public static Spectrum GetDeconvolutedSpectrum(Spectrum spec, int minCharge, int maxCharge, Tolerance tolerance, double corrThreshold,
int isotopeOffsetTolerance, double filteringWindowSize = 1.1)
{
var deconvolutedPeaks = Deconvoluter.GetDeconvolutedPeaks(spec, minCharge, maxCharge, isotopeOffsetTolerance, filteringWindowSize, tolerance, corrThreshold);
var peakList = new List<Peak>();
var binHash = new HashSet<int>();
foreach (var deconvolutedPeak in deconvolutedPeaks)
{
var mass = deconvolutedPeak.Mass;
var binNum = GetBinNumber(mass);
if (!binHash.Add(binNum)) continue;
peakList.Add(new Peak(mass, deconvolutedPeak.Intensity));
}
var productSpec = spec as ProductSpectrum;
if (productSpec != null)
{
return new ProductSpectrum(peakList, spec.ScanNum)
{
MsLevel = spec.MsLevel,
ActivationMethod = productSpec.ActivationMethod,
IsolationWindow = productSpec.IsolationWindow
};
}
return new Spectrum(peakList, spec.ScanNum);
}
public static Spectrum GetFilteredSpectrum(Spectrum spectrum, double windowWidth=100.0, int retentionCount=6)
{
windowWidth = windowWidth/2;
var peaks = spectrum.Peaks;
var filteredPeaks = new List<Peak>();
for (var peakIndex = 0; peakIndex < peaks.Count(); peakIndex++)
{
var rank = 1;
var peak = peaks[peakIndex];
var prevIndex = peakIndex - 1;
while (prevIndex >= 0)
{
var prevPeak = peaks[prevIndex];
if ((peak.Mz - prevPeak.Mz) > windowWidth) break;
if (prevPeak.Intensity > peak.Intensity) rank++;
prevIndex--;
}
var nextIndex = peakIndex + 1;
while (nextIndex < peaks.Length)
{
var nextPeak = peaks[nextIndex];
if ((nextPeak.Mz - peak.Mz) > windowWidth) break;
if (nextPeak.Intensity > peak.Intensity) rank++;
nextIndex++;
}
if (rank <= retentionCount) filteredPeaks.Add(peak);
}
var filteredSpectrum = new Spectrum(filteredPeaks.ToArray(), spectrum.ScanNum);
return filteredSpectrum;
}
internal RankedSpectrum(Spectrum spec)
{
Peaks = spec.Peaks.OrderByDescending(p => p.Intensity)
.Select((p, i) => new RankedPeak(p.Mz, p.Intensity, i))
.OrderBy(p => p.Mz)
.ToArray();
}
public SpectrumMatch Filter(SpectrumMatch specMatch)
{
var charge = specMatch.PrecursorCharge;
var peaks = specMatch.Spectrum.Peaks;
var indexes = new List<int>();
for (int i = 0; i < _offsets.Charge; i++)
{
var ion = _precursorIonTypes[i].GetIon(specMatch.PeptideComposition);
var mz = ion.GetMonoIsotopicMz();
var offsets = _offsets.GetChargeOffsets(i + 1);
foreach (var offset in offsets)
{
var offsetMz = mz + offset;
var peakIndex = specMatch.Spectrum.FindPeakIndex(offsetMz, _tolerance);
if (peakIndex > 0) indexes.Add(peakIndex);
}
}
indexes = indexes.Distinct().ToList();
var filteredPeaks = peaks.Where((t, i) => !indexes.Contains(i)).ToList();
var spectrum = new Spectrum(filteredPeaks, specMatch.Spectrum.ScanNum);
var filteredSpecMatch = new SpectrumMatch(specMatch.Sequence, spectrum, specMatch.ScanNum, charge, specMatch.Decoy);
return filteredSpecMatch;
}
public static List<DeconvolutedPeak> GetDeconvolutedPeaks(
Spectrum spec, int minCharge, int maxCharge,
int isotopeOffsetTolerance, double filteringWindowSize,
Tolerance tolerance, double corrScoreThreshold)
{
return GetDeconvolutedPeaks(spec.Peaks, minCharge, maxCharge, isotopeOffsetTolerance, filteringWindowSize,
tolerance, corrScoreThreshold);
}
public ScoredSpectrum(Spectrum spec, RankScore scorer, int charge, double massWithH2O, Tolerance tolerance)
{
_rankedSpec = new RankedSpectrum(spec);
_scorer = scorer;
_charge = charge;
_sequenceMass = massWithH2O;
_tolerance = tolerance;
}
public static DeconvolutedSpectrum GetDeconvolutedSpectrum(
