private PsmRetTimeDistribution getPercolatorResults() {
List<PercolatorFilteredPsmResult> filteredResults = null;
if (scoreCutoff == 0.01) {
// Look in the database first for pre-calculated results
PercolatorFilteredPsmResultDAO dao =
DAOFactory.instance().getPrecolatorFilteredPsmResultDAO();
filteredResults = dao.loadForAnalysis(analysisId);
}
if (filteredResults == null || filteredResults.size() == 0) {
// Results were not found in the database; calculate now
PercolatorFilteredPsmDistributionCalculator calc =
new PercolatorFilteredPsmDistributionCalculator(analysisId, scoreCutoff);
calc.calculate();
filteredResults = calc.getFilteredResults();
}
if (filteredResults == null || filteredResults.size() == 0) {
log.error("No results for searchAnalysisID: " + analysisId);
return null;
}
PercolatorFilteredPsmResultDAO statsDao =
DAOFactory.instance().getPrecolatorFilteredPsmResultDAO();
double populationMax = RoundingUtils.getInstance().roundOne(statsDao.getPopulationMax());
double populationMin = RoundingUtils.getInstance().roundOne(statsDao.getPopulationMin());
double populationMean =
RoundingUtils.getInstance().roundOne(statsDao.getPopulationAvgFilteredPercent());
double populationStddev =
RoundingUtils.getInstance().roundOne(statsDao.getPopulationStdDevFilteredPercent());
int[] allPsmCounts = null;
int[] filteredPsmCounts = null;
int maxPsmCount = 0;
double maxRt = 0;
List<FileStats> fileStats = new ArrayList<FileStats>(filteredResults.size());
MsRunSearchAnalysisDAO rsaDao = DAOFactory.instance().getMsRunSearchAnalysisDAO();
maxRt = getMaxRtForPercolatorResults(filteredResults.get(0).getBinnedResults());
int binIncr = getBinIncrement(maxRt);
int numBins = (int) Math.ceil(maxRt / binIncr);
for (PercolatorFilteredPsmResult res : filteredResults) {
int runSearchAnalysisId = res.getRunSearchAnalysisId();
String filename = rsaDao.loadFilenameForRunSearchAnalysis(runSearchAnalysisId);
FileStats stats = new FileStats(res.getRunSearchAnalysisId(), filename);
stats.setGoodCount(res.getFiltered());
stats.setTotalCount(res.getTotal());
if (scoreCutoff == 0.01) {
stats.setPopulationMean(populationMean);
stats.setPopulationStandardDeviation(populationStddev);
stats.setPopulationMin(populationMin);
stats.setPopulationMax(populationMax);
}
fileStats.add(stats);
List<PercolatorBinnedPsmResult> binnedResults = res.getBinnedResults();
Collections.sort(
binnedResults,
new Comparator<PercolatorBinnedPsmResult>() {
@Override
public int compare(PercolatorBinnedPsmResult o1, PercolatorBinnedPsmResult o2) {
return Double.valueOf(o1.getBinStart()).compareTo(o2.getBinStart());
}
});
if (allPsmCounts == null) allPsmCounts = new int[numBins];
if (filteredPsmCounts == null) filteredPsmCounts = new int[numBins];
maxRt = Math.max(maxRt, binnedResults.get(binnedResults.size() - 1).getBinEnd());
int idx = 0;
for (int j = 0; j < binnedResults.size(); j++) {
PercolatorBinnedPsmResult binned = binnedResults.get(j);
allPsmCounts[idx] += binned.getTotal();
filteredPsmCounts[idx] += binned.getFiltered();
maxPsmCount = Math.max(allPsmCounts[idx], maxPsmCount);
if (j > 0 && j % binIncr == 0) idx++;
}
}
PsmRetTimeDistribution distribution = new PsmRetTimeDistribution(Program.PERCOLATOR);
distribution.setScoreCutoff(this.scoreCutoff);
distribution.setNumBins(allPsmCounts.length);
distribution.setMaxRT(maxRt);
distribution.setMaxPsmCount(maxPsmCount);
distribution.setFilteredPsmCounts(filteredPsmCounts);
distribution.setAllPsmCounts(allPsmCounts);
distribution.setBinSize(binIncr);
distribution.setFileStatsList(fileStats);
return distribution;
}
private PsmRetTimeDistribution getPeptideProphetResults() {
List<ProphetFilteredPsmResult> filteredResults = null;
if (scoreCutoff == QCStatsGetter.PEPPROPHET_ERR_RATE_DEFAULT) {
// Look in the database first for pre-calculated results
