/// <summary>
/// Get the number of proteoform sequences possible with the provided number of dynamic modifications
/// </summary>
/// <param name="nDynMods"></param>
/// <returns></returns>
public double GetNumProteoformSequencesByNumMods(int nDynMods)
{
if (nDynMods < 1 || nDynMods > ModificationParams.MaxNumDynModsPerSequence)
{
return(0);
}
var numModCombs = _graph[_index].Length;
var nProteoforms = 0d;
for (var modIndex = 0; modIndex < numModCombs; modIndex++)
{
var modCombs = ModificationParams.GetModificationCombination(_graph[_index][modIndex].ModificationCombinationIndex);
if (modCombs.GetNumModifications() < nDynMods)
{
continue;
}
if (modCombs.GetNumModifications() > nDynMods)
{
break;
}
nProteoforms += GetNumProteoformSequences(modIndex);
//if (nProteoforms > double.MaxValue) return double.MaxValue; // overflow
}
return(nProteoforms);
}
/// <summary>
/// Get the modification combinations in the current sequence graph
/// </summary>
/// <returns></returns>
public ModificationCombination[] GetModificationCombinations()
{
var numModCombs = _graph[_index].Length;
var modCombs = new ModificationCombination[numModCombs];
for (var modIndex = 0; modIndex < numModCombs; modIndex++)
{
modCombs[modIndex] = ModificationParams.GetModificationCombination(_graph[_index][modIndex].ModificationCombinationIndex);
}
return(modCombs);
}
/// <summary>
/// Generate an amino acid map with 20 standard amino acids
/// </summary>
public AminoAcidSet()
{
// Initialization
_modificationParams = new ModificationParams();
SearchModifications = new List <SearchModification>();
_locationSpecificResidueMap = new Dictionary <SequenceLocation, Dictionary <char, AminoAcid> >();
_locationSpecificResidueModMap = new Dictionary <SequenceLocation, Dictionary <char, int[]> >();
foreach (var loc in AllSequenceLocations)
{
_locationSpecificResidueModMap[loc] = new Dictionary <char, int[]>();
}
// Initialize standard amino acid set
foreach (var loc in AllSequenceLocations)
{
var residueMap = new Dictionary <char, AminoAcid>();
switch (loc)
{
case SequenceLocation.PeptideNTerm:
residueMap.Add(AminoAcid.PeptideNTerm.Residue, AminoAcid.PeptideNTerm);
break;
case SequenceLocation.PeptideCTerm:
residueMap.Add(AminoAcid.PeptideCTerm.Residue, AminoAcid.PeptideCTerm);
break;
case SequenceLocation.ProteinNTerm:
residueMap.Add(AminoAcid.ProteinNTerm.Residue, AminoAcid.ProteinNTerm);
break;
case SequenceLocation.ProteinCTerm:
residueMap.Add(AminoAcid.ProteinCTerm.Residue, AminoAcid.ProteinCTerm);
break;
}
// loc is SequenceLocation.Everywhere
foreach (var aa in AminoAcid.StandardAminoAcidArr)
{
residueMap.Add(aa.Residue, aa);
}
_locationSpecificResidueMap[loc] = residueMap;
}
}
// Generate an amino acid map with 20 standard amino acids
public AminoAcidSet()
{
// Initialization
_modificationParams = new ModificationParams();
_locationSpecificResidueMap = new Dictionary<SequenceLocation, Dictionary<char, AminoAcid>>();
_locationSpecificResidueModMap = new Dictionary<SequenceLocation, Dictionary<char, int[]>>();
foreach (var loc in AllSequenceLocations)
{
_locationSpecificResidueModMap[loc] = new Dictionary<char, int[]>();
}
// Initialize standard amino acid set
foreach (var loc in AllSequenceLocations)
{
var residueMap = new Dictionary<char, AminoAcid>();
switch (loc)
{
case SequenceLocation.PeptideNTerm:
residueMap.Add(AminoAcid.PeptideNTerm.Residue, AminoAcid.PeptideNTerm);
break;
case SequenceLocation.PeptideCTerm:
residueMap.Add(AminoAcid.PeptideCTerm.Residue, AminoAcid.PeptideCTerm);
break;
case SequenceLocation.ProteinNTerm:
residueMap.Add(AminoAcid.ProteinNTerm.Residue, AminoAcid.ProteinNTerm);
break;
case SequenceLocation.ProteinCTerm:
residueMap.Add(AminoAcid.ProteinCTerm.Residue, AminoAcid.ProteinCTerm);
break;
}
foreach (var aa in AminoAcid.StandardAminoAcidArr)
{
residueMap.Add(aa.Residue, aa);
}
_locationSpecificResidueMap[loc] = residueMap;
}
}
/// <summary>
/// Add an amino acid residue to this generator.
