From 77d10953cf7a97065f7334d0ea14f2431b5b9b6f Mon Sep 17 00:00:00 2001
From: Olivier Langella <Olivier.Langella@moulon.inra.fr>
Date: Wed, 19 Apr 2017 20:55:54 +0200
Subject: [PATCH] store identification engine and version
---
.../identificationdatasource.cpp | 9 +
.../identificationdatasource.h | 13 ++
.../identificationxtandemfile.cpp | 2 +
src/input/xtandemsaxhandler.cpp | 155 +++++++++++++++++-
src/input/xtandemsaxhandler.h | 2 +
src/utils/types.h | 9 +
6 files changed, 189 insertions(+), 1 deletion(-)
diff --git a/src/core/identification_sources/identificationdatasource.cpp b/src/core/identification_sources/identificationdatasource.cpp
index a90445286..f3cc31f80 100644
--- a/src/core/identification_sources/identificationdatasource.cpp
+++ b/src/core/identification_sources/identificationdatasource.cpp
@@ -63,6 +63,15 @@ MsRunSp IdentificationDataSource::getMsRunSp () const {
return (_ms_run_sp);
}
+IdentificationEngine IdentificationDataSource::getIdentificationEngine() const {
+ return _engine;
+}
+const QString& IdentificationDataSource::getIdentificationEngineVersion() const {
+ return _version;
+}
+void IdentificationDataSource::setIdentificationEngineVersion(const QString& version) {
+ _version = version;
+}
pappso::SpectrumSp IdentificationDataSource::getSpectrumSp(unsigned int scan_number) const {
pappso::SpectrumSp spectrum_sp = SpectrumStore::getSpectrumSpFromMsRunSp(_ms_run_sp, scan_number);
diff --git a/src/core/identification_sources/identificationdatasource.h b/src/core/identification_sources/identificationdatasource.h
index 1319b0b1e..02d1032a5 100644
--- a/src/core/identification_sources/identificationdatasource.h
+++ b/src/core/identification_sources/identificationdatasource.h
@@ -54,11 +54,24 @@ public:
/** \brief read source content to store it in project
*/
virtual void parseTo(Project* p_project)=0;
+
+ /** \brief identification engine
+ */
+ virtual IdentificationEngine getIdentificationEngine() const;
+
+ /** \brief identification engine version
+ */
+ virtual const QString& getIdentificationEngineVersion() const;
+ /** \brief set identification engine version
+ */
+ virtual void setIdentificationEngineVersion(const QString& version);
protected :
QString _resource_name;
+ IdentificationEngine _engine = IdentificationEngine::unknown;
private :
//static std::map<QString, pappso::MsRunIdSp> _map_msrunidsp;
+ QString _version;
MsRunSp _ms_run_sp = nullptr;
};
diff --git a/src/core/identification_sources/identificationxtandemfile.cpp b/src/core/identification_sources/identificationxtandemfile.cpp
index 2fb6f1506..dd394f345 100644
--- a/src/core/identification_sources/identificationxtandemfile.cpp
+++ b/src/core/identification_sources/identificationxtandemfile.cpp
@@ -27,10 +27,12 @@
IdentificationXtandemFile::IdentificationXtandemFile(const QFileInfo & xtandem_file) : IdentificationDataSource(xtandem_file.absoluteFilePath()), _xtandem_file(xtandem_file)
{
+ _engine = IdentificationEngine::XTandem;
}
IdentificationXtandemFile::IdentificationXtandemFile(const IdentificationXtandemFile& other) : IdentificationDataSource(other),_xtandem_file (other._xtandem_file)
{
+ _engine = IdentificationEngine::XTandem;
}
IdentificationXtandemFile::~IdentificationXtandemFile()
diff --git a/src/input/xtandemsaxhandler.cpp b/src/input/xtandemsaxhandler.cpp
index c616abcf6..b32ac882d 100644
--- a/src/input/xtandemsaxhandler.cpp
+++ b/src/input/xtandemsaxhandler.cpp
@@ -134,20 +134,24 @@ bool XtandemSaxHandler::startElement_group(QXmlAttributes attrs) {
_charge = attrs.value("z").toUInt();
_retention_time = attrs.value("rt").replace("PT","").replace("S","").toDouble();
