public void GenerateComparison(SrmDocument docOriginal, IProgressMonitor progressMonitor)
{
DocOriginal = docOriginal;
_docCompare = docOriginal;
if (IsModel)
{
var handler = new MProphetResultsHandler(DocOriginal, PeakScoringModel);
handler.ScoreFeatures(progressMonitor, true);
_docCompare = handler.ChangePeaks(progressMonitor);
}
else
{
// Add in the necessary annotation definitions so that ImportPeakBoundaries can read them
var annotationNamesToStrip = new List <string>
{
MProphetResultsHandler.AnnotationName,
MProphetResultsHandler.MAnnotationName,
APEX_ANNOTATION,
START_TIME_ANNOTATION,
END_TIME_ANNOTATION
};
foreach (var annotationName in annotationNamesToStrip)
{
_docCompare = AddAnnotationIfMissing(_docCompare, annotationName);
}
// But strip the values of these annotations if there were any
var annotationNamesToKeep = _docCompare.Settings.DataSettings.AnnotationDefs
.Where(def => !annotationNamesToStrip.Contains(def.Name))
.Select(def => def.Name).ToList();
_docCompare = (SrmDocument)_docCompare.StripAnnotationValues(annotationNamesToKeep);
Importer = new PeakBoundaryImporter(_docCompare);
long lineCount = Helpers.CountLinesInFile(FilePath);
// Peek at the file to see if it has a column called "Apex"
using (var reader = new StreamReader(FilePath))
{
var line = reader.ReadLine();
var fieldNames = PeakBoundaryImporter.FIELD_NAMES.ToList();
int fieldsTotal;
int[] fieldIndices;
var separator = PeakBoundaryImporter.DetermineCorrectSeparator(line, fieldNames,
PeakBoundaryImporter.REQUIRED_NO_CHROM, out fieldIndices, out fieldsTotal);
ApexPresent = separator != null && fieldIndices[(int)PeakBoundaryImporter.Field.apex_time] != -1;
}
_docCompare = Importer.Import(FilePath, progressMonitor, lineCount, true, !ApexPresent);
if (progressMonitor.IsCanceled)
{
return;
}
}
var matches = new List <PeakBoundsMatch>();
var truePeptides = DocOriginal.Molecules.ToList();
var pickedPeptides = _docCompare.Molecules.ToList();
int nTruePeptides = truePeptides.Count;
var chromatogramSets = DocOriginal.Settings.MeasuredResults.Chromatograms;
for (int i = 0; i < chromatogramSets.Count; i++)
{
var chromSet = chromatogramSets[i];
var set = chromatogramSets[i];
for (int j = 0; j < nTruePeptides; j++)
{
var truePeptide = truePeptides[j];
var pickedPeptide = pickedPeptides[j];
// We don't care about peak picking performance for decoys
if (truePeptide.IsDecoy)
{
continue;
}
// We don't care about peak picking performance for standard peptides
if (truePeptide.GlobalStandardType != null)
{
continue;
}
var trueGroups = truePeptide.TransitionGroups.ToList();
var pickedGroups = pickedPeptide.TransitionGroups.ToList();
for (int k = 0; k < trueGroups.Count; k++)
{
var trueGroup = trueGroups[k];
var pickedGroup = pickedGroups[k];
var trueResults = trueGroup.Results[i];
var pickedResults = pickedGroup.Results[i];
if (trueResults.IsEmpty)
{
continue;
}
int nChromInfos = trueResults.Count;
for (int m = 0; m < nChromInfos; m++)
{
var trueInfo = trueResults[m];
var pickedInfo = pickedResults[m];
// For TRIC testing at the moment
// double? qvaluePicked = PeakBoundsMatch.GetScoreValue(pickedInfo,
// MProphetResultsHandler.AnnotationName, ci => pickedInfo.QValue);
// if (qvaluePicked.HasValue && qvaluePicked.Value > 0.01)
