private void handleException(Realm realm, OutOfMemoryError e) {
// Try to recover after OOM.
Runtime rt = Runtime.getRuntime();
long beforeGc = rt.freeMemory();
rt.gc();
long afterGc = rt.freeMemory();
if (afterGc > beforeGc && (afterGc - beforeGc) < 50_000_000) {
// Calling gc() cleared less than 50MB, assume unrecoverable OOM and rethrow error.
throw e;
}
// Create script exception with stacktrace from oom-error.
ScriptException exception =
newInternalError(realm.defaultContext(), Messages.Key.OutOfMemoryVM);
exception.setStackTrace(e.getStackTrace());
handleException(realm, exception);
}
/**
* Process spectra in the following way: Fragment each hit in KEGG starting with the lowest
* collision energy If hits were found they are used for the next measurement as input and the
* original molecule is not used.
*
* @param folder the folder
* @param file the file
*/
private void processSpectra(String folder, String file, int treeDepth) {
this.scoreMap = new HashMap<Integer, ArrayList<String>>();
if (blackList.contains(file)) {
completeLog += "Blacklisted Molecule: " + file;
histogramReal += "\n" + file + "\tBLACKLIST\t";
histogram += "\n" + file + "\tBLACKLIST\t";
histogramCompare += "\n" + file + "\tBLACKLIST\t";
return;
}
double exactMass = spectra.get(0).getExactMass();
// timing
long timeStart = System.currentTimeMillis();
HashMap<Double, Vector<String>> realScoreMap = new HashMap<Double, Vector<String>>();
int mode = spectra.get(0).getMode();
// instantiate and read in CID-KEGG.txt
String keggIdentifier = spectra.get(0).getKEGG();
completeLog +=
"\n\n============================================================================";
completeLog +=
"\nFile: " + spectra.get(0).getTrivialName() + " (KEGG Entry: " + keggIdentifier + ")";
// get candidates from kegg webservice...with with a given mzppm and mzabs
Vector<String> candidates =
KeggWebservice.KEGGbyMass(exactMass, (mzabs + PPMTool.getPPMDeviation(exactMass, mzppm)));
try {
GetKEGGIdentifier keggID = new GetKEGGIdentifier(folder + "CID-KEGG/CID-KEGG.txt");
// now find the corresponding KEGG entry
if (keggID.existInKEGG(spectra.get(0).getCID()))
keggIdentifier = keggID.getKEGGID(spectra.get(0).getCID());
} catch (IOException e) {
System.out.println(e.getMessage());
completeLog += "Error! Message: " + e.getMessage();
}
// comparison histogram
if (keggIdentifier.equals("none"))
histogramCompare += "\n" + file + "\t" + keggIdentifier + "\t\t" + exactMass;
else
histogramCompare +=
"\n" + file + "\t" + keggIdentifier + "\t" + candidates.size() + "\t" + exactMass;
// list of peaks which are contained in the real molecule
Vector<Peak> listOfPeaksCorresponding = new Vector<Peak>();
// list of peaks which are not contained in the real molecule
Vector<Peak> listOfPeaks = new Vector<Peak>();
// loop over all hits
for (int c = 0; c < candidates.size(); c++) {
this.foundHits = new Vector<IAtomContainer>();
// get mol file from kegg....remove "cpd:"
String candidate = KeggWebservice.KEGGgetMol(candidates.get(c).substring(4), this.keggPath);
IAtomContainer molecule = null;
try {
// write string to disk
new File(folder + file + "_Mol").mkdir();
File outFile = new File(folder + file + "_Mol/" + candidates.get(c).substring(4) + ".mol");
FileWriter out = new FileWriter(outFile);
out.write(candidate);
out.close();
// now fragment the retrieved molecule
molecule = Molfile.Read(folder + file + "_Mol/" + candidates.get(c).substring(4) + ".mol");
// now create a new folder to write the .mol files into
new File(folder + file).mkdir();
boolean status = new File(folder + file + "/" + candidates.get(c).substring(4)).mkdir();
} catch (IOException e) {
completeLog += "Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
} catch (CDKException e) {
