public IdentifiedSpectrumBuilderResult Build(string parameterFile)
{
Options = new BuildSummaryOptions(parameterFile);
Options.DatasetList.RemoveDisabled();
IIdentifiedProteinBuilder proteinBuilder = new IdentifiedProteinBuilder();
IIdentifiedProteinGroupBuilder groupBuilder = new IdentifiedProteinGroupBuilder();
var fdrCalc = Options.FalseDiscoveryRate.GetFalseDiscoveryRateCalculator();
BuildResult = new DatasetList();
//从配置进行初始化
BuildResult.InitFromOptions(Options.DatasetList, this.Progress, parameterFile);
var totalCount = BuildResult.GetOptimalSpectrumCount();
string optimalResultFile = FileUtils.ChangeExtension(parameterFile, ".optimal");
using (var sw = new StreamWriter(optimalResultFile))
{
new OptimalFileTextWriter().WriteToStream(sw, BuildResult);
UniformProteinFdrOptimalResultCalculator proteinCalc = new UniformProteinFdrOptimalResultCalculator(fdrCalc, Options.GetDecoyGroupFilter())
{
Progress = this.Progress
};
Progress.SetMessage("Filtering PSMs by protein fdr {0}, using peptide fdr {1}...", Options.FalseDiscoveryRate.FdrValue, Options.FalseDiscoveryRate.MaxPeptideFdr);
var groupFilter = Options.FalseDiscoveryRate.FilterOneHitWonder ? new IdentifiedProteinGroupSingleWonderPeptideCountFilter(Options.FalseDiscoveryRate.MinOneHitWonderPeptideCount) : null;
var ret = proteinCalc.GetOptimalResultForGroupFilter(BuildResult, Options.FalseDiscoveryRate.MaxPeptideFdr, Options.FalseDiscoveryRate.FdrValue, groupFilter);
//只保留通过筛选的蛋白质包含的PSMs。
BuildResult.KeepOptimalResultInSetOnly(ret.AcceptedSpectra);
GC.Collect();
GC.WaitForPendingFinalizers();
sw.WriteLine("After SimpleProteinFDR filter {0} with condition {1}, required peptide fdr = {2} ", ret.ProteinFdr, ret.ProteinCondition, ret.PeptideFdr);
BuildResult.ForEach(ds =>
{
sw.WriteLine("Dataset {0}", ds.Options.Name);
OptimalResultConditionUtils.WriteSpectrumBin(sw, ds, f1, f2);
});
//sw.WriteLine();
//new OptimalFileTextWriter().WriteToStream(sw, BuildResult);
return(new IdentifiedSpectrumBuilderResult()
{
Spectra = ret.AcceptedSpectra.ToList(),
PeptideFDR = ret.PeptideFdr,
ProteinFDR = ret.ProteinFdr
});
}
}
public void BuildResult(Dictionary <string, MaxQuantModificationItem> shortModMap, string noredundantFile, string iniFile, List <IIdentifiedSpectrum> spectra)
{
var proteins = new IdentifiedProteinBuilder().Build(spectra);
var groups = new IdentifiedProteinGroupBuilder().Build(proteins);
var ir = new IdentifiedResultBuilder(DefaultAccessNumberParser.GetInstance(), "").Build(groups);
new MascotResultTextFormat().WriteToFile(noredundantFile, ir);
Aminoacids samAminoacids, refAminoacids;
InitializeAminoacids(out samAminoacids, out refAminoacids);
using (var sw = new StreamWriter(iniFile))
{
for (var idx = 0; idx < samAminoacids.Count; idx++)
{
var samAA = samAminoacids[idx];
var refAA = refAminoacids[idx];
if (samAA.Visible)
{
sw.WriteLine("<{0}>\tSAM\tREF", samAA.OneName);
WriteAtom(sw, samAA, refAA, new[] { Atom.C }, "C");
WriteAtom(sw, samAA, refAA, new[] { Atom.H }, "H");
WriteAtom(sw, samAA, refAA, new[] { Atom.O }, "O");
WriteAtom(sw, samAA, refAA, new[] { Atom.N }, "N");
WriteAtom(sw, samAA, refAA, new[] { Atom.S }, "S");
WriteAtom(sw, samAA, refAA, new[] { Atom.P }, "P");
WriteAtom(sw, samAA, refAA, new[] { Atom.N15, Atom.Nx }, "15N");
