void CompNovoIdentificationBase::getCIDSpectrumLight_(PeakSpectrum & spec, const String & sequence, DoubleReal prefix, DoubleReal suffix)
{
static DoubleReal h2o_mass = EmpiricalFormula("H2O").getMonoWeight();
Peak1D p;
DoubleReal b_pos(0.0 + prefix);
DoubleReal y_pos(h2o_mass + suffix);
for (Size i = 0; i != sequence.size() - 1; ++i)
{
char aa(sequence[i]);
b_pos += aa_to_weight_[aa];
char aa2(sequence[sequence.size() - i - 1]);
y_pos += aa_to_weight_[aa2];
if (b_pos > min_mz_ && b_pos < max_mz_)
{
p.setPosition(b_pos + Constants::PROTON_MASS_U);
p.setIntensity(1.0f);
spec.push_back(p);
}
if (y_pos > min_mz_ && y_pos < max_mz_)
{
p.setPosition(y_pos + Constants::PROTON_MASS_U);
p.setIntensity(1.0f);
spec.push_back(p);
}
}
spec.sortByPosition();
return;
}
void SuffixArrayPeptideFinder::getCandidates(vector<vector<pair<SuffixArrayPeptideFinder::FASTAEntry, String> > >& candidates, const String& DTA_file)
{
DTAFile dta_file;
PeakSpectrum s;
dta_file.load(DTA_file, s);
s.sortByPosition();
PeakSpectrum::ConstIterator it(s.begin());
vector<double> spec;
for (; it != s.end(); ++it)
{
spec.push_back(it->getPosition()[0]);
}
const vector<double> specc(spec);
getCandidates(candidates, specc);
return;
}
void CompNovoIdentificationBase::windowMower_(PeakSpectrum & spec, DoubleReal windowsize, Size no_peaks)
{
PeakSpectrum copy(spec);
vector<Peak1D> to_be_deleted;
for (Size i = 0; i < spec.size(); ++i)
{
PeakSpectrum sub_spec;
bool end(false);
for (Size j = i; spec[j].getPosition()[0] - spec[i].getPosition()[0] < windowsize; )
{
sub_spec.push_back(spec[j]);
if (++j == spec.size())
{
end = true;
break;
}
}
sub_spec.sortByIntensity(true);
for (Size k = no_peaks; k < sub_spec.size(); ++k)
{
Peak1D p(sub_spec[k]);
to_be_deleted.push_back(p);
}
if (end)
{
break;
}
}
spec.clear(false);
for (PeakSpectrum::ConstIterator it = copy.begin(); it != copy.end(); ++it)
{
if (find(to_be_deleted.begin(), to_be_deleted.end(), *it) == to_be_deleted.end())
{
spec.push_back(*it);
}
}
spec.sortByPosition();
}
PeptideHit AScore::compute(const PeptideHit & hit, PeakSpectrum & real_spectrum, double fragment_mass_tolerance, bool fragment_mass_unit_ppm, Size max_peptide_len, Size max_num_perm)
{
PeptideHit phospho = hit;
//reset phospho
phospho.setScore(-1);
if (real_spectrum.empty())
{
return phospho;
}
String sequence_str = phospho.getSequence().toString();
Size number_of_phosphorylation_events = numberOfPhosphoEvents_(sequence_str);
AASequence seq_without_phospho = removePhosphositesFromSequence_(sequence_str);
if (seq_without_phospho.toUnmodifiedString().size() > max_peptide_len)
{
LOG_DEBUG << "\tcalculation aborted: peptide too long: " << seq_without_phospho.toString() << std::endl;
return phospho;
}
// determine all phospho sites
vector<Size> sites(getSites_(seq_without_phospho));
Size number_of_STY = sites.size();
if (number_of_phosphorylation_events == 0 || number_of_STY == 0 || number_of_STY == number_of_phosphorylation_events)
{
return phospho;
}
vector<vector<Size> > permutations(computePermutations_(sites, (Int)number_of_phosphorylation_events));
LOG_DEBUG << "\tnumber of permutations: " << permutations.size() << std::endl;
// TODO: using a heuristic to calculate the best phospho sites if the number of permutations are exceeding the maximum.