completeLog += "Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
}
try {
// add hydrogens
CDKAtomTypeMatcher matcher = CDKAtomTypeMatcher.getInstance(molecule.getBuilder());
for (IAtom atom : molecule.atoms()) {
IAtomType type = matcher.findMatchingAtomType(molecule, atom);
AtomTypeManipulator.configure(atom, type);
}
CDKHydrogenAdder hAdder = CDKHydrogenAdder.getInstance(molecule.getBuilder());
hAdder.addImplicitHydrogens(molecule);
AtomContainerManipulator.convertImplicitToExplicitHydrogens(molecule);
}
// there is a bug in cdk?? error happens when there is a S or Ti in the molecule
catch (IllegalArgumentException e) {
completeLog += "Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
// skip it
continue;
} catch (CDKException e) {
completeLog += "Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
// skip it
continue;
}
// get peak list....it is reset if this is not the first run
Vector<Peak> peakList = spectra.get(0).getPeakList();
// create a new instance
Fragmenter fragmenter =
new Fragmenter(
(Vector<Peak>) peakList.clone(),
mzabs,
mzppm,
mode,
breakAromaticRings,
quickRedundancy,
false,
false);
double combinedScore = 0;
int combinedHits = 0;
int combinedPeakCount = 0;
String peaks = "";
combinedPeakCount = peakList.size();
int count = 0;
boolean first = true;
// loop over the different collision energies
for (WrapperSpectrum spectrum : spectra) {
List<IAtomContainer> l = null;
long start = System.currentTimeMillis();
try {
if (first) {
try {
l = fragmenter.generateFragmentsInMemory(molecule, true, treeDepth);
count++;
} catch (OutOfMemoryError e) {
System.out.println("OUT OF MEMORY ERROR! " + candidates.get(c).substring(4));
completeLog +=
"Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
continue;
}
first = false;
} else {
peakList = spectrum.getPeakList();
combinedPeakCount += peakList.size();
// set the current peak list...
fragmenter.setPeakList((Vector<Peak>) peakList.clone());
try {
// l = fragmenter.generateFragmentsHierarchical(foundHits, true);
System.err.println("REMOVED....no improvement!");
System.exit(1);
count++;
} catch (OutOfMemoryError e) {
System.out.println("OUT OF MEMORY ERROR! " + candidates.get(c).substring(4));
completeLog +=
"Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
continue;
}
}
long time = System.currentTimeMillis() - start;
System.out.println("Benötigte Zeit: " + time);
System.out.println("Got " + l.size() + " fragments");
System.out.println("Needed " + fragmenter.getNround() + " calls to generateFragments()");
new File(
folder
+ file
+ "/"
+ candidates.get(c).substring(4)
+ "/"
+ spectrum.getCollisionEnergy())
.mkdir();
for (int i = 0; i < l.size(); i++) {
try {
// write fragments to disk
FileWriter w =
new FileWriter(
new File(
folder
+ file
+ "/"
+ candidates.get(c).substring(4)
+ "/"
+ spectrum.getCollisionEnergy()
+ "/frag_"
+ i
+ ".mol"));
MDLWriter mw = new MDLWriter(w);
mw.write(new Molecule(l.get(i)));
mw.close();
} catch (IOException e) {
System.out.println("IOException: " + e.toString());
completeLog +=
"Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
} catch (Exception e) {
System.out.println(e.toString());
completeLog +=
"Error: " + candidates.get(c).substring(4) + " Message: " + e.getMessage();
}
}
// Draw molecule and its fragments
if (showDiagrams) Render.Draw(molecule, l, "Original Molecule");
if (pdf) {
// Create PDF Output
l.add(0, molecule);
DisplayStructure ds1 = null;
// create pdf subfolder
new File(folder + file + "/" + candidates.get(c) + "pdf/").mkdir();
ds1 =
new WritePDFTable(
true,
300,
300,
0.9,
2,
false,
false,
folder + file + "/" + candidates.get(c) + "pdf/");
for (int i = 0; i < l.size(); i++) {
// ds = new displayStructure(false, 300, 300, 0.9, false, "PDF",