WriteAtom(sw, samAA, refAA, new[] { Atom.H2, Atom.Hx }, "2H");
WriteAtom(sw, samAA, refAA, new[] { Atom.C13, Atom.Cx }, "13C");
WriteAtom(sw, samAA, refAA, new[] { Atom.O18, Atom.Ox }, "18O", false);
sw.WriteLine();
}
}
var terms = new[] { new MaxQuantModificationItem()
{
Symbol = "NTERM", Composition = new AtomComposition("H")
},
new MaxQuantModificationItem()
{
Symbol = "CTERM", Composition = new AtomComposition("OH")
} };
var usedMods = (from v in shortModMap.Values
orderby v.Symbol
select v).ToList();
usedMods.AddRange(terms);
foreach (var mod in usedMods)
{
mod.WriteToSilacINI(sw);
}
}
}
public override IEnumerable <string> Process(string fileName)
{
var proteins = ParseProteins(fileName);
Progress.SetMessage("Building protein groups ...");
var groups = new IdentifiedProteinGroupBuilder().Build(proteins);
Progress.SetMessage("Building result ...");
var ir = new IdentifiedResultBuilder(null, null).Build(groups);
var result = FileUtils.ChangeExtension(fileName, ".noredundant");
RefineModifications(ir);
Progress.SetMessage("Saving result ...");
new SequestResultTextFormat(SequestHeader.SEQUEST_PROTEIN_HEADER, SequestHeader.SEQUEST_PEPTIDE_HEADER + "\tModification").WriteToFile(result, ir);
Progress.SetMessage("Finished.");
return(new string[] { result });
}
public void TestBuild()
{
List<IIdentifiedSpectrum> spectra = new SequestPeptideTextFormat().ReadFromFile(@"../../../data/TestBuilder.peptides");
Assert.AreEqual(4, spectra.Count);
IAccessNumberParser parser = AccessNumberParserFactory.FindOrCreateParser(@"(IPI\d+)", "IPI");
List<IIdentifiedProtein> proteins = new IdentifiedProteinBuilder().Build(spectra);
Assert.AreEqual(4, proteins.Count);
List<IIdentifiedProteinGroup> groups = new IdentifiedProteinGroupBuilder().Build(proteins);
Assert.AreEqual(2, groups.Count);
Assert.AreEqual(1, groups[0].Count);
Assert.AreEqual("IPI:IPI00784154.1|SW", groups[0][0].Name);
Assert.AreEqual(2, groups[1].Count);
Assert.AreEqual("REVERSED_00000001", groups[1][0].Name);
Assert.AreEqual("REVERSED_00000002", groups[1][1].Name);
IIdentifiedResult result = new IdentifiedResultBuilder(parser,"").Build(groups);
}
public void TestBuild()
{
var pep1 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("A", 1, 1, 1, ".dta"))) { Sequence = "A" };
var pep2 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("B", 1, 1, 1, ".dta"))) { Sequence = "B" };
var pep3 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("C", 1, 1, 1, ".dta"))) { Sequence = "C" };
var pep4 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("D", 1, 1, 1, ".dta"))) { Sequence = "D" };
var pep5 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("E", 1, 1, 1, ".dta"))) { Sequence = "E" };
var pep6 = new IdentifiedPeptide(new IdentifiedSpectrum(new SequestFilename("F", 1, 1, 1, ".dta"))) { Sequence = "F" };
var protein1 = new IdentifiedProtein()
{
Peptides = new IIdentifiedPeptide[] { pep1, pep3, pep5, pep6 }.ToList()
};
var protein2 = new IdentifiedProtein()
{
Peptides = new IIdentifiedPeptide[] { pep2, pep3, pep4 }.ToList()
};
//should be removed from final result since all peptides has been included in protein1 and protein2, even one protein contains both peptides
var protein3 = new IdentifiedProtein()
{
Peptides = new IIdentifiedPeptide[] { pep1, pep2 }.ToList()
};
