public void ReadInMLNFile(string filename, int _numdatasets)
{
// Read in MultiAlign file
// Note : sep 26, 2011 - This will need to be updated when the new MLN format comes out
CSVFileHandler CsvFileHandler2 = new CSVFileHandler(filename, CSVFileHandler.READ_ONLY);
CsvFileHandler2.openFile();
String[] Attributes2;
CsvFileHandler2.readLine();
while ((Attributes2 = CsvFileHandler2.readLine()) != null)
{
MultiAlignRecord mrecord = new MultiAlignRecord();
mrecord.ID = int.Parse(Attributes2[0]);
mrecord.Size = int.Parse(Attributes2[1]);
mrecord.Mass = double.Parse(Attributes2[2]);
mrecord.NET = double.Parse(Attributes2[3]);
mrecord.AllocateNumberDatasets(_numdatasets);
int n_dataset = 0;
for (int i = 4; i < Attributes2.Length; i += 4)
{
UMCRecord u = new UMCRecord();
if (Attributes2[i] != "")
{
u.ScanRep = int.Parse(Attributes2[i]);
u.Abundance = double.Parse(Attributes2[i + 1]);
u.ScanStart = int.Parse(Attributes2[i + 2]);
u.ScanEnd = int.Parse(Attributes2[i + 3]);
}
mrecord._AssociatedUMCRecords.Add(u);
n_dataset++;
}
AddRecord(mrecord);
}
}
public void AddRecord(MultiAlignRecord m_record)
{
_AllMLNRecords.Add(m_record);
}
public MultiAlignRecord(MultiAlignRecord m)
{
_ID = m._ID;
_Size = m._Size;
_Mass = m._Mass;
_NET = m._NET;
_AssociatedUMCRecords = new List<UMCRecord>();
if ((m._CandidatePeptideSeq != null) && (m._CandidatePeptideSeq.Length > 0))
{
_CandidatePeptideSeq = m._CandidatePeptideSeq;
_CandidateProteinName = m._CandidateProteinName;
_CandidatePeptideMass = m._CandidatePeptideMass;
_CandidateGlycanMass = m._CandidateGlycanMass;
_CandidateGlycanComposition = m._CandidateGlycanComposition;
_CandidateNGlycoSite = m._CandidateNGlycoSite;
}
AllocateNumberDatasets(m._AssociatedUMCRecords.Count);
for (int i = 0; i < m._AssociatedUMCRecords.Count; i++)
{
_AssociatedUMCRecords.Add(m._AssociatedUMCRecords[i]);
}
}
public void ReanInMLNV2File(string filename, int _numdatasets)
{
// Read in new version of MLN file
CSVFileHandler CsvFileHandler2 = new CSVFileHandler(filename, CSVFileHandler.READ_ONLY);
CsvFileHandler2.openFile();
String[] Attributes2;
CsvFileHandler2.readLine();
while ((Attributes2 = CsvFileHandler2.readLine()) != null)
{
MultiAlignRecord mrecord = new MultiAlignRecord();
mrecord.ID = int.Parse(Attributes2[0]);
mrecord.Mass = double.Parse(Attributes2[1]);
mrecord.NET = double.Parse(Attributes2[2]);
int numDatasetsToAllocate = 0;
if (mrecord.Mass < _MAP_MIN_MASS)
_MAP_MIN_MASS = mrecord.Mass;
if (mrecord.Mass > _MAP_MAX_MASS)
_MAP_MAX_MASS = mrecord.Mass;
if ((Attributes2.Length-5) / 8 != _numdatasets)
numDatasetsToAllocate = (Attributes2.Length -5) / 8;
else
numDatasetsToAllocate = _numdatasets;
mrecord.AllocateNumberDatasets(numDatasetsToAllocate);
for (int i = 5; i < Attributes2.Length; i += 8)
{
UMCRecord u = new UMCRecord();
if (Attributes2[i] != "")
{
