private string PeptideToString(proteome.Peptide peptide, IList <phosphoRS.PTMSiteProbability> localizationProbabilities, PhosphoRSConfig config)
{
var probabilityMap = localizationProbabilities.ToDictionary(o => o.SequencePosition, o => o.Probability);
string format = String.Format("[{{0:f{0}}}]", 0);
StringBuilder sb = new StringBuilder();
if (peptide.modifications().ContainsKey(proteome.ModificationMap.NTerminus()))
{
sb.AppendFormat(format, peptide.modifications()[proteome.ModificationMap.NTerminus()].monoisotopicDeltaMass());
}
for (int i = 0; i < peptide.sequence.Length; ++i)
{
sb.Append(peptide.sequence[i]);
if (probabilityMap.ContainsKey(i + 1))
{
if (probabilityMap[i + 1] > 0)
{
sb.AppendFormat("[{0:f0}({1:f0}%)]", config.scoredAA.MassDelta, probabilityMap[i + 1] * 100);
}
//else
// sb.AppendFormat("({0:f0})", config.scoredAA.MassDelta, probabilityMap[i + 1]);
}
else if (peptide.modifications().ContainsKey(i))
{
double modMass = peptide.modifications()[i].monoisotopicDeltaMass();
sb.AppendFormat(format, modMass);
}
}
if (peptide.modifications().ContainsKey(proteome.ModificationMap.CTerminus()))
{
sb.AppendFormat(format, peptide.modifications()[proteome.ModificationMap.CTerminus()].monoisotopicDeltaMass());
}
return(sb.ToString());
}
public static void CreateTestData (NHibernate.ISession session, IList<SpectrumTuple> testPsmSummary)
{
var dbGroups = new Map<string, SpectrumSourceGroup>();
foreach (var ssg in session.Query<SpectrumSourceGroup>())
dbGroups[ssg.Name] = ssg;
var dbSources = new Map<long, SpectrumSource>();
foreach (var ss in session.Query<SpectrumSource>())
dbSources[ss.Id.Value] = ss;
var dbAnalyses = new Map<long, Analysis>();
foreach (var a in session.Query<Analysis>())
dbAnalyses[a.Id.Value] = a;
var dbPeptides = new Map<string, Peptide>();
foreach (var pep in session.Query<Peptide>())
dbPeptides[pep.Sequence] = pep;
var bulkInserter = new BulkInserter(session.Connection);
long lastPsmId = session.CreateQuery("SELECT MAX(Id) FROM PeptideSpectrumMatch").UniqueResult<long?>().GetValueOrDefault();
long lastModId = session.CreateQuery("SELECT MAX(Id) FROM Modification").UniqueResult<long?>().GetValueOrDefault();
long lastPmId = session.CreateQuery("SELECT MAX(Id) FROM PeptideModification").UniqueResult<long?>().GetValueOrDefault();
long lastGroupId = session.CreateQuery("SELECT MAX(Id) FROM SpectrumSourceGroup").UniqueResult<long?>().GetValueOrDefault();
long lastSourceId = session.CreateQuery("SELECT MAX(Id) FROM SpectrumSource").UniqueResult<long?>().GetValueOrDefault();
long lastSglId = session.CreateQuery("SELECT MAX(Id) FROM SpectrumSourceGroupLink").UniqueResult<long?>().GetValueOrDefault();
foreach (SpectrumTuple row in testPsmSummary)
{
string groupName = row.Group;
string sourceName = "Source " + row.Source;
string analysisId = "Engine " + row.Analysis;
string peptideTuples = row.PeptideTuples;
SpectrumSourceGroup group = dbGroups[groupName];
if (String.IsNullOrEmpty(group.Name))
{
group.Id = ++lastGroupId;
group.Name = groupName;
bulkInserter.Add(group);
}
SpectrumSource source = dbSources[row.Source];
if (String.IsNullOrEmpty(source.Name))
{
source.Id = ++lastSourceId;
source.Name = sourceName;
source.Group = group;
source.Spectra = new List<Spectrum>();
bulkInserter.Add(source);
// add a source group link for the source's immediate group
bulkInserter.Add(new SpectrumSourceGroupLink() { Id = ++lastSglId, Group = group, Source = source });
#region add source group links for all of the immediate group's parent groups
if (groupName != "/")
{
string parentGroupName = groupName.Substring(0, groupName.LastIndexOf("/"));
while (true)