}
+ //label="input parameters" type="parameters"
return is_ok;
}
+
bool XtandemSaxHandler::startElement_note(QXmlAttributes attributes) {
//<note label="description">GRMZM2G083841_P01 P04711 Phosphoenolpyruvate carboxylase 1 (PEPCase 1)(PEPC 1)(EC //4.1.1.31) seq=translation; coord=9:61296279..61301686:1; parent_transcript=GRMZM2G083841_T01;
////parent_gene=GRMZM2G083841</note>
bool is_ok = true;
+ _current_note_label = attributes.value("label");
+ _current_note_type = attributes.value("type");
_is_protein_description = false;
if (attributes.value("label") == "description") {
if (_tag_stack[_tag_stack.size() - 2] == "protein") {
_is_protein_description = true;
}
}
- return is_ok;
+ return is_ok;
}
bool XtandemSaxHandler::startElement_protein(QXmlAttributes attributes) {
@@ -327,6 +331,155 @@ bool XtandemSaxHandler::endElement_note() {
if (_is_protein_description) {
_p_protein_match->getProteinXtpSp().get()->setDescription(_current_text.section(" ",1));
}
+ else {
+
+//<group label="input parameters" type="parameters">
+ /*
+ * <note type="input" label="list path, default parameters">/gorgone/pappso/tmp/temp_condor_job8533994640337729751189420695540169/QExactive_analysis_FDR_nosemi.xml</note>
+ <note type="input" label="list path, taxonomy information">/gorgone/pappso/tmp/temp_condor_job8533994640337729751189420695540169/database.xml</note>
+ <note type="input" label="output, histogram column width">30</note>
+ <note type="input" label="output, histograms">yes</note>
+ <note type="input" label="output, maximum valid expectation value">0.05</note>
+ <note type="input" label="output, maximum valid protein expectation value">0.05</note>
+ <note type="input" label="output, one sequence copy">yes</note>
+ <note type="input" label="output, parameters">yes</note>
+ <note type="input" label="output, path">/gorgone/pappso/formation/TD/xml_tandem/20120906_balliau_extract_1_A02_urzb-1.xml</note>
+ <note type="input" label="output, path hashing">no</note>
+ <note type="input" label="output, performance">yes</note>
+ <note type="input" label="output, proteins">yes</note>
+ <note type="input" label="output, results">valid</note>
+ <note type="input" label="output, sequences">yes</note>
+ <note type="input" label="output, sort results by">spectrum</note>
+ <note type="input" label="output, spectra">yes</note>
+ <note type="input" label="output, xsl path">tandem-style.xsl</note>
+ <note type="input" label="protein, C-terminal residue modification mass">0.0</note>
+ <note type="input" label="protein, N-terminal residue modification mass">0.0</note>
+ <note type="input" label="protein, cleavage C-terminal mass change">+17.00305</note>
+ <note type="input" label="protein, cleavage N-terminal mass change">+1.00794</note>
+ <note type="input" label="protein, cleavage semi">no</note>
+ <note type="input" label="protein, cleavage site">[RK]|{P}</note>
+ <note type="input" label="protein, modified residue mass file"></note>
+ <note type="input" label="protein, quick acetyl">yes</note>
+ <note type="input" label="protein, quick pyrolidone">yes</note>
+ <note type="input" label="protein, stP bias">yes</note>
+ <note type="input" label="protein, taxon">usedefined</note>
+ <note type="input" label="refine">yes</note>
+ <note type="input" label="refine, cleavage semi">no</note>
+ <note type="input" label="refine, maximum valid expectation value">0.01</note>
+ <note type="input" label="refine, modification mass">57.02146@C</note>
+ <note type="input" label="refine, modification mass 1"></note>
+ <note type="input" label="refine, point mutations">no</note>
