public ScanBasedTagSearchEngine( LcMsRun run, ISequenceTagFinder seqTagFinder, LcMsPeakMatrix featureFinder, FastaDatabase fastaDb, Tolerance tolerance, AminoAcidSet aaSet, CompositeScorerFactory ms2ScorerFactory = null, int minMatchedTagLength = DefaultMinMatchedTagLength, double maxSequenceMass = 50000.0, int minProductIonCharge = 1, int maxProductIonCharge = 20) { _run = run; _featureFinder = featureFinder; _searchableDb = new SearchableDatabase(fastaDb); _tolerance = tolerance; _aaSet = aaSet; _minMatchedTagLength = minMatchedTagLength; _maxSequenceMass = maxSequenceMass; _minProductIonCharge = minProductIonCharge; _maxProductIonCharge = maxProductIonCharge; MinScan = int.MinValue; MaxScan = int.MaxValue; _ms2ScorerFactory = ms2ScorerFactory; _seqTagFinder = seqTagFinder; }
public TagMatchFinder( ProductSpectrum spec, IScorer ms2Scorer, LcMsPeakMatrix featureFinder, string proteinSequence, Tolerance tolerance, AminoAcidSet aaSet, double maxSequenceMass) { _spec = spec; _ms2Scorer = ms2Scorer; _featureFinder = featureFinder; _proteinSequence = proteinSequence; _tolerance = tolerance; _aaSet = aaSet; _maxSequenceMass = maxSequenceMass; }
/// <summary> /// Find features in the data file /// </summary> /// <param name="rawFile">Data file (either a pbf file or a file type from which a pbf file can be auto-created)</param> /// <returns>0 if success; negative number on error</returns> private int ProcessFile(string rawFile) { var outDirectory = GetOutputDirectory(rawFile); if (string.IsNullOrEmpty(outDirectory)) return -1; var baseName = Path.GetFileName(MassSpecDataReaderFactory.RemoveExtension(rawFile)); var ms1FeaturesFilePath = Path.Combine(outDirectory, baseName + "." + FileExtension); var outCsvFilePath = Path.Combine(outDirectory, baseName + "_" + FileExtension + ".csv"); var pngFilePath = Path.Combine(outDirectory, baseName + "_" + FileExtension + ".png"); if (File.Exists(ms1FeaturesFilePath)) { Console.WriteLine(@"ProMex output already exists: {0}", ms1FeaturesFilePath); return -2; } if (!File.Exists(rawFile)) { ShowErrorMessage(@"Cannot find input file: " + rawFile); return -3; } var stopwatch = Stopwatch.StartNew(); Console.WriteLine(@"Start loading MS1 data from {0}", rawFile); var run = PbfLcMsRun.GetLcMsRun(rawFile); var featureFinder = new LcMsPeakMatrix(run, _likelihoodScorer, 1, 60, Parameters.MaxThreads); Console.WriteLine(@"Complete loading MS1 data. Elapsed Time = {0:0.000} sec", (stopwatch.ElapsedMilliseconds) / 1000.0d); if (run.GetMs1ScanVector().Length == 0) { ShowErrorMessage(@"Data file has no MS1 spectra: " + Path.GetFileName(rawFile)); return -4; } var comparer = featureFinder.Comparer; var container = new LcMsFeatureContainer(featureFinder.Ms1Spectra, _likelihoodScorer, new LcMsFeatureMergeComparer(new Tolerance(10))); var minSearchMassBin = comparer.GetBinNumber(Parameters.MinSearchMass); var maxSearchMassBin = comparer.GetBinNumber(Parameters.MaxSearchMass); double totalMassBin = maxSearchMassBin - minSearchMassBin + 1; Console.WriteLine(@"Start MS1 feature extraction."); stopwatch.Restart(); for (var binNum = minSearchMassBin; binNum <= maxSearchMassBin; binNum++) { var clusters = featureFinder.FindFeatures(binNum); container.Add(clusters); if (binNum > minSearchMassBin && (binNum - minSearchMassBin) % 1000 == 0) { var elapsed = (stopwatch.ElapsedMilliseconds) / 1000.0d; var processedBins = binNum - minSearchMassBin; var