{
if (String.IsNullOrEmpty(parentGroupName))
parentGroupName = "/";
// add the parent group if it doesn't exist yet
SpectrumSourceGroup parentGroup = session.UniqueResult<SpectrumSourceGroup>(o => o.Name == parentGroupName);
if (parentGroup == null)
{
parentGroup = new SpectrumSourceGroup() { Id = ++lastGroupId, Name = parentGroupName };
bulkInserter.Add(parentGroup);
}
bulkInserter.Add(new SpectrumSourceGroupLink() { Id = ++lastSglId, Group = parentGroup, Source = source });
if (parentGroupName == "/")
break;
parentGroupName = parentGroupName.Substring(0, parentGroupName.LastIndexOf("/"));
}
}
#endregion
}
Spectrum spectrum = source.Spectra.SingleOrDefault(o => o.Source.Id == source.Id &&
o.Index == row.Spectrum - 1);
if (spectrum == null)
{
spectrum = new Spectrum()
{
Id = source.Id * 10000 + row.Spectrum,
Index = row.Spectrum - 1,
NativeID = "scan=" + row.Spectrum,
Source = source,
PrecursorMZ = 42
};
source.Spectra.Add(spectrum);
bulkInserter.Add(spectrum);
}
Analysis analysis = dbAnalyses[row.Analysis];
if (String.IsNullOrEmpty(analysis.Name))
{
analysis.Id = dbAnalyses.Max(o => o.Value.Id).GetValueOrDefault() + 1;
analysis.Name = analysisId + " 1.0";
analysis.Software = new AnalysisSoftware() {Name = analysisId, Version = "1.0"};
analysis.StartTime = DateTime.Today.AddHours(row.Analysis);
analysis.Type = AnalysisType.DatabaseSearch;
analysis.Parameters = new SortedSet<AnalysisParameter>()
{
new AnalysisParameter()
{
Id = analysis.Id * 10000,
Analysis = analysis,
Name = "Parameter 1",
Value = "Value 1"
}
};
bulkInserter.Add(analysis);
}
// make sure peptides are sorted by their score divider (which will determine rank)
var peptideList = new SortedList<int, List<PeptideTuple>>();
foreach (string tuple in peptideTuples.Split(" ".ToCharArray(), StringSplitOptions.RemoveEmptyEntries))
{
var peptideTuple = new PeptideTuple()
{
Sequence = tuple.Split('@', '/')[0],
Charge = Convert.ToInt32(tuple.Split('@', '/')[1]),
ScoreDivider = Convert.ToInt32(tuple.Split('@', '/')[2])
};
if (!peptideList.ContainsKey(peptideTuple.ScoreDivider))
peptideList[peptideTuple.ScoreDivider] = new List<PeptideTuple>();
peptideList[peptideTuple.ScoreDivider].Add(peptideTuple);
}
int rank = 1;
int lastDivider = 1;
foreach (var peptideTupleList in peptideList.Values)
foreach (var peptideTuple in peptideTupleList)
{
using (PwizPeptide pwizPeptide = new PwizPeptide(peptideTuple.Sequence, ModParsing.ModificationParsing_Auto, ModDelimiter.ModificationDelimiter_Brackets))
{
Peptide peptide = dbPeptides[pwizPeptide.sequence];
if (String.IsNullOrEmpty(peptide.Sequence))
{
peptide = new TestPeptide(pwizPeptide.sequence);
peptide.Id = dbPeptides.Max(o => o.Value.Id).GetValueOrDefault() + 1;
peptide.MonoisotopicMass = pwizPeptide.monoisotopicMass(false);
peptide.MolecularWeight = pwizPeptide.molecularWeight(false);
dbPeptides[pwizPeptide.sequence] = peptide;
bulkInserter.Add(peptide);
createTestPeptideInstances(session, bulkInserter, peptide);
}
double neutralPrecursorMass = (spectrum.PrecursorMZ*peptideTuple.Charge) - (peptideTuple.Charge*Proton.Mass);
var psm = new PeptideSpectrumMatch()
{
Id = ++lastPsmId,
Peptide = peptide,
Spectrum = spectrum,
Analysis = analysis,
ObservedNeutralMass = neutralPrecursorMass,
MonoisotopicMassError = neutralPrecursorMass - pwizPeptide.monoisotopicMass(),
MolecularWeightError = neutralPrecursorMass - pwizPeptide.molecularWeight(),
Charge = peptideTuple.Charge,
Rank = (peptideTuple.ScoreDivider == lastDivider ? rank : ++rank),
QValue = (rank == 1 ? row.QValue : PeptideSpectrumMatch.DefaultQValue),
};
if (row.Score != null)
psm.Scores = new Dictionary<string, double>()
