From 944fd29a9a1d8d2466315a518738ec30666da35a Mon Sep 17 00:00:00 2001
From: Stefan Bienert <stefan.bienert@unibas.ch>
Date: Mon, 2 Jul 2012 11:33:45 +0200
Subject: [PATCH] Added bio unit id (pdbx_struct_assembly.id) to MMCIFBioUnit
class. (Was forgotten, before)
---
modules/io/doc/mmcif.rst | 24 +++++++++++++++++++++---
modules/io/pymod/export_mmcif_io.cc | 3 +++
modules/io/src/mol/mmcif_info.hh | 15 ++++++++++++++-
modules/io/src/mol/mmcif_reader.cc | 7 ++++---
modules/io/tests/test_mmcif_reader.cc | 2 ++
5 files changed, 44 insertions(+), 7 deletions(-)
diff --git a/modules/io/doc/mmcif.rst b/modules/io/doc/mmcif.rst
index 7e1c330a4..4b3245f2c 100644
--- a/modules/io/doc/mmcif.rst
+++ b/modules/io/doc/mmcif.rst
@@ -411,6 +411,14 @@ of the annotation available.
This stores information how a structure is to be assembled to form the
bio unit.
+ .. attribute:: id
+
+ The id of a bio unit as given by the original mmCIF file.
+
+ Also available as :meth:`GetID`. May also be modified by :meth:`SetID`.
+
+ :type: :class:`str`
+
.. attribute:: details
Special aspects of the biological assembly. If not provided, resembles an
@@ -434,6 +442,14 @@ of the annotation available.
Also available as :meth:`GetOperations`. May be modified by
:meth:`AddOperations`
+ .. method:: GetID()
+
+ See :attr:`id`
+
+ .. method:: SetID(id)
+
+ See :attr:`id`
+
.. method:: GetDetails()
See :attr:`details`
@@ -688,7 +704,7 @@ of the annotation available.
link of polymers in the mmCIF file to sequences stored in external databases
such as uniprot. The related categories ``struct_ref_seq`` and
``struct_ref_seq_dif`` also list differences between the sequences of the
- deposited structure and the sequences in the database. A promintent example of
+ deposited structure and the sequences in the database. A prominent example of
such differences include point mutations and/or expression tags.
.. attribute:: db_name
@@ -769,8 +785,10 @@ of the annotation available.
.. attribute:: details
A textual description of the difference, e.g. point mutation,
- expressiontag, purification artifact.
+ expression tag, purification artifact.
:type: :class:`str`
+
.. LocalWords: cas isbn pubmed asu seqres conop ConnectAll casp COMPND OBSLTE
-.. LocalWords: SPRSDE pdb func
+.. LocalWords: SPRSDE pdb func autofunction exptl attr pdbx oper conf spr dif
+.. LocalWords: biounits biounit uniprot UNP seqs
diff --git a/modules/io/pymod/export_mmcif_io.cc b/modules/io/pymod/export_mmcif_io.cc
index 70e6b4659..c94c72a67 100644
--- a/modules/io/pymod/export_mmcif_io.cc
+++ b/modules/io/pymod/export_mmcif_io.cc
@@ -181,12 +181,15 @@ void export_mmcif_io()
.def("AddOperations", &MMCifInfoBioUnit::AddOperations)
.def("GetOperations", make_function(&MMCifInfoBioUnit::GetOperations,
return_value_policy<copy_const_reference>()))
+ .def("SetID", &MMCifInfoBioUnit::SetID)
+ .def("GetID", &MMCifInfoBioUnit::GetID)
.add_property("details", &MMCifInfoBioUnit::GetDetails,
&MMCifInfoBioUnit::SetDetails)
.add_property("chains", make_function(&MMCifInfoBioUnit::GetChainList,
return_value_policy<copy_const_reference>()))
.add_property("operations", make_function(&MMCifInfoBioUnit::GetOperations,
return_value_policy<copy_const_reference>()))
+ .add_property("details", &MMCifInfoBioUnit::GetID, &MMCifInfoBioUnit::SetID)
;
class_<MMCifInfoStructRefs>("MMCifInfoStructRefs", init<>())
diff --git a/modules/io/src/mol/mmcif_info.hh b/modules/io/src/mol/mmcif_info.hh
index 38e49f58b..b240e5a1c 100644
--- a/modules/io/src/mol/mmcif_info.hh
+++ b/modules/io/src/mol/mmcif_info.hh
@@ -238,7 +238,16 @@ typedef boost::shared_ptr<MMCifInfoTransOp> MMCifInfoTransOpPtr;
class DLLEXPORT_OST_IO MMCifInfoBioUnit {
public:
/// \brief Create a biounit.