u.ID = int.Parse(Attributes2[i]);
u.DatasetID = int.Parse(Attributes2[i + 1]);
u.MW = double.Parse(Attributes2[i + 2]);
u.Abundance = double.Parse(Attributes2[i + 4]);
u.ScanRep = int.Parse(Attributes2[i + 5]);
u.ScanStart = int.Parse(Attributes2[i + 6]);
u.ScanEnd = int.Parse(Attributes2[i + 7]);
mrecord._AssociatedUMCRecords.Add(u);
}
}
AddRecord(mrecord);
if (mrecord._AssociatedUMCRecords.Count > 10)
{
bool debig = true;
}
}
}
/* public void CalculateCIDRepresentativeFragmentationSpectra(ref Classes.UMCRecord _u, ref Classes.Params _params)
{
Classes.UMCRecord _tempU = new UMCRecord();
int spectra_num = 0;
string cum_spectra_names = null;
List<string> cluster_names = new List<string>();
List<int> orphan_ids = new List<int>();
for (int i = 0; i < _u._AssociatedFragEvents.Count; i++)
{
if (_u._AssociatedFragEvents[i].CIDPeaks[0] != null)
{
string spectra_name = "Spectra_" + _u.DatasetID + "_" + spectra_num + "_" + _u._AssociatedFragEvents[i].CIDScan;
GlypID.Peaks.clsPeak[] thisCIDPeaks = new GlypID.Peaks.clsPeak[_u._AssociatedFragEvents[i].CIDPeaks.Length];
Array.Copy(_u._AssociatedFragEvents[i].CIDPeaks, thisCIDPeaks, _u._AssociatedFragEvents[i].CIDPeaks.Length);
_CIDSpectralUtilities.AddPeaksToList(ref thisCIDPeaks, spectra_name);
spectra_num++;
if (cum_spectra_names != null)
cum_spectra_names = cum_spectra_names + "-" + spectra_name;
else
cum_spectra_names = spectra_name;
}
else
{
// TO check
// This happens which means there was no fragmentation event then (or) HCD score was bad
if (_u._AssociatedFragEvents[i].HCDScore < 1)
orphan_ids.Add(_u.DatasetID);
}
}
}*/
public void AssignFDR(ref Classes.MapRecord _glycoMap, ref Classes.Params _params)
{
bool use_combined_score = false; // To do change this based on _params
Classes.FragEvents e = new FragEvents();
for (int i = 0 ; i < _glycoMap._AllMLNRecords.Count ; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
for (int j = 0; j < m._ClusterRepFragEvents.Count; j++)
{
if (m._ClusterRepFragEvents[j].ETDScore > 0)
{
if (m._ClusterRepFragEvents[j].GP_Record.IsDecoy)
_glycoMap._AllFalseHitsFDRScore.Add(m._ClusterRepFragEvents[j].ETDScore);
else
_glycoMap._AllTrueHitsFDRScore.Add(m._ClusterRepFragEvents[j].ETDScore);
}
}
}
// Assign ETD score based on ETD type
for (int i = 0; i < _glycoMap._AllMLNRecords.Count; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
for (int j = 0; j < m._ClusterRepFragEvents.Count; j++)
{
e = new FragEvents();
e = m._ClusterRepFragEvents[j];
if (e.ETDScore > 0)
{
int num_false_hits = 0;
int num_true_hits = 0;
foreach (double score in _glycoMap._AllFalseHitsFDRScore)
{
if (score > e.ETDScore)
num_false_hits++;
}
foreach (double score in _glycoMap._AllTrueHitsFDRScore)
{
if (score > e.ETDScore)
num_true_hits++;
}
if (num_true_hits == 0)
num_true_hits = 1; // to avoid divide by 0 for the highest.
