public IndexedMassSpectralPeak GetIndexedPeak(double theorMass, int zeroBasedScanIndex, Tolerance tolerance, int chargeState)
{
IndexedMassSpectralPeak bestPeak = null;
int ceilingMz = (int)Math.Ceiling(tolerance.GetMaximumValue(theorMass).ToMz(chargeState) * BinsPerDalton);
int floorMz = (int)Math.Floor(tolerance.GetMinimumValue(theorMass).ToMz(chargeState) * BinsPerDalton);
for (int j = floorMz; j <= ceilingMz; j++)
{
if (j < _indexedPeaks.Length && _indexedPeaks[j] != null)
{
List <IndexedMassSpectralPeak> bin = _indexedPeaks[j];
int index = BinarySearchForIndexedPeak(bin, zeroBasedScanIndex);
for (int i = index; i < bin.Count; i++)
{
IndexedMassSpectralPeak peak = bin[i];
if (peak.ZeroBasedMs1ScanIndex > zeroBasedScanIndex)
{
break;
}
double expMass = peak.Mz.ToMass(chargeState);
if (tolerance.Within(expMass, theorMass) && peak.ZeroBasedMs1ScanIndex == zeroBasedScanIndex &&
(bestPeak == null || Math.Abs(expMass - theorMass) < Math.Abs(bestPeak.Mz.ToMass(chargeState) - theorMass)))
{
bestPeak = peak;
}
}
}
}
return(bestPeak);
}
public void ToleranceMinMaxTest()
{
var tol = new Tolerance(ToleranceUnit.Absolute, 9, 10);
Assert.AreEqual(2, tol.GetMaximumValue(1));
Assert.AreEqual(0, tol.GetMinimumValue(1));
}
public void MassToleranceImplicitValue()
{
var tol = new Tolerance("10 ppm");
Assert.AreEqual(10, tol.Value);
Assert.AreEqual(1 + 1e-5 / 2, tol.GetMaximumValue(1));
Assert.AreEqual(1 - 1e-5 / 2, tol.GetMinimumValue(1));
}
private (List <LabeledMs2DataPoint>, int, int, int) SearchMS2Spectrum(IMsDataScanWithPrecursor <IMzSpectrum <IMzPeak> > ms2DataScan, Proteomics.Peptide peptide, int peptideCharge, PeptideSpectralMatch identification)
{
List <LabeledMs2DataPoint> result = new List <LabeledMs2DataPoint>();
int numMs2MassChargeCombinationsConsidered = 0;
int numMs2MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks = 0;
int numFragmentsIdentified = 0;
if (ms2DataScan.MassSpectrum.Size == 0)
{
return(result, numMs2MassChargeCombinationsConsidered, numMs2MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks, numFragmentsIdentified);
}
// Key: mz value, Value: error
var addedPeaks = new Dictionary <double, double>();
var countForThisMS2 = 0;
var countForThisMS2a = 0;
var scanWindowRange = ms2DataScan.ScanWindowRange;
IHasChemicalFormula[] fragmentList = peptide.Fragment(fragmentTypesForCalibration, true).OfType <IHasChemicalFormula>().ToArray();
foreach (var fragment in fragmentList)
{
bool fragmentIdentified = false;
bool computedIsotopologues = false;
double[] masses = new double[0];
double[] intensities = new double[0];
// First look for monoisotopic masses, do not compute distribution spectrum!