{
{"score1", (double) row.Score/peptideTuple.ScoreDivider},
{"score2", 1/((double) row.Score/peptideTuple.ScoreDivider)}
};
bulkInserter.Add(psm);
lastDivider = peptideTuple.ScoreDivider;
// add PeptideModifications and Modifications
foreach (KeyValuePair<int, ModList> itr in pwizPeptide.modifications())
{
foreach (PwizMod pwizMod in itr.Value)
{
Modification mod = session.UniqueResult<Modification>(o => o.Formula == pwizMod.formula());
if (mod == null)
{
mod = new Modification()
{
Id = ++lastModId,
Formula = pwizMod.formula(),
MonoMassDelta = pwizMod.monoisotopicDeltaMass(),
AvgMassDelta = pwizMod.averageDeltaMass(),
Name = pwizMod.formula()
};
bulkInserter.Add(mod);
}
bulkInserter.Add(new PeptideModification()
{
Id = ++lastPmId,
PeptideSpectrumMatch = psm,
Modification = mod,
Offset = itr.Key == ModMap.NTerminus() ? int.MinValue
: itr.Key == ModMap.CTerminus() ? int.MaxValue
: itr.Key
});
}
}
}
}
}
bulkInserter.Execute();
bulkInserter.Reset("");
}
public PhosphoPeptideAttestationRow(object[] queryRow)
{
PSMId = (long)queryRow[0];
SpectrumId = (long)queryRow[1];
SourceName = (string)queryRow[2];
SpectrumNativeID = (string)queryRow[3];
PrecursorMZ = Convert.ToDouble(queryRow[4]);
Charge = Convert.ToInt32(queryRow[5]);
// Build the peptide sequence with modifications. Leave the phospho sites out of the string. They
// will reunited with the string right before the PSM is submitted to phosphoRS. This is necessary
// because phosphoRS requires all phospho sites marked with a single numerical representation across all
// PSMs.
OriginalPhosphoSites = new SortedDictionary<int, long>();
var mods = new Dictionary<int, List<double>>();
string peptideSequence = (string)queryRow[7];
Peptide = new proteome.Peptide(peptideSequence);
var pwizMods = Peptide.modifications();
if (!String.IsNullOrEmpty((string)queryRow[6]))
{
var IdMassDeltaAndOffsetTriplets = ((string)queryRow[6]).Split(',');
foreach (var triplet in IdMassDeltaAndOffsetTriplets)
{
var tokens = triplet.Split(':');
long pmId = Convert.ToInt64(tokens[0]);
double deltaMass = Convert.ToDouble(tokens[1]);
int roundedDeltaMass = (int) Math.Round(deltaMass);
int offset = Convert.ToInt32(tokens[2]);
pwizMods[offset].Add(new proteome.Modification(deltaMass, deltaMass));
if (roundedDeltaMass == 80 && (peptideSequence[offset] == 'S' || peptideSequence[offset] == 'T' || peptideSequence[offset] == 'Y'))
OriginalPhosphoSites[offset] = pmId;
else
{
if (!mods.ContainsKey(offset))
mods[offset] = new List<double>();
mods[offset].Add(deltaMass);
}
}
}
string format = String.Format("[{{0:f{0}}}]", 4);
StringBuilder sb = new StringBuilder(peptideSequence);
foreach (var mod in (from m in mods orderby m.Key descending select m))
foreach (var massDelta in mod.Value)
if (mod.Key == int.MinValue)
sb.Insert(0, String.Format(format, massDelta));
else if (mod.Key == int.MaxValue || mod.Key >= sb.Length)
sb.AppendFormat(format, massDelta);
else
sb.Insert(mod.Key + 1, String.Format(format, massDelta));
UnphosphorylatedSequence = sb.ToString();
DecoyState = Convert.ToInt16(queryRow[8]);
// Determine the location of phosphorylation sites
PossiblePhosphoSites = new List<int>();
for (int residueIndex = 0; residueIndex < peptideSequence.Length; ++residueIndex)
if (peptideSequence[residueIndex] == 'S' || peptideSequence[residueIndex] == 'T' || peptideSequence[residueIndex] == 'Y')
PossiblePhosphoSites.Add(residueIndex);
}
private phosphoRS.PeptideSpectrumMatch getPhosphoRS_PSM(PhosphoRSConfig config, PhosphoPeptideAttestationRow variant)
{
// Get the phosphorylated peptide and add all modifications to the base sequence.