// A heuristic could be to calculate the best site for the first phosphorylation and based on this the best site for the second
// phosphorylation and so on until every site is determined
if (permutations.size() > max_num_perm)
{
LOG_DEBUG << "\tcalculation aborted: number of permutations exceeded" << std::endl;
return phospho;
}
vector<PeakSpectrum> th_spectra(createTheoreticalSpectra_(permutations, seq_without_phospho));
// prepare real spectrum windows
if (!real_spectrum.isSorted())
{
real_spectrum.sortByPosition();
}
vector<PeakSpectrum> windows_top10(peakPickingPerWindowsInSpectrum_(real_spectrum));
// calculate peptide score for each possible phospho site permutation
vector<vector<double> > peptide_site_scores(calculatePermutationPeptideScores_(th_spectra, windows_top10, fragment_mass_tolerance, fragment_mass_unit_ppm));
// rank peptide permutations ascending
multimap<double, Size> ranking(rankWeightedPermutationPeptideScores_(peptide_site_scores));
multimap<double, Size>::reverse_iterator rev = ranking.rbegin();
String seq1 = th_spectra[rev->second].getName();
phospho.setSequence(AASequence::fromString(seq1));
phospho.setMetaValue("search_engine_sequence", hit.getSequence().toString());
double peptide1_score = rev->first;
phospho.setMetaValue("AScore_pep_score", peptide1_score); // initialize score with highest peptide score (aka highest weighted score)
++rev;
String seq2 = th_spectra[rev->second].getName();
double peptide2_score = rev->first;
vector<ProbablePhosphoSites> phospho_sites;
determineHighestScoringPermutations_(peptide_site_scores, phospho_sites, permutations, ranking);
Int rank = 1;
double best_Ascore = std::numeric_limits<double>::max(); // the lower the better
for (vector<ProbablePhosphoSites>::iterator s_it = phospho_sites.begin(); s_it != phospho_sites.end(); ++s_it)
{
double Ascore = 0;
if (peptide1_score == peptide2_score) // set Ascore = 0 for each phosphorylation site
{
LOG_DEBUG << "\tscore of best (" << seq1 << ") and second best peptide (" << seq2 << ") are equal (" << peptide1_score << ")" << std::endl;
}
else
{
vector<PeakSpectrum> site_determining_ions;
computeSiteDeterminingIons_(th_spectra, *s_it, site_determining_ions, fragment_mass_tolerance, fragment_mass_unit_ppm);
Size N = site_determining_ions[0].size(); // all possibilities have the same number so take the first one
double p = static_cast<double>(s_it->peak_depth) / 100.0;
Size n_first = 0; // number of matching peaks for first peptide
for (Size window_idx = 0; window_idx != windows_top10.size(); ++window_idx) // for each 100 m/z window
{
n_first += numberOfMatchedIons_(site_determining_ions[0], windows_top10[window_idx], s_it->peak_depth, fragment_mass_tolerance, fragment_mass_unit_ppm);
}
double P_first = computeCumulativeScore_(N, n_first, p);
Size n_second = 0; // number of matching peaks for second peptide
for (Size window_idx = 0; window_idx < windows_top10.size(); ++window_idx) //each 100 m/z window
{
n_second += numberOfMatchedIons_(site_determining_ions[1], windows_top10[window_idx], s_it->peak_depth, fragment_mass_tolerance, fragment_mass_unit_ppm);
}
Size N2 = site_determining_ions[1].size(); // all possibilities have the same number so take the first one
double P_second = computeCumulativeScore_(N2, n_second, p);
//abs is used to avoid -0 score values
double score_first = abs(-10 * log10(P_first));
double score_second = abs(-10 * log10(P_second));
LOG_DEBUG << "\tfirst - N: " << N << ",p: " << p << ",n: " << n_first << ", score: " << score_first << std::endl;
LOG_DEBUG << "\tsecond - N: " << N2 << ",p: " << p << ",n: " << n_second << ", score: " << score_second << std::endl;