for (int chargeToLookAt = 1; chargeToLookAt <= peptideCharge; chargeToLookAt++)
{
var monoisotopicMZ = fragment.MonoisotopicMass.ToMz(chargeToLookAt);
if (monoisotopicMZ > scanWindowRange.Maximum)
{
continue;
}
if (monoisotopicMZ < scanWindowRange.Minimum)
{
break;
}
var closestPeakMZ = ms2DataScan.MassSpectrum.GetClosestPeakXvalue(monoisotopicMZ);
if (mzToleranceForMs2Search.Within(closestPeakMZ.Value, monoisotopicMZ) && !computedIsotopologues)
{
var dist = IsotopicDistribution.GetDistribution(fragment.ThisChemicalFormula, fineResolutionForIsotopeDistCalculation, 0.001);
masses = dist.Masses.ToArray();
intensities = dist.Intensities.ToArray();
Array.Sort(intensities, masses, Comparer <double> .Create((x, y) => y.CompareTo(x)));
computedIsotopologues = true;
break;
}
}
if (computedIsotopologues)
{
bool startingToAdd = false;
for (int chargeToLookAt = 1; chargeToLookAt <= peptideCharge; chargeToLookAt++)
{
if (masses.First().ToMz(chargeToLookAt) > scanWindowRange.Maximum)
{
continue;
}
if (masses.Last().ToMz(chargeToLookAt) < scanWindowRange.Minimum)
{
break;
}
var trainingPointsToAverage = new List <LabeledMs2DataPoint>();
foreach (double a in masses)
{
double theMZ = a.ToMz(chargeToLookAt);
var npwr = ms2DataScan.MassSpectrum.NumPeaksWithinRange(mzToleranceForMs2Search.GetMinimumValue(theMZ), mzToleranceForMs2Search.GetMaximumValue(theMZ));
if (npwr == 0)
{
break;
}
numMs2MassChargeCombinationsConsidered++;
if (npwr > 1)
{
numMs2MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks++;
continue;
}
var closestPeakIndex = ms2DataScan.MassSpectrum.GetClosestPeakIndex(theMZ);
var closestPeakMZ = ms2DataScan.MassSpectrum.XArray[closestPeakIndex.Value];
if (!addedPeaks.ContainsKey(closestPeakMZ))
{
addedPeaks.Add(closestPeakMZ, Math.Abs(closestPeakMZ - theMZ));
trainingPointsToAverage.Add(new LabeledMs2DataPoint(closestPeakMZ, double.NaN, double.NaN, double.NaN, Math.Log(ms2DataScan.MassSpectrum.YArray[closestPeakIndex.Value]), theMZ, null));
}
}
// If started adding and suddnely stopped, go to next one, no need to look at higher charges
if (trainingPointsToAverage.Count == 0 && startingToAdd)
{
break;
}
if (trainingPointsToAverage.Count < Math.Min(minMS2isotopicPeaksNeededForConfirmedIdentification, intensities.Count()))
{
}
else
{
startingToAdd = true;
if (!fragmentIdentified)
{
fragmentIdentified = true;
numFragmentsIdentified += 1;
}
countForThisMS2 += trainingPointsToAverage.Count;
countForThisMS2a++;
result.Add(new LabeledMs2DataPoint(trainingPointsToAverage.Select(b => b.mz).Average(),
ms2DataScan.RetentionTime,
Math.Log(ms2DataScan.TotalIonCurrent),
ms2DataScan.InjectionTime.HasValue ? Math.Log(ms2DataScan.InjectionTime.Value) : double.NaN,
trainingPointsToAverage.Select(b => b.logIntensity).Average(),
trainingPointsToAverage.Select(b => b.expectedMZ).Average(),
identification));
}
}
}
}
return(result, numMs2MassChargeCombinationsConsidered, numMs2MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks, numFragmentsIdentified);
}
private (List <LabeledMs1DataPoint>, int, int) SearchMS1Spectra(double[] theoreticalMasses, double[] theoreticalIntensities, int ms2spectrumIndex, int direction, int peptideCharge, PeptideSpectralMatch identification)
{
List <LabeledMs1DataPoint> result = new List <LabeledMs1DataPoint>();
int numMs1MassChargeCombinationsConsidered = 0;
int numMs1MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks = 0;
int theIndex;
theIndex = direction == 1 ? ms2spectrumIndex : ms2spectrumIndex - 1;
bool addedAscan = true;
int highestKnownChargeForThisPeptide = peptideCharge;
