public static IEnumerable <IHashableFeature> GetNFlankAny1AndNTermEAny2(NEC nec, string nflankAny, string nTermEAny)
{
NFlank nFlankFeature = NFlank.GetInstance();
string nFlankRegion = nFlankFeature.FeatureFunction(nec);
if (-1 == nFlankRegion.IndexOfAny(nflankAny.ToCharArray()))
{
yield break;
}
NTermE nTermEFeature = NTermE.GetInstance();
string nTermERegion = nTermEFeature.FeatureFunction(nec);
if (-1 == nTermERegion.IndexOfAny(nTermEAny.ToCharArray()))
{
yield break;
}
AnyIn anyFeature1 = AnyIn.GetInstance(nflankAny, nFlankFeature);
AnyIn anyFeature2 = AnyIn.GetInstance(nTermEAny, nTermEFeature);
And andFeature = And.GetInstance(anyFeature1, anyFeature2);
Debug.Assert((bool)andFeature.Evaluate(nec)); // real assert - must only generate true features
yield return(andFeature);
}
private static void AddFeatureWithOptionalAndHlaAndST(IHashableFeature feature, bool includeAndHlaAndST, IEnumerable <IHashableFeature> hlaishEnumeration, bool checkThatNew, ref Set <IHashableFeature> featureSet)
{
featureSet.Add(feature, checkThatNew);
if (includeAndHlaAndST)
{
foreach (IHashableFeature hlaishFeature in hlaishEnumeration)
{
And featureAndHlaish = And.GetInstance(hlaishFeature, feature);
featureSet.Add(featureAndHlaish, checkThatNew);
}
}
}
private static void AddEiFeatures(bool includeChemicalProperties, bool includeAAFeatures, bool substractChemAACrissCrossFeatures, NEC nec, Pair <NEC, Hla> necAndHla, Set <IHashableFeature> hlaishFeatureSet, Set <IHashableFeature> featureSet)
{
for (int i = 0; i < nec.E.Length; ++i)
{
IHashableFeature featureE = E.GetInstance(i + 1);
string aminoAcid = GetAminoAcidFromEpitopePosition(nec, i);
if (includeAAFeatures)
{
IsAA featureAA = IsAA.GetInstance(aminoAcid, featureE);
featureSet.AddNew(featureAA);
Debug.Assert((bool)featureAA.Evaluate(necAndHla)); // real assert - must only generate true features
foreach (IHashableFeature hlaishFeature in hlaishFeatureSet)
{
if (substractChemAACrissCrossFeatures && hlaishFeature is HasAAProp)
{
continue;
}
And featureHlaishAndAA = And.GetInstance(hlaishFeature, featureAA);
featureSet.AddNew(featureHlaishAndAA);
Debug.Assert((bool)featureHlaishAndAA.Evaluate(necAndHla)); // real assert - must only generate true features
}
}
if (includeChemicalProperties)
{
foreach (string property in VirusCount.KmerProperties.AaToPropList[aminoAcid])
{
HasAAProp featureAAProp = HasAAProp.GetInstance(property, featureE);
featureSet.AddNew(featureAAProp);
Debug.Assert((bool)featureAAProp.Evaluate(necAndHla)); // real assert - must only generate true features
foreach (IHashableFeature hlaishFeature in hlaishFeatureSet)
{
if (substractChemAACrissCrossFeatures && hlaishFeature is IsAA)
{
continue;
}
And featureHlaishAndAAProb = And.GetInstance(hlaishFeature, featureAAProp);
featureSet.AddNew(featureHlaishAndAAProb);
Debug.Assert((bool)featureHlaishAndAAProb.Evaluate(necAndHla)); // real assert - must only generate true features
}
}
}
}
//All of the above with AA replaced by chemical property of AA
}
public static IEnumerable <IHashableFeature> GetAndNFlankNotNTermEEnumeration(NEC nec)
{
NFlank nFlankFeature = NFlank.GetInstance();
string nFlankRegion = nFlankFeature.FeatureFunction(nec);
Set <string> aaSeqSet = Set <string> .GetInstance(SpecialFunctions.SubstringEnumeration(nFlankRegion, 1));
foreach (string aaSeq in aaSeqSet)
{