Ascore = score_first - score_second;
LOG_DEBUG << "\tAscore_" << rank << ": " << Ascore << std::endl;
}
if (Ascore < best_Ascore)
{
best_Ascore = Ascore;
}
phospho.setMetaValue("AScore_" + String(rank), Ascore);
++rank;
}
phospho.setScore(best_Ascore);
return phospho;
}
void CompNovoIdentificationBase::getCIDSpectrum_(PeakSpectrum & spec, const String & sequence, Size charge, DoubleReal prefix, DoubleReal suffix)
{
static DoubleReal h2o_mass = EmpiricalFormula("H2O").getMonoWeight();
static DoubleReal nh3_mass = EmpiricalFormula("NH3").getMonoWeight();
static DoubleReal co_mass = EmpiricalFormula("CO").getMonoWeight();
Peak1D p;
DoubleReal b_pos(0 + prefix);
DoubleReal y_pos(h2o_mass + suffix);
bool b_H2O_loss(false), b_NH3_loss(false), y_NH3_loss(false);
for (Size i = 0; i != sequence.size() - 1; ++i)
{
char aa(sequence[i]);
b_pos += aa_to_weight_[aa];
char aa2(sequence[sequence.size() - i - 1]);
y_pos += aa_to_weight_[aa2];
for (Size z = 1; z <= charge && z < 3; ++z)
{
// b-ions
if (b_pos >= min_mz_ && b_pos <= max_mz_)
{
for (Size j = 0; j != max_isotope_; ++j)
{
if (z == 1 /*|| b_pos > MIN_DOUBLE_MZ*/)
{
p.setPosition((b_pos + (DoubleReal)z * Constants::PROTON_MASS_U + (DoubleReal)j + Constants::NEUTRON_MASS_U) / (DoubleReal)z);
p.setIntensity(isotope_distributions_[(Size)b_pos][j] * 0.8 / (z * z));
spec.push_back(p);
}
}
}
// b-ion losses
if (b_pos - h2o_mass > min_mz_ && b_pos - h2o_mass < max_mz_)
{
if (b_H2O_loss || aa == 'S' || aa == 'T' || aa == 'E' || aa == 'D')
{
b_H2O_loss = true;
p.setPosition((b_pos + z * Constants::PROTON_MASS_U - h2o_mass) / z);
p.setIntensity(0.02 / (DoubleReal)(z * z));
if (z == 1 /* || b_pos > MIN_DOUBLE_MZ*/)
{
spec.push_back(p);
}
}
if (b_NH3_loss || aa == 'Q' || aa == 'N' || aa == 'R' || aa == 'K')
{
b_NH3_loss = true;
p.setPosition((b_pos + z * Constants::PROTON_MASS_U - nh3_mass) / z);
p.setIntensity(0.02 / (DoubleReal)(z * z));
if (z == 1 /* || b_pos > MIN_DOUBLE_MZ*/)
{
spec.push_back(p);
}
}
}
// a-ions only for charge 1
if (z == 1)
{
if (b_pos - co_mass > min_mz_ && b_pos - co_mass < max_mz_)
{
// a-ions
p.setPosition((b_pos + z * Constants::PROTON_MASS_U - co_mass) / (DoubleReal)z);
p.setIntensity(0.1f);
spec.push_back(p);
}
}
if (y_pos > min_mz_ && y_pos < max_mz_)
{
// y-ions
for (Size j = 0; j != max_isotope_; ++j)
{
if (z == 1 /* || y_pos > MIN_DOUBLE_MZ*/)
{
p.setPosition((y_pos + (DoubleReal)z * Constants::PROTON_MASS_U + (DoubleReal)j * Constants::NEUTRON_MASS_U) / (DoubleReal)z);
p.setIntensity(isotope_distributions_[(Size)y_pos][j] / (DoubleReal) (z * z));
spec.push_back(p);
}
}
// H2O loss
p.setPosition((y_pos + z * Constants::PROTON_MASS_U - h2o_mass) / (DoubleReal)z);
p.setIntensity(0.1 / (DoubleReal)(z * z));
if (aa2 == 'Q') // pyroglutamic acid formation
{
p.setIntensity(0.5f);
}
if (z == 1 /* || y_pos > MIN_DOUBLE_MZ*/)
{
spec.push_back(p);
}
// NH3 loss
if (y_NH3_loss || aa2 == 'Q' || aa2 == 'N' || aa2 == 'R' || aa2 == 'K')
{
y_NH3_loss = true;
p.setPosition((y_pos + z * Constants::PROTON_MASS_U - nh3_mass) / (DoubleReal)z);
p.setIntensity(0.1 / (DoubleReal)(z * z));
if (z == 1 /*|| y_pos > MIN_DOUBLE_MZ*/)
{
spec.push_back(p);
}
}
}
}
}
// if Q1 abundant loss of water -> pyroglutamic acid formation
if (sequence[0] == 'Q' && prefix == 0 && suffix == 0)
{
/*
for (PeakSpectrum::Iterator it = spec.begin(); it != spec.end(); ++it)
{
it->setIntensity(it->getIntensity() * 0.5);
}*/
/*
for (Size j = 0; j != max_isotope; ++j)
{
p.setPosition((precursor_weight + charge - 1 + j)/(DoubleReal)charge);
p.setIntensity(isotope_distributions_[(Int)p.getPosition()[0]][j] * 0.1);
spec.push_back(p);
}
*/
}
spec.sortByPosition();
return;
}