while (theIndex >= 1 && theIndex <= myMsDataFile.NumSpectra && addedAscan)
{
int countForThisScan = 0;
if (myMsDataFile.GetOneBasedScan(theIndex).MsnOrder > 1)
{
theIndex += direction;
continue;
}
addedAscan = false;
var fullMS1scan = myMsDataFile.GetOneBasedScan(theIndex);
var scanWindowRange = fullMS1scan.ScanWindowRange;
var fullMS1spectrum = fullMS1scan.MassSpectrum;
if (fullMS1spectrum.Size == 0)
{
break;
}
bool startingToAddCharges = false;
int chargeToLookAt = 1;
do
{
if (theoreticalMasses[0].ToMz(chargeToLookAt) > scanWindowRange.Maximum)
{
chargeToLookAt++;
continue;
}
if (theoreticalMasses[0].ToMz(chargeToLookAt) < scanWindowRange.Minimum)
{
break;
}
var trainingPointsToAverage = new List <LabeledMs1DataPoint>();
foreach (double a in theoreticalMasses)
{
double theMZ = a.ToMz(chargeToLookAt);
var npwr = fullMS1spectrum.NumPeaksWithinRange(mzToleranceForMs1Search.GetMinimumValue(theMZ), mzToleranceForMs1Search.GetMaximumValue(theMZ));
if (npwr == 0)
{
break;
}
numMs1MassChargeCombinationsConsidered++;
if (npwr > 1)
{
numMs1MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks++;
continue;
}
var closestPeakIndex = fullMS1spectrum.GetClosestPeakIndex(theMZ);
var closestPeakMZ = fullMS1spectrum.XArray[closestPeakIndex.Value];
highestKnownChargeForThisPeptide = Math.Max(highestKnownChargeForThisPeptide, chargeToLookAt);
trainingPointsToAverage.Add(new LabeledMs1DataPoint(closestPeakMZ, double.NaN, double.NaN, double.NaN, Math.Log(fullMS1spectrum.YArray[closestPeakIndex.Value]), theMZ, null));
}
// If started adding and suddnely stopped, go to next one, no need to look at higher charges
if (trainingPointsToAverage.Count == 0 && startingToAddCharges)
{
break;
}
if ((trainingPointsToAverage.Count == 0 || (trainingPointsToAverage.Count == 1 && theoreticalIntensities[0] < 0.65)) && (peptideCharge <= chargeToLookAt))
{
break;
}
if ((trainingPointsToAverage.Count == 1 && theoreticalIntensities[0] < 0.65) ||
trainingPointsToAverage.Count < Math.Min(minMS1isotopicPeaksNeededForConfirmedIdentification, theoreticalIntensities.Count()))
{
}
else
{
addedAscan = true;
startingToAddCharges = true;
countForThisScan++;
result.Add(new LabeledMs1DataPoint(trainingPointsToAverage.Select(b => b.mz).Average(),
fullMS1scan.RetentionTime,
Math.Log(fullMS1scan.TotalIonCurrent),
fullMS1scan.InjectionTime.HasValue ? Math.Log(fullMS1scan.InjectionTime.Value) : double.NaN,
trainingPointsToAverage.Select(b => b.logIntensity).Average(),
trainingPointsToAverage.Select(b => b.expectedMZ).Average(),
identification));
}
chargeToLookAt++;
} while (chargeToLookAt <= highestKnownChargeForThisPeptide + 1);
theIndex += direction;
}
return(result, numMs1MassChargeCombinationsConsidered, numMs1MassChargeCombinationsThatAreIgnoredBecauseOfTooManyPeaks);
}
public static void MatchIons(IMsDataScan <IMzSpectrum <IMzPeak> > thisScan, Tolerance productMassTolerance, double[] sortedTheoreticalProductMassesForThisPeptide, List <double> matchedIonMassesList, List <double> productMassErrorDa, List <double> productMassErrorPpm, double precursorMass, List <DissociationType> dissociationTypes, bool addCompIons)
{
var TotalProductsHere = sortedTheoreticalProductMassesForThisPeptide.Length;
if (TotalProductsHere == 0)
{
return;
}
int currentTheoreticalIndex = -1;
double currentTheoreticalMass;
do
{
currentTheoreticalIndex++;
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
} while (double.IsNaN(currentTheoreticalMass) && currentTheoreticalIndex < sortedTheoreticalProductMassesForThisPeptide.Length - 1);
if (double.IsNaN(currentTheoreticalMass))
{
return;
}