// "/home/basti/WorkspaceJava/TandemMSLookup/fragmenter/Test");
assert ds1 != null;
ds1.drawStructure(l.get(i), i);
}
if (ds1 != null) ds1.close();
}
// now read the saved mol files
List<IAtomContainer> fragments =
Molfile.Readfolder(
folder
+ file
+ "/"
+ candidates.get(c).substring(4)
+ "/"
+ spectrum.getCollisionEnergy());
List<IAtomContainer> fragmentsList = new ArrayList<IAtomContainer>();
for (int i = 0; i < fragments.size(); i++) {
fragmentsList.add(fragments.get(i));
}
// get the original peak list again
// spectrum = new SpectrumWrapper(folder + file + ".txt");
peakList = (Vector<Peak>) spectrum.getPeakList().clone();
// clean up peak list
CleanUpPeakList cList = new CleanUpPeakList(peakList);
Vector<Peak> cleanedPeakList = cList.getCleanedPeakList(spectrum.getExactMass());
// now find corresponding fragments to the mass
AssignFragmentPeak afp = new AssignFragmentPeak();
afp.setHydrogenTest(hydrogenTest);
afp.assignFragmentPeak(
fragments, cleanedPeakList, mzabs, mzppm, spectrum.getMode(), false);
Vector<PeakMolPair> hits = afp.getHits();
Vector<PeakMolPair> hitsAll = afp.getAllHits();
// add them to the list of already found fragments....they were found in the previous run
// with a smaller collission energy
for (PeakMolPair peakMolPair : hitsAll) {
foundHits.add(peakMolPair.getFragment());
}
combinedHits += ((combinedHits + hits.size()) - combinedHits);
// Render.Draw(molecule, foundHits, "Found Hits");
// now "real" scoring
Scoring score = new Scoring(cleanedPeakList);
double currentScore = score.computeScoring(afp.getHitsMZ());
combinedScore += currentScore;
// check for last run and create the data for the log file
if (count == spectra.size()) {
// save score in hashmap...if there are several hits with the same score --> vector of
// strings
if (realScoreMap.containsKey(combinedScore)) {
Vector<String> tempList = realScoreMap.get(combinedScore);
tempList.add(candidates.get(c).substring(4));
realScoreMap.put(combinedScore, tempList);
} else {
Vector<String> temp = new Vector<String>();
temp.add(candidates.get(c).substring(4));
realScoreMap.put(combinedScore, temp);
}
// save score in hashmap...if there are several hits with the same
// amount of identified peaks --> ArrayList
if (scoreMap.containsKey(combinedHits)) {
ArrayList<String> tempList = scoreMap.get(combinedHits);
tempList.add(candidates.get(c).substring(4));
scoreMap.put(combinedHits, tempList);
} else {
ArrayList<String> temp = new ArrayList<String>();
temp.add(candidates.get(c).substring(4));
scoreMap.put(combinedHits, temp);
}
}
// get all the identified peaks
for (int i = 0; i < hits.size(); i++) {
peaks += hits.get(i).getPeak().getMass() + " ";
listOfPeaks.add(hits.get(i).getPeak());
if (keggIdentifier.equals(candidates.get(c).substring(4)))
listOfPeaksCorresponding.add(hits.get(i).getPeak());
}
List<IAtomContainer> hitsListTest = new ArrayList<IAtomContainer>();
for (int i = 0; i < hits.size(); i++) {
List<IAtomContainer> hitsList = new ArrayList<IAtomContainer>();
hitsList.add(AtomContainerManipulator.removeHydrogens(hits.get(i).getFragment()));
hitsListTest.add(hits.get(i).getFragment());
// Render.Highlight(AtomContainerManipulator.removeHydrogens(molecule), hitsList ,
// Double.toString(hits.get(i).getPeak()));
}
if (showDiagrams)
Render.Draw(molecule, hitsListTest, "Fragmente von: " + candidates.get(c));
} catch (CDKException e) {
System.out.println("CDK error!" + e.getMessage());
completeLog += "CDK Error! " + e.getMessage() + "File: " + candidates.get(c).substring(4);
} catch (FileNotFoundException e) {
System.out.println("File not found" + e.getMessage());
completeLog +=
"File not found error! " + e.getMessage() + "File: " + candidates.get(c).substring(4);
} catch (IOException e) {