// continue;
var fileInfo = set.GetFileInfo(trueInfo.FileId);
var filePath = fileInfo.FilePath;
var key = new MatchKey(trueGroup.Id.GlobalIndex, trueInfo.FileId.GlobalIndex);
matches.Add(new PeakBoundsMatch(trueInfo, pickedInfo, key, chromSet, filePath,
trueGroup, truePeptide, HasNoQValues, HasNoScores, ApexPresent));
}
}
}
}
Matches = matches;
// Detect if the apex time is in seconds, and if so adjust it to minutes
if (ApexPresent)
{
// ReSharper disable PossibleMultipleEnumeration
var matchesWithApex = matches.Where(match => match.PickedApex.HasValue).ToArray();
double maxTime = 0;
if (matchesWithApex.Length > 0)
{
maxTime = matchesWithApex.Select(match => match.PickedApex.Value).Max();
}
if (maxTime > PeakBoundaryImporter.GetMaxRt(DocOriginal))
{
foreach (var match in matchesWithApex)
{
// If the end is greater than the Apex, then it must also be in seconds (otherwise minutes)
if (match.PickedEndBoundary.HasValue && match.PickedEndBoundary > match.PickedApex)
{
match.PickedEndBoundary = match.PickedEndBoundary / 60;
}
match.PickedApex = match.PickedApex / 60;
// If the start is now greater than the apex, then it must also be in seconds (otherwise minutes)
if (match.PickedStartBoundary.HasValue && match.PickedStartBoundary > match.PickedApex)
{
match.PickedStartBoundary = match.PickedStartBoundary / 60;
}
}
}
// ReSharper restore PossibleMultipleEnumeration
}
}
public void GenerateComparison(SrmDocument docOriginal, IProgressMonitor progressMonitor)
{
DocOriginal = docOriginal;
_docCompare = docOriginal;
if (IsModel)
{
var handler = new MProphetResultsHandler(DocOriginal, PeakScoringModel)
{
AddAnnotation = true,
AddMAnnotation = true
};
handler.ScoreFeatures(progressMonitor);
_docCompare = handler.ChangePeaks(progressMonitor);
}
else
{
// Add in the necessary annotation definitions so that ImportPeakBoundaries can read them
_docCompare = AddAnnotationIfMissing(_docCompare, MProphetResultsHandler.AnnotationName);
_docCompare = AddAnnotationIfMissing(_docCompare, MProphetResultsHandler.MAnnotationName);
_docCompare = AddAnnotationIfMissing(_docCompare, APEX_ANNOTATION);
// But strip the values of these annotations if there were any
var annotationNamesToStrip = new List <string>
{
MProphetResultsHandler.AnnotationName,
MProphetResultsHandler.MAnnotationName,
APEX_ANNOTATION
};
var annotationNamesToKeep = _docCompare.Settings.DataSettings.AnnotationDefs
.Where(def => !annotationNamesToStrip.Contains(def.Name))
.Select(def => def.Name).ToList();
_docCompare = (SrmDocument)_docCompare.StripAnnotationValues(annotationNamesToKeep);
Importer = new PeakBoundaryImporter(_docCompare);
long lineCount = Helpers.CountLinesInFile(FilePath);
// Peek at the file to see if it has a column called "Apex"
using (var reader = new StreamReader(FilePath))
{
var line = reader.ReadLine();
var fieldNames = PeakBoundaryImporter.FIELD_NAMES.ToList();
fieldNames.Add(new [] { APEX_ANNOTATION });
int fieldsTotal;
int[] fieldIndices;
var separator = PeakBoundaryImporter.DetermineCorrectSeparator(line, fieldNames,
PeakBoundaryImporter.REQUIRED_NO_CHROM, out fieldIndices, out fieldsTotal);
ApexPresent = separator != null && fieldIndices.Last() != -1;
}