double currentTheoreticalMz = currentTheoreticalMass + Constants.protonMass;
int testTheoreticalIndex;
double testTheoreticalMass;
double testTheoreticalMz;
// speed optimizations
double[] experimental_mzs = thisScan.MassSpectrum.XArray;
double[] experimental_intensities = thisScan.MassSpectrum.YArray;
int numExperimentalPeaks = experimental_mzs.Length;
// Loop over all experimental indices
for (int experimentalIndex = 0; experimentalIndex < numExperimentalPeaks; experimentalIndex++)
{
double currentExperimentalMz = experimental_mzs[experimentalIndex];
// If found match
if (productMassTolerance.Within(currentExperimentalMz, currentTheoreticalMz))
{
matchedIonMassesList.Add(currentTheoreticalMass);
double currentExperimentalMass = currentExperimentalMz - Constants.protonMass;
productMassErrorDa.Add(currentExperimentalMass - currentTheoreticalMass);
productMassErrorPpm.Add((currentExperimentalMass - currentTheoreticalMass) * 1000000 / currentTheoreticalMass);
currentTheoreticalIndex++;
if (currentTheoreticalIndex == TotalProductsHere)
{
break;
}
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
currentTheoreticalMz = currentTheoreticalMass + Constants.protonMass;
}
// Else if for sure did not reach the next theoretical yet
else if (currentExperimentalMz > currentTheoreticalMz)
{
// Move on to next index and never come back!
currentTheoreticalIndex++;
if (currentTheoreticalIndex == TotalProductsHere)
{
break;
}
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
currentTheoreticalMz = currentTheoreticalMass + Constants.protonMass;
// Start with the current ones
testTheoreticalIndex = currentTheoreticalIndex;
testTheoreticalMass = currentTheoreticalMass;
testTheoreticalMz = currentTheoreticalMz;
// Mark the skipped theoreticals as not found. The last one is not for sure, might be flipped!
while (currentExperimentalMz > testTheoreticalMz)
{
// Store old info for possible reuse
currentTheoreticalMz = testTheoreticalMz;
currentTheoreticalMass = testTheoreticalMass;
currentTheoreticalIndex = testTheoreticalIndex;
// Update test stuff!
testTheoreticalIndex++;
if (testTheoreticalIndex == TotalProductsHere)
{
break;
}
testTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[testTheoreticalIndex];
testTheoreticalMz = testTheoreticalMass + Constants.protonMass;
}
experimentalIndex--;
}
}
if (addCompIons)
{
double[] complementaryMasses = new double[numExperimentalPeaks];
double[] complementaryIntensities = new double[numExperimentalPeaks];
foreach (DissociationType dissociationType in dissociationTypes)
{
if (complementaryIonConversionDictionary.TryGetValue(dissociationType, out double protonMassShift))
{
currentTheoreticalIndex = -1;
do
{
currentTheoreticalIndex++;
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
} while (double.IsNaN(currentTheoreticalMass) && currentTheoreticalIndex < sortedTheoreticalProductMassesForThisPeptide.Length - 1);
double massShiftForComplementaryConversion = precursorMass + protonMassShift; //mass shift needed to reobtain the original product ion for calculating tolerance
for (int i = numExperimentalPeaks - 1; i >= 0; i--)
{
complementaryMasses[numExperimentalPeaks - i - 1] = massShiftForComplementaryConversion - experimental_mzs[i];
complementaryIntensities[numExperimentalPeaks - i - 1] = experimental_intensities[i];
}
// Loop over all experimental indices
for (int experimentalIndex = 0; experimentalIndex < numExperimentalPeaks; experimentalIndex++)
{
double currentExperimentalMass = complementaryMasses[experimentalIndex];
double originalExperimentalMass = massShiftForComplementaryConversion - currentExperimentalMass;