System.out.println("IO error: " + e.getMessage());
completeLog += "IO Error! " + e.getMessage() + "File: " + candidates.get(c).substring(4);
} catch (Exception e) {
System.out.println("Error" + e.getMessage());
completeLog += "Error! " + e.getMessage() + "File: " + candidates.get(c).substring(4);
} catch (OutOfMemoryError e) {
System.out.println("Out of memory: " + e.getMessage() + "\n" + e.getStackTrace());
System.gc();
completeLog +=
"Out of memory! " + e.getMessage() + "File: " + candidates.get(c).substring(4);
}
}
// write things to log file
foundPeaks += combinedHits;
allPeaks += combinedPeakCount;
completeLog +=
"\nFile: "
+ candidates.get(c).substring(4)
+ "\t #Peaks: "
+ combinedPeakCount
+ "\t #Found: "
+ combinedHits;
completeLog += "\tPeaks: " + peaks;
}
// easy scoring
Integer[] keylist = new Integer[scoreMap.keySet().size()];
Object[] keys = scoreMap.keySet().toArray();
for (int i = 0; i < keys.length; i++) {
keylist[i] = Integer.parseInt(keys[i].toString());
}
Arrays.sort(keylist);
String scoreList = "";
int rank = 0;
for (int i = keylist.length - 1; i >= 0; i--) {
boolean check = false;
for (int j = 0; j < scoreMap.get(keylist[i]).size(); j++) {
scoreList += "\n" + keylist[i] + " - " + scoreMap.get(keylist[i]).get(j);
if (keggIdentifier.equals(scoreMap.get(keylist[i]).get(j))) {
check = true;
}
// worst case: count all which are better or have a equal position
rank++;
}
if (check) {
histogram += "\n" + file + "\t" + keggIdentifier + "\t" + rank + "\t" + exactMass;
}
}
if (keggIdentifier.equals("none")) {
histogram += "\n" + file + "\t" + keggIdentifier + "\t\t" + exactMass;
}
completeLog += "\n\n*****************Scoring*****************************";
completeLog += "Supposed to be: " + keggIdentifier;
completeLog += scoreList;
completeLog += "\n*****************************************************\n\n";
// easy scoring end
// real scoring
Double[] keysScore = new Double[realScoreMap.keySet().size()];
keysScore = realScoreMap.keySet().toArray(keysScore);
Arrays.sort(keysScore);
String scoreListReal = "";
rank = 0;
for (int i = keysScore.length - 1; i >= 0; i--) {
boolean check = false;
for (int j = 0; j < realScoreMap.get(keysScore[i]).size(); j++) {
scoreListReal += "\n" + keysScore[i] + " - " + realScoreMap.get(keysScore[i]).get(j);
if (keggIdentifier.compareTo(realScoreMap.get(keysScore[i]).get(j)) == 0) {
check = true;
}
// worst case: count all which are better or have a equal position
rank++;
}
if (check) {
histogramReal += "\n" + file + "\t" + keggIdentifier + "\t" + rank + "\t" + exactMass;
}
}
if (keggIdentifier.equals("none")) {
histogramReal += "\n" + file + "\t" + keggIdentifier + "\t\t" + exactMass;
}
// timing
long timeEnd = System.currentTimeMillis() - timeStart;
sumTime += timeEnd;
completeLog += "\n\n*****************Scoring(Real)*****************************";
completeLog += "Supposed to be: " + keggIdentifier;
completeLog += "\nTime: " + timeEnd;
completeLog += scoreListReal;
completeLog += "\n*****************************************************\n\n";
// write the data for peak histogram to log file
for (int i = 0; i < listOfPeaks.size(); i++) {
histogramPeaksAll += listOfPeaks.get(i) + "\n";
}
// filter the peaks which are contained in the all peaks list. (exclusive)
for (int i = 0; i < listOfPeaksCorresponding.size(); i++) {
for (int j = 0; j < listOfPeaks.size(); j++) {
Double valueA = listOfPeaks.get(j).getMass();
Double valueB = listOfPeaksCorresponding.get(i).getMass();
if (valueA.compareTo(valueB) == 0) {
listOfPeaks.remove(j);
}
}
}
for (int i = 0; i < listOfPeaks.size(); i++) {
histogramPeaks += listOfPeaks.get(i) + " ";
}
for (int i = 0; i < listOfPeaksCorresponding.size(); i++) {
histogramPeaksReal += listOfPeaksCorresponding.get(i) + " ";
}
}