Spectrum spec, int minCharge, int maxCharge,
int isotopeOffsetTolerance, double filteringWindowSize,
Tolerance tolerance, double corrScoreThreshold = 0.7)
{
var peaks = GetDeconvolutedPeaks(spec.Peaks, minCharge, maxCharge, isotopeOffsetTolerance, filteringWindowSize,
tolerance, corrScoreThreshold);
return new DeconvolutedSpectrum(spec, peaks.ToArray());
}
/// <summary>
/// MaxEnt deconvolution algorithm constructor
/// </summary>
/// <param name="tolerance">tolerance</param>
/// <param name="massBinning">mass binning interface</param>
/// <param name="minCharge">maximum charge to be considered</param>
/// <param name="maxCharge">minimum charge to be considered</param>
/// <param name="minMass">minimum mass to be considered</param>
/// <param name="maxMass">maximum mass to be considered</param>
public MaxEntDeconvoluter(Spectrum spec, Tolerance tolerance, MzComparerWithBinning massBinning, int minCharge = 1, int maxCharge = 100, double minMass = 10000, double maxMass = 100000)
{
_minCharge = minCharge;
_maxCharge = maxCharge;
_minMass = minMass;
_maxMass = maxMass;
_tolerance = tolerance;
_massBinning = massBinning;
_spectrum = spec;
_deconvolutedSpectrum = null;
}
public SpectrumMatch(Sequence sequence, Spectrum spectrum, int scanNum=0, int precursorCharge=1, bool decoy=false)
{
Peptide = "";
foreach (var aa in sequence) Peptide += aa.Residue;
_spectrum = spectrum;
ScanNum = scanNum;
PrecursorCharge = precursorCharge;
Decoy = decoy;
Sequence = sequence;
if (decoy) CreateDecoy();
}
/// <summary>
/// Filter out all peaks in a spectrum that are not explained by certain ion types.
/// </summary>
/// <param name="sequence">Sequence to calculate ions from.</param>
/// <param name="spectrum">Spectrum to filter.</param>
/// <param name="ionTypes">Ion types to find peaks for.</param>
/// <param name="tolerance"></param>
/// <returns>Filtered Peptide Spectrum Match</returns>
public static Spectrum FilterIonPeaks(Sequence sequence, Spectrum spectrum, IonType[] ionTypes, Tolerance tolerance)
{
var filteredPeaks = new List<Peak>();
var specMatch = new SpectrumMatch(sequence, spectrum);
foreach (var ionType in ionTypes)
{
var ions = specMatch.GetCleavageIons(ionType);
foreach (var ion in ions)
{
var peak = spectrum.FindPeak(ion.GetMonoIsotopicMz(), tolerance);
if (peak != null) filteredPeaks.Add(peak);
}
}
filteredPeaks.Sort();
return new Spectrum(filteredPeaks, spectrum.ScanNum) {MsLevel = 2};
}
public DeconvolutedSpectrum(Spectrum originalSpec, DeconvolutedPeak[] peaks)
: base(originalSpec.ScanNum)
{
Peaks = new DeconvolutedPeak[peaks.Length];
peaks.CopyTo(Peaks, 0);
MsLevel = originalSpec.MsLevel;
var ms2Spec = originalSpec as ProductSpectrum;
if (ms2Spec == null)
{
ActivationMethod = ActivationMethod.Unknown;
}
else
{
ActivationMethod = ms2Spec.ActivationMethod;
}
MsLevel = originalSpec.MsLevel;
}
protected AbstractFragmentScorer(Spectrum spec, Tolerance tol, int minCharge = 1, int maxCharge = 20, double relativeIsotopeIntensityThreshold = 0.7)
{
Ms2Spectrum = spec;
Tolerance = tol;
MinProductCharge = minCharge;
MaxProductCharge = maxCharge;
var productSpectrum = spec as ProductSpectrum;
if (productSpectrum != null)
BaseIonTypes = productSpectrum.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
else
{
var spectrum = spec as DeconvolutedSpectrum;
if (spectrum != null)
BaseIonTypes = spectrum.ActivationMethod != ActivationMethod.ETD ? BaseIonTypesCID : BaseIonTypesETD;
else
BaseIonTypes = BaseIonTypesCID;
}
RelativeIsotopeIntensityThreshold = relativeIsotopeIntensityThreshold;
}
public SpectrumMatch(string peptide, DataFileFormat sequenceFormat,