e.FDR = (float)num_false_hits / num_true_hits;
}
}
}
/* int num_records = m._AssociatedUMCRecords.Count;
for (int j = 0; j < num_records; j++)
{
for (int k = 0; k < m._AssociatedUMCRecords[j]._AssociatedFragEvents.Count; k++)
{
e = new FragEvents();
e = m._AssociatedUMCRecords[j]._AssociatedFragEvents[k];
if (e.ETDScore > 0)
{
int num_false_hits = 0;
int num_true_hits = 0;
foreach (double score in _glycoMap._AllFalseHitsFDRScore)
{
if (score > e.ETDScore)
num_false_hits++;
}
foreach (double score in _glycoMap._AllTrueHitsFDRScore)
{
if (score > e.ETDScore)
num_true_hits++;
}
e.FDR = (float)num_false_hits / num_true_hits;
}
}
}*/
}
public void SequenceCIDPeaks(ref Classes.MapRecord _glycoMap, ref Classes.Params _params, string sequencingFolder)
{
Classes.MapRecord _tempMap = new MapRecord();
_tempMap._AssociatedDatasetNames = _glycoMap._AssociatedDatasetNames;
_tempMap._IsCID = _glycoMap._IsCID;
_tempMap._IsETD = _glycoMap._IsETD;
_tempMap._IsHCD = _glycoMap._IsHCD;
Classes.FragEvents e = new FragEvents();
if (sequencingFolder == "")
sequencingFolder = @"c:\sequencing\";
for (int i = 0; i < _glycoMap._AllMLNRecords.Count; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
int num_records = m._AssociatedUMCRecords.Count;
if (m.ID == 2290)
{
bool test = true;
}
//for (int j = 0; j < num_records; j++)
//{
// for (int k = 0; k < m._AssociatedUMCRecords[j]._AssociatedFragEvents.Count; k++)
for (int k = 0; k < m._ClusterRepFragEvents.Count; k++)
{
e = new FragEvents();
e = m._ClusterRepFragEvents[k]; // m._AssociatedUMCRecords[j]._AssociatedFragEvents[k];
if (e.ETDScore > 0)
{
float PeptideMass = 0;
float GlcNAcMass = 0;
COL.MassLib.MSScan _msScan = new COL.MassLib.MSScan(e.CIDMzs, e.CIDIntensities, Convert.ToSingle(e.TransformResult.mdbl_mz),
Convert.ToSingle(e.TransformResult.mdbl_mono_mw), Convert.ToSingle(e.TransformResult.mdbl_average_mw), Convert.ToInt32(e.TransformResult.mshort_cs));
if (e.TransformResult.mdbl_average_mw > 0)
{
PeptideMass = Convert.ToSingle(e.GP_Record.SequenceAverageMass); // this means proper deisotoping has occured so use average compostion
GlcNAcMass = COL.GlycoLib.GlycanMass.GetGlycanAVGMass(COL.GlycoLib.Glycan.Type.HexNAc);
}
else
{
PeptideMass = Convert.ToSingle(e.GP_Record.SequenceMonoMass); // CS has been assigned, in which case both y1 and precursor should be just mono
GlcNAcMass = COL.GlycoLib.GlycanMass.GetGlycanMass(COL.GlycoLib.Glycan.Type.HexNAc);
}
short y1cs = e.TransformResult.mshort_cs;
y1cs--;
while (y1cs > 0)
{
float y1Mz = ((PeptideMass + GlcNAcMass) + (float)_utils._CC_MASS * y1cs) / y1cs;
COL.GlycoSequence.GlycanSequencing _Gs = new COL.GlycoSequence.GlycanSequencing(_msScan, y1Mz, y1cs, e.GP_Record.Glycan.numHex,
e.GP_Record.Glycan.numHexNAc, e.GP_Record.Glycan.numDeHex, e.GP_Record.Glycan.numNeuAc, 0, sequencingFolder, true, 0.8f, 60);
_Gs.NumbersOfPeaksForSequencing = 140;
_Gs.CreatePrecursotMZ = true;
_Gs.RewardForCompleteStructure = 3;
if (e.TransformResult.mdbl_average_mw > 0)
