From 00780ed46f29c5e7be028af883b91f8c0824c780 Mon Sep 17 00:00:00 2001
From: Gerardo Tauriello <gerardo.tauriello@unibas.ch>
Date: Thu, 27 Jul 2017 18:55:34 +0200
Subject: [PATCH] SCHWED-2592: Extract rFree and rWork from mmCIF files
---
modules/io/doc/mmcif.rst | 22 ++++++++++++++--
modules/io/pymod/export_mmcif_io.cc | 6 +++++
modules/io/src/mol/mmcif_info.hh | 24 ++++++++++++++++-
modules/io/src/mol/mmcif_reader.cc | 26 ++++++++++++++-----
modules/io/src/mol/mmcif_reader.hh | 4 ++-
modules/io/tests/test_mmcif_reader.cc | 8 +++---
.../io/tests/testfiles/mmcif/atom_site.mmcif | 8 +++---
7 files changed, 80 insertions(+), 18 deletions(-)
diff --git a/modules/io/doc/mmcif.rst b/modules/io/doc/mmcif.rst
index 78e126def..bed692ff4 100644
--- a/modules/io/doc/mmcif.rst
+++ b/modules/io/doc/mmcif.rst
@@ -25,7 +25,8 @@ The following categories of a mmCIF file are considered by the reader:
* ``citation``: Goes into :class:`MMCifInfoCitation`
* ``citation_author``: Goes into :class:`MMCifInfoCitation`
* ``exptl``: Goes into :class:`MMCifInfo` as :attr:`~MMCifInfo.method`.
-* ``refine``: Goes into :class:`MMCifInfo` as :attr:`~MMCifInfo.resolution`.
+* ``refine``: Goes into :class:`MMCifInfo` as :attr:`~MMCifInfo.resolution`,
+ :attr:`~MMCifInfo.r_free` and :attr:`~MMCifInfo.r_work`.
* ``pdbx_struct_assembly``: Used for :class:`MMCifInfoBioUnit`.
* ``pdbx_struct_assembly_gen``: Used for :class:`MMCifInfoBioUnit`.
* ``pdbx_struct_oper_list``: Used for :class:`MMCifInfoBioUnit`.
@@ -93,11 +94,28 @@ of the annotation available.
.. attribute:: resolution
- Stores the resolution of the crystal structure.
+ Stores the resolution of the crystal structure. Set to 0 if no value in
+ loaded mmCIF file.
Also available as :meth:`GetResolution`. May also be modified by
:meth:`SetResolution`.
+ .. attribute:: r_free
+
+ Stores the R-free value of the crystal structure. Set to 0 if no value in
+ loaded mmCIF file.
+
+ Also available as :meth:`GetRFree`. May also be modified by
+ :meth:`SetRFree`.
+
+ .. attribute:: r_work
+
+ Stores the R-work value of the crystal structure. Set to 0 if no value in
+ loaded mmCIF file.
+
+ Also available as :meth:`GetRWork`. May also be modified by
+ :meth:`SetRWork`.
+
.. attribute:: operations
Stores the operations needed to transform a crystal structure into a
diff --git a/modules/io/pymod/export_mmcif_io.cc b/modules/io/pymod/export_mmcif_io.cc
index 447761e76..dd90843e3 100644
--- a/modules/io/pymod/export_mmcif_io.cc
+++ b/modules/io/pymod/export_mmcif_io.cc
@@ -310,6 +310,10 @@ void export_mmcif_io()
.def("GetMethod", &MMCifInfo::GetMethod)
.def("SetResolution", &MMCifInfo::SetResolution)
.def("GetResolution", &MMCifInfo::GetResolution)
+ .def("SetRFree", &MMCifInfo::SetRFree)
+ .def("GetRFree", &MMCifInfo::GetRFree)
+ .def("SetRWork", &MMCifInfo::SetRWork)
+ .def("GetRWork", &MMCifInfo::GetRWork)
.def("AddAuthorsToCitation", &MMCifInfo::AddAuthorsToCitation)
.def("AddOperation", &MMCifInfo::AddOperation)
.def("GetOperations", make_function(&MMCifInfo::GetOperations,
@@ -334,6 +338,8 @@ void export_mmcif_io()
.add_property("method", &MMCifInfo::GetMethod, &MMCifInfo::SetMethod)
.add_property("resolution", &MMCifInfo::GetResolution,
&MMCifInfo::SetResolution)
+ .add_property("r_free", &MMCifInfo::GetRFree, &MMCifInfo::SetRFree)
+ .add_property("r_work", &MMCifInfo::GetRWork, &MMCifInfo::SetRWork)
.add_property("operations", make_function(&MMCifInfo::GetOperations,
return_value_policy<copy_const_reference>()))
.add_property("struct_details", &MMCifInfo::GetStructDetails,
diff --git a/modules/io/src/mol/mmcif_info.hh b/modules/io/src/mol/mmcif_info.hh
index 49cd8e16c..0a00a8382 100644
--- a/modules/io/src/mol/mmcif_info.hh
+++ b/modules/io/src/mol/mmcif_info.hh
@@ -869,7 +869,7 @@ private:
class DLLEXPORT_OST_IO MMCifInfo {
public:
/// \brief Create an info object.
