diff --git a/modules/io/doc/mmcif.rst b/modules/io/doc/mmcif.rst
index da194092927732e1430ca7a63db1b50e45330ac2..7fba2f5c6df41b711f56bc8e8426c99b174ce5a2 100644
--- a/modules/io/doc/mmcif.rst
+++ b/modules/io/doc/mmcif.rst
@@ -44,14 +44,18 @@ The following categories of a mmCIF file are considered by the reader:
Notes:
-* structures in mmCIF format can have two chain names. The "new" chain name
+* Structures in mmCIF format can have two chain names. The "new" chain name
extracted from ``atom_site.label_asym_id`` is used to name the chains in the
:class:`~ost.mol.EntityHandle`. The "old" (author provided) chain name is
extracted from ``atom_site.auth_asym_id`` for the first atom of the chain.
It is added as string property named "pdb_auth_chain_name" to the
- :class:`~ost.mol.ChainHandle` and mapped into :class:`MMCifInfo` as
- :meth:`~MMCifInfo.GetMMCifPDBChainTr` & :meth:`~MMCifInfo.GetPDBMMCifChainTr`.
-
+ :class:`~ost.mol.ChainHandle`. The mapping is also stored in
+ :class:`MMCifInfo` as :meth:`~MMCifInfo.GetMMCifPDBChainTr` and
+ :meth:`~MMCifInfo.GetPDBMMCifChainTr` if SEQRES records are read in
+ :func:`~ost.io.LoadMMCIF` and a non-empty SEQRES record exists for that chain
+ (this should exclude ligands and water).
+* Molecular entities in mmCIF are identified by an ``entity.id``. Each chain is
+ mapped to an ID in :class:`MMCifInfo` as :meth:`~MMCifInfo.GetMMCifEntityIdTr`.
Info Classes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -204,15 +208,15 @@ of the annotation available.
:param cif_chain_id: atom_site.label_asym_id
:type cif_chain_id: :class:`str`
- :returns: atom_site.auth_asym_id as :class:`str`
+ :returns: atom_site.auth_asym_id as :class:`str` (empty if no mapping)
.. method:: AddPDBMMCifChainTr(pdb_chain_id, cif_chain_id)
Set up a translation for a certain PDB chain name to the mmCIF chain name.
- :param pdb_chain_id: atom_site.label_asym_id
+ :param pdb_chain_id: atom_site.auth_asym_id
:type pdb_chain_id: :class:`str`
- :param cif_chain_id: atom_site.auth_asym_id
+ :param cif_chain_id: atom_site.label_asym_id
:type cif_chain_id: :class:`str`
.. method:: GetPDBMMCifChainTr(pdb_chain_id)
@@ -221,7 +225,24 @@ of the annotation available.
:param pdb_chain_id: atom_site.auth_asym_id
:type pdb_chain_id: :class:`str`
- :returns: atom_site.label_asym_id as :class:`str`
+ :returns: atom_site.label_asym_id as :class:`str` (empty if no mapping)
+
+ .. method:: AddMMCifEntityIdTr(cif_chain_id, entity_id)
+
+ Set up a translation for a certain mmCIF chain name to the mmCIF entity ID.
+
+ :param cif_chain_id: atom_site.label_asym_id
+ :type cif_chain_id: :class:`str`
+ :param entity_id: atom_site.label_entity_id
+ :type entity_id: :class:`str`
+
+ .. method:: GetMMCifEntityIdTr(cif_chain_id)
+
+ Get the translation of a certain mmCIF chain name to the mmCIF entity ID.
