diff --git a/actions/ost-compare-structures b/actions/ost-compare-structures
index 9dacd7786f13195ca42451ac4b569a01a502cdbc..2e4c324792a614591a2aa6ae73e882b5cd5c2750 100644
--- a/actions/ost-compare-structures
+++ b/actions/ost-compare-structures
@@ -463,6 +463,31 @@ def _ParseArgs():
"that number, the chain mapping will naively enumerate all "
"possible mappings. A heuristic is used otherwise."))
+ parser.add_argument(
+ "--dump-aligned-residues",
+ dest="dump_aligned_residues",
+ default=False,
+ action="store_true",
+ help=("Dump additional info on aligned model and reference residues."))
+
+ parser.add_argument(
+ "--dump-pepnuc-alns",
+ dest="dump_pepnuc_alns",
+ default=False,
+ action="store_true",
+ help=("Dump alignments of mapped chains but with sequences that did "
+ "not undergo Molck preprocessing in the scorer. Sequences are "
+ "extracted from model/target after undergoing selection for "
+ "peptide and nucleotide residues."))
+
+ parser.add_argument(
+ "--dump-pepnuc-aligned-residues",
+ dest="dump_pepnuc_aligned_residues",
+ default=False,
+ action="store_true",
+ help=("Dump additional info on model and reference residues that occur "
+ "in pepnuc alignments."))
+
return parser.parse_args()
def _Rename(ent):
@@ -558,7 +583,7 @@ def _AlnToFastaStr(aln):
"""
s1 = aln.GetSequence(0)
s2 = aln.GetSequence(1)
- return f">reference:{s1.name}\n{str(s1)}\nmodel:{s2.name}\n{str(s2)}"
+ return f">reference:{s1.name}\n{str(s1)}\n>model:{s2.name}\n{str(s2)}"
def _GetInconsistentResidues(alns):
lst = list()
@@ -609,6 +634,41 @@ def _PatchScoresToJSONList(interface_dict, score_dict):
json_list += score_dict[ch]
return json_list
+def _GetAlignedResidues(aln):
+ aligned_residues = list()
+ for a in aln:
+ mdl_lst = list()
+ ref_lst = list()
+ for c in a:
+ mdl_r = c.GetResidue(1)
+ ref_r = c.GetResidue(0)
+ if mdl_r.IsValid():
+ olc = mdl_r.one_letter_code
+ num = mdl_r.GetNumber().num
+ ins_code = mdl_r.GetNumber().ins_code.strip("\u0000")
+ mdl_lst.append({"olc": olc,
+ "num": f"{num}.{ins_code}"})
+ else:
+ mdl_lst.append(None)
+
+ if ref_r.IsValid():
+ olc = ref_r.one_letter_code
+ num = ref_r.GetNumber().num
+ ins_code = ref_r.GetNumber().ins_code.strip("\u0000")
+ ref_lst.append({"olc": olc,
+ "num": f"{num}.{ins_code}"})
+ else:
+ ref_lst.append(None)
+
+ mdl_dct = {"chain": a.GetSequence(1).GetName(),
+ "residues": mdl_lst}
+ ref_dct = {"chain": a.GetSequence(0).GetName(),
+ "residues": ref_lst}
+
+ aligned_residues.append({"model": mdl_dct,
+ "reference": ref_dct})
+ return aligned_residues
+
def _Process(model, reference, args):
mapping = None
@@ -637,6 +697,15 @@ def _Process(model, reference, args):
out["aln"] = [_AlnToFastaStr(aln) for aln in scorer.aln]
out["inconsistent_residues"] = ir
+ if args.dump_aligned_residues:
+ out["aligned_residues"] = _GetAlignedResidues(scorer.aln)
+
+ if args.dump_pepnuc_alns:
+ out["pepnuc_aln"] = [_AlnToFastaStr(aln) for aln in scorer.pepnuc_aln]
+
+ if args.dump_pepnuc_aligned_residues:
+ out["pepnuc_aligned_residues"] = _GetAlignedResidues(scorer.pepnuc_aln)
+
if args.lddt:
out["lddt"] = scorer.lddt
diff --git a/modules/mol/alg/pymod/scoring.py b/modules/mol/alg/pymod/scoring.py