Not notFeature = Not.GetInstance(In.GetInstance(aaSeq, NTermE.GetInstance()));
//Only generate the feature when the amino acid is not in NTermE
if ((bool)notFeature.Evaluate(nec))
{
In inFeature = In.GetInstance(aaSeq, nFlankFeature);
And andFeature = And.GetInstance(inFeature, notFeature);
Debug.Assert((bool)andFeature.Evaluate(nec)); // real assert - must only generate true features
yield return(andFeature);
}
}
}
new public static KeepTest <Dictionary <string, string> > GetInstance(string inputDirectory,
string binarySeqFileName, string hlaFileName, string keepTestName, int merSize, Dictionary <int, string> pidToCaseName)
{
//!!!would be nice of classes could parse themselves
if (keepTestName.StartsWith(KeepEndOfGag.Prefix))
{
bool keepIt = bool.Parse(keepTestName.Substring(KeepEndOfGag.Prefix.Length));
return(KeepEndOfGag.GetInstance(keepIt));
}
else if (keepTestName.StartsWith(K1.Prefix))
{
int k1 = int.Parse(keepTestName.Substring(K1.Prefix.Length));
return(K1.GetInstance(k1));
}
//else if (keepTestName.StartsWith(K2.Prefix))
//{
// int k2 = int.Parse(keepTestName.Substring(K2.Prefix.Length));
// return K2.GetInstance(k2);
//}
else if (keepTestName.StartsWith(KeepNonOverlappingAA.Prefix))
{
return(KeepNonOverlappingAA.GetInstance());
}
else if (keepTestName.StartsWith(KeepGene.Prefix))
{
string geneRange = keepTestName.Substring(KeepGene.Prefix.Length);
return(KeepGene.GetInstance(geneRange));
}
else if (keepTestName.StartsWith(KeepSpecificRows.Prefix))
{
return(KeepSpecificRows.GetInstance(keepTestName.Substring(KeepSpecificRows.Prefix.Length)));
}
else if (keepTestName.StartsWith(KeepSpecificRow.Prefix))
{
return(KeepSpecificRow.GetInstance(keepTestName.Substring(KeepSpecificRow.Prefix.Length)));
}
else if (keepTestName.StartsWith(KeepSpecificGenes.Prefix))
{
return(KeepSpecificGenes.GetInstance(keepTestName.Substring(KeepSpecificGenes.Prefix.Length)));
}
else if (keepTestName.StartsWith(KeepOneOfAAPair.Prefix))
{
return(KeepOneOfAAPair.GetInstance());
}
else if (keepTestName.StartsWith(KeepAllButSamePosition.Prefix))
{
return(KeepAllButSamePosition.GetInstance());
}
else if (keepTestName.StartsWith(KeepAllButSameDeletion.Prefix))
{
return(KeepAllButSameDeletion.GetInstance());
}
else if (keepTestName.StartsWith(KeepNonTrivialRows.Prefix))
{
return(new KeepNonTrivialRows());
}
else if (keepTestName.StartsWith(KeepTestTemp.Prefix))
{
return(KeepTestTemp.GetInstance());
}
//else if (keepTestName.StartsWith(KeepPollockOneDirection.Prefix))
//{
// return KeepPollockOneDirection.GetInstance(keepTestName.Substring(KeepPollockOneDirection.Prefix.Length));
//}
//else if (keepTestName.StartsWith(KeepFisherOneDirection.Prefix))
//{
// return KeepFisherOneDirection.GetInstance(keepTestName.Substring(KeepFisherOneDirection.Prefix.Length));
//}
else if (keepTestName.StartsWith(KeepNonRare.Prefix))
{
return(KeepNonRare.GetInstance(keepTestName.Substring(KeepNonRare.Prefix.Length)));
}
else if (keepTestName.StartsWith(KeepPredictorTargetPairs.Prefix))
{
return(KeepPredictorTargetPairs.GetInstance(keepTestName.Substring(KeepPredictorTargetPairs.Prefix.Length)));
}
else if (keepTestName.StartsWith("JointGagPolTest"))
{
return(And <Dictionary <string, string> > .GetInstance(
//KeepRandom<Dictionary<string,string>>.GetInstance(0, 0.001), // how do we make it the same when we count and when we really run through it?