/// </summary>
/// <param name="index">index to add the amino acid. 0 is C-term. 1 is the C-term amino acid.</param>
/// <param name="residue">amino acid residue to add.</param>
/// <returns>true if residue is a valid amino acid; false otherwise.</returns>
private bool PutAminoAcid(int index, char residue)
{
_index = index + 1;
SequenceLocation?location = null;
if (_index == 1) // C-term residue
{
if (residue == AminoAcid.PeptideCTerm.Residue)
{
location = SequenceLocation.PeptideCTerm;
}
else if (residue == AminoAcid.ProteinCTerm.Residue)
{
location = SequenceLocation.ProteinCTerm;
}
}
else if (_index == _aminoAcidSequence.Length - 1 - NumNTermCleavages) // N-term residue
{
if (residue == AminoAcid.PeptideNTerm.Residue)
{
location = SequenceLocation.PeptideNTerm;
}
else if (residue == AminoAcid.ProteinNTerm.Residue)
{
location = SequenceLocation.ProteinNTerm;
}
}
else if (_index == 2) // Amino acid at the C-term
{
if (_aminoAcidSequence[1] == AminoAcid.PeptideCTerm)
{
location = SequenceLocation.PeptideCTerm;
}
else if (_aminoAcidSequence[1] == AminoAcid.ProteinCTerm)
{
location = SequenceLocation.ProteinCTerm;
}
}
else if (_index == _aminoAcidSequence.Length - 2 - NumNTermCleavages) // Amino acid at the N-term
{
if (_aminoAcidSequence[_aminoAcidSequence.Length - 1] == AminoAcid.PeptideNTerm)
{
location = SequenceLocation.PeptideNTerm;
}
else if (_aminoAcidSequence[_aminoAcidSequence.Length - 1] == AminoAcid.ProteinNTerm)
{
location = SequenceLocation.ProteinNTerm;
}
}
else
{
location = SequenceLocation.Everywhere;
}
if (location == null)
{
return(false);
}
var loc = (SequenceLocation)location;
var aminoAcid = AminoAcidSet.GetAminoAcid(residue, loc);
if (aminoAcid == null) // residue is not valid
{
return(false);
}
_aminoAcidSequence[_index] = aminoAcid;
_suffixComposition[_index] = _suffixComposition[_index - 1] + aminoAcid.Composition;
var modIndices = AminoAcidSet.GetModificationIndices(residue, loc);
if (!modIndices.Any()) // No modification
{
_graph[_index] = new Node[_graph[_index - 1].Length];
for (var i = 0; i < _graph[_index - 1].Length; i++)
{
_graph[_index][i] = new Node(_graph[_index - 1][i].ModificationCombinationIndex, i);
}
}
else
{
var modCombIndexToNodeMap = new Dictionary <int, Node>();
for (var i = 0; i < _graph[_index - 1].Length; i++)
{
var prevNodeIndex = i;
var prevNode = _graph[_index - 1][i];
var prevModCombIndex = prevNode.ModificationCombinationIndex;
// unmodified edge
if (modCombIndexToNodeMap.TryGetValue(prevModCombIndex, out var unmodifiedEdgeNode))
{
unmodifiedEdgeNode.AddPrevNodeIndex(prevNodeIndex);
}
else
{
modCombIndexToNodeMap.Add(prevModCombIndex, new Node(prevModCombIndex, prevNodeIndex));
}
// modified edge
foreach (var modIndex in modIndices)
{
var modCombIndex = ModificationParams.GetModificationCombinationIndex(
prevNode.ModificationCombinationIndex, modIndex);
if (modCombIndex < 0) // too many modifications
{
continue;
}
if (modCombIndexToNodeMap.TryGetValue(modCombIndex, out var modifiedEdgeNode))
{
modifiedEdgeNode.AddPrevNodeIndex(prevNodeIndex);
}
else
{
modCombIndexToNodeMap.Add(modCombIndex, new Node(modCombIndex, prevNodeIndex));
}
}
_graph[_index] = modCombIndexToNodeMap.Values.ToArray();
}
}
return(true);
}
private Tuple <double, string> GetFragmentScoreAndModifications(int seqIndex, int modIndex, IScorer scorer, double?[][] nodeScore, Tuple <double, string>[][] maxScoreAndMods)
{
var scoreAndMods = maxScoreAndMods[seqIndex][modIndex];
if (scoreAndMods != null)
{
return(scoreAndMods);
}
var node = _graph[seqIndex][modIndex];