Spectrum spectrum, int scanNum=0, int precursorCharge=1, bool decoy=false, string formula="")
{
Peptide = peptide;
_spectrum = spectrum;
_lcms = null;
ScanNum = scanNum;
PrecursorCharge = precursorCharge;
Decoy = decoy;
var sequenceReader = new SequenceReader(sequenceFormat);
Sequence = sequenceReader.GetSequence(peptide);
if (decoy) CreateDecoy();
else if (formula.Length > 0)
{
var composition = Composition.Parse(formula);
if (!composition.Equals(Sequence.Composition + Composition.H2O))
{
throw new MismatchException();
}
}
}
private double GetScore(Ion ion, Spectrum spectrum, Tolerance tolerance, double relativeIntenistyThreshold)
{
double score;
switch (_method)
{
case ScoreMethod.Cosine:
score = spectrum.GetConsineScore(ion, tolerance, relativeIntenistyThreshold);
break;
case ScoreMethod.FitScore:
score = spectrum.GetFitScore(ion, tolerance, relativeIntenistyThreshold);
break;
case ScoreMethod.Pearson:
score = spectrum.GetCorrScore(ion, tolerance, relativeIntenistyThreshold);
break;
default:
score = spectrum.GetConsineScore(ion, tolerance, relativeIntenistyThreshold);
break;
}
return score;
}
private List<DeconvolutedPeak> CollectMultiplyChargedPeaks(double mass, Spectrum spectrum)
{
var deconvPeaks = new List<DeconvolutedPeak>();
var minMz = spectrum.Peaks.First().Mz;
var maxMz = spectrum.Peaks.Last().Mz;
var minCharge = (int)Math.Max(_minCharge, Math.Ceiling(mass / (maxMz - Constants.Proton)));
var maxCharge = (int)Math.Min(_maxCharge, Math.Floor(mass / (minMz - Constants.Proton)));
for (var charge = minCharge; charge < maxCharge; charge++)
{
var mz = mass / charge + Constants.Proton;
var peak = spectrum.FindPeak(mz, _tolerance);
if (peak == null) continue;
deconvPeaks.Add(new DeconvolutedPeak(peak, charge));
}
return deconvPeaks;
}
// Select the best peak within +/- filteringWindowSize
public static List <DeconvolutedPeak> GetDeconvolutedPeaks(
Peak[] peaks, int minCharge, int maxCharge,
int isotopeOffsetTolerance, double filteringWindowSize,
Tolerance tolerance, double corrScoreThreshold)
{
var monoIsotopePeakList = new List <DeconvolutedPeak>();
for (var peakIndex = 0; peakIndex < peaks.Length; peakIndex++)
{
var peak = peaks[peakIndex];
// Check whether peak has the maximum intensity within the window
var isBest = true;
var prevIndex = peakIndex - 1;
while (prevIndex >= 0)
{
var prevPeak = peaks[prevIndex];
if ((peak.Mz - prevPeak.Mz) > filteringWindowSize)
{
break;
}
if (prevPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
prevIndex--;
}
if (!isBest)
{
continue;
}
var nextIndex = peakIndex + 1;
while (nextIndex < peaks.Length)
{
var nextPeak = peaks[nextIndex];
if ((nextPeak.Mz - peak.Mz) > filteringWindowSize)
{
break;
}
if (nextPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
nextIndex++;
}
if (!isBest)
{
continue;
}
// peak has the maximum intensity, window = [prevIndex+1,nextIndex-1]
var window = new Peak[nextIndex - prevIndex - 1];
Array.Copy(peaks, prevIndex + 1, window, 0, window.Length);
var windowSpectrum = new Spectrum(window, 1);
var peakMz = peak.Mz;
for (var charge = maxCharge; charge >= minCharge; charge--)
{
var mass = peak.Mz * charge;
var mostAbundantIsotopeIndex = Averagine.GetIsotopomerEnvelope(mass).MostAbundantIsotopeIndex;
for (var isotopeIndex = mostAbundantIsotopeIndex - isotopeOffsetTolerance; isotopeIndex <= mostAbundantIsotopeIndex + isotopeOffsetTolerance; isotopeIndex++)
{
var monoIsotopeMass = Ion.GetMonoIsotopicMass(peakMz, charge, isotopeIndex);
var isotopomerEnvelope = Averagine.GetIsotopomerEnvelope(monoIsotopeMass);
