int n = 0;
foreach (var list in finiteScoresByChr.Values)
{
n += list.Length;
}
if (n == 0)
{
this.SegmentationResults = this.GetDummySegmentationResults();
return;
}
double trimmedSD = Math.Sqrt(ChangePoint.TrimmedVariance(finiteScoresByChr, trim: trim));
Dictionary <string, Segment[]> segmentByChr = new Dictionary <string, Segment[]>();
// when parallelizing we need an RNG for each chromosome to get deterministic results
Random seedGenerator = new MersenneTwister(0);
Dictionary <string, Random> perChromosomeRandom = new Dictionary <string, Random>();
foreach (string chr in this.ScoreByChr.Keys)
{
perChromosomeRandom[chr] = new MersenneTwister(seedGenerator.NextFullRangeInt32(), true);
}
tasks = new List <ThreadStart>();
foreach (string chr in ScoreByChr.Keys)
{
tasks.Add(new ThreadStart(() =>
{
int[] ina = inaByChr[chr];
int[] lengthSeg;
double[] segmentMeans;
ChangePoint.ChangePoints(this.ScoreByChr[chr], sbdry, out lengthSeg, out segmentMeans, perChromosomeRandom[chr],
dataType: this.DataType, alpha: this.Alpha, nPerm: nPerm,
pMethod: pMethod, minWidth: minWidth, kMax: kMax, nMin: nMin, trimmedSD: trimmedSD,
undoSplits: this.UndoMethod, undoPrune: undoPrune, undoSD: undoSD, verbose: verbose);
Segment[] segments = new Segment[lengthSeg.Length];
int cs1 = 0, cs2 = -1; // cumulative sum
for (int i = 0; i < lengthSeg.Length; i++)
{
cs2 += lengthSeg[i];
int start = ina[cs1];
int end = ina[cs2];
segments[i] = new Segment();
segments[i].start = this.StartByChr[chr][start]; // Genomic start
segments[i].end = this.EndByChr[chr][end]; // Genomic end
segments[i].nMarkers = lengthSeg[i];
segments[i].mean = segmentMeans[i];
cs1 += lengthSeg[i];
}
lock (segmentByChr)
{
segmentByChr[chr] = segments;
}
}));
}
//Parallel.ForEach(tasks, t => { t.Invoke(); });
Isas.Shared.Utilities.DoWorkParallelThreads(tasks);
this.SegmentationResults = new GenomeSegmentationResults(segmentByChr);
}
public Dictionary <string, List <SegmentWithBins> > PostProcessSegments(
GenomeSegmentationResults segmentationResults,
PloidyInfo referencePloidy, Dictionary <string, List <SampleGenomicBin> > excludedIntervals, CoverageInfo coverageInfo)
{
var starts = new Dictionary <string, bool>();
var stops = new Dictionary <string, bool>();
foreach (string chr in segmentationResults.SegmentByChr.Keys)
{
for (int segmentIndex = 0; segmentIndex < segmentationResults.SegmentByChr[chr].Length; segmentIndex++)
{
var segment = segmentationResults.SegmentByChr[chr][segmentIndex];
starts[chr + ":" + segment.start] = true;
stops[chr + ":" + segment.end] = true;
}
}
int segmentNum = -1;
var segmentsByChromosome = new Dictionary <string, List <SegmentWithBins> >();
foreach (string chr in coverageInfo.StartByChr.Keys)
{
segmentsByChromosome.Add(chr, new List <SegmentWithBins>());
SegmentWithBins currentSegment = null;
List <SampleGenomicBin> excludeIntervals = null;
if (excludedIntervals.ContainsKey(chr))
{
excludeIntervals = excludedIntervals[chr];
}
var excludeIndex = 0; // Points to the first interval which *doesn't* end before our current position
uint previousBinEnd = 0;
for (int binIndex = 0; binIndex < coverageInfo.StartByChr[chr].Length; binIndex++)
{
uint start = coverageInfo.StartByChr[chr][binIndex];
uint end = coverageInfo.EndByChr[chr][binIndex];
bool newSegment = IsNewSegment(starts, chr, excludeIntervals, previousBinEnd, end, start, ref excludeIndex, referencePloidy);
var bin = new Bin(start, end, coverageInfo.CoverageByChr[chr][binIndex]);
if (newSegment)
{
segmentNum++;
currentSegment = new SegmentWithBins(segmentNum, bin);
segmentsByChromosome[chr].Add(currentSegment);
}
else
{
if (currentSegment == null)
{
currentSegment = new SegmentWithBins(segmentNum, bin);
segmentsByChromosome[chr].Add(currentSegment);
}
else
{
currentSegment.AddBin(bin);
