Function postProcessHits_

src/topp/RNPxlSearch.cpp:1275–1446

Filter by top scoring hits, reconstruct original peptide from memory efficient structure, and add additional meta information.

  /// Filter by top scoring hits, reconstruct original peptide from memory efficient structure, and add additional meta information.
  void postProcessHits_(const PeakMap& exp,
    vector<vector<AnnotatedHit> >& annotated_hits,
    vector<ProteinIdentification>& protein_ids,
    vector<PeptideIdentification>& peptide_ids,
    Size top_hits,
    const RNPxlModificationMassesResult& mm,
    const ModifiedPeptideGenerator::MapToResidueType& fixed_modifications,
    const ModifiedPeptideGenerator::MapToResidueType& variable_modifications,
    Size max_variable_mods_per_peptide,
    const map<String, PrecursorPurity::PurityScores>& purities)
  {
      // remove all but top n scoring (Note: this is currently necessary as postScoreHits_ might reintroduce nucleotide specific hits for fast scoring)
#ifdef _OPENMP
#pragma omp parallel for
#endif
      for (SignedSize scan_index = 0; scan_index < (SignedSize)annotated_hits.size(); ++scan_index)
      {
        // sort and keeps n best elements according to score
        Size topn = top_hits > annotated_hits[scan_index].size() ? annotated_hits[scan_index].size() : top_hits;
        std::partial_sort(annotated_hits[scan_index].begin(), annotated_hits[scan_index].begin() + topn, annotated_hits[scan_index].end(), AnnotatedHit::hasBetterScore);
        annotated_hits[scan_index].resize(topn);
        annotated_hits.shrink_to_fit();
      }

#ifdef _OPENMP
#pragma omp parallel for
#endif
    for (SignedSize scan_index = 0; scan_index < (SignedSize)annotated_hits.size(); ++scan_index)
    {
      if (!annotated_hits[scan_index].empty())
      {
        // create empty PeptideIdentification object and fill meta data
        PeptideIdentification pi;
        pi.setMetaValue("scan_index", static_cast<unsigned int>(scan_index));
        pi.setScoreType("hyperscore");
        pi.setHigherScoreBetter(true);
        pi.setRT(exp[scan_index].getRT());
        pi.setMZ(exp[scan_index].getPrecursors()[0].getMZ());
        pi.setMetaValue("precursor_intensity", exp[scan_index].getPrecursors()[0].getIntensity());
        Size charge = exp[scan_index].getPrecursors()[0].getCharge();

        // create full peptide hit structure from annotated hits
        vector<PeptideHit> phs;
        size_t rank(0);
        for (auto const & ah : annotated_hits[scan_index])
        {
          PeptideHit ph;
          ph.setCharge(charge);

          // get unmodified string
          const String & s = ah.sequence.getString();

          OPENMS_POSTCONDITION(!s.empty(), "Error: empty sequence in annotated hits.");
          AASequence aas = AASequence::fromString(s);

          // reapply modifications (because for memory reasons we only stored the index and recreation is fast)
          vector<AASequence> all_modified_peptides;
          ModifiedPeptideGenerator::applyFixedModifications(fixed_modifications, aas);