ProphetFilteredPsmResultDAO dao = DAOFactory.instance().getProphetFilteredPsmResultDAO();
filteredResults = dao.loadForAnalysis(analysisId);
}
if (filteredResults == null || filteredResults.size() == 0) {
// Results were not found in the database; calculate now
PeptideProphetRocDAO rocDao = DAOFactory.instance().getPeptideProphetRocDAO();
PeptideProphetROC roc = rocDao.loadRoc(analysisId);
double probability = roc.getMinProbabilityForError(scoreCutoff);
log.info("Probability for error rate of " + scoreCutoff + " is: " + probability);
ProphetFilteredPsmDistributionCalculator calc =
new ProphetFilteredPsmDistributionCalculator(analysisId, probability);
calc.calculate();
filteredResults = calc.getFilteredResults();
}
if (filteredResults == null || filteredResults.size() == 0) {
log.error("No results for searchAnalysisID: " + analysisId);
return null;
}
ProphetFilteredPsmResultDAO statsDao = DAOFactory.instance().getProphetFilteredPsmResultDAO();
double populationMax = RoundingUtils.getInstance().roundOne(statsDao.getPopulationMax());
double populationMin = RoundingUtils.getInstance().roundOne(statsDao.getPopulationMin());
double populationMean =
RoundingUtils.getInstance().roundOne(statsDao.getPopulationAvgFilteredPercent());
double populationStddev =
RoundingUtils.getInstance().roundOne(statsDao.getPopulationStdDevFilteredPercent());
int[] allPsmCounts = null;
int[] filteredPsmCounts = null;
int maxPsmCount = 0;
double maxRt = 0;
List<FileStats> fileStats = new ArrayList<FileStats>(filteredResults.size());
MsRunSearchAnalysisDAO rsaDao = DAOFactory.instance().getMsRunSearchAnalysisDAO();
maxRt = getMaxRtForProphetResults(filteredResults.get(0).getBinnedResults());
int binIncr = getBinIncrement(maxRt);
int numBins = (int) Math.ceil(maxRt / binIncr);
double actualScoreCutoff =
-1.0; // actual probability cutoff used (that corresponds to the given error rate)
for (ProphetFilteredPsmResult res : filteredResults) {
actualScoreCutoff = res.getProbability();
int runSearchAnalysisId = res.getRunSearchAnalysisId();
String filename = rsaDao.loadFilenameForRunSearchAnalysis(runSearchAnalysisId);
FileStats stats = new FileStats(res.getRunSearchAnalysisId(), filename);
stats.setGoodCount(res.getFiltered());
stats.setTotalCount(res.getTotal());
if (scoreCutoff == 0.01) {
stats.setPopulationMean(populationMean);
stats.setPopulationStandardDeviation(populationStddev);
stats.setPopulationMin(populationMin);
stats.setPopulationMax(populationMax);
}
fileStats.add(stats);
List<ProphetBinnedPsmResult> binnedResults = res.getBinnedResults();
Collections.sort(
binnedResults,
new Comparator<ProphetBinnedPsmResult>() {
@Override
public int compare(ProphetBinnedPsmResult o1, ProphetBinnedPsmResult o2) {
return Double.valueOf(o1.getBinStart()).compareTo(o2.getBinStart());
}
});
if (allPsmCounts == null) allPsmCounts = new int[numBins];
if (filteredPsmCounts == null) filteredPsmCounts = new int[numBins];
maxRt = Math.max(maxRt, binnedResults.get(binnedResults.size() - 1).getBinEnd());
int idx = 0;
for (int j = 0; j < binnedResults.size(); j++) {
ProphetBinnedPsmResult binned = binnedResults.get(j);
allPsmCounts[idx] += binned.getTotal();
filteredPsmCounts[idx] += binned.getFiltered();
maxPsmCount = Math.max(allPsmCounts[idx], maxPsmCount);
if (j > 0 && j % binIncr == 0) idx++;
}
}
PsmRetTimeDistribution distribution = new PsmRetTimeDistribution(Program.PEPTIDE_PROPHET);
distribution.setScoreCutoff(actualScoreCutoff);
distribution.setNumBins(allPsmCounts.length);
distribution.setMaxRT(maxRt);
distribution.setMaxPsmCount(maxPsmCount);
distribution.setFilteredPsmCounts(filteredPsmCounts);
distribution.setAllPsmCounts(allPsmCounts);
distribution.setBinSize(binIncr);
distribution.setFileStatsList(fileStats);
return distribution;
}