private static float GetSSRCalcHydrophobicityZScore(PeptideSpectralMatch psm, PeptideWithSetModifications Peptide, Dictionary <string, Dictionary <int, Tuple <double, double> > > d)
{
//Using SSRCalc3 but probably any number of different calculators could be used instead. One could also use the CE mobility.
SSRCalc3 calc = new SSRCalc3("SSRCalc 3.0 (300A)", SSRCalc3.Column.A300);
double hydrophobicityZscore = double.NaN;
if (d.ContainsKey(psm.FullFilePath))
{
int time = (int)Math.Round(psm.ScanRetentionTime, 0);
if (d[psm.FullFilePath].Keys.Contains(time))
{
double predictedHydrophobicity = calc.ScoreSequence(Peptide);
hydrophobicityZscore = Math.Abs(d[psm.FullFilePath][time].Item1 - predictedHydrophobicity) / d[psm.FullFilePath][time].Item2;
}
}
if (double.IsNaN(hydrophobicityZscore) || double.IsInfinity(hydrophobicityZscore))
{
hydrophobicityZscore = 100;
}
return((float)hydrophobicityZscore);
}
public static PsmData CreateOnePsmDataEntry(PeptideSpectralMatch psm, Dictionary <string, int> sequenceToPsmCount, Dictionary <string, Dictionary <int, Tuple <double, double> > > timeDependantHydrophobicityAverageAndDeviation_unmodified, Dictionary <string, Dictionary <int, Tuple <double, double> > > timeDependantHydrophobicityAverageAndDeviation_modified, int chargeStateMode, PeptideWithSetModifications selectedPeptide, string[] trainingVariables, int notchToUse, bool label)
{
float totalMatchingFragmentCount = 0;
if (trainingVariables.Contains("TotalMatchingFragmentCount"))
{
totalMatchingFragmentCount = (float)Math.Floor(psm.Score);
}
float ambiguity = 0;
if (trainingVariables.Contains("Ambiguity"))
{
ambiguity = Math.Min((float)(psm.PeptidesToMatchingFragments.Keys.Count - 1), 10);
}
float intensity = 0;
if (trainingVariables.Contains("Intensity"))
{
intensity = (float)(psm.Score - (int)psm.Score);
}
float chargeDifference = 0;
if (trainingVariables.Contains("PrecursorChargeDiffToMode"))
{
chargeDifference = -Math.Abs(chargeStateMode - psm.ScanPrecursorCharge);
}
float deltaScore = 0;
if (trainingVariables.Contains("DeltaScore"))
{
deltaScore = (float)psm.DeltaScore;
}
float psmCount = 1;
if (trainingVariables.Contains("PsmCount"))
{
psmCount = sequenceToPsmCount[String.Join("|", psm.BestMatchingPeptides.Select(p => p.Peptide.FullSequence).ToList())];
//grouping psm counts as follows is done for stability. you get very nice numbers at low psms to get good statistics. But you get a few peptides with high psm counts that could be either targets or decoys and the values swing between extremes. So grouping psms in bundles really adds stability.
List <int> psmCountList = new List <int> {
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 75, 100, 200, 300, 400, 500
};
int closest = psmCountList.OrderBy(item => Math.Abs(psmCount - item)).First();
psmCount = closest;
}
int notch = 0;
if (trainingVariables.Contains("Notch"))
{
notch = notchToUse;
}
float modCount = 0;
if (trainingVariables.Contains("ModsCount"))
{
modCount = Math.Min((float)selectedPeptide.AllModsOneIsNterminus.Keys.Count(), 10);
}
float missedCleavages = 0;
if (trainingVariables.Contains("MissedCleavagesCount"))
{
missedCleavages = selectedPeptide.MissedCleavages;
}
float longestSeq = 0;
if (trainingVariables.Contains("LongestFragmentIonSeries"))
{
longestSeq = psm.GetLongestIonSeriesBidirectional(selectedPeptide);
}
float hydrophobicityZscore = float.NaN;
if (selectedPeptide.BaseSequence.Equals(selectedPeptide.FullSequence) && trainingVariables.Contains("HydrophobicityZScore"))
{
hydrophobicityZscore = GetSSRCalcHydrophobicityZScore(psm, selectedPeptide, timeDependantHydrophobicityAverageAndDeviation_unmodified);
}
else if (trainingVariables.Contains("HydrophobicityZScore"))
{
hydrophobicityZscore = GetSSRCalcHydrophobicityZScore(psm, selectedPeptide, timeDependantHydrophobicityAverageAndDeviation_modified);
}
bool isVariantPeptide = PeptideIsVariant(selectedPeptide);
if (psm.IsDecoy)
{
label = false;
}
else
{
label = true;
}
psm.PsmData_forPEPandPercolator = new PsmData
{
TotalMatchingFragmentCount = totalMatchingFragmentCount,
Intensity = intensity,
PrecursorChargeDiffToMode = chargeDifference,
DeltaScore = deltaScore,
Notch = notch,
PsmCount = psmCount,
ModsCount = modCount,
MissedCleavagesCount = missedCleavages,
Ambiguity = ambiguity,
LongestFragmentIonSeries = longestSeq,
HydrophobicityZScore = hydrophobicityZscore,
IsVariantPeptide = Convert.ToSingle(isVariantPeptide),
Label = label
};
return(psm.PsmData_forPEPandPercolator);
}
/// <summary>