var observedPeaks = windowSpectrum.GetAllIsotopePeaks(monoIsotopeMass, charge, isotopomerEnvelope, tolerance, 0.1);
if (observedPeaks == null)
{
continue;
}
var envelop = isotopomerEnvelope.Envolope;
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 sim = FitScoreCalculator.GetDistanceAndCorrelation(envelop, observedIntensities);
var bcDist = sim.Item1;
var corr = sim.Item2;
if (corr < corrScoreThreshold && bcDist > 0.03)
{
continue;
}
// monoIsotopeMass is valid
var deconvPeak = new DeconvolutedPeak(monoIsotopeMass, observedIntensities[mostAbundantIsotopeIndex], charge, corr, bcDist, observedPeaks);
monoIsotopePeakList.Add(deconvPeak);
}
}
}
monoIsotopePeakList.Sort();
return(monoIsotopePeakList);
}
public RankedPeaks(Spectrum spectrum)
: base(spectrum.Peaks, spectrum.ScanNum)
{
_spectrum = spectrum;
Array.Sort(Peaks, new IntensityComparer());
}
public Spectrum GetDeconvolutedSpectrum(Spectrum spec, int minCharge, int maxCharge, Tolerance tolerance, double corrThreshold)
{
return GetDeconvolutedSpectrum(spec, minCharge, maxCharge, tolerance, corrThreshold, IsotopeOffsetTolerance, FilteringWindowSize);
}
/// <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;
}
/// <summary>
/// Get the deconvoluted peaks that correspond to the provided peak list
/// </summary>
/// <param name="peaks"></param>
/// <param name="minCharge"></param>
/// <param name="maxCharge"></param>
/// <param name="isotopeOffsetTolerance"></param>
/// <param name="tolerance"></param>
/// <param name="corrScoreThreshold"></param>
/// <returns></returns>
public static List <DeconvolutedPeak> GetDeconvolutedPeaks_new(
Peak[] peaks,
int minCharge,
int maxCharge,
int isotopeOffsetTolerance,
Tolerance tolerance,
double corrScoreThreshold)
{
var spectrum = new Spectrum(peaks, 0);
var monoIsotopePeakList = new List <DeconvolutedPeak>();
var sortedPeaks = peaks.OrderByDescending(peak => peak.Intensity).ToArray();
var peakUsed = new bool[peaks.Length];
foreach (var peak in sortedPeaks)
{
var peakIndex = Array.BinarySearch(peaks, peak);
if (peakUsed[peakIndex])
{
continue;
}
var bestScore = 0.0;
DeconvolutedPeak bestPeak = null;
Tuple <Peak, int>[] bestObservedPeaks = null;
for (var charge = minCharge; charge <= maxCharge; charge++)
{
var mass = peak.Mz * charge - (charge * Constants.Proton);
if (mass > MaxMass)
{
continue;
}
var isotopomerEnvelope = Averagine.GetIsotopomerEnvelope(mass);
var mostAbundantIsotopeIndex = isotopomerEnvelope.MostAbundantIsotopeIndex;
var offsetTolerance = isotopeOffsetTolerance;
if (isotopeOffsetTolerance < 0)
{
offsetTolerance = isotopomerEnvelope.Envelope.Length;
}
for (var isotopeIndex = mostAbundantIsotopeIndex - offsetTolerance;
isotopeIndex <= mostAbundantIsotopeIndex + offsetTolerance;
isotopeIndex++)
{
var monoIsotopeMass = Ion.GetMonoIsotopicMass(peak.Mz, charge, isotopeIndex);
var observedPeaks = GetAllIsotopePeaks(spectrum, monoIsotopeMass, charge, isotopomerEnvelope, tolerance, 0.1);
if (observedPeaks == null)
{
continue;
}
var envelop = isotopomerEnvelope.Envelope;
var observedIntensities = new double[observedPeaks.Length];
var observedPeakCount = 0;
for (var i = 0; i < observedPeaks.Length; i++)
{
var observedPeak = observedPeaks[i];
if (observedPeak != null && peakUsed[observedPeak.Item2])
{
observedPeak = null;
observedPeaks[i] = null;
}
observedPeakCount += observedPeak != null ? 1 : 0;
observedIntensities[i] = observedPeak != null ? (float)observedPeak.Item1.Intensity : 0.0;
}
var sim = FitScoreCalculator.GetDistanceAndCorrelation(envelop, observedIntensities);
var bcDist = sim.Item1;
