protected override MpileupResult GetMpileupResult()
{
Progress.SetMessage("Multiple thread mode, parallel by chromosome ...");
_threadCount = 0;
var chromosomes = new ConcurrentQueue <string>();
foreach (var chr in _options.ChromosomeNames)
{
chromosomes.Enqueue(chr);
}
var cts = new CancellationTokenSource();
var maxThreadCount = Math.Min(_options.ThreadCount, _options.ChromosomeNames.Count);
for (int i = 0; i < maxThreadCount; i++)
{
ThreadPool.QueueUserWorkItem(ParallelChromosome, new Tuple <CancellationTokenSource, ConcurrentQueue <string> >(cts, chromosomes));
}
Thread.Sleep(5000);
while (_threadCount > 0)
{
Thread.Sleep(100);
}
Progress.SetMessage("After thread finished ...");
var result = new MpileupResult(string.Empty, _options.CandidatesDirectory);
Progress.SetMessage("Merging summary information ...");
foreach (var chr in _options.ChromosomeNames)
{
var summaryFile = new MpileupResult(chr, _options.CandidatesDirectory).CandidateSummary;
var summary = new MpileupResultCountFormat(_options, false).ReadFromFile(summaryFile);
result.MergeWith(summary);
}
Progress.SetMessage("Loading candidates ...");
foreach (var file in Directory.GetFiles(_options.CandidatesDirectory, "*.wsm"))
{
var res = MpileupFisherResultFileFormat.ParseString(Path.GetFileNameWithoutExtension(file));
res.CandidateFile = file;
result.Results.Add(res);
}
return(result);
}
public MpileupFisherResult Parse(string line, bool writeCandidateFile = true)
{
MpileupFisherResult result = new MpileupFisherResult();
var item = _parser.GetValue(line);
item.Samples[0].SampleName = "NORMAL";
item.Samples[1].SampleName = "TUMOR";
var events = item.GetPairedEvent();
var fisherresult = item.InitializeTable(events);
result.Item = item;
result.Group = fisherresult;
if (!_rdFilter.Accept(item))
{
result.FailedReason = _rdFilter.RejectReason;
_result.MinimumReadDepthFailed++;
return(result);
}
//If the bases from all samples are same, ignore the entry.
if (item.OnlyOneEvent())
{
result.FailedReason = "Only one allele detected";
_result.OneEventFailed++;
return(result);
}
if (fisherresult.Sample1.FailedPercentage > fisherresult.Sample2.FailedPercentage)
{
result.FailedReason = "MAF decreased in tumor";
_result.MinorAlleleDecreasedFailed++;
return(result);
}
if (!_normalTest.Accept(fisherresult))
{
result.FailedReason = _normalTest.RejectReason;
_result.MinorAlleleFailedInNormalSample++;
return(result);
}
if (!_tumorTest.Accept(fisherresult))
{
result.FailedReason = _tumorTest.RejectReason;
_result.MinorAlleleFailedInTumorSample++;
return(result);
}
//group fisher exact test
fisherresult.CalculateTwoTailPValue();
if (fisherresult.PValue > _options.FisherPvalue)
{
result.FailedReason = string.Format("Fisher pvalue > {0}", _options.FisherPvalue);
_result.GroupFisherFailed++;
return(result);
}
if (writeCandidateFile)
{
//get major and second alleles
var bases = new HashSet <string>(new[] { events.MajorEvent, events.MinorEvent });
//save to file
var piFile = new PileupItemFile(bases);
result.CandidateFile = string.Format("{0}/{1}.wsm", _options.CandidatesDirectory, MpileupFisherResultFileFormat.GetString(result, '_', false));
piFile.WriteToFile(result.CandidateFile, item);
}
Progress.SetMessage("{0}\t{1}\t{2}", item.SequenceIdentifier, item.Position, fisherresult);
return(result);
}
public override IEnumerable <string> Process()
{
if (!File.Exists(_options.BaseFilename) || new FileInfo(_options.BaseFilename).Length == 0)
{
base.Process();
}
else
{
Progress.SetMessage("Base file {0} exists, ignore pileup ...", _options.BaseFilename);
}
var filterOptions = options.GetFilterOptions();
if (new FileInfo(_options.BaseFilename).Length > 0)
{
if (!filterOptions.PrepareOptions())
{
throw new Exception("Filter options failed: " + filterOptions.ParsingErrors.Merge("\n"));
}
new FilterProcessor(filterOptions).Process();
var lines = File.ReadAllLines(filterOptions.ROutputFile).Skip(1).ToArray();
var glmfailed = lines.Count(m => m.Contains("GLM_PVALUE"));
var summarylines = File.ReadAllLines(_options.SummaryFilename).ToList();
if (summarylines.Last().StartsWith("glm pvalue"))
{
summarylines.RemoveAt(summarylines.Count - 1);
}
summarylines.Add(string.Format("glm pvalue > {0}\t{1}\t{2}", options.GlmPvalue, glmfailed, lines.Length - glmfailed));
File.WriteAllLines(_options.SummaryFilename, summarylines);
}
var mutationList = GetValidationList();
var candidates = new MpileupFisherResultFileFormat().ReadFromFile(options.CandidatesFilename).ToDictionary(m => GenomeUtils.GetKey(m.Item.SequenceIdentifier, m.Item.Position));