var curNodeScore = nodeScore[seqIndex][modIndex] ??
(nodeScore[seqIndex][modIndex] = scorer.GetFragmentScore(GetComplementaryComposition(seqIndex, modIndex), GetComposition(seqIndex, modIndex)));
var bestPrevNodeIndex = -1;
var bestPrevNodeScore = double.NegativeInfinity;
string bestPrevSequence = null;
foreach (var prevNodeIndex in node.GetPrevNodeIndices())
{
var prevNodeScoreAndSequence = GetFragmentScoreAndModifications(seqIndex - 1, prevNodeIndex, scorer,
nodeScore, maxScoreAndMods);
var prevNodeScore = prevNodeScoreAndSequence.Item1;
if (prevNodeScore > bestPrevNodeScore)
{
bestPrevNodeIndex = prevNodeIndex;
bestPrevNodeScore = prevNodeScore;
bestPrevSequence = prevNodeScoreAndSequence.Item2;
}
}
if (bestPrevNodeIndex < 0) // source
{
return(maxScoreAndMods[seqIndex][modIndex] = new Tuple <double, string>((double)curNodeScore, ""));
}
var modPos = _index - seqIndex;
var aminoAcid = _aminoAcidSequence[seqIndex];
var modAa = aminoAcid as ModifiedAminoAcid;
if (modAa != null)
{
var modificationName = modAa.Modification.Name;
if (string.IsNullOrEmpty(bestPrevSequence))
{
bestPrevSequence = modificationName + " " + modPos;
}
else
{
bestPrevSequence = modificationName + " " + modPos + "," + bestPrevSequence;
}
}
var prevModCombIndex = _graph[seqIndex - 1][bestPrevNodeIndex].ModificationCombinationIndex;
var curModCombIndex = node.ModificationCombinationIndex;
if (prevModCombIndex != curModCombIndex) // modified
{
var modificationName = ModificationParams.GetModificationIndexBetween(prevModCombIndex, curModCombIndex).Name;
string newModSequence;
if (string.IsNullOrEmpty(bestPrevSequence))
{
newModSequence = modificationName + " " + modPos;
}
else
{
newModSequence = modificationName + " " + modPos + ",";
}
return(maxScoreAndMods[seqIndex][modIndex] = new Tuple <double, string>
((double)curNodeScore + bestPrevNodeScore, newModSequence + bestPrevSequence));
}
return(maxScoreAndMods[seqIndex][modIndex] = new Tuple <double, string>
((double)curNodeScore + bestPrevNodeScore, bestPrevSequence));
}
public AminoAcidSet(IEnumerable <SearchModification> searchModifications, int maxNumModsPerSequence) : this()
{
if (searchModifications == null)
{
return;
}
var modifications = searchModifications as SearchModification[] ?? searchModifications.ToArray();
// apply fixed modifications
foreach (var searchModification in modifications)
{
if (!searchModification.IsFixedModification)
{
continue;
}
var location = searchModification.Location;
var targetResidue = searchModification.TargetResidue;
foreach (var loc in AffectedLocations[location])
{
var appliedResidue = targetResidue != '*' ? targetResidue : SequenceLocationToLocationResidue[loc];
var residueMap = _locationSpecificResidueMap[loc];
residueMap[appliedResidue] = new ModifiedAminoAcid(residueMap[appliedResidue], searchModification.Modification);
}
}
if (maxNumModsPerSequence <= 0)
{
return;
}
// apply dynamic modifications
var dynamicModifications = new HashSet <Modification>();
var locationSpecificResidueVariableModMap = new Dictionary <SequenceLocation, Dictionary <char, List <Modification> > >();
foreach (var loc in AllSequenceLocations)
{
locationSpecificResidueVariableModMap[loc] = new Dictionary <char, List <Modification> >();
}
foreach (var searchModification in modifications)
{
if (searchModification.IsFixedModification)
{
continue;
}
dynamicModifications.Add(searchModification.Modification);
var location = searchModification.Location;
var targetResidue = searchModification.TargetResidue;
foreach (var loc in AffectedLocations[location])