+ <note type="input" label="refine, potential C-terminus modifications"></note>
+ <note type="input" label="refine, potential N-terminus modifications">+42.01056@[</note>
+ <note type="input" label="refine, potential modification mass">15.99491@M</note>
+ <note type="input" label="refine, potential modification mass 1"></note>
+ <note type="input" label="refine, potential modification motif"></note>
+ <note type="input" label="refine, potential modification motif 1"></note>
+ <note type="input" label="refine, spectrum synthesis">yes</note>
+ <note type="input" label="refine, unanticipated cleavage">no</note>
+ <note type="input" label="refine, use potential modifications for full refinement">yes</note>
+ <note type="input" label="residue, modification mass">57.02146@C</note>
+ <note type="input" label="residue, modification mass 1"></note>
+ <note type="input" label="residue, potential modification mass">15.99491@M</note>
+ <note type="input" label="residue, potential modification motif"></note>
+ <note type="input" label="scoring, a ions">no</note>
+ <note type="input" label="scoring, b ions">yes</note>
+ <note type="input" label="scoring, c ions">no</note>
+ <note type="input" label="scoring, cyclic permutation">yes</note>
+ <note type="input" label="scoring, include reverse">yes</note>
+ <note type="input" label="scoring, maximum missed cleavage sites">1</note>
+ <note type="input" label="scoring, minimum ion count">4</note>
+ <note type="input" label="scoring, x ions">no</note>
+ <note type="input" label="scoring, y ions">yes</note>
+ <note type="input" label="scoring, z ions">no</note>
+ <note type="input" label="spectrum, dynamic range">100.0</note>
+ <note type="input" label="spectrum, fragment mass type">monoisotopic</note>
+ <note type="input" label="spectrum, fragment monoisotopic mass error">0.02</note>
+ <note type="input" label="spectrum, fragment monoisotopic mass error units">Daltons</note>
+ <note type="input" label="spectrum, maximum parent charge">4</note>
+ <note type="input" label="spectrum, minimum fragment mz">150.0</note>
+ <note type="input" label="spectrum, minimum parent m+h">500.0</note>
+ <note type="input" label="spectrum, minimum peaks">15</note>
+ <note type="input" label="spectrum, neutral loss mass">18.01057</note>
+ <note type="input" label="spectrum, neutral loss window">0.02</note>
+ <note type="input" label="spectrum, parent monoisotopic mass error minus">10</note>
+ <note type="input" label="spectrum, parent monoisotopic mass error plus">10</note>
+ <note type="input" label="spectrum, parent monoisotopic mass error units">ppm</note>
+ <note type="input" label="spectrum, parent monoisotopic mass isotope error">yes</note>
+ */
+ //<note type="input" label="spectrum, path">/gorgone/pappso/formation/TD/mzXML/20120906_balliau_extract_1_A02_urzb-1.mzXML</note>
+
+ if (_current_note_label == "spectrum, path") {
+ _sp_msrun.get()->setFilename(_current_text);
+ }
+
+ /*
+ <note type="input" label="spectrum, sequence batch size">1000</note>
+ <note type="input" label="spectrum, threads">1</note>
+ <note type="input" label="spectrum, total peaks">100</note>
+ <note type="input" label="spectrum, use contrast angle">no</note>
+ <note type="input" label="spectrum, use neutral loss window">yes</note>
+ <note type="input" label="spectrum, use noise suppression">yes</note>
+ </group>
+
+ */
+
+//<group label="unused input parameters" type="parameters">
+
+/*
+ <note type="input" label="protein, use minimal annotations">yes</note>
+ <note type="input" label="refine, modification mass 2"></note>
+ <note type="input" label="refine, potential modification mass 2"></note>
