public FitBasedLogLikelihoodRatioScorer(ProductSpectrum ms2Spec, Tolerance tolerance, int minCharge, int maxCharge)
{
_ms2Spec = ms2Spec;
_tolerance = tolerance;
_minCharge = minCharge;
_maxCharge = maxCharge;
_baseIonTypes = ms2Spec.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
}
public LikelihoodScorer(LikelihoodScoringModel model, ProductSpectrum ms2Spec, Tolerance tolerance, int minCharge, int maxCharge, bool massErrorScore = true)
: base(ms2Spec, tolerance, minCharge, maxCharge, 0.1)
{
//_model = model;
//var refIntensity = ms2Spec.Peaks.Max(p => p.Intensity) * 0.1;
//var medIntensity = ms2Spec.Peaks.Select(p => p.Intensity).Median();
//_refIntensity = Math.Min(medIntensity, refIntensity);
_refIntensity = GetRefIntensity(ms2Spec.Peaks);
_includeMassErrorScore = massErrorScore;
}
public void OutputStatistics(ProductSpectrum spectrum, Sequence sequence)
{
var baseIonTypes = spectrum.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCid : BaseIonTypesEtd;
var cleavages = sequence.GetInternalCleavages().ToArray();
var tolerance = new Tolerance(10);
var maxIntensity = spectrum.Peaks.Max(p => p.Intensity);
foreach (var c in cleavages)
{
foreach (var baseIonType in baseIonTypes)
{
var fragmentComposition = baseIonType.IsPrefix
? c.PrefixComposition + baseIonType.OffsetComposition
: c.SuffixComposition + baseIonType.OffsetComposition;
for (int charge = MinCharge; charge <= MaxCharge; charge++)
{
var ion = new Ion(fragmentComposition, charge);
var observedPeaks = spectrum.GetAllIsotopePeaks(ion, tolerance, RelativeIsotopeIntensityThreshold);
if (observedPeaks == null) continue;
var mostAbundantIsotopeIndex = ion.Composition.GetMostAbundantIsotopeZeroBasedIndex();
// representative peak intensity
var ionPeakIntensity = observedPeaks[mostAbundantIsotopeIndex].Intensity;
// calc. correlation
var isotopomerEnvelope = ion.Composition.GetIsotopomerEnvelopeRelativeIntensities();
var observedIntensities = new double[observedPeaks.Length];
for (var i = 0; i < observedPeaks.Length; i++)
{
var observedPeak = observedPeaks[i];
observedIntensities[i] = observedPeak != null ? (float)observedPeak.Intensity : 0.0;
}
var corrCoeff = FitScoreCalculator.GetPearsonCorrelation(isotopomerEnvelope, observedIntensities);
// mz error
var mostAbundantIsotopeMz = ion.GetIsotopeMz(mostAbundantIsotopeIndex);
var errorPpm = ((observedPeaks[mostAbundantIsotopeIndex].Mz - mostAbundantIsotopeMz)/
mostAbundantIsotopeMz)*1e6;
}
}
}
}
public double GetScoreTest(Sequence sequence, ProductSpectrum spectrum)
{
var score = 0d;
var tol = new Tolerance(10);
var matchCounter = new CorrMatchedPeakCounter(spectrum,tol,1,20);
var prefixCompArr = sequence.GetPrefixCompositions().ToArray();
foreach (var c in prefixCompArr)
{
if(c.Equals(Composition.Zero)) Console.WriteLine("found zero");
}
var suffixCompArr = sequence.GetSuffixCompositions().ToArray();
for (int i = 0; i < prefixCompArr.Length; i++)
{
score += matchCounter.GetFragmentScore(prefixCompArr[i], suffixCompArr[i]);
}
return score;
}
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;
}
public CompositeScorerBasedOnDeconvolutedSpectrum(DeconvolutedSpectrum deconvolutedSpectrum, ProductSpectrum spec, Tolerance productTolerance, IMassBinning comparer)
: base(deconvolutedSpectrum, productTolerance)
{
ReferencePeakIntensity = GetRefIntensity(spec.Peaks);
_comparer = comparer;
_massBinToPeakMap = new Dictionary<int, DeconvolutedPeak>();
foreach (var p in deconvolutedSpectrum.Peaks)
{
var mass = p.Mz;
var deltaMass = productTolerance.GetToleranceAsDa(mass, 1);
var minMass = mass - deltaMass;
var maxMass = mass + deltaMass;
var binNum = comparer.GetBinNumber(mass);
if (binNum < 0)
{
binNum = comparer.GetBinNumber(minMass);
if (binNum < 0) binNum = comparer.GetBinNumber(maxMass);
}
// filter out
if (binNum < 0) continue;
UpdateDeconvPeak(binNum, p as DeconvolutedPeak);
// going up
for (var nextBinNum = binNum + 1; nextBinNum < comparer.NumberOfBins; nextBinNum++)
{
var nextBinMass = comparer.GetMassStart(nextBinNum);
if (minMass < nextBinMass && nextBinMass < maxMass) UpdateDeconvPeak(nextBinNum, p as DeconvolutedPeak); //_ionMassChkBins[nextBinNum] = true;
else break;
}
// going down
for (var prevBinNum = binNum - 1; prevBinNum < comparer.NumberOfBins; prevBinNum--)
{
var prevBinMass = comparer.GetMassEnd(prevBinNum);
if (minMass < prevBinMass && prevBinMass < maxMass) UpdateDeconvPeak(prevBinNum, p as DeconvolutedPeak); //_ionMassChkBins[prevBinNum] = true;
else break;
}
}
}
public double ComputeScore(ProductSpectrum ms2Spec, Sequence sequence)
{
_baseIonTypes = ms2Spec.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
_sequence = sequence;
_ms2Spec = ms2Spec;
if (!_rankingInfo.TryGetValue(ms2Spec.ScanNum, out _peakRanking))
{
_peakRanking = ArrayUtil.GetRankings(ms2Spec.Peaks.Select(peak => peak.Intensity));
_rankingInfo.Add(ms2Spec.ScanNum, _peakRanking);
}
FindMatchedPeaks();