double minBoundary = currentExperimentalMass - originalExperimentalMass + productMassTolerance.GetMinimumValue(originalExperimentalMass);
double maxBoundary = currentExperimentalMass - originalExperimentalMass + productMassTolerance.GetMaximumValue(originalExperimentalMass);
// If found match
if (minBoundary < currentTheoreticalMass && maxBoundary > currentTheoreticalMass)
{
matchedIonMassesList.Add(currentTheoreticalMass);
productMassErrorDa.Add(currentExperimentalMass - currentTheoreticalMass);
productMassErrorPpm.Add((currentExperimentalMass - currentTheoreticalMass) * 1000000 / currentTheoreticalMass);
currentTheoreticalIndex++;
if (currentTheoreticalIndex == TotalProductsHere)
{
break;
}
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
}
// Else if for sure passed a theoretical
else if (currentExperimentalMass > currentTheoreticalMass)
{
// Move on to next index and never come back!
currentTheoreticalIndex++;
if (currentTheoreticalIndex == TotalProductsHere)
{
break;
}
currentTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[currentTheoreticalIndex];
// Start with the current ones
testTheoreticalIndex = currentTheoreticalIndex;
testTheoreticalMass = currentTheoreticalMass;
// Mark the skipped theoreticals as not found. The last one is not for sure, might be flipped!
while (currentExperimentalMass > testTheoreticalMass)
{
// Store old info for possible reuse
currentTheoreticalMass = testTheoreticalMass;
currentTheoreticalIndex = testTheoreticalIndex;
// Update test stuff!
testTheoreticalIndex++;
if (testTheoreticalIndex == TotalProductsHere)
{
break;
}
testTheoreticalMass = sortedTheoreticalProductMassesForThisPeptide[testTheoreticalIndex];
}
experimentalIndex--;
}
}
}
else
{
throw new NotImplementedException();
}
}
}
}
//private double Search(ref double[] eMasses, ref double[] eIntenisties, double[] tMasses, double productTolerance, double tic, ref Dictionary<double, double> scores)
//{
// double score = 0.0;
// int eLength = eMasses.Length;
// int tLength = tMasses.Length;
// int e = 0;
// foreach (double t in tMasses)
// {
// double storedScore;
// if (scores.TryGetValue(t, out storedScore))
// {
// score += storedScore;
// continue;
// }
// double minMZ = t - productTolerance;
// double maxMZ = t + productTolerance;
// while (e < eLength && eMasses[e] < minMZ)
// e++;
// if (e >= eLength)
// break;
// if (eMasses[e] > maxMZ)
// continue;
// double intensities = 0;
// int index = e; // switch variables to keep e the same for the next loop around
// do
// {
// intensities += eIntenisties[index];
// index++;
// } while (index < eLength && eMasses[index] < maxMZ);
// storedScore = 1 + intensities/tic;
// score += storedScore;
// scores[t] = storedScore;
// }
// return score;
//}
/// <summary>
/// The main searching algorithm of Morpheus
/// </summary>
/// <param name="eMasses">The experimental masses</param>
/// <param name="eIntenisties">The experimental intensities</param>
/// <param name="tMasses">The theoretical masses</param>
/// <param name="productTolerance">The product mass tolerance</param>
/// <param name="tic">The total ion current of the experimental peaks</param>
/// <returns></returns>
private double Search(double[] eMasses, double[] eIntenisties, double[] tMasses, Tolerance productTolerance, double tic)
{
double score = 0.0;
double intensities = 0.0;
int eLength = eMasses.Length;
int tLength = tMasses.Length;
int e = 0;
bool forceCheck = productTolerance.GetMinimumValue(tMasses[tLength - 1]) >= eMasses[eLength - 1];
if (forceCheck)
{