_docCompare = Importer.Import(FilePath, progressMonitor, lineCount, true, !ApexPresent);
}
var matches = new List <PeakBoundsMatch>();
var truePeptides = DocOriginal.Molecules.ToList();
var pickedPeptides = _docCompare.Molecules.ToList();
int nTruePeptides = truePeptides.Count;
var chromatogramSets = DocOriginal.Settings.MeasuredResults.Chromatograms;
for (int i = 0; i < chromatogramSets.Count; i++)
{
var set = chromatogramSets[i];
for (int j = 0; j < nTruePeptides; j++)
{
var truePeptide = truePeptides[j];
var pickedPeptide = pickedPeptides[j];
// We don't care about peak picking performance for decoys
if (truePeptide.IsDecoy)
{
continue;
}
// We don't care about peak picking performance for standard peptides
if (truePeptide.GlobalStandardType != null)
{
continue;
}
var trueGroups = truePeptide.TransitionGroups.ToList();
var pickedGroups = pickedPeptide.TransitionGroups.ToList();
for (int k = 0; k < trueGroups.Count; k++)
{
var trueGroup = trueGroups[k];
var pickedGroup = pickedGroups[k];
var trueResults = trueGroup.Results[i];
var pickedResults = pickedGroup.Results[i];
if (trueResults == null)
{
continue;
}
int nChromInfos = trueResults.Count;
for (int m = 0; m < nChromInfos; m++)
{
var trueInfo = trueResults[m];
var pickedInfo = pickedResults[m];
var fileInfo = set.GetFileInfo(trueInfo.FileId);
var filePath = fileInfo.FilePath;
var key = new MatchKey(trueGroup.Id.GlobalIndex, trueInfo.FileId.GlobalIndex);
matches.Add(new PeakBoundsMatch(trueInfo, pickedInfo, key, filePath, trueGroup,
HasNoQValues, HasNoScores, ApexPresent));
}
}
}
}
Matches = matches;
// Detect if the apex time is in seconds, and if so adjust it to minutes
if (ApexPresent)
{
double maxTime = matches.Where(match => match.PickedApex.HasValue)
.Select(match => match.PickedApex.Value).Max();
if (maxTime > PeakBoundaryImporter.GetMaxRt(DocOriginal))
{
foreach (var match in matches.Where(match => match.PickedApex.HasValue))
{
match.PickedApex = match.PickedApex / 60;
}
}
}
}
private static void ImportThrowsException(SrmDocument docResults, string importText, string message, bool isMinutes = true, bool removeMissing = false, bool changePeaks = true)
{
var peakBoundaryImporter = new PeakBoundaryImporter(docResults);
using (var readerPeakBoundaries = new StringReader(importText))
{
long lineCount = Helpers.CountLinesInString(importText);
AssertEx.ThrowsException<IOException>(() => peakBoundaryImporter.Import(readerPeakBoundaries, null, lineCount, isMinutes, removeMissing, changePeaks), message);
}
}
private static SrmDocument ImportFileToDoc(SrmDocument docOld, string importFile)
{
var peakBoundaryImporter = new PeakBoundaryImporter(docOld);
long lineCount = Helpers.CountLinesInFile(importFile);
SrmDocument docNew = peakBoundaryImporter.Import(importFile, null, lineCount);
return docNew;
}
private void ImportPeakBoundaries(string fileName, long lineCount, string description)
{
var docCurrent = DocumentUI;
SrmDocument docNew = null;
using (var longWaitDlg = new LongWaitDlg(this) { Text = description })
{
var peakBoundaryImporter = new PeakBoundaryImporter(docCurrent);
longWaitDlg.PerformWork(this, 1000, longWaitBroker =>
docNew = peakBoundaryImporter.Import(fileName, longWaitBroker, lineCount));
if (docNew == null)
return;