var score = GetRankSumScore();
return score;
}
public IcScores GetScores(ProductSpectrum spec, string seqStr, Composition composition, int charge, int ms2ScanNum)
{
if (spec == null) return null;
var scorer = new CompositeScorer(spec, Tolerance, MinProductCharge, Math.Min(MaxProductCharge, charge));
var seqGraph = SequenceGraph.CreateGraph(AminoAcidSet, AminoAcid.ProteinNTerm, seqStr, AminoAcid.ProteinCTerm);
if (seqGraph == null) return null;
var bestScore = double.NegativeInfinity;
Tuple<double, string> bestScoreAndModifications = null;
var protCompositions = seqGraph.GetSequenceCompositions();
for (var modIndex = 0; modIndex < protCompositions.Length; modIndex++)
{
seqGraph.SetSink(modIndex);
var protCompositionWithH2O = seqGraph.GetSinkSequenceCompositionWithH2O();
if (!protCompositionWithH2O.Equals(composition)) continue;
var curScoreAndModifications = seqGraph.GetFragmentScoreAndModifications(scorer);
var curScore = curScoreAndModifications.Item1;
if (!(curScore > bestScore)) continue;
bestScoreAndModifications = curScoreAndModifications;
bestScore = curScore;
}
if (bestScoreAndModifications == null) return null;
var modifications = bestScoreAndModifications.Item2;
var seqObj = Sequence.CreateSequence(seqStr, modifications, AminoAcidSet);
double score;
int nMatchedFragments;
GetCompositeScores(seqObj, charge, ms2ScanNum, out score, out nMatchedFragments);
return new IcScores(nMatchedFragments, score, modifications);
}
/// <summary>
/// Handle a single spectrum element and child nodes
/// Called by ReadSpectrumList (xml hierarchy)
/// </summary>
/// <param name="reader">XmlReader that is only valid for the scope of the single spectrum element</param>
/// <param name="includePeaks">Whether to read binary data arrays</param>
private Spectrum ReadSpectrum(XmlReader reader, bool includePeaks = true)
{
reader.MoveToContent();
string index = reader.GetAttribute("index");
//Console.WriteLine("Reading spectrum indexed by " + index);
// This is correct for Thermo files converted by msConvert, but need to implement for others as well
string spectrumId = reader.GetAttribute("id"); // Native ID in mzML_1.1.0; unique identifier in mzML_1.0.0, often same as nativeID
string nativeId = spectrumId;
if (_version == MzML_Version.mzML1_0_0)
{
nativeId = reader.GetAttribute("nativeID"); // Native ID in mzML_1.0.0
}
int scanNum = -1;
// If a random access reader, there is already a scan number stored, based on the order of the index. Use it instead.
if (_randomAccess)
{
scanNum = (int) (_spectrumOffsets.NativeToIdMap[nativeId]);
}
else
{
scanNum = (int)(_artificialScanNum++);
// Interpret the NativeID (if the format has an interpreter) and use it instead of the artificial number.
// TODO: Better handling than the artificial ID for other nativeIDs (ones currently not supported)
int num = 0;
if (NativeIdConversion.TryGetScanNumberInt(nativeId, out num))
{
scanNum = num;
}
}
int defaultArraySize = Convert.ToInt32(reader.GetAttribute("defaultArrayLength"));
reader.ReadStartElement("spectrum"); // Throws exception if we are not at the "spectrum" tag.
bool is_ms_ms = false;
int msLevel = 0;
bool centroided = false;
double tic = 0;
List<Precursor> precursors = new List<Precursor>();
List<ScanData> scans = new List<ScanData>();
List<BinaryDataArray> bdas = new List<BinaryDataArray>();
while (reader.ReadState == ReadState.Interactive)
{
// Handle exiting out properly at EndElement tags
if (reader.NodeType != XmlNodeType.Element)
{
reader.Read();
continue;
}
//////////////////////////////////////////////////////////////////////////////////////
///
/// MS1 Spectra: only need Spectrum data: scanNum, MSLevel, ElutionTime, mzArray, IntensityArray
///
/// MS2 Spectra: use ProductSpectrum; adds ActivationMethod and IsolationWindow
///
//////////////////////////////////////////////////////////////////////////////////////
switch (reader.Name)
{
case "referenceableParamGroupRef":
// Schema requirements: zero to many instances of this element
reader.Skip();
break;
case "cvParam":
// Schema requirements: zero to many instances of this element
/* MAY supply a *child* term of MS:1000465 (scan polarity) only once
* e.g.: MS:1000129 (negative scan)
* e.g.: MS:1000130 (positive scan)
* MUST supply a *child* term of MS:1000559 (spectrum type) only once
* e.g.: MS:1000322 (charge inversion mass spectrum)
* e.g.: MS:1000325 (constant neutral gain spectrum)
* e.g.: MS:1000326 (constant neutral loss spectrum)
* e.g.: MS:1000328 (e/2 mass spectrum)
* e.g.: MS:1000341 (precursor ion spectrum)
* e.g.: MS:1000579 (MS1 spectrum)
* e.g.: MS:1000580 (MSn spectrum)
* e.g.: MS:1000581 (CRM spectrum)
* e.g.: MS:1000582 (SIM spectrum)
* e.g.: MS:1000583 (SRM spectrum)
* e.g.: MS:1000620 (PDA spectrum)
* e.g.: MS:1000627 (selected ion current chromatogram)
* e.g.: MS:1000789 (enhanced multiply charged spectrum)
* e.g.: MS:1000790 (time-delayed fragmentation spectrum)
* et al.