KeepOneOfAAPair.GetInstance(),
KeepNonOverlappingAA.GetInstance(),
KeepSpecificGenes.GetInstance(keepTestName.Substring("JointGagPolTest".Length))));
}
else
{
return(KeepTest <TRow> .GetInstance(inputDirectory, binarySeqFileName, null, keepTestName, merSize, pidToCaseName));
}
}
//static private Dictionary<Pair<NEC, Hla>, bool> CloseHuman = null;
//GeneratorType.Hla | GeneratorType.Position | GeneratorType.Property | GeneratorType.AndHla | GeneratorType.Zero6Supertype | GeneratorType.AndZero6Supertype
private static Set <IHashableFeature> GenerateFeatureSet(
object entity, string supertypeTableSource,
int?flankSizeOrNull,
bool includeFlankNECFeatures,
bool includeChemicalProperties, bool includeAAFeatures,
bool addEiFeatures
)
{
bool includeAndHlaAndSTWithEpitopeAdjFeatures = false;
bool subtractSupertypeFeatures = false;
bool subtractHlaFeatures = false;
bool substractChemAACrissCrossFeatures = false;
SpecialFunctions.CheckCondition(!includeAndHlaAndSTWithEpitopeAdjFeatures || includeFlankNECFeatures);
Pair <NEC, Hla> necAndHlaX = (Pair <NEC, Hla>)entity;
NEC nec = (null == flankSizeOrNull) ? necAndHlaX.First : NEC.GetInstance(necAndHlaX.First, (int)flankSizeOrNull);
Hla hla = necAndHlaX.Second;
Debug.Assert(nec.N.Length == nec.C.Length); // real assert
Pair <NEC, Hla> necAndHla = new Pair <NEC, Hla>(nec, hla);
Set <IHashableFeature> hlaishFeatureSet = new Set <IHashableFeature>();
CreateAndAddHlaFeature(subtractHlaFeatures, hla, necAndHla, ref hlaishFeatureSet);
CreateAndAddFeatureSupertype(supertypeTableSource, subtractSupertypeFeatures, hla, necAndHla, ref hlaishFeatureSet, Assembly.GetExecutingAssembly(), Predictor.ResourceString);
Set <IHashableFeature> featureSet = Set <IHashableFeature> .GetInstance(hlaishFeatureSet);
if (addEiFeatures)
{
AddEiFeatures(includeChemicalProperties, includeAAFeatures, substractChemAACrissCrossFeatures, nec, necAndHla, hlaishFeatureSet, featureSet);
}
if (includeFlankNECFeatures)
{
List <IHashableFeature> aaInNFlankFeatureList = new List <IHashableFeature>(In.GetAASeqInRegionInstance(1, necAndHla, NFlank.GetInstance()));
DebugCheckThatEvaluatesToTrue(necAndHla, aaInNFlankFeatureList);
if (includeAAFeatures)
{
featureSet.AddNewRange(aaInNFlankFeatureList); //AA in N flank
featureSet.AddNewRange(In.GetAASeqInRegionInstance(2, necAndHla, NFlank.GetInstance())); //AA1-AA2 in Nflank
featureSet.AddNewRange(SubSeq.GetInSubSeqEnumeration(NFlank.GetInstance(), false, 1, necAndHla)); //AA@x in N flank (numbering is 5 4 3 2 1)
featureSet.AddNewRange(SubSeq.GetInSubSeqEnumeration(NFlank.GetInstance(), false, 2, necAndHla)); //AA1-AA2@x in Nflank (x is position of AA2, i.e., the smaller number)
featureSet.AddNewRange(In.GetAASeqInRegionInstance(1, necAndHla, CFlank.GetInstance())); //AA in Cflank
featureSet.AddNewRange(In.GetAASeqInRegionInstance(2, necAndHla, CFlank.GetInstance())); //AA1-AA2 in Cflank
featureSet.AddNewRange(SubSeq.GetInSubSeqEnumeration(CFlank.GetInstance(), true, 1, necAndHla)); //AA@x in C flank (numbering is 1 2 3 4 5)
featureSet.AddNewRange(SubSeq.GetInSubSeqEnumeration(CFlank.GetInstance(), true, 2, necAndHla)); //AA1-AA2@x in Cflank (x is position of AA1, i.e., the smaller number)
}
if (includeChemicalProperties)
{
featureSet.AddNewOrOldRange(InProperty.GetPropertySeqInRegionInstance(1, necAndHla, NFlank.GetInstance()));
featureSet.AddNewOrOldRange(SubSeq.GetInPropertySubSeqEnumeration(NFlank.GetInstance(), false, 1, necAndHla));
featureSet.AddNewOrOldRange(InProperty.GetPropertySeqInRegionInstance(1, necAndHla, CFlank.GetInstance()));
featureSet.AddNewOrOldRange(SubSeq.GetInPropertySubSeqEnumeration(CFlank.GetInstance(), true, 1, necAndHla));
featureSet.AddNewOrOldRange(InProperty.GetPropertySeqInRegionInstance(2, necAndHla, NFlank.GetInstance()));
featureSet.AddNewOrOldRange(SubSeq.GetInPropertySubSeqEnumeration(NFlank.GetInstance(), false, 2, necAndHla));
featureSet.AddNewOrOldRange(InProperty.GetPropertySeqInRegionInstance(2, necAndHla, CFlank.GetInstance()));
featureSet.AddNewOrOldRange(SubSeq.GetInPropertySubSeqEnumeration(CFlank.GetInstance(), true, 2, necAndHla));
}
}
if (includeFlankNECFeatures)
{
if (includeAAFeatures)
{