proteome.Peptide phosphoPeptide = new proteome.Peptide(variant.UnphosphorylatedSequence, proteome.ModificationParsing.ModificationParsing_Auto, proteome.ModificationDelimiter.ModificationDelimiter_Brackets);
proteome.ModificationMap variantPeptideMods = phosphoPeptide.modifications();
variant.OriginalPhosphoSites.Keys.ToList().ForEach(location => { variantPeptideMods[location].Add(config.pwizMod); });
// This modification ID is used to tell phosphoRS how to modify the sequence.
int modificationID = config.phosphorylationSymbol + 1;
// Build a string representation of all modificaitons in a peptide for phospoRS
// "0.00011000000000.0" : 1 is the ID of the modification. All phosphos in a data
// set need to have one ID. This ID is used by the PhosphoRS to figure out which
// mods need to be scored.
var ptmRepresentation = new StringBuilder();
// Store all modifications in phosphoRS modification objects
var modifications = new List<phosphoRS.AminoAcidModification>();
// Get the n-terminal modifications.
if (variantPeptideMods.ContainsKey(proteome.ModificationMap.NTerminus()))
{
phosphoRS.AminoAcidModification otherMod = new phosphoRS.AminoAcidModification('2', "unknown", "unk", "none", variantPeptideMods[proteome.ModificationMap.NTerminus()].monoisotopicDeltaMass(), 0.0, null);
modifications.Add(otherMod);
ptmRepresentation.Append(modificationID.ToString() + ".");
//++modificationID;
}
else
{
ptmRepresentation.Append("0.");
}
// Process all other modifications.
for (int aaIndex = 0; aaIndex < phosphoPeptide.sequence.Length; ++aaIndex)
{
// If phosphorylation, use the existing scoredAA variable.
if (variantPeptideMods.ContainsKey(aaIndex))
{
if (variant.OriginalPhosphoSites.Keys.Contains(aaIndex))
{
modifications.Add(config.scoredAA);
ptmRepresentation.Append(config.phosphorylationSymbol.ToString()[0]);
}
else
{
// Otherwise, make an "unknown" modification with a separate modification ID.
var otherMod = new phosphoRS.AminoAcidModification(modificationID.ToString()[0], "unknown", "unk", "none", variantPeptideMods[aaIndex].monoisotopicDeltaMass(), 0.0, phosphoRS.AminoAcidSequence.ParseAASequence("" + phosphoPeptide.sequence[aaIndex]));
modifications.Add(otherMod);
ptmRepresentation.Append(modificationID.ToString());
//++modificationID;
}
}
else
{
ptmRepresentation.Append("0");
}
}
// Process any c-terminal modifications.
if (variantPeptideMods.ContainsKey(proteome.ModificationMap.CTerminus()))
{
var otherMod = new phosphoRS.AminoAcidModification(modificationID.ToString()[0], "unknown", "unk", "none", variantPeptideMods[proteome.ModificationMap.CTerminus()].monoisotopicDeltaMass(), 0.0, null);
modifications.Add(otherMod);
ptmRepresentation.Append("." + modificationID.ToString());
}
else
{
ptmRepresentation.Append(".0");
}
// Get the phosphoRS peptide sequence.