* MUST supply term MS:1000525 (spectrum representation) or any of its children only once
* e.g.: MS:1000127 (centroid spectrum)
* e.g.: MS:1000128 (profile spectrum)
* MAY supply a *child* term of MS:1000499 (spectrum attribute) one or more times
* e.g.: MS:1000285 (total ion current)
* e.g.: MS:1000497 (zoom scan)
* e.g.: MS:1000504 (base peak m/z)
* e.g.: MS:1000505 (base peak intensity)
* e.g.: MS:1000511 (ms level)
* e.g.: MS:1000527 (highest observed m/z)
* e.g.: MS:1000528 (lowest observed m/z)
* e.g.: MS:1000618 (highest observed wavelength)
* e.g.: MS:1000619 (lowest observed wavelength)
* e.g.: MS:1000796 (spectrum title)
* et al.
*/
switch (reader.GetAttribute("accession"))
{
case "MS:1000127":
// name="centroid spectrum"
centroided = true;
break;
case "MS:1000128":
// name="profile spectrum"
centroided = false;
break;
case "MS:1000511":
// name="ms level"
msLevel = Convert.ToInt32(reader.GetAttribute("value"));
break;
case "MS:1000579":
// name="MS1 spectrum"
is_ms_ms = false;
break;
case "MS:1000580":
// name="MSn spectrum"
is_ms_ms = true;
break;
case "MS:1000285":
// name="total ion current"
tic = Convert.ToDouble(reader.GetAttribute("value"));
break;
}
reader.Read(); // Consume the cvParam element (no child nodes)
break;
case "userParam":
// Schema requirements: zero to many instances of this element
reader.Skip();
break;
case "spectrumDescription": // mzML_1.0.0 compatibility
// Schema requirements: one instance of this element
ReadSpectrumDescription(reader.ReadSubtree(), ref scans, ref precursors, out centroided);
reader.ReadEndElement(); // "spectrumDescription" must have child nodes
break;
case "scanList":
// Schema requirements: zero to one instances of this element
scans.AddRange(ReadScanList(reader.ReadSubtree()));
reader.ReadEndElement(); // "scanList" must have child nodes
break;
case "precursorList":
// Schema requirements: zero to one instances of this element
precursors.AddRange(ReadPrecursorList(reader.ReadSubtree()));
reader.ReadEndElement(); // "precursorList" must have child nodes
break;
case "productList":
// Schema requirements: zero to one instances of this element
reader.Skip();
break;
case "binaryDataArrayList":
// Schema requirements: zero to one instances of this element
if (includePeaks)
{
bdas.AddRange(ReadBinaryDataArrayList(reader.ReadSubtree(), defaultArraySize));
reader.ReadEndElement(); // "binaryDataArrayList" must have child nodes
}
else
{
reader.Skip();
}
break;
default:
reader.Skip();
break;
}
}
reader.Close();
// Process the spectrum data
ScanData scan = new ScanData();
Spectrum spectrum;
BinaryDataArray mzs = new BinaryDataArray();
BinaryDataArray intensities = new BinaryDataArray();
foreach (var bda in bdas)
{
if (bda.ArrayType == ArrayType.m_z_array)
{
mzs = bda;
}
else if (bda.ArrayType == ArrayType.intensity_array)
{
intensities = bda;
}
}
if (!centroided && includePeaks)
{
// Centroid spectrum
// ProteoWizard
var centroider = new Centroider(mzs.Data, intensities.Data);
double[] centroidedMzs, centroidedIntensities;
centroider.GetCentroidedData(out centroidedMzs, out centroidedIntensities);
mzs.Data = centroidedMzs;
intensities.Data = centroidedIntensities;
}
if (scans.Count == 1)
{
scan = scans[0];
}
else if (scans.Count > 1)
{
// TODO: Should do something else to appropriately handle combinations...
scan = scans[0];
}
if (is_ms_ms)
{
Precursor precursor = new Precursor();
if (precursors.Count == 1)
{
precursor = precursors[0];
}
else if (precursors.Count > 1)
{
// TODO: Should do something else to appropriately handle multiple precursors...
precursor = precursors[0];
}
SelectedIon ion = new SelectedIon();
if (precursor.Ions.Count == 1)
{
ion = precursor.Ions[0];
}
else if (precursor.Ions.Count > 1)
{
// TODO: Should do something else to appropriately handle multiple selected ions...
ion = precursor.Ions[0];
}
var pspectrum = new ProductSpectrum(mzs.Data, intensities.Data, scanNum);
pspectrum.ActivationMethod = precursor.Activation;
// Select mz value to use based on presence of a Thermo-specific user param.