var corr = sim.Item2;
var foundPeakRatio = observedPeakCount / ((double)envelop.Length);
var interferenceScore = 10.0;
var filteredObserved = observedPeaks.Where(p => p != null).ToArray();
if (filteredObserved.Length >= 2)
{
var allPeaks =
spectrum.Peaks.Where(p => p.Mz >= filteredObserved[0].Item1.Mz && p.Mz <= filteredObserved[filteredObserved.Length - 1].Item1.Mz).ToArray();
interferenceScore = CalculateInterferenceScore(allPeaks, filteredObserved);
}
bcDist = Math.Max(bcDist, double.Epsilon);
if (corr < corrScoreThreshold && bcDist > 0.1)
{
continue;
}
var score = (foundPeakRatio * corr) / (bcDist * (Math.Abs(mostAbundantIsotopeIndex - isotopeIndex) + 1) * interferenceScore);
//if (corr < corrScoreThreshold) continue;
// monoIsotopeMass is valid
if (score >= bestScore)
{
bestScore = score;
bestPeak = new DeconvolutedPeak(monoIsotopeMass, observedIntensities[mostAbundantIsotopeIndex], charge, corr, bcDist, observedPeaks.Where(p => p != null).Select(p => p.Item1).ToArray());
bestObservedPeaks = observedPeaks;
}
}
}
if (bestPeak != null)
{
monoIsotopePeakList.Add(bestPeak);
foreach (var p in bestObservedPeaks)
{
if (p != null)
{
bestPeak.ObservedPeakIndices.Add(p.Item2);
peakUsed[p.Item2] = true;
}
}
}
}
monoIsotopePeakList.Sort();
return(monoIsotopePeakList);
}
private void HandleSpectrum(
ref SpectrumTrackingInfo trackingInfo,
Dictionary<int, List<int>> isolationMzBinToScanNums,
Spectrum spec)
{
trackingInfo.SpecRead += 1;
//Console.WriteLine("Reading Scan {0}; {1} peaks", spec.ScanNum, spec.Peaks.Length);
ScanNumToMsLevel[spec.ScanNum] = spec.MsLevel;
ScanNumElutionTimeMap[spec.ScanNum] = spec.ElutionTime;
if (spec.MsLevel == 1)
{
if (trackingInfo.PrecursorSignalToNoiseRatioThreshold > 0.0)
spec.FilterNoise(trackingInfo.PrecursorSignalToNoiseRatioThreshold);
//foreach (var peak in spec.Peaks)
//{
// _ms1PeakList.Add(new LcMsPeak(peak.Mz, peak.Intensity, spec.ScanNum));
//}
_ms1PeakList.AddRange(spec.Peaks.Select(peak => new LcMsPeak(peak.Mz, peak.Intensity, spec.ScanNum)));
_scanNumSpecMap.Add(spec.ScanNum, spec);
}
else if (spec.MsLevel == 2)
{
var productSpec = spec as ProductSpectrum;
if (productSpec != null)
{
if (trackingInfo.ProductSignalToNoiseRatioThreshold > 0.0)
productSpec.FilterNoise(trackingInfo.ProductSignalToNoiseRatioThreshold);
var isolationWindow = productSpec.IsolationWindow;
var minBinNum = (int)Math.Round(isolationWindow.MinMz * IsolationWindowBinningFactor);
var maxBinNum = (int)Math.Round(isolationWindow.MaxMz * IsolationWindowBinningFactor);
for (var binNum = minBinNum; binNum <= maxBinNum; binNum++)
{
List<int> scanNumList;
if (!isolationMzBinToScanNums.TryGetValue(binNum, out scanNumList))
{
scanNumList = new List<int>();
isolationMzBinToScanNums[binNum] = scanNumList;
}
scanNumList.Add(productSpec.ScanNum);
}
_scanNumSpecMap.Add(spec.ScanNum, productSpec);
}
}
if (spec.ScanNum < trackingInfo.MinScanNum) trackingInfo.MinScanNum = spec.ScanNum;
if (spec.ScanNum > trackingInfo.MaxScanNum) trackingInfo.MaxScanNum = spec.ScanNum;
if (spec.MsLevel < trackingInfo.MinMsLevel) trackingInfo.MinMsLevel = spec.MsLevel;
if (spec.MsLevel > trackingInfo.MaxMsLevel) trackingInfo.MaxMsLevel = spec.MsLevel;
}
public CompositeScorer(Spectrum ms2Spec, Tolerance tol, int minCharge, int maxCharge, double relativeIsotopeIntensityThreshold = 0.1)
: base(ms2Spec, tol, minCharge, maxCharge, relativeIsotopeIntensityThreshold)
{
ReferencePeakIntensity = GetRefIntensity(ms2Spec.Peaks);
}
/// <summary>
/// Initializes a new instance of the <see cref="TargetedDeconvolutedSpectrum" /> class.