if (!peakBoundaryImporter.UnrecognizedPeptidesCancel(this))
return;
if (longWaitDlg.IsDocumentChanged(docCurrent))
{
MessageDlg.Show(this, Resources.SkylineWindow_ImportPeakBoundaries_Unexpected_document_change_during_operation);
return;
}
}
ModifyDocument(description, doc =>
{
if (!ReferenceEquals(doc, docCurrent))
throw new InvalidDataException(Resources.SkylineWindow_ImportPeakBoundaries_Unexpected_document_change_during_operation);
return docNew;
});
}
public void GenerateComparison(SrmDocument docOriginal, IProgressMonitor progressMonitor)
{
DocOriginal = docOriginal;
_docCompare = docOriginal;
if (IsModel)
{
var handler = new MProphetResultsHandler(DocOriginal, PeakScoringModel)
{
AddAnnotation = true,
AddMAnnotation = true
};
handler.ScoreFeatures(progressMonitor);
_docCompare = handler.ChangePeaks(progressMonitor);
}
else
{
// Add in the necessary annotation definitions so that ImportPeakBoundaries can read them
_docCompare = AddAnnotationIfMissing(_docCompare, MProphetResultsHandler.AnnotationName);
_docCompare = AddAnnotationIfMissing(_docCompare, MProphetResultsHandler.MAnnotationName);
_docCompare = AddAnnotationIfMissing(_docCompare, APEX_ANNOTATION);
// But strip the values of these annotations if there were any
var annotationNamesToStrip = new List<string>
{
MProphetResultsHandler.AnnotationName,
MProphetResultsHandler.MAnnotationName,
APEX_ANNOTATION
};
var annotationNamesToKeep = _docCompare.Settings.DataSettings.AnnotationDefs
.Where(def => !annotationNamesToStrip.Contains(def.Name))
.Select(def => def.Name).ToList();
_docCompare = (SrmDocument)_docCompare.StripAnnotationValues(annotationNamesToKeep);
Importer = new PeakBoundaryImporter(_docCompare);
long lineCount = Helpers.CountLinesInFile(FilePath);
// Peek at the file to see if it has a column called "Apex"
using (var reader = new StreamReader(FilePath))
{
var line = reader.ReadLine();
var fieldNames = PeakBoundaryImporter.FIELD_NAMES.ToList();
fieldNames.Add(new [] {APEX_ANNOTATION});
int fieldsTotal;
int[] fieldIndices;
var separator = PeakBoundaryImporter.DetermineCorrectSeparator(line, fieldNames,
PeakBoundaryImporter.REQUIRED_NO_CHROM, out fieldIndices, out fieldsTotal);
ApexPresent = separator != null && fieldIndices.Last() != -1;
}
_docCompare = Importer.Import(FilePath, progressMonitor, lineCount, true, !ApexPresent);
}
var matches = new List<PeakBoundsMatch>();
var truePeptides = DocOriginal.Molecules.ToList();
var pickedPeptides = _docCompare.Molecules.ToList();
int nTruePeptides = truePeptides.Count;
var chromatogramSets = DocOriginal.Settings.MeasuredResults.Chromatograms;
for (int i = 0; i < chromatogramSets.Count; i++)
{
var set = chromatogramSets[i];
for (int j = 0; j < nTruePeptides; j++)
{
var truePeptide = truePeptides[j];
var pickedPeptide = pickedPeptides[j];
// We don't care about peak picking performance for decoys
if (truePeptide.IsDecoy)
continue;
// We don't care about peak picking performance for standard peptides
if (truePeptide.GlobalStandardType != null)
continue;
var trueGroups = truePeptide.TransitionGroups.ToList();
var pickedGroups = pickedPeptide.TransitionGroups.ToList();
for (int k = 0; k < trueGroups.Count; k++)
{
var trueGroup = trueGroups[k];
var pickedGroup = pickedGroups[k];