// Assign spectrum ID, amino acid sequence, list of all modifications, a so-called 'modification position string' (here every digit represents an amino acid within the peptide sequence
// '0' indicates not modified, values != '0' indicate the unique identifier of the amino acid's modification the first digit represents the n-terminus the last digit represents the c-terminus)
var AAS = phosphoRS.AminoAcidSequence.Create((int)variant.SpectrumId, phosphoPeptide.sequence, modifications, ptmRepresentation.ToString());
// Make a phosphoRS peptide-spectrum match.
return new phosphoRS.PeptideSpectrumMatch((int)variant.PSMId, variant.SpectrumType, variant.Charge, variant.PrecursorMZ, variant.Peaks, AAS);
}
public PhosphoPeptideAttestationRow(object[] queryRow)
{
PSMId = (long)queryRow[0];
SpectrumId = (long)queryRow[1];
SourceName = (string)queryRow[2];
SpectrumNativeID = (string)queryRow[3];
PrecursorMZ = Convert.ToDouble(queryRow[4]);
Charge = Convert.ToInt32(queryRow[5]);
// Build the peptide sequence with modifications. Leave the phospho sites out of the string. They
// will reunited with the string right before the PSM is submitted to phosphoRS. This is necessary
// because phosphoRS requires all phospho sites marked with a single numerical representation across all
// PSMs.
OriginalPhosphoSites = new SortedDictionary <int, long>();
var mods = new Dictionary <int, List <double> >();
string peptideSequence = (string)queryRow[7];
Peptide = new proteome.Peptide(peptideSequence);
var pwizMods = Peptide.modifications();
if (!String.IsNullOrEmpty((string)queryRow[6]))
{
var IdMassDeltaAndOffsetTriplets = ((string)queryRow[6]).Split(Properties.Settings.Default.GroupConcatSeparator[0]);
foreach (var triplet in IdMassDeltaAndOffsetTriplets)
{
var tokens = triplet.Split(':');
long pmId = Convert.ToInt64(tokens[0]);
double deltaMass = Convert.ToDouble(tokens[1]);
int roundedDeltaMass = (int)Math.Round(deltaMass);
int offset = Convert.ToInt32(tokens[2]);
pwizMods[offset].Add(new proteome.Modification(deltaMass, deltaMass));
if (roundedDeltaMass == 80 && (peptideSequence[offset] == 'S' || peptideSequence[offset] == 'T' || peptideSequence[offset] == 'Y'))
{
OriginalPhosphoSites[offset] = pmId;
}
else
{
if (!mods.ContainsKey(offset))
{
mods[offset] = new List <double>();
}
mods[offset].Add(deltaMass);
}
}
}
string format = String.Format("[{{0:f{0}}}]", 4);
StringBuilder sb = new StringBuilder(peptideSequence);
foreach (var mod in (from m in mods orderby m.Key descending select m))
{
foreach (var massDelta in mod.Value)
{
if (mod.Key == int.MinValue)
{
sb.Insert(0, String.Format(format, massDelta));
}
else if (mod.Key == int.MaxValue || mod.Key >= sb.Length)
{
sb.AppendFormat(format, massDelta);
}
else
{
sb.Insert(mod.Key + 1, String.Format(format, massDelta));
}
}
}
UnphosphorylatedSequence = sb.ToString();
DecoyState = Convert.ToInt16(queryRow[8]);
// Determine the location of phosphorylation sites
PossiblePhosphoSites = new List <int>();
for (int residueIndex = 0; residueIndex < peptideSequence.Length; ++residueIndex)
{
if (peptideSequence[residueIndex] == 'S' || peptideSequence[residueIndex] == 'T' || peptideSequence[residueIndex] == 'Y')
{
PossiblePhosphoSites.Add(residueIndex);
}
}
}
private phosphoRS.PeptideSpectrumMatch getPhosphoRS_PSM(PhosphoRSConfig config, PhosphoPeptideAttestationRow variant)
{
// Get the phosphorylated peptide and add all modifications to the base sequence.