- MMCifInfo(): exptl_method_(""), resolution_(0.0f) {};
+ MMCifInfo(): exptl_method_(""), resolution_(0), r_free_(0), r_work_(0) { }
/// \brief Add an item to the list of citations
///
@@ -916,6 +916,26 @@ public:
/// \return experiment resolution
Real GetResolution() const { return resolution_; }
+ /// \brief Set R-free value.
+ ///
+ /// \param res experiment R-free value
+ void SetRFree(Real r_free) { r_free_ = r_free; }
+
+ /// \brief Get R-free value.
+ ///
+ /// \return experiment R-free value
+ Real GetRFree() const { return r_free_; }
+
+ /// \brief Set R-work value.
+ ///
+ /// \param r_work experiment R-work value
+ void SetRWork(Real r_work) { r_work_ = r_work; }
+
+ /// \brief Get R-work value.
+ ///
+ /// \return experiment R-work value
+ Real GetRWork() const { return r_work_; }
+
/// \brief Add a new mmCIF/ PDB chain name tuple.
///
/// \param cif chain name as used by the mmCIF file (label_asym_id)
@@ -1049,6 +1069,8 @@ private:
// members
String exptl_method_;
Real resolution_;
+ Real r_free_;
+ Real r_work_;
MMCifInfoStructDetails struct_details_; ///< mmCIF struct category
MMCifInfoObsolete obsolete_; ///< obsolete/ superseded entry
MMCifInfoRevisions revisions_; ///< list of revisions
diff --git a/modules/io/src/mol/mmcif_reader.cc b/modules/io/src/mol/mmcif_reader.cc
index abd02fe6e..1c971ee2f 100644
--- a/modules/io/src/mol/mmcif_reader.cc
+++ b/modules/io/src/mol/mmcif_reader.cc
@@ -219,9 +219,12 @@ bool MMCifReader::OnBeginLoop(const StarLoopDesc& header)
} else if (header.GetCategory() == "refine") {
category_ = REFINE;
// mandatory items
- this->TryStoreIdx(REFINE_ENTRY_ID, "entry_id", header);
- this->TryStoreIdx(LS_D_RES_HIGH, "ls_d_res_high", header);
- this->TryStoreIdx(LS_D_RES_LOW, "ls_d_res_low", header);
+ this->TryStoreIdx(LS_D_RES_HIGH, "ls_d_res_high", header);
+ // optional items
+ indices_[REFINE_ENTRY_ID] = header.GetIndex("entry_id");
+ indices_[LS_D_RES_LOW] = header.GetIndex("ls_d_res_low");
+ indices_[LS_R_FACTOR_R_WORK] = header.GetIndex("ls_R_factor_R_work");
+ indices_[LS_R_FACTOR_R_FREE] = header.GetIndex("ls_R_factor_R_free");
cat_available = true;
} else if (header.GetCategory() == "pdbx_struct_assembly") {
category_ = PDBX_STRUCT_ASSEMBLY;
@@ -909,10 +912,21 @@ void MMCifReader::ParseExptl(const std::vector<StringRef>& columns)
void MMCifReader::ParseRefine(const std::vector<StringRef>& columns)
{
- StringRef col=columns[indices_[LS_D_RES_HIGH]];
+ StringRef col = columns[indices_[LS_D_RES_HIGH]];
if (col.size()!=1 || (col[0]!='?' && col[0]!='.')) {