+
+ :param cif_chain_id: atom_site.label_asym_id
+ :type cif_chain_id: :class:`str`
+ :returns: atom_site.label_entity_id as :class:`str` (empty if no mapping)
.. method:: AddRevision(num, date, status)
diff --git a/modules/io/pymod/export_mmcif_io.cc b/modules/io/pymod/export_mmcif_io.cc
index 7c6a5b69750a362f01842ad13c52b555353b017a..447761e76bebe047f6fb1fdc674ca8f2cd6d38fa 100644
--- a/modules/io/pymod/export_mmcif_io.cc
+++ b/modules/io/pymod/export_mmcif_io.cc
@@ -322,6 +322,8 @@ void export_mmcif_io()
.def("GetMMCifPDBChainTr", &MMCifInfo::GetMMCifPDBChainTr)
.def("AddPDBMMCifChainTr", &MMCifInfo::AddPDBMMCifChainTr)
.def("GetPDBMMCifChainTr", &MMCifInfo::GetPDBMMCifChainTr)
+ .def("AddMMCifEntityIdTr", &MMCifInfo::AddMMCifEntityIdTr)
+ .def("GetMMCifEntityIdTr", &MMCifInfo::GetMMCifEntityIdTr)
.def("SetRevisionsDateOriginal", &MMCifInfo::SetRevisionsDateOriginal)
.def("AddRevision", &MMCifInfo::AddRevision)
.def("GetRevisions", &MMCifInfo::GetRevisions)
diff --git a/modules/io/src/mol/mmcif_info.cc b/modules/io/src/mol/mmcif_info.cc
index 27934ad09079feee4a2b7e1e031ad6182dc472d9..3e0fa35370df8d80480920e3f14f5d0f95ddd7ff 100644
--- a/modules/io/src/mol/mmcif_info.cc
+++ b/modules/io/src/mol/mmcif_info.cc
@@ -58,6 +58,24 @@ String MMCifInfo::GetPDBMMCifChainTr(String pdb) const
return tr_it->second;
}
+void MMCifInfo::AddMMCifEntityIdTr(String cif, String ent_id)
+{
+ std::map<String, String>::iterator tr_it = cif_2_entity_id_.find(cif);
+ if (tr_it != cif_2_entity_id_.end()) {
+ throw IOException("mmCIF chain id '" + cif + "' is already mapped to "
+ "entity id '" + tr_it->second + "'.");
+ }
+ cif_2_entity_id_.insert(std::pair<String, String>(cif, ent_id));
+}
+
+String MMCifInfo::GetMMCifEntityIdTr(String cif) const
+{
+ std::map<String, String>::const_iterator tr_it =
+ cif_2_entity_id_.find(cif);
+ if (tr_it == cif_2_entity_id_.end()) { return ""; }
+ return tr_it->second;
+}
+
void MMCifInfo::AddAuthorsToCitation(StringRef id, std::vector<String> list)
{
// find citation
diff --git a/modules/io/src/mol/mmcif_info.hh b/modules/io/src/mol/mmcif_info.hh
index 8087e7e0718972ddc1343899af4d69696fda1f3e..8054c160f5386130bc94d181b7a04c81258c6b22 100644
--- a/modules/io/src/mol/mmcif_info.hh
+++ b/modules/io/src/mol/mmcif_info.hh
@@ -940,6 +940,18 @@ public:
/// \return chain name as used in the PDB file (label_asym_id)
String GetPDBMMCifChainTr(String pdb) const;
+ /// \brief Add a new mmCIF chain name / entity ID tuple.
+ ///
+ /// \param cif chain name as used by the mmCIF file (label_asym_id)
+ /// \param ent_id entity ID as used by the mmCIF file (label_entity_id)
+ void AddMMCifEntityIdTr(String cif, String ent_id);
+
+ /// \brief Get the entity ID for a CIF chain name
+ ///
+ /// \param cif chain name as used by the mmCIF file (label_asym_id)
+ /// \return entity ID as used by the mmCIF file (label_entity_id)
+ String GetMMCifEntityIdTr(String cif) const;
+
/// \brief Add a biounit
///
/// \param bu biounit to be added
@@ -1043,6 +1055,7 @@ private:
MMCifInfoStructRefs struct_refs_;
std::map<String, String> cif_2_pdb_chain_id_;
std::map<String, String> pdb_2_cif_chain_id_;
+ std::map<String, String> cif_2_entity_id_;