{
var residueModMap = locationSpecificResidueVariableModMap[loc];
var appliedResidue = targetResidue != '*' ? targetResidue : SequenceLocationToLocationResidue[loc];
List <Modification> modList;
if (residueModMap.TryGetValue(appliedResidue, out modList))
{
modList.Add(searchModification.Modification);
}
else
{
residueModMap.Add(appliedResidue, new List <Modification> {
searchModification.Modification
});
}
}
}
var dynModArray = dynamicModifications.ToArray();
_modificationParams = new ModificationParams(dynModArray, maxNumModsPerSequence);
foreach (var loc in AllSequenceLocations)
{
var residueModMap = locationSpecificResidueVariableModMap[loc];
foreach (var entry in residueModMap)
{
var residue = entry.Key;
var modList = entry.Value;
var modIndexArr = new int[modList.Count];
var index = -1;
foreach (var mod in modList)
{
var modIndex = Array.IndexOf(dynModArray, mod);
modIndexArr[++index] = modIndex;
}
_locationSpecificResidueModMap[loc].Add(residue, modIndexArr);
}
}
}
public AminoAcidSet(IEnumerable<SearchModification> searchModifications, int maxNumModsPerSequence): this()
{
if (searchModifications == null) return;
var modifications = searchModifications as SearchModification[] ?? searchModifications.ToArray();
// apply fixed modifications
foreach (var searchModification in modifications)
{
if (!searchModification.IsFixedModification) continue;
var location = searchModification.Location;
var targetResidue = searchModification.TargetResidue;
foreach (var loc in AffectedLocations[location])
{
var appliedResidue = targetResidue != '*' ? targetResidue : SequenceLocationToLocationResidue[loc];
var residueMap = _locationSpecificResidueMap[loc];
residueMap[appliedResidue] = new ModifiedAminoAcid(residueMap[appliedResidue], searchModification.Modification);
}
}
if (maxNumModsPerSequence <= 0) return;
// apply dynamic modifications
var dynamicModifications = new HashSet<Modification>();
var locationSpecificResidueVariableModMap = new Dictionary<SequenceLocation,Dictionary<char, List<Modification>>>();
foreach (var loc in AllSequenceLocations)
{
locationSpecificResidueVariableModMap[loc] = new Dictionary<char, List<Modification>>();
}
foreach (var searchModification in modifications)
{
if (searchModification.IsFixedModification) continue;
dynamicModifications.Add(searchModification.Modification);
var location = searchModification.Location;
var targetResidue = searchModification.TargetResidue;
foreach (var loc in AffectedLocations[location])
{
var residueModMap = locationSpecificResidueVariableModMap[loc];
var appliedResidue = targetResidue != '*' ? targetResidue : SequenceLocationToLocationResidue[loc];
List<Modification> modList;
if (residueModMap.TryGetValue(appliedResidue, out modList))
{
modList.Add(searchModification.Modification);
}
else
{
residueModMap.Add(appliedResidue, new List<Modification> { searchModification.Modification });
}
}
}
var dynModArray = dynamicModifications.ToArray();
_modificationParams = new ModificationParams(dynModArray, maxNumModsPerSequence);
foreach (var loc in AllSequenceLocations)
{
var residueModMap = locationSpecificResidueVariableModMap[loc];
foreach (var entry in residueModMap)
{
var residue = entry.Key;
var modList = entry.Value;
var modIndexArr = new int[modList.Count];
var index = -1;
foreach (var mod in modList)
{
var modIndex = Array.IndexOf(dynModArray, mod);
modIndexArr[++index] = modIndex;
}
_locationSpecificResidueModMap[loc].Add(residue, modIndexArr);
}
}
}
/// <summary>
/// Add an amino acid residue to this generator.