- info_.SetResolution(this->TryGetReal(columns[indices_[LS_D_RES_HIGH]],
- "refine.ls_d_res_high"));
+ info_.SetResolution(this->TryGetReal(col, "refine.ls_d_res_high"));
+ }
+ if (indices_[LS_R_FACTOR_R_WORK] != -1) {
+ col = columns[indices_[LS_R_FACTOR_R_WORK]];
+ if (col.size()!=1 || (col[0]!='?' && col[0]!='.')) {
+ info_.SetRWork(this->TryGetReal(col, "refine.ls_R_factor_R_work"));
+ }
+ }
+ if (indices_[LS_R_FACTOR_R_FREE] != -1) {
+ col = columns[indices_[LS_R_FACTOR_R_FREE]];
+ if (col.size()!=1 || (col[0]!='?' && col[0]!='.')) {
+ info_.SetRFree(this->TryGetReal(col, "refine.ls_R_factor_R_free"));
+ }
}
}
diff --git a/modules/io/src/mol/mmcif_reader.hh b/modules/io/src/mol/mmcif_reader.hh
index 55a195652..3503223eb 100644
--- a/modules/io/src/mol/mmcif_reader.hh
+++ b/modules/io/src/mol/mmcif_reader.hh
@@ -427,7 +427,9 @@ private:
typedef enum {
REFINE_ENTRY_ID, ///< id
LS_D_RES_HIGH, ///< crystal resolution
- LS_D_RES_LOW
+ LS_D_RES_LOW,
+ LS_R_FACTOR_R_WORK, ///< R-work value
+ LS_R_FACTOR_R_FREE ///< R-free value
} RefineItems;
/// \enum items of the pdbx_struct_assembly category
diff --git a/modules/io/tests/test_mmcif_reader.cc b/modules/io/tests/test_mmcif_reader.cc
index 83cba518f..fb87382d1 100644
--- a/modules/io/tests/test_mmcif_reader.cc
+++ b/modules/io/tests/test_mmcif_reader.cc
@@ -678,11 +678,9 @@ BOOST_AUTO_TEST_CASE(mmcif_refine_tests)
IOProfile profile;
MMCifReader mmcif_p(s, eh, profile);
BOOST_REQUIRE_NO_THROW(mmcif_p.Parse());
- #if OST_DOUBLE_PRECISION
- BOOST_CHECK_CLOSE(mmcif_p.GetInfo().GetResolution(), 2.0, 0.001);
- #else
- BOOST_CHECK_CLOSE(mmcif_p.GetInfo().GetResolution(), 2.0f, 0.001f);
- #endif
+ BOOST_CHECK_CLOSE(mmcif_p.GetInfo().GetResolution(), Real(2), Real(0.001));
+ BOOST_CHECK_CLOSE(mmcif_p.GetInfo().GetRFree(), Real(0.229), Real(0.01));
+ BOOST_CHECK_CLOSE(mmcif_p.GetInfo().GetRWork(), Real(0.200), Real(0.01));
}
BOOST_TEST_MESSAGE(" done.");
BOOST_TEST_MESSAGE(" capturing fishy data lines...");
diff --git a/modules/io/tests/testfiles/mmcif/atom_site.mmcif b/modules/io/tests/testfiles/mmcif/atom_site.mmcif
index 68b2fbb57..9bd8159d6 100644
--- a/modules/io/tests/testfiles/mmcif/atom_site.mmcif
+++ b/modules/io/tests/testfiles/mmcif/atom_site.mmcif
@@ -51,9 +51,11 @@ _citation_author.name
_exptl.entry_id experiment1
_exptl.method 'Deep-fry'
-_refine.entry_id '1BAR'
-_refine.ls_d_res_high 2.0
-_refine.ls_d_res_low 1.5
+_refine.entry_id '1BAR'
+_refine.ls_d_res_high 2.0
+_refine.ls_d_res_low 1.5
+_refine.ls_R_factor_R_work 0.200
+_refine.ls_R_factor_R_free 0.229
# biounit begin
loop_
--
GitLab