public ProteinGroup2(MyProtein prot)
{
MyProteins = new List <MyProtein>()
{
prot
};
MyPeptides = prot.DistinctPeptides;
}
public List <MyProtein> MaxParsimonyList_Old()
{
//int proteinCounter = 0;
//Lets generate an array that represents a list of proteins according to the MaxParsimony
List <MyProtein> MaxParsimony = new List <MyProtein>(MyProteins.TheGroups.Count);
foreach (int groupNumber in MyProteins.TheGroups)
{
List <MyProtein> proteins = MyProteins.MyProteinList.FindAll(a => a.GroupNo == groupNumber);
if (proteins[0].MyGroupType == ProteinGroupType.All || proteins[0].MyGroupType == ProteinGroupType.Unique || proteins[0].MyGroupType == ProteinGroupType.Single)
{
//proteinCounter++;
MaxParsimony.Add(proteins[0]);
continue;
}
if (proteins[0].MyGroupType == ProteinGroupType.Some && proteins.Count == 2)
{
//proteinCounter += 2;
MaxParsimony.AddRange(proteins);
continue;
}
//We are dealing with a "some group"; lets generate the bipartite graph
List <string> peptides = (from p in proteins
from s in p.DistinctPeptides
select s).Distinct().ToList();
proteins = (from p in proteins
select p).OrderByDescending(p => p.DistinctPeptides).ThenByDescending(c => c.Length).ToList();
//proteins.Sort((a,b) => b.DistinctPeptides.Count.CompareTo(a.DistinctPeptides.Count));
//At each step, we will chose the item that contains the largest number of uncovered results
// http://en.wikipedia.org/wiki/Set_cover_problem
List <MyProtein> minumumCoverSet = new List <MyProtein>();
//minumumCoverSet.Add(proteins[0]);
MaxParsimony.Add(proteins[0]);
proteins.RemoveAt(0);
//Now, lets cycle until we have the full set complete
try
{
while (true)
{
//Get the covered peptides
List <string> coveredPeptides = (from p in minumumCoverSet
from pep in p.DistinctPeptides
select pep).Distinct().ToList();
//Find the maximum coverage possible
int maxCoverage = proteins.Max(a => a.DistinctPeptides.Union(coveredPeptides).Distinct().Count());
//find the item from the proteinNodes that will provide the greatest additional cover
MyProtein nextProtein = proteins.Find(a => a.DistinctPeptides.Union(coveredPeptides).Distinct().Count() == maxCoverage);
//add it to the minimum cover set
MaxParsimony.Add(nextProtein);
//minumumCoverSet.Add(nextProtein);
//remove it from the protein nodes
proteins.Remove(nextProtein);
//check if we still need to converge
if (proteins.Count == 0 || coveredPeptides.Count == peptides.Count)
{
break;
}
}
}
catch (Exception e)
{
Console.WriteLine("Max parsimony problem on group number: " + groupNumber + "\n" + e.Message);
}
//proteinCounter += minumumCoverSet.Count;
}
return(MaxParsimony);
//return proteinCounter;
}
public List <MyProtein> MaxParsimonyList()
{
//int proteinCounter = 0;
//Lets generate an array that represents a list of proteins according to the MaxParsimony
List <MyProtein> MaxParsimony = new List <MyProtein>(MyProteins.TheGroups.Count);
List <MyProtein> ProteinsTMP = PatternTools.ObjectCopier.Clone(MyProteins.MyProteinList);
foreach (int groupNumber in MyProteins.TheGroups)
{
List <MyProtein> proteins = ProteinsTMP.FindAll(a => a.GroupNo == groupNumber);
List <MyProtein> minumumCoverSet = new List <MyProtein>();
proteins.Sort((a, b) => b.Length.CompareTo(a.Length));
if (proteins[0].MyGroupType == ProteinGroupType.All || proteins[0].MyGroupType == ProteinGroupType.Unique || proteins[0].MyGroupType == ProteinGroupType.Single)
{
//proteinCounter++;
minumumCoverSet.Add(proteins[0]);
proteins.RemoveAt(0);
//Add for remaining proteins
if (proteins.Count > 0)
{
string add = "; Additional IDs concatenated into MaxParsimony group: ";
add += string.Join(", ", proteins.Select(a => a.Locus).ToList());
minumumCoverSet[0].Description += add;
}
MaxParsimony.AddRange(minumumCoverSet);
continue;
}
//We are dealing with a "some group"; lets generate the bipartite graph
List <string> peptides = (from p in proteins
from s in p.DistinctPeptides
select s).Distinct().ToList();
proteins = (from p in proteins
select p).OrderByDescending(p => p.DistinctPeptides.Count).ThenByDescending(p => p.Length).ToList();
//proteins.Sort((a,b) => b.DistinctPeptides.Count.CompareTo(a.DistinctPeptides.Count));
//At each step, we will chose the item that contains the largest number of uncovered results
// http://en.wikipedia.org/wiki/Set_cover_problem
//minumumCoverSet.Add(proteins[0]);
//MaxParsimony.Add(proteins[0]);
//proteins.RemoveAt(0);
//Now, lets cycle until we have the full set complete
try
{
while (true)
{
//Get the covered peptides
List <string> coveredPeptides = (from p in minumumCoverSet
from pep in p.DistinctPeptides
select pep).Distinct().ToList();
//Find the maximum coverage possible
int maxCoverage = proteins.Max(a => a.DistinctPeptides.Intersect(peptides).Distinct().Count());
//find the item from the proteinNodes that will provide the greatest additional cover
MyProtein nextProtein = proteins.Find(a => a.DistinctPeptides.Intersect(peptides).Distinct().Count() == maxCoverage);
//add it to the minimum cover set
//MaxParsimony.Add(nextProtein);
minumumCoverSet.Add(nextProtein);
//remove it from the protein nodes
proteins.Remove(nextProtein);
//check if we still need to converge
//Get the covered peptides
coveredPeptides = (from p in minumumCoverSet
from pep in p.DistinctPeptides
select pep).Distinct().ToList();
peptides = peptides.Except(coveredPeptides).ToList();
if (proteins.Count == 0 || coveredPeptides.Count == peptides.Count || peptides.Count == 0)
{
break;
}
}
}
catch (Exception e)
{
Console.WriteLine("Max parsimony problem on group number: " + groupNumber + "\n" + e.Message);
}
//Add for remaining proteins
if (proteins.Count > 0)
{
string add = "; Additional IDs concatenated into MaxParsimony group: ";
add += string.Join(", ", proteins.Select(a => a.Locus).ToList());
minumumCoverSet[0].Description += add;
}
MaxParsimony.AddRange(minumumCoverSet);
}
return(MaxParsimony);
//return proteinCounter;
}