index 45ccb439e860377dcc672c9fa32516de13c6c749..af70aca279de11ec5a615d6bfa92a794a17bf665 100644
--- a/modules/mol/alg/pymod/scoring.py
+++ b/modules/mol/alg/pymod/scoring.py
@@ -147,15 +147,18 @@ class Scorer:
lddt_no_stereochecks=False, n_max_naive=12,
oum=False):
- if isinstance(model, mol.EntityView):
- model = mol.CreateEntityFromView(model, False)
+ self._target_orig = target
+ self._model_orig = model
+
+ if isinstance(self._model_orig, mol.EntityView):
+ self._model = mol.CreateEntityFromView(self._model_orig, False)
else:
- model = model.Copy()
+ self._model = self._model_orig.Copy()
- if isinstance(target, mol.EntityView):
- target = mol.CreateEntityFromView(target, False)
+ if isinstance(self._target_orig, mol.EntityView):
+ self._target = mol.CreateEntityFromView(self._target_orig, False)
else:
- target = target.Copy()
+ self._target = self._target_orig.Copy()
if molck_settings is None:
molck_settings = MolckSettings(rm_unk_atoms=True,
@@ -166,8 +169,8 @@ class Scorer:
colored=False,
map_nonstd_res=True,
assign_elem=True)
- Molck(model, conop.GetDefaultLib(), molck_settings)
- Molck(target, conop.GetDefaultLib(), molck_settings)
+ Molck(self._model, conop.GetDefaultLib(), molck_settings)
+ Molck(self._target, conop.GetDefaultLib(), molck_settings)
self._model = model.Select("peptide=True or nucleotide=True")
self._target = target.Select("peptide=True or nucleotide=True")
@@ -245,6 +248,7 @@ class Scorer:
self._target_interface_residues = None
self._aln = None
self._stereochecked_aln = None
+ self._pepnuc_aln = None
# lazily constructed scorer objects
self._lddt_scorer = None
@@ -326,6 +330,14 @@ class Scorer:
"""
return self._model
+ @property
+ def model_orig(self):
+ """ The original model passed at object construction
+
+ :type: :class:`ost.mol.EntityHandle`/:class:`ost.mol.EntityView`
+ """
+ return self._model_orig
+
@property
def target(self):
""" Target with Molck cleanup
@@ -334,6 +346,14 @@ class Scorer:
"""
return self._target
+ @property
+ def target_orig(self):
+ """ The original target passed at object construction
+
+ :type: :class:`ost.mol.EntityHandle`/:class:`ost.mol.EntityView`
+ """
+ return self._target_orig
+
@property
def aln(self):
""" Alignments of :attr:`model`/:attr:`target` chains
@@ -353,12 +373,26 @@ class Scorer:
The alignments may differ, as stereochecks potentially remove residues
- :type: :class:``
+ :type: :class:`list` of :class:`ost.seq.AlignmentHandle`
"""
if self._stereochecked_aln is None:
self._compute_stereochecked_aln()
return self._stereochecked_aln
+ @property
+ def pepnuc_aln(self):
+ """ Alignments of :attr:`model_orig`/:attr:`target_orig` chains
+
+ Selects for peptide and nucleotide residues before sequence
+ extraction. Includes residues that would be removed by molck in
+ structure preprocessing.
+
+ :type: :class:`list` of :class:`ost.seq.AlignmentHandle`
+ """
+ if self._pepnuc_aln is None:
+ self._compute_pepnuc_aln()
+ return self._pepnuc_aln
+
@property
def stereochecked_model(self):
""" View of :attr:`~model` that has stereochemistry checks applied
@@ -638,11 +672,15 @@ class Scorer:
""" Interfaces in :attr:`~target` with non-zero contribution to
:attr:`~qs_global`/:attr:`~qs_best`
+ Chain names are lexicographically sorted.