}
}
previousBinEnd = end;
}
}
return(segmentsByChromosome);
}
static int Main(string[] args)
{
CanvasCommon.Utilities.LogCommandLine(args);
List <string> cleanedFiles = new List <string>();
List <string> outPartitionedFiles = new List <string>();
List <string> vafFiles = new List <string>();
bool needHelp = false;
bool isGermline = false;
string filterBedFile = null;
string referenceFolder = null;
string commonCNVsbedPath = null;
string evennessMetricFile = null;
SegmentSplitUndo undoMethod = SegmentSplitUndo.None;
SegmentationInput.SegmentationMethod partitionMethod = SegmentationInput.SegmentationMethod.Wavelets;
string parameterconfigPath = Path.Combine(Isas.Framework.Utilities.Utilities.GetAssemblyFolder(typeof(CanvasPartition)), "CanvasPartitionParameters.json");
string ploidyVcfPath = null;
OptionSet p = new OptionSet()
{
{ "i|infile=", "input file - usually generated by CanvasClean", v => cleanedFiles.Add(v) },
{ "v|vaffile=", "variant frequencyfiles - usually generated by CanvasSNV", v => vafFiles.Add(v) },
{ "o|outfile=", "text file to output", v => outPartitionedFiles.Add(v) },
{ "m|method=", "segmentation method (Wavelets/CBS). Default: " + partitionMethod, v => partitionMethod = (SegmentationInput.SegmentationMethod)Enum.Parse(typeof(SegmentationInput.SegmentationMethod), v) },
{ "r|reference=", "folder that contains both genome.fa and GenomeSize.xml", v => referenceFolder = v },
{ "s|split=", "CBS split method (None/Prune/SDUndo). Default: " + undoMethod, v => undoMethod = (SegmentSplitUndo)Enum.Parse(typeof(SegmentSplitUndo), v) },
{ "b|bedfile=", "bed file to exclude (don't span these intervals)", v => filterBedFile = v },
{ "c|commoncnvs=", "bed file with common CNVs (always include these intervals into segmentation results)", v => commonCNVsbedPath = v },
{ "g|germline", "flag indicating that input file represents germline genome", v => isGermline = v != null },
{ $"{CommandLineOptions.EvennessMetricFile}=", "output file for evenness metric (optional)", v => evennessMetricFile = v },
{ "p|ploidyVcfFile=", "vcf file specifying reference ploidy (e.g. for sex chromosomes) (optional)", v => ploidyVcfPath = v },
{ "config=", "parameter configuration path (default {parameterconfigPath})", v => parameterconfigPath = v },
{ "h|help", "show this message and exit", v => needHelp = v != null }
};
List <string> extraArgs = p.Parse(args);
if (extraArgs.Any())
{
throw new IlluminaException($"Unknown arguments: {string.Join(",", extraArgs)}");
}
if (needHelp)
{
ShowHelp(p);
return(0);
}
if (!cleanedFiles.Any() || !outPartitionedFiles.Any() || referenceFolder == null)
{
ShowHelp(p);
return(0);
}
if (cleanedFiles.Any(inFile => !File.Exists(inFile)))
{
Console.WriteLine("CanvasPartition.exe: File {0} does not exist! Exiting.", cleanedFiles);
return(1);
}
if (!string.IsNullOrEmpty(filterBedFile) && !File.Exists(filterBedFile))
{
Console.WriteLine("CanvasPartition.exe: File {0} does not exist! Exiting.", filterBedFile);
return(1);
}
if (!File.Exists(parameterconfigPath))
{
Console.WriteLine($"CanvasPedigreeCaller.exe: File {parameterconfigPath} does not exist! Exiting.");
return(1);
}
if (!string.IsNullOrEmpty(ploidyVcfPath) && !File.Exists(ploidyVcfPath))
{
Console.WriteLine("CanvasPartition.exe: File {0} does not exist! Exiting.", ploidyVcfPath);
return(1);
}
var parameterconfigFile = new FileLocation(parameterconfigPath);
var canvasPartitionParameters = Deserialize <CanvasPartitionParameters>(parameterconfigFile);
ILogger logger = new Logger(Console.Out.ToEnumerable(), Console.Error.ToEnumerable());
var processor = new SegmentationResultsProcessor(canvasPartitionParameters.MaxInterBinDistInSegment);
var segmentationInputs = vafFiles.Count > 0 && vafFiles.Count == cleanedFiles.Count ?