_Gs.UseAVGMass = true;
else
_Gs.UseAVGMass = false;
int structure_count = _Gs.StartSequencing();
if (structure_count > 0)
{
List<COL.GlycoLib.GlycanStructure> topstructures = _Gs.GetTopRankScoreStructre(1);
e.CIDSequencingScore = topstructures[0].Score;
e.GP_Record.GlycanSequence = topstructures[0].IUPACString;
// Printing out the sequences
string opfile = sequencingFolder + m.ID.ToString() + "_" + e.CIDScan.ToString() + ".txt";
Utils.CSVFileHandler cidString = new CSVFileHandler(opfile, CSVFileHandler.WRITE_ONLY);
cidString.openFile();
for (int zz = 0; zz < topstructures.Count; zz++)
{
cidString.writeLine(topstructures[zz].IUPACString);
/*COL.GlycoLib.GlycansDrawer _gdraw = new COL.GlycoLib.GlycansDrawer(topstructures[zz].IUPACString, false);
System.Drawing.Image img1 = _gdraw.GetImage();*/
}
cidString.closeFile();
break;
}
y1cs--;
}
}
}
}
}
public void ResolveIdentificationsAcrossClusters(ref Classes.MapRecord _glycoMap, ref Classes.Params _params)
{
Classes.MapRecord _tempMap = new MapRecord();
_tempMap._AssociatedDatasetNames = _glycoMap._AssociatedDatasetNames;
_tempMap._IsCID = _glycoMap._IsCID;
_tempMap._IsETD = _glycoMap._IsETD;
_tempMap._IsHCD = _glycoMap._IsHCD;
_tempMap._AllFalseHitsFDRScore = _glycoMap._AllFalseHitsFDRScore;
_tempMap._AllTrueHitsFDRScore = _glycoMap._AllTrueHitsFDRScore;
int num_clashes = 0;
for (int i = 0; i < _glycoMap._AllMLNRecords.Count; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
if (_glycoMap._IsETD.Contains(true))
{
if (m._ClusterRepFragEvents.Count > 0)
{
if (m.HasUniformIDsAcrossClusters())
{
m.SetIdsBasedOnHighestETD();
if (m.IDLabel == _ID_label.Unverified)
m.SetIdsBasedOnMassAndGlycanType(_params.ScoringParams.PPMTolerance, true);
}
else
{
// Need to think about this
bool debug = true;
num_clashes++;
}
}
else
{
m.SetIdsBasedOnMassAndGlycanType(_params.ScoringParams.PPMTolerance, false);
}
}
else
{
//m.
}
}
}
public void GetRepresentatives(ref Classes.MapRecord _glycoMap, ref Classes.Params _params)
{
Classes.MapRecord _tempMap = new MapRecord();
_tempMap._AssociatedDatasetNames = _glycoMap._AssociatedDatasetNames;
_tempMap._IsCID = _glycoMap._IsCID;
_tempMap._IsETD = _glycoMap._IsETD;
_tempMap._IsHCD = _glycoMap._IsHCD;
_tempMap._AllFalseHitsFDRScore = _glycoMap._AllFalseHitsFDRScore;
_tempMap._AllTrueHitsFDRScore = _glycoMap._AllTrueHitsFDRScore;
bool use_etd = false ;
if (_glycoMap._IsETD.Contains(true))
use_etd = true ;
FragEvents tempE = new FragEvents();
for (int i = 0; i < _glycoMap._AllMLNRecords.Count; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
int num_clusters = m._ClusterNames.Count;
if (m.ID == 2290)
{
bool test = true;
}
for (int c = 0; c < num_clusters; c++)
{
string clustername = m._ClusterNames[c];
string[] spectra = clustername.Split('-');
Dictionary <int, List<int>> umcs_frag_ids = new Dictionary<int, List<int>>() ;
List<int>frag_to_look = new List<int> () ;
foreach (string s in spectra)
{
string[] parts = s.Split('_');