// The user param has a slightly higher precision, if that matters.
double mz = scan.MonoisotopicMz == 0.0 ? ion.SelectedIonMz : scan.MonoisotopicMz;
pspectrum.IsolationWindow = new IsolationWindow(precursor.IsolationWindowTargetMz, precursor.IsolationWindowLowerOffset, precursor.IsolationWindowUpperOffset, mz, ion.Charge);
//pspectrum.IsolationWindow.OldCharge = ion.OldCharge;
//pspectrum.IsolationWindow.SelectedIonMz = ion.SelectedIonMz;
spectrum = pspectrum;
}
else
{
spectrum = new Spectrum(mzs.Data, intensities.Data, scanNum);
}
spectrum.MsLevel = msLevel;
spectrum.ElutionTime = scan.StartTime;
spectrum.NativeId = nativeId;
spectrum.TotalIonCurrent = tic;
return spectrum;
}
private void GetNodeStatistics(bool isDecoy, ProductSpectrum ms2Spec, Sequence sequence, StreamWriter writer) //, StreamWriter mzErrorWriter)
{
if (ms2Spec == null) return;
if (sequence == null) return;
//var refIntensity = ms2Spec.Peaks.Max(p => p.Intensity) * 0.01;
//refIntensity = Math.Min(ms2Spec.Peaks.Select(p => p.Intensity).Median(), refIntensity);
var BaseIonTypesCID = new[] { BaseIonType.B, BaseIonType.Y };
var BaseIonTypesETD = new[] { BaseIonType.C, BaseIonType.Z };
var tolerance = new Tolerance(15);
var minCharge = 1;
var maxCharge = 20;
var baseIonTypes = ms2Spec.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
var refIntensity = ms2Spec.Peaks.Max(p => p.Intensity);
var activationMethodFlag = ms2Spec.ActivationMethod == ActivationMethod.ETD ? 1 : 2;
var cleavages = sequence.GetInternalCleavages();
var prevPrefixFragMass = 0d;
var prevPrefixObsIonMass = 0d;
var prevPrefixObsIonCharge = 0;
var prevPrefixObsIonIntensity = 0d;
var prevSuffixFragMass = 0d;
var prevSuffixObsIonMass = 0d;
var prevSuffixObsIonCharge = 0;
var prevSuffixObsIonIntensity = 0d;
var nComplementaryFrags = 0;
var cleavageIndex = 0;
foreach (var c in cleavages)
{
var bothObs = true;
foreach (var baseIonType in baseIonTypes)
{
var peakType = baseIonType.IsPrefix ? 1 : 2; // unexplained
var fragmentComposition = baseIonType.IsPrefix
? c.PrefixComposition + baseIonType.OffsetComposition
: c.SuffixComposition + baseIonType.OffsetComposition;
var curFragMass = fragmentComposition.Mass;
var curObsIonMass = 0d;
var curObsIonCharge = 0;
var curObsIonDist = 1.0d;
var curObsIonCorr = 0d;
var curObsIonIntensity = 0d;
var ionMatch = false;
for (var charge = minCharge; charge <= maxCharge; charge++)
{
var ion = new Ion(fragmentComposition, charge);
var isotopePeaks = ms2Spec.GetAllIsotopePeaks(ion, tolerance, 0.1);
if (isotopePeaks == null) continue;
var distCorr = AbstractFragmentScorer.GetDistCorr(ion, isotopePeaks);
if (distCorr.Item2 < 0.7 && distCorr.Item1 > 0.07) continue;
var mostAbundantIsotopeIndex = ion.Composition.GetMostAbundantIsotopeZeroBasedIndex();
var mostAbuPeak = isotopePeaks[mostAbundantIsotopeIndex];
var summedIntensity = isotopePeaks.Where(p => p != null).Sum(p => p.Intensity);
var intScore = summedIntensity/refIntensity;
//var intScore = mostAbuPeak.Intensity / medIntensity;
//var intScore = summedIntensity / refIntensity;
if (ionMatch == false || curObsIonIntensity < intScore)
{
curObsIonMass = Ion.GetMonoIsotopicMass(mostAbuPeak.Mz, charge, mostAbundantIsotopeIndex);
curObsIonCharge = charge;
curObsIonCorr = distCorr.Item2;
curObsIonDist = distCorr.Item1;
curObsIonIntensity = intScore;
}
ionMatch = true;
}
if (!ionMatch)
{
bothObs = false;
continue;
}
writer.Write(activationMethodFlag);
writer.Write("\t");
writer.Write(peakType);
writer.Write("\t");
writer.Write("{0:0.000}", curFragMass);
writer.Write("\t");
writer.Write("{0}", curObsIonCharge);
writer.Write("\t");
writer.Write("{0:0.000}", curObsIonDist);
writer.Write("\t");
writer.Write("{0:0.000}", curObsIonCorr);
writer.Write("\t");
writer.Write("{0:0.000}", curObsIonIntensity);
writer.Write("\t");
writer.Write("{0:0.000}", (Math.Abs(curFragMass - curObsIonMass)/curFragMass)*1e6);
writer.Write("\n");
// mz error output
/*
if (baseIonType.IsPrefix && prevPrefixFragMass > 0 & prevPrefixObsIonMass > 0)
{
var aaMass = Math.Abs(prevPrefixFragMass - curFragMass);
var massError = Math.Abs(Math.Abs(prevPrefixObsIonMass - curObsIonMass) - aaMass);
var massErrorPpm = (massError / curObsIonMass) * 1e6;
mzErrorWriter.WriteLine("{0}\t{1:0.000}\t{2}", activationMethodFlag, massErrorPpm, Math.Abs(prevPrefixObsIonCharge - curObsIonCharge));
}
else if (prevSuffixFragMass > 0 & prevSuffixObsIonMass > 0)
{
var aaMass = Math.Abs(prevSuffixFragMass - curFragMass);