//should be removed from final result since all peptides has been included in protein1
var protein4 = new IdentifiedProtein()
{
Peptides = new IIdentifiedPeptide[] { pep1, pep5 }.ToList()
};
var actual = new IdentifiedProteinGroupBuilder().Build(new IIdentifiedProtein[] { protein1, protein2, protein3 }.ToList());
Assert.AreEqual(2, actual.Count);
Assert.AreSame(protein1, actual[0][0]);
Assert.AreSame(protein2, actual[1][0]);
}
public IdentifiedSpectrumBuilderResult Build(string parameterFile)
{
Options = new BuildSummaryOptions(parameterFile);
Options.DatasetList.RemoveDisabled();
IIdentifiedProteinGroupFilter conFilter = Options.Database.GetNotContaminationDescriptionFilter(this.Progress);
var fdrCalc = Options.FalseDiscoveryRate.GetFalseDiscoveryRateCalculator();
BuildResult = new DatasetList();
//从配置进行初始化
BuildResult.InitFromOptions(Options.DatasetList, this.Progress, parameterFile);
var totalCount = BuildResult.GetOptimalSpectrumCount();
string optimalResultFile = FileUtils.ChangeExtension(parameterFile, ".optimal");
using (var sw = new StreamWriter(optimalResultFile))
{
new OptimalFileTextWriter().WriteToStream(sw, BuildResult);
UniformProteinFdrOptimalResultCalculator proteinCalc = new UniformProteinFdrOptimalResultCalculator(fdrCalc, Options.GetDecoyGroupFilter())
{
Progress = this.Progress
};
sw.WriteLine(OptimalFilteredItem.GetHeader());
var uniqueFilter = new IdentifiedProteinGroupUniquePeptideCountFilter(2);
OptimalFilteredItem finalItem = null;
List <IIdentifiedSpectrum> allSpectrum = Options.PeptideRetrieval ? BuildResult.GetSpectra() : null;
int fdrPeptideCount = Options.FalseDiscoveryRate.FdrPeptideCount > 2 ? Options.FalseDiscoveryRate.FdrPeptideCount : 2;
double firstStepFdr = Options.FalseDiscoveryRate.MaxPeptideFdr;
bool bFirst = true;
for (int curPeptideCount = fdrPeptideCount; curPeptideCount >= 2; curPeptideCount--)
{
//重新根据保留的Spectra构建SpectrumBin。
if (!bFirst)
{
BuildResult.BuildSpectrumBin();
}
bFirst = false;
var curItem = new OptimalFilteredItem();
IIdentifiedProteinGroupFilter groupFilter;
bool bNeedFirstStep = curPeptideCount > 2;
if (bNeedFirstStep)
{
Progress.SetMessage("Filtering PSMs by protein fdr {0}, unique peptide count >= 2 and peptide count >= {1} using peptide fdr {2}...", Options.FalseDiscoveryRate.FdrValue, curPeptideCount, firstStepFdr);
//第一步，根据UniquePeptideCount和PeptideCount进行筛选，得到满足蛋白质Fdr要求所对应的肽段fdr。
var countFilter = new IdentifiedProteinGroupPeptideCountFilter(curPeptideCount);
if (conFilter != null)
{
groupFilter = new AndIdentifiedProteinGroupFilter(new IIdentifiedProteinGroupFilter[] { conFilter, uniqueFilter, countFilter });
}
else
{
groupFilter = new AndIdentifiedProteinGroupFilter(new IIdentifiedProteinGroupFilter[] { uniqueFilter, countFilter });
}
curItem.Unique2CountResult = proteinCalc.GetOptimalResultForGroupFilter(BuildResult, firstStepFdr, Options.FalseDiscoveryRate.FdrValue, groupFilter);
firstStepFdr = curItem.Unique2CountResult.PeptideFdr;
//只保留没有被通过筛选的蛋白质包含的PSMs。
BuildResult.KeepOptimalResultInSetOnly(new HashSet <IIdentifiedSpectrum>(curItem.Unique2CountResult.RejectedSpectra));
GC.Collect();
GC.WaitForPendingFinalizers();
}
else
{
curItem.Unique2CountResult = new ProteinFdrFilteredItem();
}
Progress.SetMessage("Filtering PSMs by protein fdr {0}, unique peptide count >= 2 using peptide fdr {1}...", Options.FalseDiscoveryRate.FdrValue, firstStepFdr);
