From 98ebf05e210265b04ff515fe44e93474429603c5 Mon Sep 17 00:00:00 2001
From: Xavier Robin <xavier.robin@unibas.ch>
Date: Mon, 13 Mar 2023 16:54:47 +0100
Subject: [PATCH] refactor: SCHWED-5481 -ec instead of -cr, report
inconsistencies
---
actions/ost-compare-ligand-structures | 28 +++++++++++++++-------
modules/mol/alg/pymod/chain_mapping.py | 32 +++++++++++++++++++++++--
modules/mol/alg/pymod/ligand_scoring.py | 16 +++++++++++++
3 files changed, 66 insertions(+), 10 deletions(-)
diff --git a/actions/ost-compare-ligand-structures b/actions/ost-compare-ligand-structures
index b04f1be17..ce0decf6a 100644
--- a/actions/ost-compare-ligand-structures
+++ b/actions/ost-compare-ligand-structures
@@ -148,13 +148,19 @@ def _ParseArgs():
"a global BLOSUM62-based alignment (NUC44 for nucleotides)."))
parser.add_argument(
- "-cr",
- "--check-resnames",
- dest="check_resnames",
- default=True,
- action="store_false",
- help=("Enforce residue name matches between mapped model and target"
- "residues."))
+ "-ec",
+ "--enforce-consistency",
+ dest="enforce_consistency",
+ default=False,
+ action="store_true",
+ help=("Enforce consistency of residue names between the reference "
+ "binding site and the model. By default residue name "
+ "discrepancies are reported but the program proceeds. "
+ "If this is set to True, the program will fail with an error "
+ "message if the residues names differ. "
+ "Note: more binding site mappings may be explored during "
+ "scoring, but only inconsistencies in the selected mapping are "
+ "reported."))
parser.add_argument(
"-sm",
@@ -309,7 +315,7 @@ def _Process(model, model_ligands, reference, reference_ligands, args):
model_ligands=model_ligands,
target_ligands=reference_ligands,
resnum_alignments=args.residue_number_alignment,
- check_resnames=args.check_resnames,
+ check_resnames=args.enforce_consistency,
substructure_match=args.substructure_match,
radius=args.radius,
lddt_pli_radius=args.lddt_pli_radius,
@@ -360,6 +366,9 @@ def _Process(model, model_ligands, reference, reference_ligands, args):
lddt_pli["bs_ref_res_mapped"]]
lddt_pli["bs_mdl_res_mapped"] = [r.qualified_name for r in
lddt_pli["bs_mdl_res_mapped"]]
+ lddt_pli["inconsistent_residues"] = ["%s-%s" %(
+ x.qualified_name, y.qualified_name) for x,y in lddt_pli[
+ "inconsistent_residues"]]
out["lddt_pli"][model_key] = lddt_pli
if args.rmsd:
@@ -379,6 +388,9 @@ def _Process(model, model_ligands, reference, reference_ligands, args):
rmsd["bs_ref_res_mapped"]]
rmsd["bs_mdl_res_mapped"] = [r.qualified_name for r in
rmsd["bs_mdl_res_mapped"]]
+ rmsd["inconsistent_residues"] = ["%s-%s" %(
+ x.qualified_name, y.qualified_name) for x,y in rmsd[
+ "inconsistent_residues"]]
out["rmsd"][model_key] = rmsd
return out
diff --git a/modules/mol/alg/pymod/chain_mapping.py b/modules/mol/alg/pymod/chain_mapping.py
index 37c36a132..74dfc115f 100644
--- a/modules/mol/alg/pymod/chain_mapping.py
+++ b/modules/mol/alg/pymod/chain_mapping.py
@@ -197,6 +197,7 @@ class ReprResult:
self._gdt_1 = None
self._ost_query = None
self._flat_mapping = None
+ self._inconsistent_residues = None
@property
def lDDT(self):
@@ -228,7 +229,7 @@ class ReprResult:
@property
def mdl_view(self):
- """ The :attr:`ref_view` represention in the model
+ """ The :attr:`ref_view` representation in the model
:type: :class:`ost.mol.EntityView`
"""
@@ -249,7 +250,22 @@ class ReprResult:
:type: :class:`mol.ResidueViewList`
"""
return self.mdl_view.residues
-
+
+ @property
+ def inconsistent_residues(self):
+ """ A list of mapped residue whose names do not match (eg. ALA in the
+ reference and LEU in the model).