var trueResults = trueGroup.Results[i];
var pickedResults = pickedGroup.Results[i];
if (trueResults == null)
continue;
int nChromInfos = trueResults.Count;
for (int m = 0; m < nChromInfos; m++)
{
var trueInfo = trueResults[m];
var pickedInfo = pickedResults[m];
var fileInfo = set.GetFileInfo(trueInfo.FileId);
var filePath = fileInfo.FilePath;
var key = new MatchKey(trueGroup.Id.GlobalIndex, trueInfo.FileId.GlobalIndex);
matches.Add(new PeakBoundsMatch(trueInfo, pickedInfo, key, filePath, trueGroup,
HasNoQValues, HasNoScores, ApexPresent));
}
}
}
}
Matches = matches;
// Detect if the apex time is in seconds, and if so adjust it to minutes
if (ApexPresent)
{
double maxTime = matches.Where(match => match.PickedApex.HasValue)
.Select(match => match.PickedApex.Value).Max();
if (maxTime > PeakBoundaryImporter.GetMaxRt(DocOriginal))
{
foreach (var match in matches.Where(match => match.PickedApex.HasValue))
{
match.PickedApex = match.PickedApex / 60;
}
}
}
}
public void TestMProphetResultsHandler()
{
var testFilesDir = new TestFilesDir(TestContext, ZIP_FILE);
var documentFile = testFilesDir.GetTestPath("MProphetGold-trained.sky");
SrmDocument doc = ResultsUtil.DeserializeDocument(documentFile);
// Load libraries
doc = doc.ChangeSettings(doc.Settings.ChangePeptideLibraries(libraries =>
{
var lib = libraries.Libraries[0];
return libraries.ChangeLibrarySpecs(new LibrarySpec[]
{
new BiblioSpecLiteSpec(lib.Name, testFilesDir.GetTestPath(lib.FileNameHint))
});
}));
// Load an empty doc, so that we can make a change and
// cause the .skyd to be loaded
var docContainer = new ResultsTestDocumentContainer(null, documentFile);
docContainer.SetDocument(doc, null, true);
docContainer.AssertComplete();
SrmDocument docOriginal = docContainer.Document;
var peakScoringModel = docOriginal.Settings.PeptideSettings.Integration.PeakScoringModel;
var resultsHandler = new MProphetResultsHandler(docOriginal, peakScoringModel) { QValueCutoff = Q_CUTOFF };
// 1. Reintegrate and export report produces expected file
resultsHandler.ScoreFeatures();
var docNew = resultsHandler.ChangePeaks();
var reportSpec = MakeReportSpec();
if (IsSaveAll)
{
// For regenerating expected files if things change
ReportToCsv(reportSpec, docNew, testFilesDir.GetTestPath(REPORT_EXPECTED), CultureInfo.GetCultureInfo("en-US"));
ReportToCsv(reportSpec, docNew, testFilesDir.GetTestPathIntl(REPORT_EXPECTED), CultureInfo.GetCultureInfo("fr-FR"));
}
string docNewActual = testFilesDir.GetTestPath(REPORT_ACTUAL);
string docNewExpected = testFilesDir.GetTestPathLocale(REPORT_EXPECTED);
ReportToCsv(reportSpec, docNew, docNewActual, CultureInfo.CurrentCulture);
AssertEx.FileEquals(docNewExpected, docNewActual);
// 2. Reintegrating again gives no change in document
var resultsHandlerRepeat = new MProphetResultsHandler(docNew, peakScoringModel) { QValueCutoff = Q_CUTOFF };
resultsHandlerRepeat.ScoreFeatures();
var docRepeat = resultsHandlerRepeat.ChangePeaks();
Assert.AreSame(docRepeat, docNew);
Assert.AreNotSame(docOriginal, docNew);
// 3. Export mProphet results gives expected file
var calcs = peakScoringModel.PeakFeatureCalculators;
var mProphetActual = testFilesDir.GetTestPath(MPROPHET_ACTUAL);
var mProphetExpected = testFilesDir.GetTestPathLocale(MPROPHET_EXPECTED);