/// </summary>
/// <param name="index">index to add the amino acid. 0 is C-term. 1 is the C-term amino acid.</param>
/// <param name="residue">amino acid residue to add.</param>
/// <param name="loc">location of the residue</param>
/// <returns>true if residue is a valid amino acid; false otherwise.</returns>
private bool PutAminoAcid(int index, char residue, SequenceLocation loc)
{
_index = index + 1;
var aminoAcid = AminoAcidSet.GetAminoAcid(residue, loc);
if (aminoAcid == null) // residue is not valid
{
return(false);
}
var fragmentComposition = _fragmentComposition[_index - 1] + aminoAcid.Composition;
if (fragmentComposition.Mass > _maxSequenceMass)
{
return(false);
}
_aminoAcidSequence[_index] = aminoAcid;
_fragmentComposition[_index] = fragmentComposition;
var modIndices = AminoAcidSet.GetModificationIndices(residue, loc);
if (!modIndices.Any()) // No modification
{
_graph[_index] = new Node[_graph[_index - 1].Length];
for (var i = 0; i < _graph[_index - 1].Length; i++)
{
_graph[_index][i] = new Node(_graph[_index - 1][i].ModificationCombinationIndex, i);
}
}
else
{
var modCombIndexToNodeMap = new Dictionary <int, Node>();
for (var i = 0; i < _graph[_index - 1].Length; i++)
{
var prevNodeIndex = i;
var prevNode = _graph[_index - 1][i];
var prevModCombIndex = prevNode.ModificationCombinationIndex;
// unmodified edge
if (modCombIndexToNodeMap.TryGetValue(prevModCombIndex, out var unmodifiedEdgeNode))
{
unmodifiedEdgeNode.AddPrevNodeIndex(prevNodeIndex);
}
else
{
modCombIndexToNodeMap.Add(prevModCombIndex, new Node(prevModCombIndex, prevNodeIndex));
}
// modified edge
foreach (var modIndex in modIndices)
{
var modCombIndex = ModificationParams.GetModificationCombinationIndex(
prevNode.ModificationCombinationIndex, modIndex);
if (modCombIndex < 0) // too many modifications
{
continue;
}
if (modCombIndexToNodeMap.TryGetValue(modCombIndex, out var modifiedEdgeNode))
{
modifiedEdgeNode.AddPrevNodeIndex(prevNodeIndex);
}
else
{
modCombIndexToNodeMap.Add(modCombIndex, new Node(modCombIndex, prevNodeIndex));
}
}
_graph[_index] = modCombIndexToNodeMap.Values.ToArray();
}
}
return(true);
}
/// <summary>
/// Get the fragment score and modifications
/// </summary>
/// <param name="seqIndex"></param>
/// <param name="modIndex"></param>
/// <param name="scorer"></param>
/// <param name="nodeScore"></param>
/// <param name="maxScoreAndMods"></param>
/// <returns></returns>
private Tuple <double, LinkedList <ModificationInstance> > GetFragmentScoreAndModifications(
int seqIndex,
int modIndex,
IScorer scorer,
IReadOnlyList <double?[]> nodeScore,
IReadOnlyList <Tuple <double, LinkedList <ModificationInstance> >[]> maxScoreAndMods)
{
var scoreAndMods = maxScoreAndMods[seqIndex][modIndex];
if (scoreAndMods != null)
{
return(scoreAndMods);
}
var node = _graph[seqIndex][modIndex];
var curNodeScore = nodeScore[seqIndex][modIndex] ??
