public void TestFitMinusOneScore(int precursor, string adduct, string commonName, string id, string rawFilePath)
{
var lipid = new Lipid()
{
AdductFull = adduct, CommonName = commonName
};
var lipidTarget = lipid.CreateLipidTarget();
var composition = lipidTarget.Composition;
var compMinus1 = new Composition(composition.C, composition.H - 1, composition.N, composition.O, composition.S, composition.P); //Subtract one hydrogen to make this a minus1 fit score
var lcmsRun = PbfLcMsRun.GetLcMsRun(rawFilePath);
var spectrum = lcmsRun.GetSpectrum(precursor);
var relativeIntensityThreshold = 0.1;
var tolerance = new Tolerance(30, ToleranceUnit.Ppm);
//Get the values to use to calculate pearson correlation
var observedPeaks = LipidUtil.GetAllIsotopePeaks(spectrum, compMinus1, tolerance,
relativeIntensityThreshold);
if (observedPeaks == null)
{
Console.WriteLine("Observed peaks is null for scan " + id);
}
var isotopomerEnvelope = IsoProfilePredictor.GetIsotopomerEnvelop(
compMinus1.C,
compMinus1.H,
compMinus1.N,
compMinus1.O,
compMinus1.S);
var observedIntensities = new double[observedPeaks.Length];
for (var i = 0; i < observedPeaks.Length; i++)
{
var observedPeak = observedPeaks[i];
observedIntensities[i] = observedPeak != null ? (float)observedPeak.Intensity : 0.0;
}
Console.WriteLine("The theoretical y values are: ");
foreach (var value in isotopomerEnvelope.Envolope)
{
Console.WriteLine(value + ", ");
}
Console.WriteLine("The observed peak intensity x values are: ");
foreach (var value in observedIntensities)
{
Console.WriteLine(value + ", ");
}
}
public void PearsonCorrelationFileCombiner(string directoryPath)
{
var dirFiles = Directory.GetFiles(directoryPath);
var correlationCalculator = new PearsonCorrelationFitUtil();
var cosineCalculator = new CosineFitUtil();
// Each dictionary corresponds to a dataset, each dictionary key corresponds to the TSV header.
var results = new List <Dictionary <string, List <string> > >();
var headers = new HashSet <string>();
foreach (var pathToResults in dirFiles.Where(path => path.EndsWith(".txt")))
{
var datasetName = Path.GetFileNameWithoutExtension(pathToResults);
var pathToRaw = GetRawFilePath(directoryPath, datasetName);
var rawName = Path.GetFileName(pathToRaw);
if (string.IsNullOrEmpty(pathToRaw))
{
continue;
}
var lcmsRun = PbfLcMsRun.GetLcMsRun(pathToRaw);
var tolerance = new Tolerance(30, ToleranceUnit.Ppm);
using (var reader = new StreamReader(new FileStream(pathToResults, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)))
{
results.Add(new Dictionary <string, List <string> >()); // Add dictionary for new dataset.
var datasetResults = results.Last(); // Results for the current dataset.
var lineCount = 0;
var headerToIndex = new Dictionary <string, int>();
while (!reader.EndOfStream)
{
var line = reader.ReadLine();
if (string.IsNullOrWhiteSpace(line))
{
continue;
}
var pieces = line.Split('\t').ToArray();
if (lineCount++ == 0)
{ // First line
for (var i = 0; i < pieces.Length; i++)
{
var header = pieces[i];
headerToIndex.Add(header, i);
datasetResults.Add(header, new List <string>());
}
datasetResults.Add("Raw File", new List <string>());
datasetResults.Add("Pearson Corr Score", new List <string>());
datasetResults.Add("Pearson Corr M-1 Score", new List <string>());
datasetResults.Add("Cosine Score", new List <string>());
datasetResults.Add("Cosine M-1 Score", new List <string>());
headers.UnionWith(datasetResults.Keys);
continue;
}
var precursor = Convert.ToInt32(pieces[headerToIndex["Precursor Scan"]]);
var commonName = pieces[headerToIndex["Common Name"]];
var adduct = pieces[headerToIndex["Adduct"]];
var spectrum = lcmsRun.GetSpectrum(precursor);
if (spectrum == null)
{
Console.WriteLine("Invalid scan number: {0}", precursor);
continue;
}
var lipid = new Lipid {
AdductFull = adduct, CommonName = commonName
};
var lipidTarget = lipid.CreateLipidTarget();
var spectrumSearchResult = new SpectrumSearchResult(null, null, spectrum, null, null, new Xic(), lcmsRun)
{
PrecursorTolerance = tolerance
};
var pearsonCorrScore = correlationCalculator.GetFitScore(spectrumSearchResult, lipidTarget.Composition);
var pearsonCorrMinus1Score = correlationCalculator.GetFitMinus1Score(spectrumSearchResult, lipidTarget.Composition);
var cosineScore = cosineCalculator.GetFitScore(spectrumSearchResult, lipidTarget.Composition);
var cosineMinus1Score = cosineCalculator.GetFitScore(
spectrumSearchResult,
lipidTarget.Composition);
// Add results to results dictionary.