var massError = Math.Abs(Math.Abs(prevSuffixObsIonMass - curObsIonMass) - aaMass);
var massErrorPpm = (massError / curObsIonMass) * 1e6;
mzErrorWriter.WriteLine("{0}\t{1:0.000}\t{2}", activationMethodFlag, massErrorPpm, Math.Abs(prevSuffixObsIonCharge - curObsIonCharge));
}
*/
if (baseIonType.IsPrefix)
{
prevPrefixFragMass = curFragMass;
prevPrefixObsIonMass = curObsIonMass;
prevPrefixObsIonCharge = curObsIonCharge;
prevPrefixObsIonIntensity = curObsIonIntensity;
//Array.Copy(curObsIonMass, prevPrefixObsIonMass, curObsIonMass.Length);
}
else
{
prevSuffixFragMass = curFragMass;
prevSuffixObsIonMass = curObsIonMass;
prevSuffixObsIonCharge = curObsIonCharge;
prevSuffixObsIonIntensity = curObsIonIntensity;
//Array.Copy(curObsIonMass, prevSuffixObsIonMass, curObsIonMass.Length);
}
}
if (bothObs)
{
//pairWriter.Write("{0}\t{1}\t", prevPrefixObsIonIntensity, prevSuffixObsIonIntensity);
nComplementaryFrags++;
}
cleavageIndex++;
}
Console.WriteLine("{0}\t{1}", nComplementaryFrags, sequence.Count);
//if (!isDecoy) Console.WriteLine("{0}", totalExplainedAbundanceRatio);
}
private void GetMatchStatistics(ProductSpectrum ms2Spec, Sequence sequence, int parentIonCharge, StreamWriter writer)
{
if (ms2Spec == null) return;
if (sequence == null) return;
var BaseIonTypesCID = new[] { BaseIonType.B, BaseIonType.Y };
var BaseIonTypesETD = new[] { BaseIonType.C, BaseIonType.Z };
var tolerance = new Tolerance(12);
var MinProductCharge = 1;
var MaxProductCharge = Math.Min(parentIonCharge+2, 20);
var baseIonTypes = ms2Spec.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
var refIntensity = CompositeScorer.GetRefIntensity(ms2Spec.Peaks);
var activationMethodFlag = ms2Spec.ActivationMethod == ActivationMethod.ETD ? 2 : 1;
var cleavages = sequence.GetInternalCleavages();
var nComplementaryFrags = 0;
var prefixStat = new FragmentStat();
var suffixStat = new FragmentStat();
var minMz = ms2Spec.Peaks.First().Mz;
var maxMz = ms2Spec.Peaks.Last().Mz;
var cleavageIndex = 0;
var preFixIonCheck = new bool[sequence.Count + 1];
var sufFixIonCheck = new bool[sequence.Count + 1];
foreach (var c in cleavages)
{
var prefixHit = false;
var suffixHit = false;
foreach (var baseIonType in baseIonTypes)
{
var stat = baseIonType.IsPrefix ? prefixStat : suffixStat;
var fragmentComposition = baseIonType.IsPrefix
? c.PrefixComposition + baseIonType.OffsetComposition
: c.SuffixComposition + baseIonType.OffsetComposition;
if (fragmentComposition.Mass < ms2Spec.Peaks[0].Mz) continue;
var curFragMass = fragmentComposition.Mass;
/*var curObsIonCharge = 0;
var curObsIonDist = 1.0d;
var curObsIonCorr = 0d;
var curObsIonIntensity = 0d;
var curObsIonMassError = 0d;*/
var mostAbundantIsotopeIndex = fragmentComposition.GetMostAbundantIsotopeZeroBasedIndex();
var fragmentIonMostAbuMass = fragmentComposition.Mass + Constants.C13MinusC12 * mostAbundantIsotopeIndex;
var maxCharge = (int)Math.Floor(fragmentIonMostAbuMass / (minMz - Constants.Proton));
var minCharge = (int)Math.Ceiling(fragmentIonMostAbuMass / (maxMz - Constants.Proton));
if (maxCharge < 1 || maxCharge > MaxProductCharge) maxCharge = MaxProductCharge;
if (minCharge < 1 || minCharge < MinProductCharge) minCharge = MinProductCharge;
//var ionMatch = false;
for (var charge = minCharge; charge <= maxCharge; charge++)
{
var ion = new Ion(fragmentComposition, charge);
var isotopePeaks = ms2Spec.GetAllIsotopePeaks(ion, tolerance, 0.1);
if (isotopePeaks == null) continue;
var distCorr = CompositeScorer.GetDistCorr(ion, isotopePeaks);
if (distCorr.Item2 < 0.7 && distCorr.Item1 > 0.03) continue;
var mostAbuPeak = isotopePeaks[mostAbundantIsotopeIndex];
var intScore = mostAbuPeak.Intensity / refIntensity;
/*
if (ionMatch == false || curObsIonIntensity < intScore)
{
curObsIonCharge = charge;
curObsIonCorr = distCorr.Item2;
curObsIonDist = distCorr.Item1;
curObsIonIntensity = intScore;
var mostAbuPeakMz = Ion.GetIsotopeMz(curFragMass, charge, mostAbundantIsotopeIndex);
curObsIonMassError = (Math.Abs(mostAbuPeak.Mz - mostAbuPeakMz) / mostAbuPeakMz) * 1e6;
//var curObsIonMass = Ion.GetMonoIsotopicMass(mostAbuPeak.Mz, charge, mostAbundantIsotopeIndex);
//curObsIonMassError = (Math.Abs(curFragMass - curObsIonMass) / curFragMass) * 1e6;
}
ionMatch = true;
*/
var mostAbuPeakMz = Ion.GetIsotopeMz(curFragMass, charge, mostAbundantIsotopeIndex);
var curObsIonMassError = (Math.Abs(mostAbuPeak.Mz - mostAbuPeakMz) / mostAbuPeakMz) * 1e6;
stat.Count++;
stat.Intensity += Math.Min(intScore, 1.0);