int umc_id = Convert.ToInt32(parts[1]);
int frag_id = Convert.ToInt32(parts[2]) ;
if (umcs_frag_ids.ContainsKey(umc_id))
{
umcs_frag_ids[umc_id].Add(frag_id);
}
else
{
frag_to_look.Clear() ;
frag_to_look.Add(frag_id) ;
umcs_frag_ids.Add(umc_id, frag_to_look);
}
}
double max_etd_score = 0;
double min_hcd_score = 0 ;
FragEvents maxFragEvent = new FragEvents();
for (int j = 0; j < m._AssociatedUMCRecords.Count; j++)
{
if (umcs_frag_ids.ContainsKey(m._AssociatedUMCRecords[j].DatasetID))
{
tempE = new FragEvents();
if (use_etd)
{
tempE = m._AssociatedUMCRecords[j].FragEventWithHighestETDScore(true, umcs_frag_ids[m._AssociatedUMCRecords[j].DatasetID]);
if (tempE.ETDScore > max_etd_score)
{
maxFragEvent = tempE;
}
}
else
{
if (!_params.ProcessOnlyNGlycopeptides)
maxFragEvent = m._AssociatedUMCRecords[j].FragEventWithLowestHCDScore(umcs_frag_ids[m._AssociatedUMCRecords[j].DatasetID]);
}
}
}
if ((maxFragEvent.ETDScore > 0) || (!_params.ProcessOnlyNGlycopeptides))
m._ClusterRepFragEvents.Add(maxFragEvent);
/*else
{
FragEvents minHCDFragEvent = new FragEvents();
double min_hcd_score = 1;
for (int j= 0 ; j < m._AssociatedUMCRecords.Count ; j++)
{
if (umcs_frag_ids.ContainsKey(m._AssociatedUMCRecords[j].DatasetID))
{
tempE = new FragEvents();
tempE = m._AssociatedUMCRecords[j].FragEventWithLowestHCDScore(umcs_frag_ids[m._AssociatedUMCRecords[j].DatasetID]);
if (tempE.HCDScore < min_hcd_score)
minHCDFragEvent = tempE;
}
}
m._ClusterRepFragEvents.Add(minHCDFragEvent);
} */
}
}
}
/// <summary>
/// Function to cluster CID spectra in each record.
/// </summary>
/// <param name="_glycoMap"></param>
/// <param name="_params"></param>
public void ClusterRecordsOnCID(ref Classes.MapRecord _glycoMap, ref Classes.Params _params)
{
Classes.MapRecord _tempMap = new MapRecord();
_tempMap._AssociatedDatasetNames = _glycoMap._AssociatedDatasetNames;
_tempMap._IsCID = _glycoMap._IsCID;
_tempMap._IsETD = _glycoMap._IsETD;
_tempMap._IsHCD = _glycoMap._IsHCD;
int Id = 0;
FragEvents tempE = new FragEvents();
for (int i = 0; i < _glycoMap._AllMLNRecords.Count; i++)
{
MultiAlignRecord m = new MultiAlignRecord();
m = _glycoMap._AllMLNRecords[i];
int num_records = m._AssociatedUMCRecords.Count;
short mincs = m.MinChargeStateObserved();
short maxcs = m.MaxChargeStateObserved();
for (short thiscs = mincs; thiscs <= maxcs; thiscs++)
{
int spectra_num = 0;
string cum_spectra_names = null;
List<string> cluster_names = new List<string>();
List<int> orphan_ids = new List<int>();
for (int j = 0; j < num_records; j++)
{
if (m._AssociatedUMCRecords[j]._AssociatedFragEvents.Count == 0)
{
// Not fragmented in this UMC but keep track of this but
// // This happens which means there was no fragmentation event then (or) HCD score was bad
if (m._AssociatedUMCRecords[j].Abundance > 0)
orphan_ids.Add(m._AssociatedUMCRecords[j].DatasetID);
}
for (int k = 0; k < m._AssociatedUMCRecords[j]._AssociatedFragEvents.Count; k++)
{
if ((m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].CIDPeaks[0] != null) &&