//EV in Epitope
AddFeatureWithOptionalAndHlaAndST(In.GetAASeqInRegionInstance(2, necAndHla, Epitope.GetInstance()), includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, false, ref featureSet);//AA1-AA2 in Epitope
//RR in Epitope[@1-2]
AddFeatureWithOptionalAndHlaAndST(SubSeq.GetInSubSeqEnumeration(Epitope.GetInstance(), true, 2, necAndHla), includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, false, ref featureSet);//AA1-AA2@x in Epitope (x is position of AA1, i.e., the smaller number)
}
if (includeChemicalProperties)
{
//polar,cyclic in Epitope
AddFeatureWithOptionalAndHlaAndST(InProperty.GetPropertySeqInRegionInstance(2, necAndHla, Epitope.GetInstance()), includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, false, ref featureSet);
//polar,large in Epitope[@8-9]
AddFeatureWithOptionalAndHlaAndST(SubSeq.GetInPropertySubSeqEnumeration(Epitope.GetInstance(), true, 2, necAndHla), includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, false, ref featureSet);
}
//AA1-AA2 in Nflank,Epitope, etc
if (null != flankSizeOrNull && (int)flankSizeOrNull > 0)
{
string epitope = (string)Epitope.GetInstance().Evaluate(entity);
SubSeq lastNAAFeature = SubSeq.GetInstance(1, 1, false, NFlank.GetInstance());
string lastNAA = (string)lastNAAFeature.Evaluate(entity);
In inLastNAA = In.GetInstance(lastNAA, lastNAAFeature);
SubSeq firstEAAFeature = SubSeq.GetInstance(1, 1, true, Epitope.GetInstance());
string firstEAA = (string)firstEAAFeature.Evaluate(entity);
Debug.Assert(firstEAA == epitope.Substring(0, 1));// real assert
In inFirstEAA = In.GetInstance(firstEAA, firstEAAFeature);
SubSeq lastEAAFeature = SubSeq.GetInstance(epitope.Length, epitope.Length, true, Epitope.GetInstance());
string lastEAA = (string)lastEAAFeature.Evaluate(entity);
In inLastEAA = In.GetInstance(lastEAA, lastEAAFeature);
SubSeq firstCAAFeature = SubSeq.GetInstance(1, 1, true, CFlank.GetInstance());
string firstCAA = (string)firstCAAFeature.Evaluate(entity);
In inFirstCAA = In.GetInstance(firstCAA, firstCAAFeature);
if (includeAAFeatures)
{
And andLastNNAAFirstEAA = And.GetInstance(inLastNAA, inFirstEAA);
AddFeatureWithOptionalAndHlaAndST(andLastNNAAFirstEAA, includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, /*checkThatNew*/ true, ref featureSet);
And andLastEAAFirstCAA = And.GetInstance(inLastEAA, inFirstCAA);
AddFeatureWithOptionalAndHlaAndST(andLastEAAFirstCAA, includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, /*checkThatNew*/ true, ref featureSet);
}
if (includeChemicalProperties)
{
foreach (string lastNProperty in KmerProperties.AaToPropList[Biology.GetInstance().OneLetterAminoAcidAbbrevTo3Letter[lastNAA[0]]])
{
InProperty inLastNProperty = InProperty.GetInstance(lastNProperty, lastNAAFeature);
foreach (string firstEProperty in KmerProperties.AaToPropList[Biology.GetInstance().OneLetterAminoAcidAbbrevTo3Letter[firstEAA[0]]])
{
InProperty inFirstEProperty = InProperty.GetInstance(firstEProperty, firstEAAFeature); //!!!get this out of the loop?
And andLastNPropertyFirstEProperty = And.GetInstance(inLastNProperty, inFirstEProperty);
AddFeatureWithOptionalAndHlaAndST(andLastNPropertyFirstEProperty, includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, /*checkThatNew*/ false, ref featureSet);
Debug.Assert((bool)andLastNPropertyFirstEProperty.Evaluate(necAndHla));
}
}
foreach (string lastEProperty in KmerProperties.AaToPropList[Biology.GetInstance().OneLetterAminoAcidAbbrevTo3Letter[lastEAA[0]]])
{
InProperty inlastEProperty = InProperty.GetInstance(lastEProperty, lastEAAFeature);
foreach (string firstCProperty in KmerProperties.AaToPropList[Biology.GetInstance().OneLetterAminoAcidAbbrevTo3Letter[firstCAA[0]]])
{
InProperty infirstCProperty = InProperty.GetInstance(firstCProperty, firstCAAFeature); //!!!get this out of the loop?
And andlastEPropertyfirstCProperty = And.GetInstance(inlastEProperty, infirstCProperty);
AddFeatureWithOptionalAndHlaAndST(andlastEPropertyfirstCProperty, includeAndHlaAndSTWithEpitopeAdjFeatures, hlaishFeatureSet, /*checkThatNew*/ false, ref featureSet);
Debug.Assert((bool)andlastEPropertyfirstCProperty.Evaluate(necAndHla));
}
}
}
}
}
return(featureSet);
}