+
:type: :class:`list` of :class:`tuple` with 2 elements each:
(trg_ch1, trg_ch2)
"""
if self._qs_target_interfaces is None:
self._qs_target_interfaces = self.qs_scorer.qsent1.interacting_chains
+ self._qs_target_interfaces = \
+ [(min(x[0],x[1]), max(x[0],x[1])) for x in self._qs_target_interfaces]
return self._qs_target_interfaces
@property
@@ -650,11 +688,16 @@ class Scorer:
""" Interfaces in :attr:`~model` with non-zero contribution to
:attr:`~qs_global`/:attr:`~qs_best`
+ Chain names are lexicographically sorted.
+
:type: :class:`list` of :class:`tuple` with 2 elements each:
(mdl_ch1, mdl_ch2)
"""
if self._qs_model_interfaces is None:
self._qs_model_interfaces = self.qs_scorer.qsent2.interacting_chains
+ self._qs_model_interfaces = \
+ [(min(x[0],x[1]), max(x[0],x[1])) for x in self._qs_model_interfaces]
+
return self._qs_model_interfaces
@property
@@ -662,6 +705,8 @@ class Scorer:
""" Interfaces in :attr:`~qs_target_interfaces` that can be mapped
to :attr:`~model`.
+ Target chain names are lexicographically sorted.
+
:type: :class:`list` of :class:`tuple` with 4 elements each:
(trg_ch1, trg_ch2, mdl_ch1, mdl_ch2)
"""
@@ -724,26 +769,32 @@ class Scorer:
def contact_target_interfaces(self):
""" Interfaces in :class:`target` which have at least one contact
- Contact as defined in :attr:`~native_contacts`
+ Contact as defined in :attr:`~native_contacts`,
+ chain names are lexicographically sorted.
:type: :class:`list` of :class:`tuple` with 2 elements each
(trg_ch1, trg_ch2)
"""
if self._contact_target_interfaces is None:
- self._contact_target_interfaces = self.contact_scorer.cent1.interacting_chains
+ tmp = self.contact_scorer.cent1.interacting_chains
+ tmp = [(min(x[0],x[1]), max(x[0],x[1])) for x in tmp]
+ self._contact_target_interfaces = tmp
return self._contact_target_interfaces
@property
def contact_model_interfaces(self):
""" Interfaces in :class:`model` which have at least one contact
- Contact as defined in :attr:`~native_contacts`
+ Contact as defined in :attr:`~native_contacts`,
+ chain names are lexicographically sorted.
:type: :class:`list` of :class:`tuple` with 2 elements each
(mdl_ch1, mdl_ch2)
"""
if self._contact_model_interfaces is None:
- self._contact_model_interfaces = self.contact_scorer.cent2.interacting_chains
+ tmp = self.contact_scorer.cent2.interacting_chains
+ tmp = [(min(x[0],x[1]), max(x[0],x[1])) for x in tmp]
+ self._contact_model_interfaces = tmp
return self._contact_model_interfaces
@property
@@ -902,7 +953,7 @@ class Scorer:
""" Interfaces in :attr:`target` that are relevant for DockQ
In principle a subset of :attr:`~contact_target_interfaces` that only
- contains peptide sequences.
+ contains peptide sequences. Chain names are lexicographically sorted.
:type: :class:`list` of :class:`tuple` with 2 elements each:
(trg_ch1, trg_ch2)
@@ -920,6 +971,8 @@ class Scorer:
""" Interfaces in :attr:`dockq_target_interfaces` that can be mapped
to model
+ Target chain names are lexicographically sorted
+
:type: :class:`list` of :class:`tuple` with 4 elements each:
(trg_ch1, trg_ch2, mdl_ch1, mdl_ch2)
"""
@@ -1322,6 +1375,12 @@ class Scorer:
alns[-1].AttachView(1, mdl_seqs[mdl_ch].GetAttachedView())
return alns
+ def _compute_pepnuc_aln(self):
+ query = "peptide=true or nucleotide=true"
+ pep_nuc_target = self.target_orig.Select(query)
+ pep_nuc_model = self.model_orig.Select(query)
+ self._pepnuc_aln = self._aln_helper(pep_nuc_target, pep_nuc_model)
+
def _compute_aln(self):
self._aln = self._aln_helper(self.target, self.model)