foreach (double t in tMasses)
{
IRange<double> range = productTolerance.GetRange(t);
double minMZ = range.Minimum;
double maxMZ = range.Maximum;
while (e < eLength && eMasses[e] < minMZ)
e++;
if (e >= eLength)
break;
if (eMasses[e] > maxMZ)
continue;
score++;
int index = e; // switch variables to keep e the same for the next loop around
do
{
intensities += eIntenisties[index];
index++;
} while (index < eLength && eMasses[index] < maxMZ);
}
}
else
{
foreach (double t in tMasses)
{
IRange<double> range = productTolerance.GetRange(t);
double minMZ = range.Minimum;
double maxMZ = range.Maximum;
while (eMasses[e] < minMZ)
e++;
if (eMasses[e] > maxMZ)
continue;
score++;
int index = e; // switch variables to keep e the same for the next loop around
do
{
intensities += eIntenisties[index];
index++;
} while (index < eLength && eMasses[index] < maxMZ);
}
}
return score + intensities/tic;
}
//private double Search(ref double[] eMasses, ref double[] eIntenisties, double[] tMasses, double productTolerance, double tic, ref Dictionary<double, double> scores)
//{
// double score = 0.0;
// int eLength = eMasses.Length;
// int tLength = tMasses.Length;
// int e = 0;
// foreach (double t in tMasses)
// {
// double storedScore;
// if (scores.TryGetValue(t, out storedScore))
// {
// score += storedScore;
// continue;
// }
// double minMZ = t - productTolerance;
// double maxMZ = t + productTolerance;
// while (e < eLength && eMasses[e] < minMZ)
// e++;
// if (e >= eLength)
// break;
// if (eMasses[e] > maxMZ)
// continue;
// double intensities = 0;
// int index = e; // switch variables to keep e the same for the next loop around
// do
// {
// intensities += eIntenisties[index];
// index++;
// } while (index < eLength && eMasses[index] < maxMZ);
// storedScore = 1 + intensities/tic;
// score += storedScore;
// scores[t] = storedScore;
// }
// return score;
//}
/// <summary>
/// The main searching algorithm of Morpheus
/// </summary>
/// <param name="eMasses">The experimental masses</param>
/// <param name="eIntenisties">The experimental intensities</param>
/// <param name="tMasses">The theoretical masses</param>
/// <param name="productTolerance">The product mass tolerance</param>
/// <param name="tic">The total ion current of the experimental peaks</param>
/// <returns></returns>
private double Search(double[] eMasses, double[] eIntenisties, double[] tMasses, Tolerance productTolerance, double tic)
{
double score = 0.0;
double intensities = 0.0;
int eLength = eMasses.Length;
int tLength = tMasses.Length;
int e = 0;
bool forceCheck = productTolerance.GetMinimumValue(tMasses[tLength - 1]) >= eMasses[eLength - 1];
if (forceCheck)
{
foreach (double t in tMasses)
{
IRange <double> range = productTolerance.GetRange(t);
double minMZ = range.Minimum;
double maxMZ = range.Maximum;
while (e < eLength && eMasses[e] < minMZ)
{
e++;
}
if (e >= eLength)
{
break;
}
if (eMasses[e] > maxMZ)
{
continue;
}
score++;
int index = e; // switch variables to keep e the same for the next loop around
do
{
intensities += eIntenisties[index];
index++;
} while (index < eLength && eMasses[index] < maxMZ);
}
}
else
{
foreach (double t in tMasses)
{
IRange <double> range = productTolerance.GetRange(t);
double minMZ = range.Minimum;
double maxMZ = range.Maximum;
while (eMasses[e] < minMZ)
{
e++;
}
if (eMasses[e] > maxMZ)
{
continue;
}
score++;
int index = e; // switch variables to keep e the same for the next loop around
do
{
intensities += eIntenisties[index];
index++;
} while (index < eLength && eMasses[index] < maxMZ);
}
}
return(score + intensities / tic);
}