processedPercentage = ((double)processedBins / totalMassBin) * 100; Console.WriteLine(@"Processing {0:0.0}% of mass bins ({1:0.0} Da); elapsed time = {2:0.000} sec; # of features = {3}", processedPercentage, featureFinder.Comparer.GetMzEnd(binNum), elapsed, container.NumberOfFeatures); } } Console.WriteLine(@"Complete MS1 feature extraction."); Console.WriteLine(@" - Elapsed time = {0:0.000} sec", (stopwatch.ElapsedMilliseconds) / 1000.0d); Console.WriteLine(@" - Number of extracted features = {0}", container.NumberOfFeatures); Console.WriteLine(@"Start selecting mutually independent features from feature network graph"); stopwatch.Restart(); // write result files var tsvWriter = new StreamWriter(ms1FeaturesFilePath); tsvWriter.WriteLine(GetHeaderString(Parameters.ScoreReport)); StreamWriter csvWriter = null; if (Parameters.CsvOutput) { csvWriter = new StreamWriter(outCsvFilePath); csvWriter.WriteLine("scan_num,charge,abundance,mz,fit,monoisotopic_mw,FeatureID"); } var filteredFeatures = container.GetFilteredFeatures(featureFinder); var featureId = 0; foreach (var feature in filteredFeatures) { featureId++; tsvWriter.WriteLine("{0}\t{1}", featureId, GetString(feature, Parameters.ScoreReport)); var mostAbuIdx = feature.TheoreticalEnvelope.IndexOrderByRanking[0]; if (csvWriter != null) { foreach (var envelope in feature.EnumerateEnvelopes()) { //var mostAbuIsotopeInternalIndex = cluster.IsotopeList.SortedIndexByIntensity[0]; var mostAbuPeak = envelope.Peaks[mostAbuIdx]; if (mostAbuPeak == null || !mostAbuPeak.Active) continue; var fitscore = 1.0 - feature.BestCorrelationScore; csvWriter.WriteLine(string.Format("{0},{1},{2},{3},{4},{5},{6}", envelope.ScanNum, envelope.Charge, envelope.Abundance, mostAbuPeak.Mz, fitscore, envelope.MonoMass, featureId)); } } } tsvWriter.Close(); Console.WriteLine(@"Complete feature filtration"); Console.WriteLine(@" - Elapsed time = {0:0.000} sec", (stopwatch.ElapsedMilliseconds) / 1000.0d); Console.WriteLine(@" - Number of filtered features = {0}", featureId); Console.WriteLine(@" - ProMex output: {0}", ms1FeaturesFilePath); if (csvWriter != null) { csvWriter.Close(); Console.WriteLine(@" - ProMex output in ICR2LS format: {0}", outCsvFilePath); } if (Parameters.FeatureMapImage) { CreateFeatureMapImage(run, ms1FeaturesFilePath, pngFilePath); } return 0; }
public void TestLcMsFeatureXic() { var methodName = MethodBase.GetCurrentMethod().Name; TestUtils.ShowStarting(methodName); const string rawFile = @"\\proto-11\MSXML_Cache\PBF_Gen_1_193\2015_1\CPTAC_Intact_rep2_15Jan15_Bane_C2-14-08-02RZ.pbf"; //const string rawFile = @"D:\MassSpecFiles\training\raw\QC_Shew_Intact_26Sep14_Bane_C2Column3.pbf"; if (!File.Exists(rawFile)) { Assert.Ignore(@"Skipping test {0} since file not found: {1}", methodName, rawFile); } var run = PbfLcMsRun.GetLcMsRun(rawFile); var scorer = new LcMsFeatureLikelihood(); var featureFinder = new LcMsPeakMatrix(run, scorer); var feature = featureFinder.GetLcMsPeakCluster(2388.278, 4, 3774, 3907); //feature = featureFinder.GetLcMsPeakCluster(8151.3706, 7, 13, 4201, 4266); //feature = featureFinder.GetLcMsPeakCluster(8151.41789, 7, 13, 2861, 2941); var ms1ScanToIndex = run.GetMs1ScanNumToIndex(); var minCol = ms1ScanToIndex[feature.MinScanNum]; var maxCol = ms1ScanToIndex[feature.MaxScanNum]; //var minRow = feature.MinCharge - LcMsPeakMatrix.MinScanCharge; //var maxRow = feature.MaxCharge - LcMsPeakMatrix.MinScanCharge; Console.WriteLine("---------------------------------------------------------------"); for (var i = 0; i < feature.Envelopes.Length; i++) { for (var j = 0; j < feature.Envelopes[i].Length; j++) { Console.Write(feature.Envelopes[i][j] != null ? feature.Envelopes[i][j].PearsonCorrelation : 0); Console.Write("\t"); } Console.Write("\n"); } Console.WriteLine("---------------------------------------------------------------"); for (var i = 0; i < feature.Envelopes.Length; i++) { for (var j = 0; j < feature.Envelopes[i].Length; j++) { Console.Write(feature.Envelopes[i][j] != null ? feature.Envelopes[i][j].BhattacharyyaDistance : 0); Console.Write("\t"); } Console.Write("\n"); } Console.WriteLine("---------------------------------------------------------------"); for (var i = 0; i < feature.Envelopes.Length; i++) { for (var j = 0; j < feature.Envelopes[i].Length; j++) { Console.Write(feature.Envelopes[i][j] != null ? feature.Envelopes[i][j].Abundance : 0); Console.Write("\t"); } Console.Write("\n"); } }
public void TestLcMsFeatureFinder() { var methodName = MethodBase.GetCurrentMethod().Name; TestUtils.ShowStarting(methodName); const string rawFile = @"D:\MassSpecFiles\training\raw\QC_Shew_Intact_26Sep14_Bane_C2Column3.pbf"; //const string rawFile = @"D:\MassSpecFiles\CompRef\CPTAC_Intact_CR_Pool_2_25Jun15_Bane_15-02-02RZ.pbf"; //const string rawFile = @"D:\MassSpecFiles\IMER\Dey_IMERblast_01_08May14_Alder_14-01-33.pbf"; //const string rawFile = @"\\proto-11\MSXML_Cache\PBF_Gen_1_193\2015_3\MZ20150729FG_WT1.pbf"; if (!File.Exists(rawFile)) { Assert.Ignore(@"Skipping test {0} since file not found: {1}", methodName, rawFile); } // var outTsvFilePath = MassSpecDataReaderFactory.ChangeExtension(rawFile, "ms1ft"); //var scoreDataPath = @"D:\MassSpecFiles\training"; var scorer = new LcMsFeatureLikelihood(); var stopwatch = Stopwatch.StartNew(); Console.WriteLine(@"Start loading MS1 data from {0}", rawFile); var run = PbfLcMsRun.GetLcMsRun(rawFile); var featureFinder = new LcMsPeakMatrix(run, scorer); Console.WriteLine(@"Complete loading MS1 data. Elapsed Time = {0:0.000} sec", (stopwatch.ElapsedMilliseconds)/1000.0d); var container = new LcMsFeatureContainer(featureFinder.Ms1Spectra, scorer, new LcMsFeatureMergeComparer(new Tolerance(10))); var minSearchMassBin = featureFinder.Comparer.GetBinNumber(11180.33677); var maxSearchMassBin = featureFinder.Comparer.GetBinNumber(11180.33677); double totalMassBin = maxSearchMassBin - minSearchMassBin + 1; Console.WriteLine(@"Start MS1 feature extraction."); stopwatch.Restart(); for (var binNum = minSearchMassBin; binNum <= maxSearchMassBin; binNum++) { var clusters = featureFinder.FindFeatures(binNum); container.Add(clusters); if (binNum > minSearchMassBin && (binNum - minSearchMassBin)%1000 == 0) { var elapsed = (stopwatch.ElapsedMilliseconds)/1000.0d; var processedBins = binNum - minSearchMassBin; var processedPercentage = ((double) processedBins/totalMassBin)*100; Console.WriteLine( @"Processing {0:0.0}% of mass bins ({1:0.0} Da); elapsed time = {2:0.000} sec; # of features = {3}", processedPercentage, featureFinder.Comparer.GetMzEnd(binNum), elapsed, container.NumberOfFeatures); } } Console.WriteLine(@"Complete MS1 feature extraction."); Console.WriteLine(@" - Elapsed time = {0:0.000} sec", (stopwatch.ElapsedMilliseconds)/1000.0d); Console.WriteLine(@" - Number of extracted features = {0}", container.NumberOfFeatures); // write result files Console.WriteLine(@"Start selecting mutually independent features from feature network graph"); stopwatch.Stop(); // Start to quantify accurate abundance stopwatch.Restart(); //var quantAnalyzer = new TargetMs1FeatureMatrix(run); //var oriResult = new List<Ms1FeatureCluster>(); //var quantResult = new List<Ms1Feature>(); var featureId = 0; var ms1ScanNums = run.GetMs1ScanVector(); //tsvWriter.WriteLine(GetHeaderString() + "\tQMinScanNum\tQMaxScanNum\tQMinCharge\tQMaxCharge\tQAbundance"); var filteredFeatures = container.GetFilteredFeatures(featureFinder); foreach (var feature in filteredFeatures) { Console.Write(featureId); Console.Write("\t"); Console.Write(feature.Mass); Console.Write("\t"); Console.Write(feature.MinScanNum); Console.Write("\t"); Console.Write(feature.MaxScanNum); Console.Write("\t"); Console.Write(feature.MinCharge); Console.Write("\t"); Console.Write(feature.MaxCharge); Console.Write("\t"); Console.Write(feature.RepresentativeScanNum); Console.Write("\t"); Console.Write(feature.RepresentativeMz); Console.Write("\t"); Console.Write(feature.RepresentativeCharge); Console.Write("\t"); //Console.Write(feature.BestSummedEnvelopeDistance); Console.Write("\t"); //Console.Write(feature.BestEnvelopeDistance); Console.Write("\t"); Console.Write(feature.BestDistanceScoreAcrossCharge[0]); Console.Write("\t"); Console.Write(feature.BestDistanceScoreAcrossCharge[1]); Console.Write("\t"); Console.Write(feature.BestCorrelationScoreAcrossCharge[0]); Console.Write("\t"); Console.Write(feature.BestCorrelationScoreAcrossCharge[1]); Console.Write("\t"); Console.Write(feature.BestIntensityScoreAcrossCharge[0]); Console.Write("\t"); Console.Write(feature.BestIntensityScoreAcrossCharge[1]); Console.Write("\t"); Console.Write(feature.AbundanceDistributionAcrossCharge[0]); Console.Write("\t"); Console.Write(feature.AbundanceDistributionAcrossCharge[1]); Console.Write("\t"); Console.Write(feature.XicCorrelationBetweenBestCharges[0]); Console.Write("\t"); Console.Write(feature.XicCorrelationBetweenBestCharges[1]); Console.Write("\t"); Console.Write(feature.Score); Console.Write("\n"); featureId++; } }
public void FeatureFind(List<ProteinSpectrumMatch> prsms, LcMsRun run, string outTsvFilePath) { var featureFinder = new LcMsPeakMatrix(run, new LcMsFeatureLikelihood()); // write result files var tsvWriter = new StreamWriter(outTsvFilePath); tsvWriter.WriteLine(LcMsFeatureFinderLauncher.GetHeaderString(false)); var featureId = 1; foreach (var match in prsms) { var minScan = run.GetPrevScanNum(match.ScanNum, 1); var maxScan = run.GetNextScanNum(match.ScanNum, 1); var feature = featureFinder.GetLcMsPeakCluster(match.Mass, match.Charge, minScan, maxScan); if (feature == null) continue; tsvWriter.WriteLine("{0}\t{1}", featureId, LcMsFeatureFinderLauncher.GetString(feature, false)); featureId++; } tsvWriter.Close(); }
public void ExtractLcMsFeaturesForTrainingSet() { var methodName = MethodBase.GetCurrentMethod().Name; TestUtils.ShowStarting(methodName); const string idFileFolder = @"D:\MassSpecFiles\training\FilteredIdResult"; if (!Directory.Exists(idFileFolder)) { Assert.Ignore(@"Skipping test {0} since folder not found: {1}", methodName, idFileFolder); } var tolerance = new Tolerance(10); var tolerance2 = new Tolerance(20); var id = 1; for(var d = 0; d < TrainSetFileLists.Length; d++) { var dataset = TrainSetFileLists[d]; var