private void LoadUpPredictor(string modelName, int eLength, int ncLength, Converter <Hla, Hla> hlaForNormalization)
{
//Load up the predictor
string featurerizerName;
switch (modelName.ToLower())
{
//!!!would be better not to have multiple of these switch statements around - looks like a job for a Class
case "lanliedb03062007":
featurerizerName = "[email protected]";
SampleNEC = NEC.GetInstance("", new string(' ', eLength), "");
HlaFactory = HlaFactory.GetFactory("MixedWithB15AndA68");
SourceDataFileName = "lanlIedb03062007.pos.source.txt";
NameList = new string[] { "LANL", "IEDB" };
break;
default:
SpecialFunctions.CheckCondition(false, "Don't know what featurerizer to use for the model");
featurerizerName = null;
SourceDataFileName = null;
NameList = null;
break;
}
Converter <object, Set <IHashableFeature> > featurizer = FeatureLib.CreateFeaturizer(featurerizerName);
//GeneratorType generatorType = GeneratorType.ComboAndZero6SuperType;
//FeatureSerializer featureSerializer = PositiveNegativeExperimentCollection.GetFeatureSerializer();
//KmerDefinition = kmerDefinition;
//HlaResolution hlaResolution = HlaResolution.ABMixed;
string resourceName = string.Format("maxentModel{0}{1}{2}{3}.xml", modelName.Split('.')[0], SampleNEC.N.Length, SampleNEC.E.Length, SampleNEC.C.Length);
EpitopeLearningDataList = new List <EpitopeLearningDataDupHlaOK>();
using (StreamReader streamReader = Predictor.OpenResource(resourceName))
{
Logistic = (Logistic)FeatureLib.FeatureSerializer.FromXmlStreamReader(streamReader);
//Logistic.FeatureGenerator = EpitopeFeatureGenerator.GetInstance(KmerDefinition, generatorType, featureSerializer).GenerateFeatureSet;
Logistic.FeatureGenerator = FeatureLib.CreateFeaturizer(featurerizerName);
foreach (string name in NameList)
{
EpitopeLearningData epitopeLearningDataX = EpitopeLearningData.GetDbWhole(HlaFactory, SampleNEC.E.Length, name, SourceDataFileName);
Debug.Assert(epitopeLearningDataX.Count > 0, "Expect given data to have some data");
//!!!combine with previous step
EpitopeLearningDataDupHlaOK epitopeLearningData = new EpitopeLearningDataDupHlaOK(epitopeLearningDataX.Name);
foreach (KeyValuePair <Pair <string, Hla>, bool> merAndHlaAndLabel in epitopeLearningDataX)
{
Hla hlaIn = merAndHlaAndLabel.Key.Second;
Hla hlaOut = hlaForNormalization(hlaIn);
Dictionary <Hla, Dictionary <Hla, bool> > hla2ToHlaToLabel = SpecialFunctions.GetValueOrDefault(epitopeLearningData, merAndHlaAndLabel.Key.First);
Dictionary <Hla, bool> hlaToLabel = SpecialFunctions.GetValueOrDefault(hla2ToHlaToLabel, hlaOut);
hlaToLabel.Add(hlaIn, merAndHlaAndLabel.Value);
}
EpitopeLearningDataList.Add(epitopeLearningData);
}
}
HlaForNormalization = hlaForNormalization;
}
//!!!very similar to other code
public static Dictionary <Pair <NEC, Hla>, bool> ReadTable(HlaFactory hlaFactory, string fileName, bool dedup)
{
Dictionary <Pair <NEC, Hla>, bool> table = new Dictionary <Pair <NEC, Hla>, bool>();
foreach (Dictionary <string, string> row in SpecialFunctions.TabFileTable(fileName, "N\tepitope\tC\thla\tlabel", false))
{
string n = row["N"];
string epitope = row["epitope"];
SpecialFunctions.CheckCondition(Biology.GetInstance().LegalPeptide(epitope), string.Format("Peptide, '{0}', contains illegal char.", epitope));
string c = row["C"];
NEC nec = NEC.GetInstance(n, epitope, c);
Hla hla = hlaFactory.GetGroundInstance(row["hla"]);
string labelString = row["label"];
SpecialFunctions.CheckCondition(labelString == "0" || labelString == "1", "Expect label to be '0' or '1'");
Pair <NEC, Hla> pair = new Pair <NEC, Hla>(nec, hla);
bool labelAsBool = (labelString == "1");
