protected DataTable ParseFromFile(string filePath, CIStringCollection items, CIStringCollection experiments)
{
DataTable results;
FileServer fServer;
string[][] fileContents;
fServer = new FileServer();
OnStatusChange("Retrieving genotypes from file");
fileContents = fServer.ReadMultipleColumns(filePath, 3);
results = new DataTable("Genotype");
results.Columns.Add("Item", Type.GetType("System.String"));
results.Columns.Add("Experiment", Type.GetType("System.String"));
results.Columns.Add("Alleles", Type.GetType("System.String"));
for (int i = 0; i < fileContents.GetLength(0); i++)
{
if (items.Contains(fileContents[i][0].Trim()) && experiments.Contains(fileContents[i][1].Trim()))
{
results.Rows.Add(new object[] { fileContents[i][0], fileContents[i][1], fileContents[i][2] });
}
if (i % 1000 == 0)
{
OnStatusChange("Finished reading line " + i.ToString());
}
}
return(results);
}
private CIStringCollection Union(CIStringCollection values1, CIStringCollection values2, out CIStringCollection values1NotIn2, out CIStringCollection values2NotIn1)
{
//Sort out values which are both in source 1 and source 2, and also save
//those which are only in source 1.
CIStringCollection valuesUnion;
valuesUnion = new CIStringCollection();
values1NotIn2 = new CIStringCollection();
values2NotIn1 = new CIStringCollection();
foreach (string tempValue in values1)
{
if (values2.Contains(tempValue))
{
if (!valuesUnion.Contains(tempValue))
{
valuesUnion.Add(tempValue);
}
}
else
{
if (!values1NotIn2.Contains(tempValue))
{
values1NotIn2.Add(tempValue);
}
}
}
//Save values which are only in source 2.
foreach (string tempValue in values2)
{
if (!values1.Contains(tempValue))
{
if (!values2NotIn1.Contains(tempValue))
{
values2NotIn1.Add(tempValue);
}
}
}
return(valuesUnion);
}
public void HarmonizePolarities(GenotypeDictionary dict1, GenotypeDictionary dict2, out GenotypeDictionary harmonizedDict1, out GenotypeDictionary harmonizedDict2)
{
//Harmonizes dict2 after dict1 and removes all entries from A/T or G/C experiments and incompatible
//SNP types and items from SNPs which are monomorphic in both source.
string tempExp, tempAlleles1, tempAlleles2;
CIStringCollection uniqueExps;
Dictionary <string, string> allelesDict1, allelesDict2;
int counter;
//Find all experiments and their SNP types (the experiment names should be the same in both dictionaries, use dict1 here).
uniqueExps = new CIStringCollection();
allelesDict1 = new Dictionary <string, string>(StringComparer.InvariantCultureIgnoreCase);
allelesDict2 = new Dictionary <string, string>(StringComparer.InvariantCultureIgnoreCase);
OnStatusChange("Extracting experiments and SNP types");
counter = 0;
foreach (string tempKey in dict1.Keys)
{
tempExp = dict1.GetExperiment(tempKey);
tempAlleles1 = dict1[tempKey].ToUpper();
tempAlleles2 = dict2[tempKey].ToUpper();
if (!uniqueExps.Contains(tempExp))
{
//This experiment has not been processed before.
uniqueExps.Add(tempExp);
//Get the unique alleles from dictionary 1.
if (tempAlleles1.Substring(0, 1) != tempAlleles1.Substring(2, 1))
{
//The alleles are different, add both.
allelesDict1.Add(tempExp, tempAlleles1.Substring(0, 1) + tempAlleles1.Substring(2, 1));
}
else
{
//The alleles are the same, add only the first.
allelesDict1.Add(tempExp, tempAlleles1.Substring(0, 1));
}
//Get the unique alleles from dictionary 2.
if (tempAlleles2.Substring(0, 1) != tempAlleles2.Substring(2, 1))
{
//The alleles are different, add both.
allelesDict2.Add(tempExp, tempAlleles2.Substring(0, 1) + tempAlleles2.Substring(2, 1));
}
else
{
//The alleles are the same, add only the first.
allelesDict2.Add(tempExp, tempAlleles2.Substring(0, 1));
}
}
else
{
//This experiment has been processed before.