dataname = Path.GetFileNameWithoutExtension(dataset); var filtedIdResultFile = string.Format(@"{0}\{1}.trainset.tsv", idFileFolder, Path.GetFileNameWithoutExtension(dataset)); var featureResult = string.Format(@"{0}\{1}.ms1ft", idFileFolder, Path.GetFileNameWithoutExtension(dataset)); if (!File.Exists(dataset)) { Console.WriteLine(@"Warning: Skipping since file not found: {0}", dataset); continue; } if (!File.Exists(filtedIdResultFile)) { Console.WriteLine(@"Warning: Skipping since file not found: {0}", filtedIdResultFile); continue; } var run = PbfLcMsRun.GetLcMsRun(dataset); var targetStatWriter = new StreamWriter(string.Format(@"D:\MassSpecFiles\training\statistics\{0}.tsv", Path.GetFileNameWithoutExtension(dataset))); var decoyStatWriter = new StreamWriter(string.Format(@"D:\MassSpecFiles\training\statistics\{0}_decoy.tsv", Path.GetFileNameWithoutExtension(dataset))); var writer = new StreamWriter(featureResult); writer.Write("Ms2MinScan\tMs2MaxScan\tMs2MinCharge\tMs2MaxCharge\tMs2Mass\t"); writer.Write("Mass\tMinScan\tMaxScan\tMinCharge\tMaxCharge\tMinTime\tMaxTime\tElution\tGood\n"); var tsvParser = new TsvFileParser(filtedIdResultFile); var featureFinder = new LcMsPeakMatrix(run); for (var i = 0; i < tsvParser.NumData; i++) { var minScan = int.Parse(tsvParser.GetData("MinScan")[i]); var maxScan = int.Parse(tsvParser.GetData("MaxScan")[i]); var minCharge = int.Parse(tsvParser.GetData("MinCharge")[i]); var maxCharge = int.Parse(tsvParser.GetData("MaxCharge")[i]); var mass = double.Parse(tsvParser.GetData("Mass")[i]); writer.Write(minScan); writer.Write("\t"); writer.Write(maxScan); writer.Write("\t"); writer.Write(minCharge); writer.Write("\t"); writer.Write(maxCharge); writer.Write("\t"); writer.Write(mass); writer.Write("\t"); var binNum = featureFinder.Comparer.GetBinNumber(mass); var binMass = featureFinder.Comparer.GetMzAverage(binNum); var binNumList = (mass < binMass) ? new int[] { binNum, binNum - 1, binNum + 1 } : new int[] { binNum, binNum + 1, binNum - 1 }; LcMsPeakCluster refinedFeature = null; foreach (var bi in binNumList) { var tempList = new List<LcMsPeakCluster>(); var features = featureFinder.FindFeatures(bi); var massTh = (mass < 2000) ? tolerance2.GetToleranceAsTh(mass) : tolerance.GetToleranceAsTh(mass); foreach (var feature in features) { if (Math.Abs(mass - feature.Mass) < massTh) tempList.Add(feature); } //var nHits = 0; var highestAbu = 0d; //var scans = Enumerable.Range(minScan, maxScan - minScan + 1); foreach (var feature in tempList) { //var scans2 = Enumerable.Range(feature.MinScanNum, feature.MaxScanNum - feature.MinScanNum + 1); //var hitScans = scans.Intersect(scans2).Count(); if (feature.MinScanNum < 0.5*(minScan + maxScan) && 0.5*(minScan + maxScan) < feature.MaxScanNum) { if (feature.Abundance > highestAbu) { refinedFeature = feature; highestAbu = feature.Abundance; } } /*if (hitScans > 0) { refinedFeature = feature; nHits = hitScans; }*/ } if (refinedFeature != null) break; } if (refinedFeature != null) { writer.Write(refinedFeature.Mass); writer.Write("\t"); writer.Write(refinedFeature.MinScanNum); writer.Write("\t"); writer.Write(refinedFeature.MaxScanNum); writer.Write("\t"); writer.Write(refinedFeature.MinCharge); writer.Write("\t"); writer.Write(refinedFeature.MaxCharge); writer.Write("\t"); writer.Write(refinedFeature.MinElutionTime); writer.Write("\t"); writer.Write(refinedFeature.MaxElutionTime); writer.Write("\t"); writer.Write(refinedFeature.MaxElutionTime - refinedFeature.MinElutionTime); writer.Write("\t"); var good = (refinedFeature.MinScanNum <= minScan && refinedFeature.MaxScanNum >= maxScan); writer.Write(good ? 