};
diff --git a/modules/io/src/mol/mmcif_reader.cc b/modules/io/src/mol/mmcif_reader.cc
index 9e2d95a21750a18f1873937c4a2807e75f8d702e..bbee2014c766b2c8b6589be70a035855c959e7ae 100644
--- a/modules/io/src/mol/mmcif_reader.cc
+++ b/modules/io/src/mol/mmcif_reader.cc
@@ -506,8 +506,10 @@ void MMCifReader::ParseAndAddAtom(const std::vector<StringRef>& columns)
curr_chain_.SetStringProp("pdb_auth_chain_name", auth_chain_name);
++chain_count_;
// store entity id
+ String ent_id = columns[indices_[LABEL_ENTITY_ID]].str();
chain_id_pairs_.push_back(std::pair<mol::ChainHandle,String>(curr_chain_,
- columns[indices_[LABEL_ENTITY_ID]].str()));
+ ent_id));
+ info_.AddMMCifEntityIdTr(cif_chain_name, ent_id);
}
assert(curr_chain_.IsValid());
} else if (chain_id_pairs_.back().second != // unit test
diff --git a/modules/io/tests/test_mmcif_info.cc b/modules/io/tests/test_mmcif_info.cc
index d37b8f396198045128e8e5acce8598cafed9adb4..6c49768d2eb0806473cefe958c7b5d9f9482e79b 100644
--- a/modules/io/tests/test_mmcif_info.cc
+++ b/modules/io/tests/test_mmcif_info.cc
@@ -269,6 +269,11 @@ BOOST_AUTO_TEST_CASE(mmcif_info)
BOOST_CHECK("B" == info.GetPDBMMCifChainTr("A"));
BOOST_CHECK("" == info.GetPDBMMCifChainTr("C"));
+ info.AddMMCifEntityIdTr("A", "1");
+ BOOST_CHECK_THROW(info.AddMMCifEntityIdTr("A", "1"), IOException);
+ BOOST_CHECK("1" == info.GetMMCifEntityIdTr("A"));
+ BOOST_CHECK("" == info.GetMMCifEntityIdTr("C"));
+
BOOST_CHECK(info.GetRevisions().GetSize() == 0);
BOOST_TEST_MESSAGE(" done.");
diff --git a/modules/io/tests/test_mmcif_reader.cc b/modules/io/tests/test_mmcif_reader.cc
index 77b2bcd14b9ae79c13a0a56def3af3aecfac7701..c5788c9238b030aa538e2629ce10504da94bb06a 100644
--- a/modules/io/tests/test_mmcif_reader.cc
+++ b/modules/io/tests/test_mmcif_reader.cc
@@ -97,6 +97,20 @@ void SetAtomSiteHeader(StarLoopDesc* mmcif_h)
mmcif_h->Add(StringRef("Cartn_z", 7));
}
+conop::CompoundLibPtr SetDefaultCompoundLib() {
+ // return NULL if not successful, else return newly set default lib
+ // REQ: OST_ROOT to be set
+ char * ost_root = getenv("OST_ROOT");
+ if (!ost_root) return conop::CompoundLibPtr();
+ SetPrefixPath(ost_root);
+ String lib_path = GetSharedDataPath() + "/compounds.chemlib";
+ conop::CompoundLibPtr compound_lib = conop::CompoundLib::Load(lib_path);
+ if (compound_lib) {
+ conop::Conopology::Instance().SetDefaultLib(compound_lib);
+ }
+ return compound_lib;
+}
+
BOOST_AUTO_TEST_SUITE( io );
BOOST_AUTO_TEST_CASE(mmcif_isvalidpdbident)
@@ -140,21 +154,13 @@ BOOST_AUTO_TEST_CASE(mmcif_trystoreidx)
BOOST_AUTO_TEST_CASE(mmcif_convert_seqres)
{
- char * ost_root=getenv("OST_ROOT");
- if(!ost_root){
- std::cout << "WARNING: skipping SEQRES import unit test. "
- << "Rule-based processor is required" << std::endl;
- return;
- }
- SetPrefixPath(ost_root);
- String lib_path=GetSharedDataPath()+"/compounds.chemlib";
- conop::CompoundLibPtr compound_lib=conop::CompoundLib::Load(lib_path);
+ conop::CompoundLibPtr compound_lib = SetDefaultCompoundLib();
if (!compound_lib) {
- std::cout << "WARNING: skipping SEQRES import unit test. Compound "
- << "library is required" << std::endl;
- return;