stat.Corr += distCorr.Item2;
stat.Dist += distCorr.Item1;
stat.MassError += curObsIonMassError;
if (baseIonType.IsPrefix) prefixHit = true;
else suffixHit = true;
}
//if (!ionMatch) continue;
}
if (prefixHit) preFixIonCheck[cleavageIndex] = true;
if (suffixHit) sufFixIonCheck[cleavageIndex] = true;
if (prefixHit && suffixHit)
{
nComplementaryFrags++;
}
cleavageIndex++;
}
var preContCount = 0;
var sufContCount = 0;
for (var i = 0; i < preFixIonCheck.Length - 1; i++)
{
if (preFixIonCheck[i] && preFixIonCheck[i + 1]) preContCount++;
if (sufFixIonCheck[i] && sufFixIonCheck[i + 1]) sufContCount++;
}
writer.Write(activationMethodFlag);
writer.Write("\t");
writer.Write(sequence.Composition.Mass);
writer.Write("\t");
writer.Write(sequence.Count);
writer.Write("\t");
writer.Write(nComplementaryFrags);
writer.Write("\t");
writer.Write("{0}\t{1}\t{2}\t{3}\t{4}\t{5}", prefixStat.Count, preContCount, prefixStat.Intensity, prefixStat.Corr, prefixStat.Dist, prefixStat.MassError);
writer.Write("\t");
writer.Write("{0}\t{1}\t{2}\t{3}\t{4}\t{5}", suffixStat.Count, sufContCount, suffixStat.Intensity, suffixStat.Corr, suffixStat.Dist, suffixStat.MassError);
writer.Write("\n");
}
public CorrMatchedPeakCounter(ProductSpectrum ms2Spec, Tolerance tolerance, int minCharge, int maxCharge, double corrScoreThreshold = 0.7):
base(ms2Spec, tolerance, minCharge, maxCharge, 0.1)
{
_corrScoreThreshold = corrScoreThreshold;
}
/// <summary>
/// Reads the mass spectrum with the specified scanNum from the raw file
/// </summary>
/// <param name="scanNum">scan number</param>
/// <param name="includePeaks">whether to include peak data</param>
/// <returns>mass spectrum</returns>
public Spectrum ReadMassSpectrum(int scanNum, bool includePeaks = true)
{
var scanInfo = GetScanInfo(scanNum);
// default empty arrays, if peak data not requested.
double[] mzArr = new double[]{};
double[] intensityArr = new double[]{};
if (includePeaks)
{
_msfileReader.GetScanData(scanNum, out mzArr, out intensityArr, 0, true);
}
var elutionTime = RtFromScanNum(scanNum);
var nativeId = "controllerType=0 controllerNumber=1 scan=" + scanNum;
// Call scanInfo.MSLevel in order to update dictionary _msLevel
var msLevel = ReadMsLevel(scanNum);
if (msLevel == 1)
return new Spectrum(mzArr, intensityArr, scanNum)
{
ElutionTime = elutionTime,
TotalIonCurrent = scanInfo.TotalIonCurrent,
NativeId = nativeId,
};
var isolationWindow = ReadPrecursorInfo(scanNum);
var productSpec = new ProductSpectrum(mzArr, intensityArr, scanNum)
{
MsLevel = scanInfo.MSLevel,
ElutionTime = elutionTime,
TotalIonCurrent = scanInfo.TotalIonCurrent,
NativeId = nativeId,
ActivationMethod = GetActivationMethod(scanNum),
IsolationWindow = isolationWindow
};
return productSpec;
}
public IList<SpectrumMatch> Read()
{
var specMatches = new List<SpectrumMatch>();
var file = File.ReadLines(_fileName);
var mgfState = MgfState.Label;
var sequence = "";
int scanNum = 0, charge = 0;
var peaks = new List<Peak>();
var peptideSet = new HashSet<string>();
foreach (var line in file)
{
switch (mgfState)
{
case MgfState.Label:
if (line == "BEGIN IONS") mgfState = MgfState.Parameter;
else throw new FormatException("Invalid MGF file.");
break;
case MgfState.Parameter:
var parameter = line.Split('=');
if (parameter.Length < 2) throw new FormatException("Invalid line in MGF file: " + line);
if (parameter[0] == "SEQ") sequence = parameter[1];
else if (parameter[0] == "SCANS") scanNum = Convert.ToInt32(parameter[1]);
else if (parameter[0] == "CHARGE")
{
var chargeStr = parameter[1].Substring(0, parameter[1].Length - 1);
charge = Convert.ToInt32(chargeStr);
mgfState = MgfState.Peak;
if (sequence == "" || scanNum == 0 || charge == 0)
throw new FormatException("Incomplete spectrum entry.");
}
break;
case MgfState.Peak:
if (line == "END IONS")
{
if (peaks.Count == 0) throw new FormatException("Empty peak list.");
mgfState = MgfState.Label;
if (peptideSet.Contains(sequence))
{
sequence = "";
scanNum = 0;
charge = 0;
peaks.Clear();
continue;
}
peptideSet.Add(sequence);
var spectrum = new ProductSpectrum(peaks, scanNum) { MsLevel = 2 };
var sequenceReader = new MgfSequenceReader();
var seq = sequenceReader.GetSequence(sequence);
var specMatch = new SpectrumMatch(seq, spectrum, scanNum, charge, _decoy);
sequence = "";
scanNum = 0;
charge = 0;
specMatches.Add(specMatch);
peaks.Clear();
}
else
{
var parts = line.Split('\t');
if (parts.Length < 2) throw new FormatException("Invalid line in MGF file: " + line);