(m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].TransformResult.mshort_cs == thiscs))
{
string spectra_name = "Spectra_" + m._AssociatedUMCRecords[j].DatasetID + "_" + m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].ID + "_" +
spectra_num + "_" + m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].CIDScan;
GlypID.Peaks.clsPeak[] thisCIDPeaks = new GlypID.Peaks.clsPeak[m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].CIDPeaks.Length];
Array.Copy(m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].CIDPeaks, thisCIDPeaks, m._AssociatedUMCRecords[j]._AssociatedFragEvents[k].CIDPeaks.Length);
_CIDSpectralUtilities.AddPeaksToList(ref thisCIDPeaks, spectra_name);
spectra_num++;
if (cum_spectra_names != null)
cum_spectra_names = cum_spectra_names + "-" + spectra_name;
else
cum_spectra_names = spectra_name;
}
}
}
int num_clusters = _CIDSpectralUtilities.ClusterSpectraInList();
_CIDSpectralUtilities.GetClusterNames(ref cluster_names);
if (num_clusters > 1)
{
// Indicates glycoforms
bool debug = true;
debug = true;
}
for (int c = 0; c < num_clusters; c++)
{
m._ClusterNames.Add(cluster_names[c]);
}
_CIDSpectralUtilities.Clear();
}
_tempMap.AddRecord(m);
}
// Restore them.
_glycoMap.ClearRecords();
for (int i = 0; i < _tempMap._AllMLNRecords.Count; i++)
{
_glycoMap.AddRecord(_tempMap._AllMLNRecords[i]);
}
/*else
{
/* MultiAlignRecord _tempM = new MultiAlignRecord(m);
_tempM.ID = Id;
_tempMap.AddRecord(_tempM);
Id++; */
/* if (cum_spectra_names != null)
{
// Choose the spectra with the greatest SNR
num_clusters = 1;
cluster_names.Add(cum_spectra_names);
_CIDSpectralUtilities.AssignClusters(ref cluster_names);
}
}*/
/* for (int k = 0; k < num_clusters; k++)
{
// ----- Get a representative spectrum index for each cluster ---//
MultiAlignRecord _tempM = new MultiAlignRecord(m);
_tempM.ID = Id;
_tempM._AssociatedUMCRecords.Clear();
GlypID.Peaks.clsPeak[] repCIDPeaks = new GlypID.Peaks.clsPeak[0];
int repOrigIndex = _CIDSpectralUtilities.GetRepresentativePeaksFromCluster(k, ref repCIDPeaks, _params.ScoringParams.MinCIDMz, _params.ScoringParams.MaxCIDMz, true);
if (repCIDPeaks.Length > 1)
{
_tempM._RepresentativeCIDPeaks = repCIDPeaks;
_tempM._RepresentativeDatasetID_CID = repOrigIndex;
}
// Attach UMCs corresponding to that cluster.
List<int> allOrigIDs = new List<int>();
_CIDSpectralUtilities.GetOriginalIDFromCluster(k, ref allOrigIDs);
for (int j = 0; j < num_records; j++)
{
int id = m._AssociatedUMCRecords[j].DatasetID;
if (allOrigIDs.Exists(element => element == id))
{
UMCRecord _tempUMC = new UMCRecord();
_tempUMC = m._AssociatedUMCRecords[j];
_tempM._AssociatedUMCRecords.Add(_tempUMC);
}
else if (orphan_ids.Exists(element => element == id)) // This takes care of non fragmentation but still has stuff present
{
UMCRecord _tempUMC = new UMCRecord();
_tempUMC = m._AssociatedUMCRecords[j];
_tempM._AssociatedUMCRecords.Add(_tempUMC);
}
}
_tempMap.AddRecord(_tempM);
Id++;
}*/
}