//Add the first allele from dictionary one if it does not already exist in the string.
if (!allelesDict1[tempExp].Contains(tempAlleles1.Substring(0, 1)))
{
allelesDict1[tempExp] = allelesDict1[tempExp] + tempAlleles1.Substring(0, 1);
}
//Add the second allele from dictionary one if it does not already exist in the string.
if (!allelesDict1[tempExp].Contains(tempAlleles1.Substring(2, 1)))
{
allelesDict1[tempExp] = allelesDict1[tempExp] + tempAlleles1.Substring(2, 1);
}
//Add the first allele from dictionary two if it does not already exist in the string.
if (!allelesDict2[tempExp].Contains(tempAlleles2.Substring(0, 1)))
{
allelesDict2[tempExp] = allelesDict2[tempExp] + tempAlleles2.Substring(0, 1);
}
//Add the second allele from dictionary one if it does not already exist in the string.
if (!allelesDict2[tempExp].Contains(tempAlleles2.Substring(2, 1)))
{
allelesDict2[tempExp] = allelesDict2[tempExp] + tempAlleles2.Substring(2, 1);
}
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting experiments and SNP types (processed genotype " + counter + ")");
}
}
//Remove data from SNPs with more than two alleles in source 1 from both dictionaries and save.
MyRemovedExpTooManyAlleles1 = new CIStringCollection();
OnStatusChange("Extracting harmonizable data step 1/6");
counter = 0;
foreach (string exp in allelesDict1.Keys)
{
if (allelesDict1[exp].Length > 2)
{
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpTooManyAlleles1.Add(exp);
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 1/6 (processed experiment " + counter + ")");
}
}
//Remove data from SNPs with more than two alleles in source 2 from both dictionaries and save.
MyRemovedExpTooManyAlleles2 = new CIStringCollection();
OnStatusChange("Extracting harmonizable data step 2/6");
counter = 0;
foreach (string exp in allelesDict2.Keys)
{
if (allelesDict2[exp].Length > 2)
{
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpTooManyAlleles2.Add(exp);
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 2/6 (processed experiment " + counter + ")");
}
}
//Remove data from SNPs which are monomorphic in both sources and save.
MyRemovedExpMonomorphicInBoth = new CIStringCollection();
OnStatusChange("Extracting harmonizable data step 3/6");
counter = 0;
foreach (string exp in allelesDict1.Keys)
{
if (allelesDict1[exp].Length == 1 && allelesDict2[exp].Length == 1)
{
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpMonomorphicInBoth.Add(exp);
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 3/6 (processed experiment " + counter + ")");
}
}
//Remove data from SNPs with incompatible SNP types from both dictionaries and save.
MyRemovedExpIncompatibleSNPTypes = new CIStringCollection();
OnStatusChange("Extracting harmonizable data step 4/6");
counter = 0;
foreach (string exp in allelesDict1.Keys)
{
if (!MyRemovedExpTooManyAlleles1.Contains(exp) && !MyRemovedExpTooManyAlleles2.Contains(exp) &&
!MyRemovedExpMonomorphicInBoth.Contains(exp))
{
if (!this.IsCompatibleSNPTypes(allelesDict1[exp], allelesDict2[exp]))
{
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpIncompatibleSNPTypes.Add(exp);
}
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 4/6 (processed experiment " + counter + ")");
}
}
//Remove data from SNPs with A/T or G/C SNPs from both dictionaries and save.