var mz = Convert.ToDouble(parts[0]);
var intensity = Convert.ToDouble(parts[1]);
peaks.Add(new Peak(mz, intensity));
}
break;
}
}
return specMatches;
}
private IEnumerable<TagSequenceMatch> GetMatches(IEnumerable<SequenceTag> tags, ProductSpectrum spec, IScorer scorer)
{
// Match tags against the database
var proteinsToTags = GetProteinToMatchedTagsMap(tags, _searchableDb, _aaSet, _tolerance, _tolerance);
//var tagSequenceMatchList = new List<TagSequenceMatch>();
// Extend matches
foreach (var entry in proteinsToTags)
{
var proteinName = entry.Key;
var matchedTagSet = entry.Value;
var proteinSequence = matchedTagSet.Sequence;
var tagFinder = new TagMatchFinder(spec, scorer, _featureFinder, proteinSequence, _tolerance, _aaSet, _maxSequenceMass);
foreach (var matchedTag in matchedTagSet.Tags)
{
if (matchedTag.Length < _minMatchedTagLength) continue;
if (matchedTag.NTermFlankingMass == null && matchedTag.CTermFlankingMass == null) continue;
var matches = tagFinder.FindMatches(matchedTag).ToArray();
//var prevScore = double.NegativeInfinity;
//foreach (var match in matches.OrderByDescending(m => m.Score))
foreach(var match in matches)
{
var sequence = proteinSequence.Substring(match.StartIndex, match.EndIndex - match.StartIndex);
//re-scoring
var sequenceObj = Sequence.CreateSequence(sequence, match.ModificationText, _aaSet);
match.Score = sequenceObj.GetInternalCleavages().Sum(c => scorer.GetFragmentScore(c.PrefixComposition, c.SuffixComposition));
//var numMatches = matchedTag.Length * 2 + match.NTermScore + match.CTermScore;
//var score = match.NTermScore + match.CTermScore;
//score += (matchedTag.NumReliableNTermFlankingMasses > 0)
// ? matchedTag.Length*CompositeScorer.ScoreParam.Prefix.ConsecutiveMatch
//: matchedTag.Length*CompositeScorer.ScoreParam.Suffix.ConsecutiveMatch;
// Poisson p-value score
//var n = (match.EndIndex - match.StartIndex - 1)*2;
//var lambda = numMatches / n;
//var pValue = 1 - Poisson.CDF(lambda, numMatches);
//var pScore = (pValue > 0) ? - Math.Log(pValue, 2) : 50.0;
//if (numMatches < 5) break;
//if (prevScore - numMatches > 2) break;
//prevScore = numMatches;
var pre = match.StartIndex == 0 ? '-' : proteinSequence[match.StartIndex - 1]; // startIndex is inclusive
var post = match.EndIndex >= proteinSequence.Length ? '-' : proteinSequence[match.EndIndex]; // endIndex is Exclusive
yield return new TagSequenceMatch(sequence, proteinName, match, pre, post);
//tagSequenceMatchList.Add(new TagSequenceMatch(sequence, proteinName, match, pre, post));
}
}
}
//return tagSequenceMatchList;
}
private FlankingMassMatch GetBestMatchInTheGraph(ShiftedSequenceGraph seqGraph, ProductSpectrum spec, double? featureMass)
{
FlankingMassMatch match = null;
var bestScore = double.NegativeInfinity;
var protCompositions = seqGraph.GetSequenceCompositions();
for (var modIndex = 0; modIndex < protCompositions.Length; modIndex++)
{
seqGraph.SetSink(modIndex);
var protCompositionWithH2O = seqGraph.GetSinkSequenceCompositionWithH2O();
var sequenceMass = protCompositionWithH2O.Mass;
if (featureMass != null && !_tolerance.IsWithin(sequenceMass, (double)featureMass)) continue;
var charge =
(int)
Math.Round(sequenceMass /
(spec.IsolationWindow.IsolationWindowTargetMz - Constants.Proton));
var mostAbundantIsotopeMz = Ion.GetIsotopeMz(sequenceMass, charge,
Averagine.GetIsotopomerEnvelope(sequenceMass).MostAbundantIsotopeIndex);
if (!spec.IsolationWindow.Contains(mostAbundantIsotopeMz)) continue;
//var feature = new TargetFeature(sequenceMass, charge, spec.ScanNum);
if (_featureFinder != null)
{
var ms1Corr = _featureFinder.GetMs1EvidenceScore(spec.ScanNum, sequenceMass, charge);
if (ms1Corr < Ms1CorrThreshold) continue;
}
var curScoreAndModifications = seqGraph.GetScoreAndModifications(_ms2Scorer);
var curScore = curScoreAndModifications.Item1;
// var curScore = seqGraph.GetFragmentScore(_ms2Scorer);
if (curScore > bestScore)
{
match = new FlankingMassMatch(curScore,
sequenceMass - Composition.H2O.Mass - seqGraph.ShiftMass, charge, curScoreAndModifications.Item2);
//match = new FlankingMassMatch(curScore,
// sequenceMass - Composition.H2O.Mass - seqGraph.ShiftMass, charge, new ModificationInstance[0]);
bestScore = curScore;
}
}
return match;
}
internal DeconvScorer(ProductSpectrum deconvolutedSpectrum, Tolerance productTolerance)
{
if (deconvolutedSpectrum.ActivationMethod != ActivationMethod.ETD)
{
_prefixOffsetMass = BaseIonType.B.OffsetComposition.Mass;
_suffixOffsetMass = BaseIonType.Y.OffsetComposition.Mass;