if (dedup && table.ContainsKey(pair))
{
SpecialFunctions.CheckCondition(table[pair] == labelAsBool, "The example " + pair.ToString() + " appears with contradictory labels.");
continue;
}
table.Add(pair, labelAsBool);
}
return(table);
}
////!!! this could be moved into a class
//private object CreateKey(Prediction prediction, Best display)
//{
// switch (display)
// {
// case Best.overall:
// return "best";
// case Best.perHla:
// return prediction.Hla;
// case Best.perPrediction:
// return prediction;
// case Best.perLength:
// return prediction.K;
// case Best.perHlaAndLength:
// return new Pair<Hla, int>(prediction.Hla, prediction.K);
// default:
// SpecialFunctions.CheckCondition(false, "Don't know how to display " + display.ToString());
// return null;
// }
//}
/// <summary>
/// HlaSetSpecification class choices:
/// HlaSetSpecification.Singleton – Means that an Hla will be given and it is the only hla to be considered
/// HlaSetSpecification.Supertype – Means that a supertype will be given and it’s hlas should be considered
/// HlaSetSpecification.All – Means to consider all known hlas
/// </summary>
/// <param name="inputPeptide">a string of amino acids</param>
/// <param name="merLength">A value from the MerLength enum, which includes MerLength.scan, MerLength.given, MerLength.Eight, etc</param>
/// <param name="hlaSetSpecification">A predefined HlaSetSpecification class.</param>
/// <param name="hlaOrSupertypeOrNull">The hla or supertype required by HlaSetSpecification, or null for HlaSetSpecification.All</param>
/// <param name="modelOnly">If should report the probability from the model, even when the epitope is on a source list.</param>
/// <returns></returns>
public IEnumerable <Prediction> PredictionEnumeration(string inputPeptide, MerLength merLength, int?dOfCenter, HlaSetSpecification hlaSetSpecification, string hlaOrSupertypeOrNull, bool modelOnly)
{
Set <Hla> hlaSet = HlaSet(hlaSetSpecification, hlaOrSupertypeOrNull);
foreach (int eLength in KEnumeration(merLength, inputPeptide.Length))
{
Predictor predictor = KToPredictor[eLength];
Dictionary <Hla, double> hlaToPriorLogOdds = KToHlaToPriorLogOdds[eLength];
int necLength = NCLength + eLength + NCLength;
foreach (int startIndex in StartIndexEnumeration(inputPeptide.Length, necLength, dOfCenter))
{
string peptide = inputPeptide.Substring(startIndex, necLength);
NEC nec = NEC.GetInstance(peptide, NCLength, eLength, NCLength);
foreach (Hla hla in hlaSet)
{
Hla hlaForNormalization = HlaForNormalization(hla);
double priorLogOddsOfThisLengthAndHla;
if (!hlaToPriorLogOdds.TryGetValue(hlaForNormalization, out priorLogOddsOfThisLengthAndHla))
{
SpecialFunctions.CheckCondition(!RaiseErrorIfNotFoundInNormalizationTable, string.Format("Hla '{0}' (which is '{1}' for the purposes of normalization) and is not found in the normalization table", hla, hlaForNormalization));
priorLogOddsOfThisLengthAndHla = SpecialFunctions.LogOdds(RatioOfTrueToFalseTrainingExample);
}
string source;
double originalP = predictor.Predict(nec, hla, modelOnly, out source);
double originalLogOdds = SpecialFunctions.LogOdds(originalP);
double correctedLogOdds = originalLogOdds + priorLogOddsOfThisLengthAndHla;
double posteriorProbability = SpecialFunctions.InverseLogOdds(correctedLogOdds);
double weightOfEvidence = correctedLogOdds - SpecialFunctions.LogOdds(RatioOfTrueToFalseTrainingExample);
Prediction prediction = Prediction.GetInstance(inputPeptide, hla, posteriorProbability, weightOfEvidence, nec, startIndex + NCLength + 1, startIndex + NCLength + eLength, source);
yield return(prediction);
}
}
}
}