MyRemovedExpATCG = new CIStringCollection();
OnStatusChange("Extracting harmonizable data step 5/6");
counter = 0;
foreach (string exp in allelesDict1.Keys)
{
if (!MyRemovedExpTooManyAlleles1.Contains(exp) && !MyRemovedExpTooManyAlleles2.Contains(exp) &&
!MyRemovedExpMonomorphicInBoth.Contains(exp) && !MyRemovedExpIncompatibleSNPTypes.Contains(exp))
{
if (allelesDict1[exp].Length > 1)
{
if (this.GetNumberOfATalleles(allelesDict1[exp]) != 1)
{
//Either 0 or 2 A or T alleles.
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpATCG.Add(exp);
}
}
else
{
if (this.GetNumberOfATalleles(allelesDict2[exp]) != 1)
{
dict1.RemoveExperiment(exp);
dict2.RemoveExperiment(exp);
MyRemovedExpATCG.Add(exp);
}
}
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 5/6 (processed experiment " + counter + ")");
}
}
//Store only data from experiments passing all tests.
harmonizedDict1 = new GenotypeDictionary();
harmonizedDict2 = new GenotypeDictionary();
OnStatusChange("Extracting harmonizable data step 6/6");
counter = 0;
foreach (string tempKey in dict1.Keys)
{
tempExp = dict1.GetExperiment(tempKey);
if (!MyRemovedExpTooManyAlleles1.Contains(tempExp) &&
!MyRemovedExpTooManyAlleles2.Contains(tempExp) &&
!MyRemovedExpMonomorphicInBoth.Contains(tempExp) &&
!MyRemovedExpIncompatibleSNPTypes.Contains(tempExp) &&
!MyRemovedExpATCG.Contains(tempExp))
{
harmonizedDict1.Add(tempKey, dict1[tempKey]);
harmonizedDict2.Add(tempKey, dict2[tempKey]);
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Extracting harmonizable data step 6/6 (processed genotype " + counter + ")");
}
}
//Go through dict1. If the SNP type for the current row is different than
//the SNP type in dict2, convert the row in dict2.
OnStatusChange("Harmonizing");
counter = 0;
foreach (string tempKey in harmonizedDict1.Keys)
{
tempExp = harmonizedDict1.GetExperiment(tempKey);
tempAlleles1 = allelesDict1[tempExp];
tempAlleles2 = allelesDict2[tempExp];
if (tempAlleles1.Length == 2 && tempAlleles2.Length == 2)
{
if ((
(tempAlleles1.Substring(0, 1) == tempAlleles2.Substring(0, 1)) &&
(tempAlleles1.Substring(1, 1) == tempAlleles2.Substring(1, 1))
) ||
(
(tempAlleles1.Substring(0, 1) == tempAlleles2.Substring(1, 1)) &&
(tempAlleles1.Substring(1, 1) == tempAlleles2.Substring(0, 1))
)
)
{
//They are the same.
}
else
{
harmonizedDict2[tempKey] = this.ReverseComplement(harmonizedDict2[tempKey]);
}
}
else if (tempAlleles1.Length == 1 && tempAlleles2.Length == 2)
{
if (tempAlleles1 != tempAlleles2.Substring(0, 1) && tempAlleles1 != tempAlleles2.Substring(1, 1))
{
harmonizedDict2[tempKey] = this.ReverseComplement(harmonizedDict2[tempKey]);
}
}
else if (tempAlleles1.Length == 2 && tempAlleles2.Length == 1)
{
if (tempAlleles2 != tempAlleles1.Substring(0, 1) && tempAlleles2 != tempAlleles1.Substring(1, 1))
{
harmonizedDict2[tempKey] = this.ReverseComplement(harmonizedDict2[tempKey]);
}
}
else
{
throw new Exception("Inappropriate number of alleles detected when converting SNP types");
}
counter++;
if (counter % 100 == 0)
{
OnStatusChange("Harmonizing (processed genotype " + counter + ")");
}
}
}