- MMCifInfoBioUnit(): details_("") {};
+ MMCifInfoBioUnit(): id_(""), details_("") {};
+
+ /// \brief Set id
+ ///
+ /// \param id id
+ void SetID(String id) { id_ = id; }
+ /// \brief Get id
+ ///
+ /// \return id
+ String GetID() const { return id_; }
/// \brief Set details
///
@@ -274,6 +283,9 @@ public:
}
bool operator==(const MMCifInfoBioUnit& bu) const {
+ if (this->id_ != bu.id_) {
+ return false;
+ }
if (this->details_ != bu.details_) {
return false;
}
@@ -314,6 +326,7 @@ public:
}
private:
+ String id_; ///< pdbx_struct_assembly.id
String details_; ///< pdbx_struct_assembly.details
std::vector<String> chains_; ///< chains involved in this assembly
std::vector<std::vector<MMCifInfoTransOpPtr> > operations_;
diff --git a/modules/io/src/mol/mmcif_reader.cc b/modules/io/src/mol/mmcif_reader.cc
index 409902ab4..2bcdee810 100644
--- a/modules/io/src/mol/mmcif_reader.cc
+++ b/modules/io/src/mol/mmcif_reader.cc
@@ -975,7 +975,7 @@ void MMCifReader::ParsePdbxStructAssemblyGen(const std::vector<StringRef>& colum
MMCifBioUAssembly assembly;
assembly.biounit = MMCifInfoBioUnit();
- assembly.biounit.SetDetails(columns[indices_[ASSEMBLY_ID]].str());
+ assembly.biounit.SetID(columns[indices_[ASSEMBLY_ID]].str());
std::vector<StringRef> tmp_chains=columns[indices_[ASYM_ID_LIST]].split(',');
std::vector<StringRef>::const_iterator tc_it;
@@ -1586,14 +1586,15 @@ void MMCifReader::OnEndData()
bua_it != bu_assemblies_.end();
++bua_it) {
// pair with pdbx_struct_assembly entry
- buom_it = bu_origin_map_.find(bua_it->biounit.GetDetails());
+ buom_it = bu_origin_map_.find(bua_it->biounit.GetID());
if (buom_it == bu_origin_map_.end()) {
throw IOException(this->FormatDiagnostic(STAR_DIAG_ERROR,
"No pdbx_struct_assembly.id '"+
- bua_it->biounit.GetDetails() +
+ bua_it->biounit.GetID() +
"' found as requested by pdbx_struct_assembly_gen."));
}
bua_it->biounit.SetDetails(buom_it->second);
+ bua_it->biounit.SetID(buom_it->first);
// pair with pdbx_struct_oper_list
for (aol_it = bua_it->operations.begin();
diff --git a/modules/io/tests/test_mmcif_reader.cc b/modules/io/tests/test_mmcif_reader.cc
index 953ac3132..b17b22ea8 100644
--- a/modules/io/tests/test_mmcif_reader.cc
+++ b/modules/io/tests/test_mmcif_reader.cc
@@ -1254,9 +1254,11 @@ BOOST_AUTO_TEST_CASE(mmcif_testreader)
BOOST_MESSAGE(" done.");
BOOST_MESSAGE(" reading data fields which should not fail...");
+ BOOST_MESSAGE(mmcif_p.GetInfo().GetBioUnits().back().GetID());
BOOST_CHECK(mmcif_p.GetInfo().GetMethod().str() == "Deep-fry");
BOOST_CHECK(mmcif_p.GetInfo().GetBioUnits().back().GetDetails() ==
"author_defined_assembly");
+ BOOST_CHECK(mmcif_p.GetInfo().GetBioUnits().back().GetID() == "2");
BOOST_CHECK(mmcif_p.GetInfo().GetBioUnits().back().GetChainList().back() ==
"F");
MMCifInfoBioUnit bu = mmcif_p.GetInfo().GetBioUnits().back();
--
GitLab