diff --git a/actions/ost-compare-ligand-structures b/actions/ost-compare-ligand-structures
index d8e55d6b5a22bdd332bf46192b9e839fb56adeec..c05c678d10df72a85116227a1e83ec226faa8d57 100644
--- a/actions/ost-compare-ligand-structures
+++ b/actions/ost-compare-ligand-structures
@@ -442,6 +442,8 @@ def _Process(model, model_ligands, reference, reference_ligands, args):
trg_lig = scorer.target.FindResidue(chain, resnum)
out["unassigned_reference_ligands"][reference_ligands_map[
trg_lig.hash_code]] = unassigned
+ out["unassigned_model_ligand_descriptions"] = scorer.unassigned_model_ligand_descriptions
+ out["unassigned_reference_ligand_descriptions"] = scorer.unassigned_target_ligand_descriptions
if args.lddt_pli:
diff --git a/modules/mol/alg/pymod/ligand_scoring.py b/modules/mol/alg/pymod/ligand_scoring.py
index e4b37ed65f2a279c466569c603bfba80eba6d079..072f60573ee67e4e0f42a3d1497740b11511c0e6 100644
--- a/modules/mol/alg/pymod/ligand_scoring.py
+++ b/modules/mol/alg/pymod/ligand_scoring.py
@@ -313,6 +313,10 @@ class LigandScorer:
self._unassigned_target_ligands = None
self._unassigned_model_ligands = None
self._unassigned_target_ligands_reason = {}
+ self._unassigned_target_ligand_short = None
+ self._unassigned_model_ligand_short = None
+ self._unassigned_target_ligand_descriptions = None
+ self._unassigned_model_ligand_descriptions = None
# Keep track of symmetries/isomorphisms
# 0.0: no isomorphism
# 1.0: isomorphic
@@ -1145,7 +1149,7 @@ class LigandScorer:
`unassigned=True`: `False`
If the scoring object was instantiated with `unassigned=True`, in
- addition the unmapped ligands will be reported with a score of `None`
+ addition the unassigned ligands will be reported with a score of `None`
and the following information:
* `unassigned`: `True`,
@@ -1253,11 +1257,13 @@ class LigandScorer:
"""Get a dictionary of target ligands not assigned to any model ligand,
keyed by target ligand (chain name, :class:`~ost.mol.ResNum`).
- Assignment is the same as for the lDDT-PLI score (and is controlled
+ The assignment for the lDDT-PLI score is used (and is controlled
by the `rmsd_assignment` argument).
- Each sub-dictionary contains a string from a controlled dictionary
+ Each item contains a string from a controlled dictionary
about the reason for the absence of assignment.
+ A human-readable description can be obtained from the
+ :attr:`unassigned_target_ligand_descriptions` property.
Currently, the following reasons are reported:
@@ -1280,24 +1286,42 @@ class LigandScorer:
:rtype: :class:`dict`
"""
- if self._unassigned_target_ligands is None:
+ if self._unassigned_target_ligand_short is None:
if self.rmsd_assignment:
self._assign_ligands_rmsd_only()
else:
self._assign_ligands_lddt_pli()
- return self._unassigned_target_ligands
+ self._unassigned_target_ligand_short = {}
+ self._unassigned_target_ligand_descriptions = {}
+ for cname, res in self._unassigned_target_ligands.items():
+ self._unassigned_target_ligand_short[cname] = {}
+ for resnum, val in res.items():
+ self._unassigned_target_ligand_short[cname][resnum] = val[0]
+ self._unassigned_target_ligand_descriptions[val[0]] = val[1]
+ return self._unassigned_target_ligand_short
+
+ @property
+ def unassigned_target_ligand_descriptions(self):
+ """Get a human-readable description of why target ligands were
+ unassigned, as a dictionary keyed by the controlled dictionary
+ from :attr:`unassigned_target_ligands`.
+ """
+ if self._unassigned_target_ligand_descriptions is None:
+ _ = self.unassigned_target_ligands # assigned there
+ return self._unassigned_target_ligand_descriptions
@property
def unassigned_model_ligands(self):
"""Get a dictionary of model ligands not assigned to any target ligand,
keyed by model ligand (chain name, :class:`~ost.mol.ResNum`).
- Assignment is the same as for the lDDT-PLI score (and is controlled
+ The assignment for the lDDT-PLI score is used (and is controlled
by the `rmsd_assignment` argument).
- Each sub-dictionary contains a tuple with information about the reason
- for the absence of assignment, in short and long format.
-
+ Each item contains a string from a controlled dictionary
+ about the reason for the absence of assignment.