/// </summary>
/// <param name="spectrum">Non-deconvoluted spectrum.</param>
/// <param name="minCharge">The minimum charge state to consider.</param>
/// <param name="maxCharge">The maximum charge state to consider.</param>
public TargetedDeconvolutedSpectrum(Spectrum spectrum, int minCharge = 1, int maxCharge = 20)
{
this.spectrum = spectrum;
this.minCharge = minCharge;
this.maxCharge = maxCharge;
}
/// <summary>
/// Get the deconvoluted peaks, selecting the best peak within +/- filteringWindowSize
/// </summary>
/// <param name="scanNum">Scan number (included in any exceptions that are caught)</param>
/// <param name="peaks"></param>
/// <param name="minCharge"></param>
/// <param name="maxCharge"></param>
/// <param name="isotopeOffsetTolerance"></param>
/// <param name="filteringWindowSize"></param>
/// <param name="tolerance"></param>
/// <param name="corrScoreThreshold"></param>
/// <returns></returns>
public static List <DeconvolutedPeak> GetDeconvolutedPeaks(
int scanNum, Peak[] peaks,
int minCharge, int maxCharge,
int isotopeOffsetTolerance, double filteringWindowSize,
Tolerance tolerance, double corrScoreThreshold)
{
try
{
var monoIsotopePeakList = new List <DeconvolutedPeak>();
for (var peakIndex = 0; peakIndex < peaks.Length; peakIndex++)
{
var peak = peaks[peakIndex];
// Check whether peak has the maximum intensity within the window
var isBest = true;
var prevIndex = peakIndex - 1;
while (prevIndex >= 0)
{
var prevPeak = peaks[prevIndex];
if ((peak.Mz - prevPeak.Mz) > filteringWindowSize)
{
break;
}
if (prevPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
prevIndex--;
}
if (!isBest)
{
continue;
}
var nextIndex = peakIndex + 1;
while (nextIndex < peaks.Length)
{
var nextPeak = peaks[nextIndex];
if ((nextPeak.Mz - peak.Mz) > filteringWindowSize)
{
break;
}
if (nextPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
nextIndex++;
}
if (!isBest)
{
continue;
}
// peak has the maximum intensity, window = [prevIndex+1,nextIndex-1]
var window = new Peak[nextIndex - prevIndex - 1];
Array.Copy(peaks, prevIndex + 1, window, 0, window.Length);
var windowSpectrum = new Spectrum(window, 1);
var peakMz = peak.Mz;
//var bestScore = 0.0;
//DeconvolutedPeak bestPeak = null;
for (var charge = maxCharge; charge >= minCharge; charge--)
{
var mass = (peak.Mz * charge) - charge * Constants.Proton;
//var isotopomerEnvelope = Averagine.GetIsotopomerEnvelope(mass);
//var mostAbundantIsotopeIndex = isotopomerEnvelope.MostAbundantIsotopeIndex;
var mostAbundantIsotopeIndex = Averagine.GetIsotopomerEnvelope(mass).MostAbundantIsotopeIndex;
for (var isotopeIndex = mostAbundantIsotopeIndex - isotopeOffsetTolerance; isotopeIndex <= mostAbundantIsotopeIndex + isotopeOffsetTolerance; isotopeIndex++)
{
var monoIsotopeMass = Ion.GetMonoIsotopicMass(peakMz, charge, isotopeIndex);
var isotopomerEnvelope = Averagine.GetIsotopomerEnvelope(monoIsotopeMass);
var observedPeaks = windowSpectrum.GetAllIsotopePeaks(monoIsotopeMass, charge, isotopomerEnvelope, tolerance, 0.1);
if (observedPeaks == null)
{
continue;
}
var envelop = isotopomerEnvelope.Envelope;
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 sim = FitScoreCalculator.GetDistanceAndCorrelation(envelop, observedIntensities);
var bcDist = sim.Item1;
var corr = sim.Item2;
//var score = corr / (bcDist * ((double)Math.Abs(isotopeIndex - mostAbundantIsotopeIndex) / envelop.Length));
if (corr < corrScoreThreshold && bcDist > 0.03)
{
continue;
}
// monoIsotopeMass is valid
//if (score >= bestScore)
//{
// bestScore = score;
// bestPeak = new DeconvolutedPeak(monoIsotopeMass, observedIntensities[mostAbundantIsotopeIndex], charge, corr, bcDist, observedPeaks);
//}
var deconvPeak = new DeconvolutedPeak(monoIsotopeMass, observedIntensities[mostAbundantIsotopeIndex], charge, corr, bcDist, observedPeaks);
monoIsotopePeakList.Add(deconvPeak);
}
}
//if (bestPeak != null)
//{
// monoIsotopePeakList.Add(bestPeak);
//}
}
monoIsotopePeakList.Sort();
return(monoIsotopePeakList);
}
catch (Exception ex)
{