cleanedFiles.Zip(vafFiles, (inFile, vafFile) => new SegmentationInput(inFile, vafFile, filterBedFile,
referenceFolder, evennessMetricFile, logger, processor)).ToList() :
cleanedFiles.Select(inFile => new SegmentationInput(inFile, null, filterBedFile,
referenceFolder, evennessMetricFile, logger, processor)).ToList();
GenomeSegmentationResults segmentationResults;
PloidyInfo referencePloidy = ploidyVcfPath != null?PloidyInfo.LoadPloidyFromVcfFileNoSampleId(ploidyVcfPath) : null;
switch (partitionMethod)
{
default: // use Wavelets if CBS is not selected
Console.WriteLine("{0} Running Wavelet Partitioning", DateTime.Now);
var waveletsRunner = new WaveletsRunner(new WaveletsRunner.WaveletsRunnerParams(isGermline,
commonCNVsbedPath, madFactor:
canvasPartitionParameters.MadFactor,
thresholdLowerMaf: canvasPartitionParameters.ThresholdLowerMaf,
evennessScoreThreshold: canvasPartitionParameters.EvennessScoreThreshold, verbose: 2));
segmentationResults = new GenomeSegmentationResults(waveletsRunner.Run(segmentationInputs.Single(),
canvasPartitionParameters.EvennessScoreWindow));
PostProcessAndWriteResults(segmentationInputs.Single(), outPartitionedFiles.Single(), referencePloidy, segmentationResults);
break;
case SegmentationInput.SegmentationMethod.CBS:
{
Console.WriteLine("{0} Running CBS Partitioning", DateTime.Now);
var cbsRunner = new CBSRunner(canvasPartitionParameters.MaxInterBinDistInSegment, undoMethod,
canvasPartitionParameters.CBSalpha);
var sampleSegmentations = new List <GenomeSegmentationResults>();
foreach (var input in segmentationInputs)
{
var segmentation = new GenomeSegmentationResults(cbsRunner.Run(input, verbose: 2));
sampleSegmentations.Add(segmentation);
}
segmentationResults = GenomeSegmentationResults.SplitOverlappingSegments(sampleSegmentations);
foreach (var(segmentationInput, outPartitionedFile) in segmentationInputs.Zip(outPartitionedFiles))
{
PostProcessAndWriteResults(segmentationInput, outPartitionedFile, referencePloidy, segmentationResults);
}
break;
}
case SegmentationInput.SegmentationMethod.HMM:
{
Console.WriteLine("{0} Running HMM Partitioning", DateTime.Now);
var hiddenMarkovModelsRunner = new HiddenMarkovModelsRunner(cleanedFiles.Count);
bool isPerSample = false;
segmentationResults =
new GenomeSegmentationResults(hiddenMarkovModelsRunner.Run(segmentationInputs, isPerSample));
for (int i = 0; i < segmentationInputs.Count; i++)
{
PostProcessAndWriteResults(segmentationInputs[i], outPartitionedFiles[i], referencePloidy, segmentationResults);
}
break;
}
case SegmentationInput.SegmentationMethod.PerSampleHMM:
{
Console.WriteLine("{0} Running Per-sample HMM Partitioning", DateTime.Now);
var hiddenMarkovModelsRunner = new HiddenMarkovModelsRunner(1);
var sampleSegmentations = new List <GenomeSegmentationResults>();
bool isPerSample = true;
foreach (var input in segmentationInputs)
{
var segmentation =
new GenomeSegmentationResults(
hiddenMarkovModelsRunner.Run(input.Yield().ToList(), isPerSample));
sampleSegmentations.Add(segmentation);
}
segmentationResults = GenomeSegmentationResults.SplitOverlappingSegments(sampleSegmentations);
foreach (var(segmentationInput, outPartitionedFile) in segmentationInputs.Zip(outPartitionedFiles))
{
PostProcessAndWriteResults(segmentationInput, outPartitionedFile, referencePloidy,
segmentationResults);
}
break;
}
}
Console.WriteLine("{0} CanvasPartition results written out", DateTime.Now);
return(0);
}