proteome.Peptide phosphoPeptide = new proteome.Peptide(variant.UnphosphorylatedSequence, proteome.ModificationParsing.ModificationParsing_Auto, proteome.ModificationDelimiter.ModificationDelimiter_Brackets);
proteome.ModificationMap variantPeptideMods = phosphoPeptide.modifications();
variant.OriginalPhosphoSites.Keys.ToList().ForEach(location => { variantPeptideMods[location].Add(config.pwizMod); });
// This modification ID is used to tell phosphoRS how to modify the sequence.
int modificationID = config.phosphorylationSymbol + 1;
// Build a string representation of all modificaitons in a peptide for phospoRS
// "0.00011000000000.0" : 1 is the ID of the modification. All phosphos in a data
// set need to have one ID. This ID is used by the PhosphoRS to figure out which
// mods need to be scored.
var ptmRepresentation = new StringBuilder();
// Store all modifications in phosphoRS modification objects
var modifications = new List <phosphoRS.AminoAcidModification>();
// Get the n-terminal modifications.
if (variantPeptideMods.ContainsKey(proteome.ModificationMap.NTerminus()))
{
phosphoRS.AminoAcidModification otherMod = new phosphoRS.AminoAcidModification('2', "unknown", "unk", "none", variantPeptideMods[proteome.ModificationMap.NTerminus()].monoisotopicDeltaMass(), 0.0, null);
modifications.Add(otherMod);
ptmRepresentation.Append(modificationID.ToString() + ".");
//++modificationID;
}
else
{
ptmRepresentation.Append("0.");
}
// Process all other modifications.
for (int aaIndex = 0; aaIndex < phosphoPeptide.sequence.Length; ++aaIndex)
{
// If phosphorylation, use the existing scoredAA variable.
if (variantPeptideMods.ContainsKey(aaIndex))
{
if (variant.OriginalPhosphoSites.Keys.Contains(aaIndex))
{
modifications.Add(config.scoredAA);
ptmRepresentation.Append(config.phosphorylationSymbol.ToString()[0]);
}
else
{
// Otherwise, make an "unknown" modification with a separate modification ID.
var otherMod = new phosphoRS.AminoAcidModification(modificationID.ToString()[0], "unknown", "unk", "none", variantPeptideMods[aaIndex].monoisotopicDeltaMass(), 0.0, phosphoRS.AminoAcidSequence.ParseAASequence("" + phosphoPeptide.sequence[aaIndex]));
modifications.Add(otherMod);
ptmRepresentation.Append(modificationID.ToString());
//++modificationID;
}
}
else
{
ptmRepresentation.Append("0");
}
}
// Process any c-terminal modifications.
if (variantPeptideMods.ContainsKey(proteome.ModificationMap.CTerminus()))
{
var otherMod = new phosphoRS.AminoAcidModification(modificationID.ToString()[0], "unknown", "unk", "none", variantPeptideMods[proteome.ModificationMap.CTerminus()].monoisotopicDeltaMass(), 0.0, null);
modifications.Add(otherMod);
ptmRepresentation.Append("." + modificationID.ToString());
}
else
{
ptmRepresentation.Append(".0");
}
// Get the phosphoRS peptide sequence.
// Assign spectrum ID, amino acid sequence, list of all modifications, a so-called 'modification position string' (here every digit represents an amino acid within the peptide sequence
// '0' indicates not modified, values != '0' indicate the unique identifier of the amino acid's modification the first digit represents the n-terminus the last digit represents the c-terminus)
var AAS = phosphoRS.AminoAcidSequence.Create((int)variant.SpectrumId, phosphoPeptide.sequence, modifications, ptmRepresentation.ToString());
// Make a phosphoRS peptide-spectrum match.
return(new phosphoRS.PeptideSpectrumMatch((int)variant.PSMId, variant.SpectrumType, variant.Charge, variant.PrecursorMZ, variant.Peaks, AAS));
}