//第二步，根据UniquePeptideCount进行筛选，计算得到满足给定蛋白质fdr的结果。
double secondStepFdr = bNeedFirstStep ? Options.FalseDiscoveryRate.MaxPeptideFdr : firstStepFdr;
if (conFilter != null)
{
groupFilter = new AndIdentifiedProteinGroupFilter(new IIdentifiedProteinGroupFilter[] { conFilter, uniqueFilter });
}
else
{
groupFilter = uniqueFilter;
}
curItem.Unique2Result = proteinCalc.GetOptimalResultForGroupFilter(BuildResult, secondStepFdr, Options.FalseDiscoveryRate.FdrValue, groupFilter);
//只保留没有被通过筛选的蛋白质包含的PSMs。
BuildResult.KeepOptimalResultInSetOnly(new HashSet <IIdentifiedSpectrum>(curItem.Unique2Result.RejectedSpectra));
GC.Collect();
GC.WaitForPendingFinalizers();
curItem.Unique1Result = FilterOneHitWonders(conFilter, proteinCalc);
GC.Collect();
GC.WaitForPendingFinalizers();
sw.WriteLine(curItem.ToString());
if (finalItem == null || finalItem.TotalProteinCount < curItem.TotalProteinCount)
{
finalItem = curItem;
}
curItem = null;
GC.Collect();
GC.WaitForPendingFinalizers();
Console.WriteLine(MyConvert.Format("Filtering PSMs by protein fdr {0}, unique peptide count >= 2 and peptide count >= {1} using peptide fdr {2}...cost {3}.", Options.FalseDiscoveryRate.FdrValue, curPeptideCount, firstStepFdr, SystemUtils.CostMemory()));
}
Progress.SetMessage("Filtering PSMs by protein fdr {0} finished, free memory...", Options.FalseDiscoveryRate.FdrValue);
if (finalItem != null)
{
sw.WriteLine();
sw.WriteLine("Final result : ");
WriteScoreMap(sw, BuildResult, finalItem.Unique2CountResult);
WriteScoreMap(sw, BuildResult, finalItem.Unique2Result);
WriteScoreMap(sw, BuildResult, finalItem.Unique1Result);
var finalSpectra = finalItem.GetSpectra();
if (Options.PeptideRetrieval)
{
Progress.SetMessage("Retrivaling peptides passed maximum peptide FDR for proteins passed protein FDR...");
var proteinBuilder = new IdentifiedProteinBuilder();
var groupBuilder = new IdentifiedProteinGroupBuilder();
List <IIdentifiedProtein> proteins = proteinBuilder.Build(finalSpectra);
List <IIdentifiedProteinGroup> groups = groupBuilder.Build(proteins);
var proteinMap = new Dictionary <string, IIdentifiedProteinGroup>();
foreach (var g in groups)
{
foreach (var p in g)
{
proteinMap[p.Name] = g;
}
}
var savedSpectra = new HashSet <IIdentifiedSpectrum>(finalItem.GetSpectra());
foreach (var spectrum in allSpectrum)
{
if (savedSpectra.Contains(spectrum))
{
continue;
}
var pgs = new HashSet <IIdentifiedProteinGroup>();
foreach (var protein in spectrum.Proteins)
{
IIdentifiedProteinGroup pg;
if (proteinMap.TryGetValue(protein, out pg))
{
pgs.Add(pg);
}
}
//if the spectrum doesn't map to protein passed FDR filter, ignore
//if the spectrum maps to multiple groups, ignore
if (pgs.Count == 0 || pgs.Count > 1)
{
continue;
}
//The spectrum should map to all proteins in the group
if (pgs.First().All(l => spectrum.Proteins.Contains(l.Name)))
{
finalSpectra.Add(spectrum);
}
}
}
BuildResult.ClearSpectra();
GC.Collect();
GC.WaitForPendingFinalizers();
return(new IdentifiedSpectrumBuilderResult()
{
Spectra = finalSpectra,
PeptideFDR = finalItem.Unique2Result.PeptideFdr,
ProteinFDR = Options.FalseDiscoveryRate.FdrValue
});
}
else
{
return(new IdentifiedSpectrumBuilderResult()
{
Spectra = new List <IIdentifiedSpectrum>()
});
}
}
}