throw new Exception(string.Format("Error getting deconvoluted peaks for scan {0} in GetDeconvolutedPeaks: {1}", scanNum, ex.Message), ex);
}
}
private Peak GetHighestPeak(Ion ion, Spectrum spectrum, Tolerance tolerance, double relativeIntensityThreshold)
{
var peaks = spectrum.GetAllIsotopePeaks(ion, tolerance, relativeIntensityThreshold);
Peak highestPeak = null;
double highestIntensity = 0.0;
if (peaks == null) return null;
foreach (var peak in peaks)
{
if (peak != null && peak.Intensity >= highestIntensity)
{
highestPeak = peak;
highestIntensity = peak.Intensity;
}
}
return highestPeak;
}
public Peak[] GetAllIsotopePeaks(Spectrum spec, Ion ion, Tolerance tolerance, double relativeIntensityThreshold, out int[] peakIndexList)
{
var mostAbundantIsotopeIndex = ion.Composition.GetMostAbundantIsotopeZeroBasedIndex();
var isotopomerEnvelope = ion.Composition.GetIsotopomerEnvelopeRelativeIntensities();
peakIndexList = new int[isotopomerEnvelope.Length];
var mostAbundantIsotopeMz = ion.GetIsotopeMz(mostAbundantIsotopeIndex);
var mostAbundantIsotopeMatchedPeakIndex = spec.FindPeakIndex(mostAbundantIsotopeMz, tolerance);
if (mostAbundantIsotopeMatchedPeakIndex < 0) return null;
var observedPeaks = new Peak[isotopomerEnvelope.Length];
observedPeaks[mostAbundantIsotopeIndex] = spec.Peaks[mostAbundantIsotopeMatchedPeakIndex];
peakIndexList[mostAbundantIsotopeIndex] = mostAbundantIsotopeMatchedPeakIndex;
// go down
var peakIndex = mostAbundantIsotopeMatchedPeakIndex - 1;
for (var isotopeIndex = mostAbundantIsotopeIndex - 1; isotopeIndex >= 0; isotopeIndex--)
{
if (isotopomerEnvelope[isotopeIndex] < relativeIntensityThreshold) break;
var isotopeMz = ion.GetIsotopeMz(isotopeIndex);
var tolTh = tolerance.GetToleranceAsTh(isotopeMz);
var minMz = isotopeMz - tolTh;
var maxMz = isotopeMz + tolTh;
for (var i = peakIndex; i >= 0; i--)
{
var peakMz = spec.Peaks[i].Mz;
if (peakMz < minMz)
{
peakIndex = i;
break;
}
if (peakMz <= maxMz) // find match, move to prev isotope
{
var peak = spec.Peaks[i];
if (observedPeaks[isotopeIndex] == null ||
peak.Intensity > observedPeaks[isotopeIndex].Intensity)
{
observedPeaks[isotopeIndex] = peak;
peakIndexList[isotopeIndex] = i;
}
}
}
}
// go up
peakIndex = mostAbundantIsotopeMatchedPeakIndex + 1;
for (var isotopeIndex = mostAbundantIsotopeIndex + 1; isotopeIndex < isotopomerEnvelope.Length; isotopeIndex++)
{
if (isotopomerEnvelope[isotopeIndex] < relativeIntensityThreshold) break;
var isotopeMz = ion.GetIsotopeMz(isotopeIndex);
var tolTh = tolerance.GetToleranceAsTh(isotopeMz);
var minMz = isotopeMz - tolTh;
var maxMz = isotopeMz + tolTh;
for (var i = peakIndex; i < spec.Peaks.Length; i++)
{
var peakMz = spec.Peaks[i].Mz;
if (peakMz > maxMz)
{
peakIndex = i;
break;
}
if (peakMz >= minMz) // find match, move to prev isotope
{
var peak = spec.Peaks[i];
if (observedPeaks[isotopeIndex] == null ||
peak.Intensity > observedPeaks[isotopeIndex].Intensity)
{
observedPeaks[isotopeIndex] = peak;
peakIndexList[isotopeIndex] = i;
}
}
}
}
return observedPeaks;
}
// Select the best peak within +/- filteringWindowSize
public static List<DeconvolutedPeak> GetDeconvolutedPeaks(
Peak[] peaks, int minCharge, int maxCharge,
int isotopeOffsetTolerance, double filteringWindowSize,
Tolerance tolerance, double corrScoreThreshold)
{
var monoIsotopePeakList = new List<DeconvolutedPeak>();
for (var peakIndex = 0; peakIndex < peaks.Length; peakIndex++)
{
var peak = peaks[peakIndex];
// Check whether peak has the maximum intensity within the window
var isBest = true;
var prevIndex = peakIndex - 1;
while (prevIndex >= 0)
{
var prevPeak = peaks[prevIndex];
if ((peak.Mz - prevPeak.Mz) > filteringWindowSize) break;
if (prevPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
prevIndex--;
}
if (!isBest) continue;
var nextIndex = peakIndex + 1;
while (nextIndex < peaks.Length)
{
var nextPeak = peaks[nextIndex];
if ((nextPeak.Mz - peak.Mz) > filteringWindowSize) break;
if (nextPeak.Intensity > peak.Intensity)
{
isBest = false;
break;
}
nextIndex++;
}
if (!isBest) continue;
