private void PrepareDeNovoRegistries()
{
deNovoRegistryList = new List <DeNovoRegistry>();
if (PeptideList == null || PeptideList.Equals(""))
{
if (Arguments.MPexResultPath.Length > 0)
{
Console.WriteLine("Loading " + Arguments.MPexResultPath);
ResultPckg2 pkg = ResultPckg2.DeserializeResultPackage(Arguments.MPexResultPath);
deNovoRegistryList = new List <DeNovoRegistry>();
foreach (KeyValuePair <string, List <DeNovoRegistry> > kvp in pkg.MyUnmappedRegistries)
{
deNovoRegistryList.AddRange(kvp.Value);
}
}
else
{
deNovoRegistryList = PreprareDeNovoRegistries();
}
//Eliminate denovo registries not having the minimum score
int removedForNotHavingMinDenovoScore = deNovoRegistryList.RemoveAll(a => a.DeNovoScore < Arguments.MinDeNovoScore);
Console.WriteLine("Removed elements for not attending the minimum denovo score: " + removedForNotHavingMinDenovoScore);
}
else
{
List <string> peptides = Regex.Split(PeptideList, "\n").ToList();
for (int i = 0; i < peptides.Count; i++)
{
string cleanSequence = PatternTools.pTools.CleanPeptide(peptides[i], true);
cleanSequence = Regex.Replace(cleanSequence, " ", "");
cleanSequence = Regex.Replace(cleanSequence, "\t", "");
deNovoRegistryList.Add(new DeNovoRegistry(0, i, 1, cleanSequence, 1, "Provided"));
}
}
//Lets make sure we are satisfying the top hit constraint
if (Arguments.DeNovoOption == DeNovoOption.PepNovoFull || Arguments.DeNovoOption == DeNovoOption.PepNovo)
{
deNovoRegistryList.RemoveAll(a => a.Rank + 1 > Arguments.TopHits);
}
else
{
deNovoRegistryList.RemoveAll(a => a.Rank > Arguments.TopHits);
}
//Finally, lets remove registries that do not satisfy the minimum left criterion
int removedForNotHavingMinimumLength = deNovoRegistryList.RemoveAll(a => a.CleanSequence.Length < Arguments.PeptideMinNumAA);
Console.WriteLine("Removed for not attending the minimum number of aa : " + removedForNotHavingMinimumLength);
}
private void FeedLocusFastaDictPex(DirectoryClassDescription myDir, FileInfo[] resultFilteredFiles)
{
foreach (FileInfo fi in resultFilteredFiles)
{
Console.WriteLine("Parsing " + fi.FullName);
ResultPckg2 thisPckg = ResultPckg2.DeserializeResultPackage(fi.FullName);
myResultPackages.Add(new FileInfoResultPackage(fi, thisPckg, myDir.ClassLabel));
foreach (FastaItem fastaItem in thisPckg.MyFasta)
{
if (!locusFastaDict.ContainsKey(fastaItem.SequenceIdentifier))
{
locusFastaDict.Add(fastaItem.SequenceIdentifier, fastaItem.Sequence);
}
}
}
}
public void ParseDirs(List <DirectoryClassDescription> dirs)
{
MyDirectoryDescriptionDictionary = dirs;
MyResultPackages = new List <ThePackage>();
foreach (DirectoryClassDescription dcd in dirs)
{
//Get all Sepro Files in this and in deeper directories
FileInfo[] fileInfo = new DirectoryInfo(dcd.MyDirectoryFullName).GetFiles("*.mpex", SearchOption.AllDirectories);
foreach (FileInfo file in fileInfo)
{
//First lets unserialize the object
Console.WriteLine("Loading .. " + file.FullName);
ResultPckg2 rp = ResultPckg2.DeserializeResultPackage(file.FullName);
MyResultPackages.Add(new ThePackage(rp, file, dcd.ClassLabel));
}
Console.WriteLine("Done loading.");
}
}
public ThePackage(ResultPckg2 rp, FileInfo file, int classLabel)
{
MyPackage = rp;
MyFileInfo = file;
MyClassLabel = classLabel;
}
/// <summary>
/// Multicore alignment method. It will feed the top alignments into the DenovoRegistryObject
/// </summary>
/// <param name="database">A list with a parsed database</param>
/// <param name="deNovoRegistryList">A lsit with parsed peptide registries</param>