}
else
{
_prefixOffsetMass = BaseIonType.C.OffsetComposition.Mass;
_suffixOffsetMass = BaseIonType.Z.OffsetComposition.Mass;
}
_ionMassBins = new HashSet<int>();
foreach (var p in deconvolutedSpectrum.Peaks)
{
var mass = p.Mz;
var deltaMass = productTolerance.GetToleranceAsDa(mass, 1);
var minMass = mass - deltaMass;
var maxMass = mass + deltaMass;
var minBinNum = GetBinNumber(minMass);
var maxBinNum = GetBinNumber(maxMass);
for (var binNum = minBinNum; binNum <= maxBinNum; binNum++)
{
_ionMassBins.Add(binNum);
}
}
}
public Tuple<int, int, int, int, string, string> GetLongestSequence(ProductSpectrum spectrum, Sequence sequence)
{
_spectrum = spectrum;
_sequence = sequence;
_baseIonTypes = _spectrum.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCid : BaseIonTypesEtd;
var cleavages = _sequence.GetInternalCleavages().ToArray();
var prefixValueArr = new int[cleavages.Length];
var suffixValueArr = new int[cleavages.Length];
var prefixPeakArr = new Peak[cleavages.Length];
var suffixPeakArr = new Peak[cleavages.Length];
//var peakList = new double[_spectrum.Peaks.Length];
int cleavageIndex = 0;
/*
for (int i = 0; i < peakList.Length; i++)
{
peakList[i] = _spectrum.Peaks[i].Intensity;
}*/
//var rankings = ArrayUtil.GetRankings(peakList);
foreach (var c in cleavages)
{
foreach (var baseIonType in _baseIonTypes)
{
var fragmentComposition = baseIonType.IsPrefix
? c.PrefixComposition + baseIonType.OffsetComposition
: c.SuffixComposition + baseIonType.OffsetComposition;
for (var charge = _minCharge; charge <= _maxCharge; charge++)
{
var ion = new Ion(fragmentComposition, charge);
if (_spectrum.GetCorrScore(ion, _tolerance, RelativeIsotopeIntensityThreshold) < .7) continue;
if (baseIonType.IsPrefix) prefixValueArr[cleavageIndex] = 1;
else suffixValueArr[cleavageIndex] = 1;
}
}
cleavageIndex++;
}
var prefixSequenceArr = new int[_sequence.Count];
var suffixSequenceArr = new int[_sequence.Count];
prefixSequenceArr[0] = prefixValueArr[0];
suffixSequenceArr[suffixSequenceArr.Length - 1] = suffixValueArr[suffixValueArr.Length - 1];
for (int i = 1; i < prefixValueArr.Length; i++)
{
if (prefixValueArr[i] == 1 && prefixValueArr[i - 1] == 1)
{
if (_sequence[i] is ModifiedAminoAcid) continue;
prefixSequenceArr[i] = 1;
}
}
for (int i = suffixValueArr.Length - 2; i >= 0; i--)
{
if (suffixValueArr[i] == 1 && suffixValueArr[i + 1] == 1)
{
if (_sequence[i + 1] is ModifiedAminoAcid) continue;
suffixSequenceArr[i + 1] = 1;
}
}
var prefixSubString = FindLongestSubstring(prefixSequenceArr);
var prefixStartIndex = -1;
var prefixEndIndex = -1;
//var prefixSequencePeaks = new List<Peak>();
//var prefixPval = -1.0;
var prefixSequence = "";
if (prefixSubString != "")
{
var prefixIndex = string.Concat(prefixSequenceArr);
prefixStartIndex = prefixIndex.IndexOf(prefixSubString) + 1;
prefixEndIndex = (prefixStartIndex == 1) ? 1 : prefixStartIndex + prefixSubString.Length - 1;
//prefixSequencePeaks = GetPrefixSequencePeaks(prefixPeakArr, prefixStartIndex, prefixEndIndex);
//var prefixRankSum = GetSequenceRankSum(prefixSequencePeaks, rankings, peakList);
//prefixPval = FitScoreCalculator.GetRankSumPvalue(peakList.Length, prefixSequencePeaks.Count, prefixRankSum);
prefixSequence = GetStringSubSequence(_sequence, prefixStartIndex, prefixEndIndex);
}
var suffixSubString = FindLongestSubstring(suffixSequenceArr);
var suffixStartIndex = -1;
var suffixEndIndex = -1;
//var suffixSequencePeaks = new List<Peak>();
//var suffixPval = -1.0;
var suffixSequence = "";
if (suffixSubString != "")
{
var suffixIndex = string.Concat(suffixSequenceArr);
suffixStartIndex = suffixIndex.IndexOf(suffixSubString) + 1;
suffixEndIndex = (suffixStartIndex == 1) ? 1 : suffixStartIndex + suffixSubString.Length - 1;
//suffixSequencePeaks = GetSuffixSequencePeaks(suffixPeakArr, suffixStartIndex, suffixEndIndex);
//var suffixRankSum = GetSequenceRankSum(suffixSequencePeaks, rankings, peakList);
//suffixPval = FitScoreCalculator.GetRankSumPvalue(peakList.Length, suffixSequencePeaks.Count, suffixRankSum);
suffixSequence = GetStringSubSequence(_sequence, suffixStartIndex, suffixEndIndex);
}
return new Tuple<int, int, int, int, string, string>(prefixStartIndex, prefixEndIndex, suffixStartIndex, suffixEndIndex, prefixSequence, suffixSequence);
}
public MatchedPeakCounter(ProductSpectrum ms2Spec, Tolerance tolerance, int minCharge, int maxCharge)
: base(ms2Spec, tolerance, minCharge, maxCharge)
{
RelativeIsotopeIntensityThreshold = 0.7;
}