1 : 0); writer.Write("\n"); //writer.Write(0); writer.Write("\t"); //writer.Write(0); writer.Write("\n"); OutputEnvelopPeakStat(id, refinedFeature, targetStatWriter); var chargeRange = featureFinder.GetDetectableMinMaxCharge(refinedFeature.RepresentativeMass, run.MinMs1Mz, run.MaxMs1Mz); refinedFeature.UpdateWithDecoyScore(featureFinder.Ms1Spectra, chargeRange.Item1, chargeRange.Item2); OutputEnvelopPeakStat(id, refinedFeature, decoyStatWriter); id++; } else { writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\t"); writer.Write(0); writer.Write("\n"); } //var feature = featureFinder.FindLcMsPeakCluster(mass, (int) scan, (int) charge); } writer.Close(); targetStatWriter.Close(); decoyStatWriter.Close(); Console.WriteLine(dataname); } }
public void TestFeatureExampleForFigure() { var methodName = MethodBase.GetCurrentMethod().Name; TestUtils.ShowStarting(methodName); const string rawFile = @"\\proto-11\MSXML_Cache\PBF_Gen_1_193\2015_1\CPTAC_Intact_rep6_15Jan15_Bane_C2-14-08-02RZ.pbf"; //const string rawFile = @"D:\MassSpecFiles\training\raw\QC_Shew_Intact_26Sep14_Bane_C2Column3.pbf"; if (!File.Exists(rawFile)) { Assert.Ignore(@"Skipping test {0} since file not found: {1}", methodName, rawFile); } var run = PbfLcMsRun.GetLcMsRun(rawFile); var scorer = new LcMsFeatureLikelihood(); var featureFinder = new LcMsPeakMatrix(run, scorer); var feature = featureFinder.GetLcMsPeakCluster(28061.6177, 20, 34, 7624, 7736); var writer = new StreamWriter(@"D:\MassSpecFiles\CPTAC_rep10\example\peaks.txt"); var envelope = feature.TheoreticalEnvelope; foreach (var e in envelope.Isotopes) Console.WriteLine(e.Ratio); foreach (var env in feature.EnumerateEnvelopes()) { var corr = env.PearsonCorrelation; for(var i = 0; i < envelope.Size; i++) { var peak = env.Peaks[i]; if (peak == null) continue; writer.Write("{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6}\n", env.ScanNum, run.GetElutionTime(env.ScanNum), env.Charge, i, peak.Mz, peak.Intensity, corr); } } writer.Close(); }
public void TestMs1EvidenceScore() { var methodName = MethodBase.GetCurrentMethod().Name; TestUtils.ShowStarting(methodName); const string TestRawFile = @"\\protoapps\UserData\Jungkap\Lewy\Lewy_intact_01.pbf"; if (!File.Exists(TestRawFile)) { Assert.Ignore(@"Skipping test {0} since file not found: {1}", methodName, TestRawFile); } const string TestResultFile = @"\\protoapps\UserData\Jungkap\Lewy\Lewy_intact_01_IcTda.tsv"; if (!File.Exists(TestResultFile)) { Assert.Ignore(@"Skipping test {0} since file not found: {1}", methodName, TestResultFile); } var run = PbfLcMsRun.GetLcMsRun(TestRawFile); var tsvParser = new TsvFileParser(TestResultFile); var featureFinder = new LcMsPeakMatrix(run); for (var i = 0; i < tsvParser.NumData; i++) { var scan = int.Parse(tsvParser.GetData("Scan")[i]); var charge = int.Parse(tsvParser.GetData("Charge")[i]); var mass = double.Parse(tsvParser.GetData("Mass")[i]); var qvalue = double.Parse(tsvParser.GetData("QValue")[i]); //var targetFeature = new TargetFeature(mass, charge, scan); var score = featureFinder.GetMs1EvidenceScore(scan, mass, charge); Console.WriteLine("{0}\t{1}\t{2}\t{3}\t{4}", scan, mass, charge, qvalue, score); } }