///
/// </summary>
/// <param name="psm"></param>
/// <param name="sequenceToPsmCount"></param>
/// <param name="selectedPeptide"></param>
/// <param name="notchToUse"></param>
/// <param name="trueOrFalse"></param>
/// <returns></returns>
public static PsmData CreateOnePsmDataEntry(PeptideSpectralMatch psm, Dictionary <string, int> sequenceToPsmCount, Dictionary <string, Dictionary <int, Tuple <double, double> > > timeDependantHydrophobicityAverageAndDeviation_unmodified, Dictionary <string, Dictionary <int, Tuple <double, double> > > timeDependantHydrophobicityAverageAndDeviation_modified, PeptideWithSetModifications selectedPeptide, string searchType, int?notchToUse, bool?trueOrFalse = null)
{
float ambiguity = (float)psm.PeptidesToMatchingFragments.Keys.Count;
float intensity = (float)(psm.Score - (int)psm.Score);
float charge = psm.ScanPrecursorCharge;
float deltaScore = (float)psm.DeltaScore;
float psmCount = sequenceToPsmCount[String.Join("|", psm.BestMatchingPeptides.Select(p => p.Peptide.FullSequence).ToList())];
int notch = 0;
if (notchToUse.HasValue)
{
notch = notchToUse.Value;
}
else if (psm.Notch.HasValue)
{
notch = psm.Notch.Value;
}
if (selectedPeptide == null)
{
selectedPeptide = psm.BestMatchingPeptides.Select(p => p.Peptide).First();
}
float modCount = selectedPeptide.AllModsOneIsNterminus.Keys.Count();
float missedCleavages = selectedPeptide.MissedCleavages;
float longestSeq = psm.GetLongestIonSeriesBidirectional(selectedPeptide);
float hydrophobicityZscore = float.NaN;
if (selectedPeptide.BaseSequence.Equals(selectedPeptide.FullSequence) && searchType == "standard")
{
hydrophobicityZscore = GetSSRCalcHydrophobicityZScore(psm, selectedPeptide, timeDependantHydrophobicityAverageAndDeviation_unmodified);
}
else if (searchType == "standard")
{
hydrophobicityZscore = GetSSRCalcHydrophobicityZScore(psm, selectedPeptide, timeDependantHydrophobicityAverageAndDeviation_modified);
}
bool label;
if (trueOrFalse != null)
{
label = trueOrFalse.Value;
}
else if (psm.IsDecoy)
{
label = false;
}
else
{
label = true;
}
return(new PsmData()
{
Intensity = intensity,
ScanPrecursorCharge = charge,
DeltaScore = deltaScore,
Notch = notch,
PsmCount = psmCount,
ModsCount = modCount,
MissedCleavagesCount = missedCleavages,
Ambiguity = ambiguity,
LongestFragmentIonSeries = longestSeq,
HydrophobicityZScore = hydrophobicityZscore,
Label = label
});
}
private static void AddMatchedIonsData(Dictionary <string, string> s, PeptideSpectralMatch peptide)
{
//sb for writing, format for double.ToString, header for input dictionary key
(StringBuilder sb, string format, string header)[] matchedIonInfo = new(StringBuilder, string, string)[]
public static PsmData CreateOnePsmDataFromPsm(PeptideSpectralMatch psm, Dictionary <string, int> accessionCounts, Dictionary <string, int> sequenceToPsmCount, bool?trueOrFalse = null)
{
// TODO: some properties like DeltaScore will need to be recalculated if we keep top N peptides per PSM
// and rerank them because the top-scoring peptide can change
float ambiguity = (float)psm.PeptidesToMatchingFragments.Count;//(psm.BaseSequence.Split('|').Count());
float intensity = (float)(psm.Score - (int)psm.Score);
float charge = psm.ScanPrecursorCharge;
float deltaScore = (float)psm.DeltaScore;
float psmCount = sequenceToPsmCount[String.Join("|", psm.BestMatchingPeptides.Select(p => p.Peptide.FullSequence).ToList())];
var firstPeptide = psm.BestMatchingPeptides.Select(p => p.Peptide).First();
float modCount = firstPeptide.AllModsOneIsNterminus.Keys.Count();
float notch = 0;
if (psm.Notch.HasValue)
{
notch = psm.Notch.Value;
}
//todo: for non-specific cleavage, ignore missed cleavages
float missedCleavages = firstPeptide.MissedCleavages;
float longestSeq = psm.GetLongestIonSeriesBidirectional(firstPeptide);
string accession = firstPeptide.Protein.Accession;
float appearances;
if (accessionCounts.Keys.Count != 0 && accessionCounts.ContainsKey(accession))
{
appearances = (float)accessionCounts[accession];
}
else
{
appearances = 1;
}
float score = (float)psm.Score;
bool label;
if (trueOrFalse != null)
{
label = trueOrFalse.Value;
}
else if (psm.IsDecoy)
{
label = false;
}
else
{
label = true;
}
return(new PsmData()
{
Intensity = intensity,
ScanPrecursorCharge = charge,
DeltaScore = deltaScore,
Notch = notch,
PsmCount = psmCount,
ModsCount = modCount,
MissedCleavagesCount = missedCleavages,
Ambiguity = ambiguity,
LongestFragmentIonSeries = longestSeq,
AccessionAppearances = appearances,
Label = label
});
}