// peak has the maximum intensity, window = [prevIndex+1,nextIndex-1]
var window = new Peak[nextIndex - prevIndex - 1];
Array.Copy(peaks, prevIndex + 1, window, 0, window.Length);
var windowSpectrum = new Spectrum(window, 1);
var peakMz = peak.Mz;
for (var charge = maxCharge; charge >= minCharge; charge--)
{
var mass = peak.Mz * charge;
var mostAbundantIsotopeIndex = Averagine.GetIsotopomerEnvelope(mass).MostAbundantIsotopeIndex;
for (var isotopeIndex = mostAbundantIsotopeIndex - isotopeOffsetTolerance; isotopeIndex <= mostAbundantIsotopeIndex + isotopeOffsetTolerance; isotopeIndex++)
{
var monoIsotopeMass = Ion.GetMonoIsotopicMass(peakMz, charge, isotopeIndex);
var isotopomerEnvelope = Averagine.GetIsotopomerEnvelope(monoIsotopeMass);
var observedPeaks = windowSpectrum.GetAllIsotopePeaks(monoIsotopeMass, charge, isotopomerEnvelope, tolerance, 0.1);
if (observedPeaks == null) continue;
var envelop = isotopomerEnvelope.Envolope;
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 sim = FitScoreCalculator.GetDistanceAndCorrelation(envelop, observedIntensities);
var bcDist = sim.Item1;
var corr = sim.Item2;
if (corr < corrScoreThreshold && bcDist > 0.03) continue;
// monoIsotopeMass is valid
var deconvPeak = new DeconvolutedPeak(monoIsotopeMass, observedIntensities[mostAbundantIsotopeIndex], charge, corr, bcDist, observedPeaks);
monoIsotopePeakList.Add(deconvPeak);
}
}
}
monoIsotopePeakList.Sort();
return monoIsotopePeakList;
}
private static Tuple <Peak, int>[] GetAllIsotopePeaks(
Spectrum spec,
double monoisotopicMass,
int charge,
IsotopomerEnvelope envelope,
Tolerance tolerance,
double relativeIntensityThreshold)
{
var mostAbundantIsotopeIndex = envelope.MostAbundantIsotopeIndex;
var isotopomerEnvelope = envelope.Envelope;
var mostAbundantIsotopeMz = Ion.GetIsotopeMz(monoisotopicMass, charge, mostAbundantIsotopeIndex);
var mostAbundantIsotopePeakIndex = spec.FindPeakIndex(mostAbundantIsotopeMz, tolerance);
if (mostAbundantIsotopePeakIndex < 0)
{
return(null);
}
var observedPeaks = new Tuple <Peak, int> [isotopomerEnvelope.Length];
observedPeaks[mostAbundantIsotopeIndex] = new Tuple <Peak, int>(spec.Peaks[mostAbundantIsotopePeakIndex], mostAbundantIsotopePeakIndex);
// go down
var peakIndex = mostAbundantIsotopePeakIndex - 1;
for (var isotopeIndex = mostAbundantIsotopeIndex - 1; isotopeIndex >= 0; isotopeIndex--)
{
if (isotopomerEnvelope[isotopeIndex] < relativeIntensityThreshold)
{
break;
}
var isotopeMz = Ion.GetIsotopeMz(monoisotopicMass, charge, isotopeIndex);
var tolTh = tolerance.GetToleranceAsMz(isotopeMz);
var minMz = isotopeMz - tolTh;
var maxMz = isotopeMz + tolTh;
for (var i = peakIndex; i >= 0; i--)
{
var peakMz = spec.Peaks[i].Mz;
if (peakMz < minMz)
{
peakIndex = i;
break;
}
if (peakMz <= maxMz) // find match, move to prev isotope
{
var peak = spec.Peaks[i];
if (observedPeaks[isotopeIndex] == null ||
peak.Intensity > observedPeaks[isotopeIndex].Item1.Intensity)
{
observedPeaks[isotopeIndex] = new Tuple <Peak, int>(peak, peakIndex);
}
}
}
}
// go up
peakIndex = mostAbundantIsotopePeakIndex + 1;
for (var isotopeIndex = mostAbundantIsotopeIndex + 1; isotopeIndex < isotopomerEnvelope.Length; isotopeIndex++)
{
if (isotopomerEnvelope[isotopeIndex] < relativeIntensityThreshold)
{
break;
}
var isotopeMz = Ion.GetIsotopeMz(monoisotopicMass, charge, isotopeIndex);
var tolTh = tolerance.GetToleranceAsMz(isotopeMz);
var minMz = isotopeMz - tolTh;
var maxMz = isotopeMz + tolTh;
for (var i = peakIndex; i < spec.Peaks.Length; i++)
{
var peakMz = spec.Peaks[i].Mz;
if (peakMz > maxMz)
{
peakIndex = i;
break;
}
if (peakMz >= minMz) // find match, move to prev isotope
{
var peak = spec.Peaks[i];
if (observedPeaks[isotopeIndex] == null ||
peak.Intensity > observedPeaks[isotopeIndex].Item1.Intensity)
{
observedPeaks[isotopeIndex] = new Tuple <Peak, int>(peak, peakIndex);
}
}
}
}
return(observedPeaks);
}
public CompositeScorer(Spectrum ms2Spec, Tolerance tol, double relativeIsotopeIntensityThreshold = 0.1)
: base(ms2Spec, tol, 1, 20, relativeIsotopeIntensityThreshold)
{
}