+
+ The mismatches are reported as a tuple of :class:`~ost.mol.ResidueView`
+ (reference, model), or as an empty list if all the residue names match.
+
+ :type: :class:`list`
+ """
+ if self._inconsistent_residues is None:
+ self._inconsistent_residues = self._GetInconsistentResidues(
+ self.ref_residues, self.mdl_residues)
+ return self._inconsistent_residues
+
@property
def ref_bb_pos(self):
""" Representative backbone positions for reference residues.
@@ -475,6 +491,18 @@ class ReprResult:
bb_pos.append(at.GetPos())
return bb_pos
+ def _GetInconsistentResidues(self, ref_residues, mdl_residues):
+ """ Helper to extract a list of inconsistent residues.
+ """
+ if len(ref_residues) != len(mdl_residues):
+ raise ValueError("Something terrible happened... Reference and "
+ "model lengths differ... RUN...")
+ inconsistent_residues = list()
+ for ref_residue, mdl_residue in zip(ref_residues, mdl_residues):
+ if ref_residue.name != mdl_residue.name:
+ inconsistent_residues.append((ref_residue, mdl_residue))
+ return inconsistent_residues
+
class ChainMapper:
""" Class to compute chain mappings
diff --git a/modules/mol/alg/pymod/ligand_scoring.py b/modules/mol/alg/pymod/ligand_scoring.py
index ac7dc060d..2298bbbbd 100644
--- a/modules/mol/alg/pymod/ligand_scoring.py
+++ b/modules/mol/alg/pymod/ligand_scoring.py
@@ -488,6 +488,7 @@ class LigandScorer:
"chain_mapping": binding_site.GetFlatChainMapping(),
"transform": binding_site.transform,
"substructure_match": substructure_match,
+ "inconsistent_residues": binding_site.inconsistent_residues,
}
LogDebug("Saved RMSD")
@@ -548,6 +549,7 @@ class LigandScorer:
"chain_mapping": binding_site.GetFlatChainMapping(),
"transform": binding_site.transform,
"substructure_match": substructure_match,
+ "inconsistent_residues": binding_site.inconsistent_residues,
}
LogDebug("Saved lDDT-PLI")
@@ -765,6 +767,13 @@ class LigandScorer:
(substructure) match. A value of `True` indicates that the target
ligand covers only part of the model, while `False` indicates a
perfect match.
+ * `inconsistent_residues`: a list of tuples of mapped residues views
+ (:class:`~ost.mol.ResidueView`) with residue names that differ
+ between the reference and the model, respectively.
+ The list is empty if all residue names match, which is guaranteed
+ if `check_resnames=True`.
+ Note: more binding site mappings may be explored during scoring,
+ but only inconsistencies in the selected mapping are reported.
:rtype: :class:`dict`
"""
@@ -820,6 +829,13 @@ class LigandScorer:
(substructure) match. A value of `True` indicates that the target
ligand covers only part of the model, while `False` indicates a
perfect match.
+ * `inconsistent_residues`: a list of tuples of mapped residues views
+ (:class:`~ost.mol.ResidueView`) with residue names that differ
+ between the reference and the model, respectively.
+ The list is empty if all residue names match, which is guaranteed
+ if `check_resnames=True`.
+ Note: more binding site mappings may be explored during scoring,
+ but only inconsistencies in the selected mapping are reported.
:rtype: :class:`dict`
"""
--
GitLab