+ A human-readable description can be obtained from the
+ :attr:`unassigned_model_ligand_descriptions` property.
Currently, the following reasons are reported:
* `no_ligand`: there was no ligand in the target.
@@ -1322,12 +1346,29 @@ class LigandScorer:
:rtype: :class:`dict`
"""
- if self._unassigned_model_ligands is None:
+ if self._unassigned_model_ligand_short is None:
if self.rmsd_assignment:
self._assign_ligands_rmsd_only()
else:
self._assign_ligands_lddt_pli()
- return self._unassigned_model_ligands
+ self._unassigned_model_ligand_short = {}
+ self._unassigned_model_ligand_descriptions = {}
+ for cname, res in self._unassigned_model_ligands.items():
+ self._unassigned_model_ligand_short[cname] = {}
+ for resnum, val in res.items():
+ self._unassigned_model_ligand_short[cname][resnum] = val[0]
+ self._unassigned_model_ligand_descriptions[val[0]] = val[1]
+ return self._unassigned_model_ligand_short
+
+ @property
+ def unassigned_model_ligand_descriptions(self):
+ """Get a human-readable description of why model ligands were
+ unassigned, as a dictionary keyed by the controlled dictionary
+ from :attr:`unassigned_model_ligands`.
+ """
+ if self._unassigned_model_ligand_descriptions is None:
+ _ = self.unassigned_model_ligands # assigned there
+ return self._unassigned_model_ligand_descriptions
def _set_custom_mapping(self, mapping):
diff --git a/modules/mol/alg/tests/test_ligand_scoring.py b/modules/mol/alg/tests/test_ligand_scoring.py
index 9a279594c5dbe60927986e4e901ac936d8e8e66e..3896aa2fdd9d3a2e85deee047ac1e48ac7b5dc7f 100644
--- a/modules/mol/alg/tests/test_ligand_scoring.py
+++ b/modules/mol/alg/tests/test_ligand_scoring.py
@@ -502,24 +502,19 @@ class TestLigandScoring(unittest.TestCase):
# Check unassigned targets
# NA: not in contact with target
trg_na = sc.target.FindResidue("L_NA", 1)
- assert sc._find_unassigned_target_ligand_reason(trg_na)[0] == "binding_site"
- assert sc.unassigned_target_ligands["L_NA"][1][0] == "binding_site"
+ assert sc.unassigned_target_ligands["L_NA"][1] == "binding_site"
# ZN: no representation
trg_zn = sc.target.FindResidue("H", 1)
- assert sc._find_unassigned_target_ligand_reason(trg_zn)[0] == "model_representation"
- assert sc.unassigned_target_ligands["H"][1][0] == "model_representation"
+ assert sc.unassigned_target_ligands["H"][1] == "model_representation"
# AFB: not identical to anything in the model
trg_afb = sc.target.FindResidue("G", 1)
- assert sc._find_unassigned_target_ligand_reason(trg_afb)[0] == "identity"
- assert sc.unassigned_target_ligands["G"][1][0] == "identity"
+ assert sc.unassigned_target_ligands["G"][1] == "identity"
# F.G3D1: J.G3D1 assigned instead
trg_fg3d1 = sc.target.FindResidue("F", 1)
- assert sc._find_unassigned_target_ligand_reason(trg_fg3d1)[0] == "stoichiometry"
- assert sc.unassigned_target_ligands["F"][1][0] == "stoichiometry"
+ assert sc.unassigned_target_ligands["F"][1] == "stoichiometry"
# CMO: disconnected
trg_cmo1 = sc.target.FindResidue("L_CMO", 1)
- sc._find_unassigned_target_ligand_reason(trg_cmo1)[0] == "disconnected"
- assert sc.unassigned_target_ligands["L_CMO"][1][0] == "disconnected"
+ assert sc.unassigned_target_ligands["L_CMO"][1] == "disconnected"
# J.G3D1: assigned to L_2.G3D1 => error
trg_jg3d1 = sc.target.FindResidue("J", 1)
with self.assertRaises(RuntimeError):
@@ -532,23 +527,19 @@ class TestLigandScoring(unittest.TestCase):
# Check unassigned models
# OXY: not identical to anything in the model
mdl_oxy = sc.model.FindResidue("L_OXY", 1)
- assert sc._find_unassigned_model_ligand_reason(mdl_oxy)[0] == "identity"
- assert sc.unassigned_model_ligands["L_OXY"][1][0] == "identity"