/// <returns>A list with aligned registries.</returns>
private ResultPckg2 AlignMAlign(List <FastaItem> database, List <DeNovoRegistry> deNovoRegistryList)
{
int removedPeptides = deNovoRegistryList.RemoveAll(a => a.Sequence.Length < Arguments.PeptideMinNumAA);
Console.WriteLine("Total denovo registries: " + deNovoRegistryList.Count);
Dictionary <string, List <DeNovoRegistry> > dnovoRegistryDict = (from registry in deNovoRegistryList
group registry by registry.CleanSequence into registryGroup
select new { Seq = registryGroup.Key, Registries = registryGroup.ToList() }).ToDictionary(a => a.Seq, a => a.Registries);
Console.WriteLine("Total denovo registries after grouping exact sequences: " + dnovoRegistryDict.Keys.Count);
//We need to feed the alignment List
KeyValuePair <int[, ], Dictionary <char, int> > m = MAligner.Utils.LoadSubstitutionMatrixFromString(PepExplorer2.Properties.Resources.PAM30MS);
Aligner mAligner = new Aligner(m);
int count = 0;
ConcurrentBag <AlignmentResult> alignments = new ConcurrentBag <AlignmentResult>();
ConcurrentBag <KeyValuePair <string, List <DeNovoRegistry> > > unmappedRegistries = new ConcurrentBag <KeyValuePair <string, List <DeNovoRegistry> > >();
Parallel.ForEach(dnovoRegistryDict, kvp =>
//foreach (KeyValuePair<string, List<DeNovoRegistry>> kvp in dnovoRegistryDict)
{
Console.WriteLine("Aligning: " + kvp.Key);
List <DeNovoR> dRegistries = (from r in kvp.Value
select new DeNovoR(r.ScanNumber, r.DeNovoScore, r.PtmSequence, r.Charge, r.FileName)).ToList();
byte[] peptideBytes = Encoding.ASCII.GetBytes(kvp.Key);
//Lets take care of palindromic hits that might spoil our result.
List <FastaItem> validItems = database.FindAll(a => ((double)mAligner.IDScore(peptideBytes, a.SequenceInBytes) / (double)peptideBytes.Length) > Arguments.MinIdentity).ToList();
if (validItems.Count(a => a.SequenceIdentifier.StartsWith(Arguments.DecoyLabel)) < validItems.Count)
{
validItems.RemoveAll(a => a.SequenceIdentifier.StartsWith(Arguments.DecoyLabel));
}
//----------------
if (validItems.Exists(a => a.SequenceIdentifier.StartsWith(Arguments.DecoyLabel)))
{
Console.WriteLine("Decoy hit included");
}
if (validItems.Count > 0)
{
//Now we need to find the best proteins
char[] peptide = kvp.Key.ToCharArray();
KeyValuePair <List <int>, List <string> > alg = mAligner.GetClosestPeptideInDB(peptide, validItems);
//Find the max value and only store the best IDs for this peptide
int bestScore = alg.Key.Max();
if (alg.Key.Count == validItems.Count)
{
List <string> protIDs = new List <string>();
for (int i = 0; i < validItems.Count; i++)
{
if (alg.Key[i] == bestScore)
{
protIDs.Add(validItems[i].SequenceIdentifier);
}
}
alignments.Add(new AlignmentResult(dRegistries, bestScore, protIDs));
}
else
{
Console.WriteLine("shouldn't be here, failed in peptide " + kvp.Key);
}
}
else
{
unmappedRegistries.Add(kvp);
}
Interlocked.Increment(ref count);
Console.WriteLine("Done aligning sequence no {0} ({1}) ", count, kvp.Key);
//Report progress
Progress = Math.Round((count / (double)dnovoRegistryDict.Keys.Count), 2) * 100;
}
);
//Lets get only the stuff we used from the DB
List <string> usedFastaIDs = (from b in alignments
from pID in b.ProtIDs
select pID).Distinct().ToList();
List <FastaItem> usedFasta = database.FindAll(a => usedFastaIDs.Contains(a.SequenceIdentifier));
ResultPckg2 resultPckg = new ResultPckg2(usedFasta, Arguments, alignments.ToList(), unmappedRegistries.ToList());
return(resultPckg);
}