datasetResults["Raw File"].Add(rawName);
foreach (var header in headerToIndex.Keys)
{
datasetResults[header].Add(pieces[headerToIndex[header]]);
}
datasetResults["Pearson Corr Score"].Add(pearsonCorrScore.ToString());
datasetResults["Pearson Corr M-1 Score"].Add(pearsonCorrMinus1Score.ToString());
datasetResults["Cosine Score"].Add(cosineScore.ToString());
datasetResults["Cosine M-1 Score"].Add(cosineMinus1Score.ToString());
}
}
}
// Write results
var outputFilePath = Path.Combine(directoryPath, "training.tsv");
using (var writer = new StreamWriter(outputFilePath))
{
// Write headers
foreach (var header in headers)
{
writer.Write("{0}\t", header);
}
writer.WriteLine();
// Write data
foreach (var datasetResults in results)
{
var fileLength = datasetResults["Pearson Corr Score"].Count;
for (var i = 0; i < fileLength; i++)
{
foreach (var header in headers)
{
var value = datasetResults.ContainsKey(header) ? datasetResults[header][i] : string.Empty;
writer.Write("{0}\t", value);
}
writer.WriteLine();
}
}
}
}
public void TestPearsonCorrelationWholeFile(string directoryPath)
{
var dirFiles = Directory.GetFiles(directoryPath);
var correlationCalculator = new PearsonCorrelationFitUtil();
foreach (var pathToResults in dirFiles.Where(path => path.EndsWith(".txt")))
{
var datasetName = Path.GetFileNameWithoutExtension(pathToResults);
var pathToRaw = GetRawFilePath(directoryPath, datasetName);
if (string.IsNullOrEmpty(pathToRaw))
{
continue;
}
var lcmsRun = PbfLcMsRun.GetLcMsRun(pathToRaw);
var tolerance = new Tolerance(30, ToleranceUnit.Ppm);
var rawFileName = Path.GetFileName(pathToRaw);
var datasetDirPath = Path.GetDirectoryName(pathToResults);
var outputFileName = string.Format("{0}_training.tsv", datasetName);
var outputPath = Path.Combine(datasetDirPath, outputFileName);
using (var writer = new StreamWriter(outputPath))
using (var reader = new StreamReader(new FileStream(pathToResults, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)))
{
var lineCount = 0;
var headerToIndex = new Dictionary <string, int>();
while (!reader.EndOfStream)
{
var line = reader.ReadLine();
if (string.IsNullOrWhiteSpace(line))
{
continue;
}
var pieces = line.Split('\t').ToArray();
if (lineCount++ == 0)
{ // First line
writer.Write("Raw File\t");
for (var i = 0; i < pieces.Length; i++)
{
headerToIndex.Add(pieces[i], i);
writer.Write("{0}\t", pieces[i]);
}
writer.WriteLine("Fit Score\tFit M-1 Score");
continue;
}
var precursor = Convert.ToInt32(pieces[headerToIndex["Precursor Scan"]]);
var commonName = pieces[headerToIndex["Common Name"]];
var adduct = pieces[headerToIndex["Adduct"]];
var spectrum = lcmsRun.GetSpectrum(precursor);
if (spectrum == null)
{
Console.WriteLine("Invalid scan number: {0}", precursor);
continue;
}
var lipid = new Lipid {
AdductFull = adduct, CommonName = commonName
};
var lipidTarget = lipid.CreateLipidTarget();
var spectrumSearchResult = new SpectrumSearchResult(null, null, spectrum, null, null, new Xic(), lcmsRun)
{
PrecursorTolerance = tolerance
};
var fitScore = correlationCalculator.GetFitScore(spectrumSearchResult, lipidTarget.Composition);
var fitMinus1Score = correlationCalculator.GetFitMinus1Score(spectrumSearchResult, lipidTarget.Composition);
writer.Write(rawFileName + "\t");
writer.Write(line);
writer.WriteLine("{0}\t{1}", fitScore, fitMinus1Score);
}
}
}
}
public void FillCompAndMassForTargetsFile()
{
const string targetsFile = @"E:\Source\Liquid\trunk\LiquidTest\testFiles\Global_LipidMaps_POS_7b_Decoys.txt";
const string outputfile = @"E:\Source\Liquid\trunk\LiquidTest\testFiles\Global_LipidMaps_POS_7b_Decoys_test.txt";
const int massCol = 6;
const int compCol = 7;
var output = new List <string>();
using (var targets = new StreamReader(new FileStream(targetsFile, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)))
{
var header = targets.ReadLine();
output.Add(header);
while (!targets.EndOfStream)
{
var target = targets.ReadLine();
if (string.IsNullOrWhiteSpace(target))
{
continue;
}
var splitTarget = target.Split('\t');
try
{
if (string.IsNullOrEmpty(splitTarget[massCol]) || string.IsNullOrEmpty(splitTarget[compCol]))
{
var lipid = new Lipid {
CommonName = splitTarget[1], AdductFull = splitTarget[2]
};
var newTarget = lipid.CreateLipidTarget();
splitTarget[massCol] = newTarget.Composition.Mass.ToString(CultureInfo.InvariantCulture);
splitTarget[compCol] = newTarget.Composition.ToPlainString();
}
var rebuilt = new StringBuilder();
rebuilt.Append(splitTarget[0]);
for (var i = 1; i < splitTarget.Length; i++)
{
rebuilt.Append("\t" + splitTarget[i]);
}
output.Add(rebuilt.ToString());
}
catch (Exception)
{
// Ignore the error
}
}
}
using (var outstream = new StreamWriter(outputfile))
{
foreach (var x in output)
{
outstream.WriteLine(x);
}
}
}