(nodeScore[seqIndex][modIndex] = _isForward
? scorer.GetFragmentScore(GetComposition(seqIndex, modIndex),
GetComplementaryComposition(seqIndex, modIndex))
: scorer.GetFragmentScore(GetComplementaryComposition(seqIndex, modIndex),
GetComposition(seqIndex, modIndex))
);
var bestPrevNodeIndex = -1;
var bestPrevNodeScore = double.NegativeInfinity;
LinkedList <ModificationInstance> bestPrevMods = null;
foreach (var prevNodeIndex in node.GetPrevNodeIndices())
{
var prevNodeScoreAndSequence = GetFragmentScoreAndModifications(seqIndex - 1, prevNodeIndex, scorer,
nodeScore, maxScoreAndMods);
var prevNodeScore = prevNodeScoreAndSequence.Item1;
if (prevNodeScore > bestPrevNodeScore)
{
bestPrevNodeIndex = prevNodeIndex;
bestPrevNodeScore = prevNodeScore;
bestPrevMods = prevNodeScoreAndSequence.Item2;
}
}
if (bestPrevNodeIndex < 0) // source
{
return(maxScoreAndMods[seqIndex][modIndex] = new Tuple <double, LinkedList <ModificationInstance> >((double)curNodeScore, new LinkedList <ModificationInstance>()));
}
var modPos = _isForward ? seqIndex : _index - seqIndex;
if (modPos <= 1)
{
--modPos;
}
var aminoAcid = _aminoAcidSequence[seqIndex];
var modAa = aminoAcid as ModifiedAminoAcid;
LinkedList <ModificationInstance> newMods = null;
if (modAa != null)
{
newMods = bestPrevMods == null ? new LinkedList <ModificationInstance>() : new LinkedList <ModificationInstance>(bestPrevMods);
var modIns = new ModificationInstance(modAa.Modification, modPos);
if (_isForward)
{
newMods.AddLast(modIns);
}
else
{
newMods.AddFirst(modIns);
}
}
var prevModCombIndex = _graph[seqIndex - 1][bestPrevNodeIndex].ModificationCombinationIndex;
var curModCombIndex = node.ModificationCombinationIndex;
if (prevModCombIndex != curModCombIndex) // modified
{
if (newMods == null)
{
newMods = bestPrevMods == null ? new LinkedList <ModificationInstance>() : new LinkedList <ModificationInstance>(bestPrevMods);
}
var modification = ModificationParams.GetModificationIndexBetween(prevModCombIndex, curModCombIndex);
var modIns = new ModificationInstance(modification, modPos);
if (_isForward)
{
newMods.AddLast(modIns);
}
else
{
newMods.AddFirst(modIns);
}
}
return(maxScoreAndMods[seqIndex][modIndex] = new Tuple <double, LinkedList <ModificationInstance> >
((double)curNodeScore + bestPrevNodeScore, newMods ?? bestPrevMods));
}
public void TestModificationParams()
{
var methodName = MethodBase.GetCurrentMethod().Name;
ShowStarting(methodName);
var modifications = new[] { Modification.Acetylation, Modification.Phosphorylation, Modification.Oxidation}; //, Modification.PyroGluQ };
var modParams = new ModificationParams(modifications, 3);
int numCombinations = modParams.NumModificationCombinations;
for (int modCombIndex = 0; modCombIndex < numCombinations; modCombIndex++)
{
var modCombination = modParams.GetModificationCombination(modCombIndex);
Console.WriteLine("{0}: {1} {2} {3}", modCombIndex, modCombination, modCombination.Composition, modCombination.Composition.Mass);
}
Console.WriteLine(modParams.GetModificationCombinationIndex(8, 0));
Console.WriteLine(modParams.GetModificationCombinationIndex(19, 1));
}