if (IsSaveAll)
{
// For regenerating files
SaveMProphetFeatures(resultsHandler, testFilesDir.GetTestPath(MPROPHET_EXPECTED), CultureInfo.GetCultureInfo("en-US"), calcs);
SaveMProphetFeatures(resultsHandler, testFilesDir.GetTestPathIntl(MPROPHET_EXPECTED), CultureInfo.GetCultureInfo("fr-FR"), calcs);
}
SaveMProphetFeatures(resultsHandler, mProphetActual, CultureInfo.CurrentCulture, calcs);
AssertEx.FileEquals(mProphetExpected, mProphetActual);
// 4. Export mProphet -> Import Peak Boundaries leads to same result as reintegrate
var resultsHandlerQAll = new MProphetResultsHandler(docOriginal, peakScoringModel) {QValueCutoff = 1.0};
resultsHandlerQAll.ScoreFeatures();
var docNewQAll = resultsHandlerQAll.ChangePeaks();
var peakBoundaryImporter = new PeakBoundaryImporter(docNewQAll);
long lineCount = Helpers.CountLinesInFile(mProphetActual);
peakBoundaryImporter.Import(mProphetActual, null, lineCount);
var docImport = peakBoundaryImporter.Document;
// Serialized documents are easier to debug when something is different
string strDocNew = SerializeDoc(docNewQAll);
string strDocImport = SerializeDoc(docImport);
AssertEx.NoDiff(strDocNew, strDocImport);
Assert.AreSame(docNewQAll, docImport);
// 5. Reintegration with q value cutoff of <0 causes all peaks set to null
var handlerAllNull = new MProphetResultsHandler(docOriginal, peakScoringModel) {QValueCutoff = -0.001};
handlerAllNull.ScoreFeatures();
var docNull = handlerAllNull.ChangePeaks();
foreach (var transitionNode in docNull.PeptideTransitions)
foreach(var chromInfo in transitionNode.ChromInfos)
Assert.IsTrue(chromInfo.IsEmpty || transitionNode.IsDecoy);
// 6. Reintegration adjusts example peak to null at q=0.005 cutoff, but adjusts it to a non-null peak at q=0.20
const int groupNum = 11;
var midQNode = resultsHandler.Document.PeptideTransitionGroups.ToList()[groupNum];
foreach (var chromInfo in midQNode.Transitions.SelectMany(transition => transition.ChromInfos))
Assert.IsTrue(chromInfo.IsEmpty);
resultsHandler.QValueCutoff = Q_CUTOFF_HIGH;
resultsHandler.ChangePeaks();
var midQNodeNew = resultsHandler.Document.PeptideTransitionGroups.ToList()[groupNum];
foreach (var chromInfo in midQNodeNew.Transitions.SelectMany(transition => transition.ChromInfos))
Assert.IsFalse(chromInfo.IsEmpty);
// 7. Labeled peptide pairs still have matching peaks
foreach (var peptideNode in resultsHandler.Document.Peptides)
{
Assert.AreEqual(2, peptideNode.TransitionGroupCount);
var groupList = peptideNode.TransitionGroups.ToList();
var lightGroup = groupList[0];
var heavyGroup = groupList[0];
var lightChromInfo = lightGroup.ChromInfos.ToList()[0];
var heavyChromInfo = heavyGroup.ChromInfos.ToList()[0];
Assert.AreEqual(lightChromInfo.StartRetentionTime, heavyChromInfo.StartRetentionTime);
Assert.AreEqual(lightChromInfo.EndRetentionTime, heavyChromInfo.EndRetentionTime);
Assert.AreEqual(lightChromInfo.RetentionTime, heavyChromInfo.RetentionTime);
}
// 8. Verify that chosen peaks and q values are the same as those in mProphet paper:
// http://www.nature.com/nmeth/journal/v8/n5/full/nmeth.1584.html#/supplementary-information
// TODO: Grab this data from the mProphet paper
// Release open streams
docContainer.Release();
}