+ <note type="input" label="refine, potential modification motif 2"></note>
+ <note type="input" label="residue, modification mass 2"></note>
+ <note type="input" label="residue, potential modification mass 1"></note>
+ <note type="input" label="residue, potential modification mass 2"></note>
+ <note type="input" label="residue, potential modification motif 1"></note>
+ <note type="input" label="residue, potential modification motif 2"></note>
+ <note type="input" label="scoring, pluggable scoring">no</note>
+</group>
+*/
+
+//<group label="performance parameters" type="parameters">
+/*
+ <note label="list path, sequence source #1">/gorgone/pappso/formation/TD/Database/Genome_Z_mays_5a.fasta</note>
+ <note label="list path, sequence source #2">/gorgone/pappso/formation/TD/Database/contaminants_standarts.fasta</note>
+ <note label="list path, sequence source description #1">no description</note>
+ <note label="list path, sequence source description #2">no description</note>
+ <note label="modelling, duplicate peptide ids">6019</note>
+ <note label="modelling, duplicate proteins">19735</note>
+ <note label="modelling, estimated false positives">18</note>
+ <note label="modelling, reversed sequence false positives">20</note>
+ <note label="modelling, spectrum noise suppression ratio">0.00</note>
+ <note label="modelling, total peptides used">96618641</note>
+ <note label="modelling, total proteins used">273656</note>
+ <note label="modelling, total spectra assigned">7464</note>
+ <note label="modelling, total spectra used">12199</note>
+ <note label="modelling, total unique assigned">6260</note>
+ <note label="process, start time">2013:12:20:16:47:19</note>
+ */
+ //<note label="process, version">X! Tandem Sledgehammer (2013.09.01.1)</note>
+ if (_current_note_label == "process, version") {
+ _p_identification_data_source->setIdentificationEngineVersion(_current_text);
+ }
+ /*
+ <note label="quality values">243 476 437 382 384 417 399 416 346 387 390 382 321 355 311 283 253 272 251 228</note>
+ <note label="refining, # input models">4893</note>
+ <note label="refining, # input spectra">5520</note>
+ <note label="refining, # partial cleavage">326</note>
+ <note label="refining, # point mutations">0</note>
+ <note label="refining, # potential C-terminii">0</note>
+ <note label="refining, # potential N-terminii">392</note>
+ <note label="refining, # unanticipated cleavage">0</note>
+ <note label="timing, initial modelling total (sec)">170.96</note>
+ <note label="timing, initial modelling/spectrum (sec)">0.0140</note>
+ <note label="timing, load sequence models (sec)">0.33</note>
+ <note label="timing, refinement/spectrum (sec)">0.0141</note>
+</group>
+*/
+
+ }
+ _current_text = "";
return is_ok;
}
diff --git a/src/input/xtandemsaxhandler.h b/src/input/xtandemsaxhandler.h
index 85babb2ec..be0448349 100644
--- a/src/input/xtandemsaxhandler.h
+++ b/src/input/xtandemsaxhandler.h
@@ -96,6 +96,8 @@ private:
QString _current_group_label;
QString _current_group_type;
+ QString _current_note_label;
+ QString _current_note_type;
unsigned int _scan;
pappso::pappso_double _mhplus_obser;
unsigned int _charge;
diff --git a/src/utils/types.h b/src/utils/types.h
index 19244ecd2..6c98714a1 100644
--- a/src/utils/types.h
+++ b/src/utils/types.h
@@ -28,6 +28,15 @@
#include <cstdint>
/*********** enumerations *********************************/
+/** \def IdentificationEngine identification engine
+ *
+ */
+enum class IdentificationEngine {
+ unknown, ///< X!Tandem
+ XTandem, ///< X!Tandem
+ peptider ///< peptider
+};
+
/** \def GroupingType list of available grouping algoritms
*
*/
--
GitLab