+ std::cout << "WARNING: skipping mmcif_convert_seqres unit test. "
+ << "Compound library is required" << std::endl;
+ return;
}
- conop::Conopology::Instance().SetDefaultLib(compound_lib);
+
mol::EntityHandle eh=mol::CreateEntity();
TestMMCifReaderProtected tmmcif_p("testfiles/mmcif/atom_site.mmcif", eh);
@@ -402,21 +408,12 @@ BOOST_AUTO_TEST_CASE(mmcif_entity_tests)
BOOST_AUTO_TEST_CASE(mmcif_entity_poly_tests)
{
- char * ost_root=getenv("OST_ROOT");
- if(!ost_root){
- std::cout << "WARNING: skipping SEQRES import unit test. "
- << "Rule-based processor is required" << std::endl;
+ if (!SetDefaultCompoundLib()) {
+ std::cout << "WARNING: skipping mmcif_entity_poly_tests unit test. "
+ << "Compound library is required" << std::endl;
return;
}
- SetPrefixPath(ost_root);
- String lib_path=GetSharedDataPath()+"/compounds.chemlib";
- conop::CompoundLibPtr compound_lib=conop::CompoundLib::Load(lib_path);
- if (!compound_lib) {
- std::cout << "WARNING: skipping SEQRES import unit test. Compound "
- << "lib is required" << std::endl;
- return;
- }
- conop::Conopology::Instance().SetDefaultLib(compound_lib);
+
BOOST_TEST_MESSAGE(" Running mmcif_entity_poly_tests...");
mol::ChainHandle ch;
IOProfile profile;
@@ -1301,4 +1298,58 @@ BOOST_AUTO_TEST_CASE(mmcif_testreader)
BOOST_TEST_MESSAGE(" done.");
}
+// helper for mmcif_test_chain_mappings
+inline void CheckChainMap(mol::EntityHandle eh, const MMCifInfo& info,
+ const String& cif_name, const String& pdb_name,
+ bool check_info_map) {
+ // check chain
+ mol::ChainHandle ch = eh.FindChain(cif_name);
+ BOOST_CHECK(ch.IsValid());
+ BOOST_CHECK(ch.HasProp("pdb_auth_chain_name"));
+ BOOST_CHECK(ch.GetStringProp("pdb_auth_chain_name") == pdb_name);
+ // info mapping
+ if (check_info_map) {
+ BOOST_CHECK(info.GetMMCifPDBChainTr(cif_name) == pdb_name);
+ BOOST_CHECK(info.GetPDBMMCifChainTr(pdb_name) == cif_name);
+ }
+}
+
+BOOST_AUTO_TEST_CASE(mmcif_test_chain_mappings)
+{
+ BOOST_TEST_MESSAGE(" Running mmcif_test_chain_mappings tests...");
+
+ // check compound lib
+ bool compound_lib_available = SetDefaultCompoundLib();
+
+ // load data
+ mol::EntityHandle eh = mol::CreateEntity();
+ std::ifstream s("testfiles/mmcif/atom_site.mmcif");
+ IOProfile profile;
+ MMCifReader mmcif_p(s, eh, profile);
+ if (compound_lib_available) {
+ mmcif_p.SetReadSeqRes(true);
+ }
+ BOOST_REQUIRE_NO_THROW(mmcif_p.Parse());
+ const MMCifInfo& info = mmcif_p.GetInfo();
+
+ // check 1-to-1 mappings
+ CheckChainMap(eh, info, "A", "A", compound_lib_available);
+ CheckChainMap(eh, info, "C", "C", compound_lib_available);
+ CheckChainMap(eh, info, "O", "B", false); // water
+ CheckChainMap(eh, info, "Z", "Z", compound_lib_available);
+
+ // check entity ID mapping
+ BOOST_CHECK(info.GetMMCifEntityIdTr("A") == "1");
+ BOOST_CHECK(info.GetMMCifEntityIdTr("C") == "1");
+ BOOST_CHECK(info.GetMMCifEntityIdTr("O") == "5");
+ BOOST_CHECK(info.GetMMCifEntityIdTr("Z") == "1");
+
+ // check non-existent mappings
+ BOOST_CHECK(info.GetMMCifPDBChainTr("B") == "");
+ BOOST_CHECK(info.GetPDBMMCifChainTr("O") == "");
+ BOOST_CHECK(info.GetMMCifEntityIdTr("B") == "");
+
+ BOOST_TEST_MESSAGE(" done.");
+}
+
BOOST_AUTO_TEST_SUITE_END();