void MascotXMLFile::initializeLookup(SpectrumMetaDataLookup& lookup, const MSExperiment<>& exp, const String& scan_regex)
{
// load spectra and extract scan numbers from the native IDs
// (expected format: "... scan=#"):
lookup.readSpectra(exp.getSpectra());
if (scan_regex.empty()) // use default formats
{
if (!lookup.empty()) // raw data given -> spectrum look-up possible
{
// possible formats and resulting scan numbers:
// - Mascot 2.3 (?):
// <pep_scan_title>scan=818</pep_scan_title> -> 818
// - ProteomeDiscoverer/Mascot 2.3 or 2.4:
// <pep_scan_title>Spectrum136 scans:712,</pep_scan_title> -> 712
// - other variants:
// <pep_scan_title>Spectrum3411 scans: 2975,</pep_scan_title> -> 2975
// <...>File773 Spectrum198145 scans: 6094</...> -> 6094
// <...>6860: Scan 10668 (rt=5380.57)</...> -> 10668
// <pep_scan_title>Scan Number: 1460</pep_scan_title> -> 1460
lookup.addReferenceFormat("[Ss]can( [Nn]umber)?s?[=:]? *(?<SCAN>\\d+)");
// - with .dta input to Mascot:
// <...>/path/to/FTAC05_13.673.673.2.dta</...> -> 673
lookup.addReferenceFormat("\\.(?<SCAN>\\d+)\\.\\d+\\.(?<CHARGE>\\d+)(\\.dta)?");
}
// title containing RT and MZ instead of scan number:
// <...>575.848571777344_5018.0811_controllerType=0 controllerNumber=1 scan=11515_EcoliMS2small</...>
lookup.addReferenceFormat("^(?<MZ>\\d+(\\.\\d+)?)_(?<RT>\\d+(\\.\\d+)?)");
}
else // use only user-defined format
{
lookup.addReferenceFormat(scan_regex);
}
}
void MascotXMLFile::load(const String& filename,
ProteinIdentification& protein_identification,
vector<PeptideIdentification>& id_data,
map<String, vector<AASequence> >& peptides,
const SpectrumMetaDataLookup& lookup)
{
//clear
protein_identification = ProteinIdentification();
id_data.clear();
Internal::MascotXMLHandler handler(protein_identification, id_data,
filename, peptides, lookup);
parse_(filename, &handler);
// since the Mascot XML can contain "peptides" without sequences,
// the identifications without any real peptide hit are removed
vector<PeptideIdentification> filtered_hits;
filtered_hits.reserve(id_data.size());
Size missing_sequence = 0; // counter
for (vector<PeptideIdentification>::iterator id_it = id_data.begin();
id_it != id_data.end(); ++id_it)
{
const vector<PeptideHit>& peptide_hits = id_it->getHits();
if (!peptide_hits.empty() &&
(peptide_hits.size() > 1 || !peptide_hits[0].getSequence().empty()))
{
filtered_hits.push_back(*id_it);
}
else if (!id_it->empty()) ++missing_sequence;
}
if (missing_sequence)
{
LOG_WARN << "Warning: Removed " << missing_sequence
<< " peptide identifications without sequence." << endl;
}
id_data.swap(filtered_hits);
// check if we have (some) RT information:
Size no_rt_count = 0;
for (vector<PeptideIdentification>::iterator id_it = id_data.begin();
id_it != id_data.end(); ++id_it)
{
if (!id_it->hasRT()) ++no_rt_count;
}
if (no_rt_count)
{
LOG_WARN << "Warning: " << no_rt_count << " (of " << id_data.size()
<< ") peptide identifications have no retention time value."
<< endl;
}
// if we have a mapping, but couldn't find any RT values, that's an error:
if (!lookup.empty() && (no_rt_count == id_data.size()))
{
throw Exception::MissingInformation(
__FILE__, __LINE__, __PRETTY_FUNCTION__,
"No retention time information for peptide identifications found");
}
// argh! Mascot 2.2 tends to repeat the first hit (yes it appears twice),
// so we delete one of them
for (vector<PeptideIdentification>::iterator it = id_data.begin();
it != id_data.end(); ++it)
{
vector<PeptideHit> peptide_hits = it->getHits();
// check if equal, except for rank
if (peptide_hits.size() > 1 &&
peptide_hits[0].getScore() == peptide_hits[1].getScore() &&
peptide_hits[0].getSequence() == peptide_hits[1].getSequence() &&
peptide_hits[0].getCharge() == peptide_hits[1].getCharge())
{
// erase first hit
peptide_hits.erase(peptide_hits.begin() + 1);
it->setHits(peptide_hits);
}
}
}
spectrum.getPrecursors().push_back(prec);
spectra.push_back(spectrum);
spectrum.setNativeID("spectrum=2");
spectrum.setRT(3.0);
spectrum.setMSLevel(2);
prec.setMZ(500.0);
prec.setCharge(3);
spectrum.getPrecursors()[0] = prec;
spectra.push_back(spectrum);
SpectrumMetaDataLookup lookup;
START_SECTION((template <typename SpectrumContainer> void readSpectra(const SpectrumContainer&, const String&, bool)))
{
lookup.readSpectra(spectra, SpectrumLookup::default_scan_regexp, true);
TEST_EQUAL(lookup.empty(), false);
}
END_SECTION
START_SECTION((void getSpectrumMetaData(Size, SpectrumMetaData&) const))
{
SpectrumMetaDataLookup::SpectrumMetaData meta;
lookup.getSpectrumMetaData(0, meta);
TEST_EQUAL(meta.rt, 1.0);
TEST_EQUAL(meta.ms_level, 1);
TEST_EQUAL(meta.native_id, "spectrum=0");
TEST_EQUAL(meta.scan_number, 0);
lookup.getSpectrumMetaData(1, meta);
TEST_EQUAL(meta.rt, 2.0);
TEST_EQUAL(meta.precursor_rt, 1.0);