var items = new FilterItemTextFormat().ReadFromFile(filterOptions.ROutputFile).ToDictionary(m => GenomeUtils.GetKey(m.Chr, m.Start));
var result = new List <FilterItem>();
foreach (var mutation in mutationList.Items)
{
var key = GenomeUtils.GetKey(mutation.Chr, mutation.Pos);
if (items.ContainsKey(key))
{
result.Add(items[key]);
}
else
{
var item = new FilterItem();
item.Chr = mutation.Chr;
item.Start = mutation.Pos.ToString();
item.End = item.Start;
item.FisherNormal = string.Empty;
item.BrglmConverged = string.Empty;
item.BrglmGroup = 1.0;
item.BrglmGroupFdr = 1.0;
item.BrglmScore = string.Empty;
item.BrglmStrand = string.Empty;
item.BrglmPosition = string.Empty;
item.Identity = string.Empty;
result.Add(item);
if (candidates.ContainsKey(key))
{
var cand = candidates[key];
item.ReferenceAllele = cand.Item.Nucleotide.ToString();
item.MajorAllele = cand.Group.SucceedName;
item.MinorAllele = cand.Group.FailedName;
item.NormalMajorCount = cand.Group.Sample1.Succeed;
item.NormalMinorCount = cand.Group.Sample1.Failed;
item.TumorMajorCount = cand.Group.Sample2.Succeed;
item.TumorMinorCount = cand.Group.Sample2.Failed;
item.FisherGroup = cand.Group.PValue;
item.Filter = cand.FailedReason;
Console.WriteLine("In candidates : " + item.Filter);
}
else
{
item.NormalMajorCount = 0;
item.NormalMinorCount = 0;
item.TumorMajorCount = 0;
item.TumorMinorCount = 0;
item.Filter = "No coverage";
Console.WriteLine("No read : " + item.Filter);
}
}
}
new FilterItemVcfWriter(filterOptions).WriteToFile(_options.OutputSuffix + ".vcf", result);
new FilterItemTextFormat().WriteToFile(_options.OutputSuffix + ".tsv", result);
return(new string[] { _options.OutputSuffix + ".tsv", _options.OutputSuffix + ".vcf" });
}
public MpileupFisherResult Parse(string line, bool writeCandidateFile = true)
{
var parts = line.Split('\t');
if (bedMap.Count > 0)
{
var sp = _parser.GetSequenceIdentifierAndPosition(parts);
if (IsIgnored(sp.SequenceIdentifier, sp.Position))
{
_result.Ignored++;
return(null);
}
}
//for debug
//var sptemp = _parser.GetSequenceIdentifierAndPosition(parts);
//if (sptemp.SequenceIdentifier == "2" && sptemp.Position == 89161431)
//{
// var debugFile = string.Format("{0}/debug.txt", _options.CandidatesDirectory);
// File.WriteAllLines(debugFile, new[] { line });
// Console.WriteLine("Catched the line");
// System.Windows.Forms.Application.Exit();
//}
//didn't consider minimum score requirement
if (!_parser.HasEnoughReads(parts))
{
_result.MinimumReadDepthFailed++;
return(null);
}
//didn't consider minimum score requirement
if (!_parser.HasMinorAllele(parts))
{
_result.OneEventFailed++;
return(null);
}
//parsing full result considering score limitation
var item = _parser.GetValue(parts);
if (item == null)
{
_result.MinimumReadDepthFailed++;
return(null);
}
//If the bases from all samples are same, ignore the entry.
if (item.OnlyOneEvent())
{
_result.OneEventFailed++;
return(null);
}
item.Samples[0].SampleName = "NORMAL";
item.Samples[1].SampleName = "TUMOR";
var events = item.GetPairedEvent();
var fisherresult = item.InitializeTable(events);
if (fisherresult.Sample1.FailedPercentage > fisherresult.Sample2.FailedPercentage)
{
_result.MinorAlleleDecreasedFailed++;
return(null);
}
if (!_tumorTest.Accept(fisherresult))
{
_result.MinorAlleleFailedInTumorSample++;
return(null);
}
if (!_normalTest.Accept(fisherresult))
{
_result.MinorAlleleFailedInNormalSample++;
return(null);
}
//group fisher exact test
fisherresult.CalculateTwoTailPValue();
if (_options.UseZeroMinorAlleleStrategy && fisherresult.Sample1.Failed == 0)
{
//Console.WriteLine("UseZeroMinorAlleleStrategy : {0}", fisherresult);
if (fisherresult.PValue > _options.ZeroMinorAlleleStrategyFisherPvalue)
{
_result.GroupFisherFailed++;
return(null);
}
}
else if (fisherresult.PValue > _options.FisherPvalue)
{
_result.GroupFisherFailed++;
return(null);
}
//passed all criteria
_result.CandidateCount++;
var result = new MpileupFisherResult
{
Item = item,
Group = fisherresult,
};
//save to file
if (writeCandidateFile)
{
//get major and second alleles
var bases = new HashSet <string>(new[] { events.MajorEvent, events.MinorEvent });
var piFile = new PileupItemFile(bases);
result.CandidateFile = string.Format("{0}/{1}.wsm", _options.CandidatesDirectory, MpileupFisherResultFileFormat.GetString(result, '_', false));
piFile.WriteToFile(result.CandidateFile, item);
}
Progress.SetMessage("{0}\t{1}\t{2}", item.SequenceIdentifier, item.Position, fisherresult);
return(result);
}