+ assert sc.unassigned_model_ligands["L_OXY"][1] == "identity"
assert sc.lddt_pli["L_OXY"][1] is None
# NA: not in contact with target
mdl_na = sc.model.FindResidue("L_NA", 1)
- assert sc._find_unassigned_model_ligand_reason(mdl_na)[0] == "binding_site"
- assert sc.unassigned_model_ligands["L_NA"][1][0] == "binding_site"
+ assert sc.unassigned_model_ligands["L_NA"][1] == "binding_site"
assert sc.lddt_pli["L_NA"][1] is None
# ZN: no representation
mdl_zn = sc.model.FindResidue("L_ZN", 1)
- assert sc._find_unassigned_model_ligand_reason(mdl_zn)[0] == "model_representation"
- assert sc.unassigned_model_ligands["L_ZN"][1][0] == "model_representation"
+ assert sc.unassigned_model_ligands["L_ZN"][1] == "model_representation"
assert sc.lddt_pli["L_ZN"][1] is None
# MG in L_MG_2 has stupid coordinates and is not assigned
mdl_mg_2 = sc.model.FindResidue("L_MG_2", 1)
- assert sc._find_unassigned_model_ligand_reason(mdl_mg_2)[0] == "stoichiometry"
- assert sc.unassigned_model_ligands["L_MG_2"][1][0] == "stoichiometry"
+ assert sc.unassigned_model_ligands["L_MG_2"][1] == "stoichiometry"
assert sc.lddt_pli["L_MG_2"][1] is None
# MG in L_MG_0: assigned to I.MG1 => error
mdl_mg_0 = sc.model.FindResidue("L_MG_0", 1)
@@ -557,8 +548,7 @@ class TestLigandScoring(unittest.TestCase):
assert "L_MG_0" not in sc.unassigned_model_ligands
# CMO: disconnected
mdl_cmo1 = sc.model.FindResidue("L_CMO", 1)
- sc._find_unassigned_model_ligand_reason(mdl_cmo1)[0] == "disconnected"
- assert sc.unassigned_model_ligands["L_CMO"][1][0] == "disconnected"
+ assert sc.unassigned_model_ligands["L_CMO"][1] == "disconnected"
# Raises with an invalid ligand
with self.assertRaises(ValueError):
sc._find_unassigned_model_ligand_reason(sc.target_ligands[0])
@@ -567,47 +557,36 @@ class TestLigandScoring(unittest.TestCase):
sc = LigandScorer(mdl, trg, None, None, unassigned=True,
rmsd_assignment=True)
assert sc.unassigned_model_ligands == {
- 'L_ZN': {1: ('model_representation',
- 'No representation of the reference binding site was found in the model')},
- 'L_NA': {1: ('binding_site',
- 'No residue in proximity of the target ligand')},
- 'L_OXY': {1: ('identity',
- 'Ligand was not found in the target (by full graph isomorphism)')},
- 'L_MG_2': {1: ('stoichiometry',
- 'Ligand was already assigned to an other model ligand (different stoichiometry)')},
- "L_CMO": {1: ('disconnected',
- 'Ligand graph is disconnected')}
+ 'L_ZN': {1: 'model_representation'},
+ 'L_NA': {1: 'binding_site'},
+ 'L_OXY': {1: 'identity'},
+ 'L_MG_2': {1: 'stoichiometry'},
+ "L_CMO": {1: 'disconnected'}
}
assert sc.unassigned_target_ligands == {
- 'G': {1: ('identity',
- 'Ligand was not found in the model (by full graph isomorphism)')},
- 'H': {1: ('model_representation',
- 'No representation of the reference binding site was found in the model')},
- 'J': {1: ('stoichiometry',
- 'Ligand was already assigned to an other target ligand (different stoichiometry)')},
- 'K': {1: ('identity',
- 'Ligand was not found in the model (by full graph isomorphism)')},
- 'L_NA': {1: ('binding_site',
- 'No residue in proximity of the target ligand')},
- "L_CMO": {1: ('disconnected',
- 'Ligand graph is disconnected')}
+ 'G': {1: 'identity'},
+ 'H': {1: 'model_representation'},
+ 'J': {1: 'stoichiometry'},
+ 'K': {1: 'identity'},
+ 'L_NA': {1: 'binding_site'},
+ "L_CMO": {1: 'disconnected'}
}
assert sc.lddt_pli["L_OXY"][1] is None
# With missing ligands
sc = LigandScorer(mdl.Select("cname=A"), trg, None, None)
- assert sc.unassigned_target_ligands["E"][1] == ('no_ligand', 'No ligand in the model')
+ assert sc.unassigned_target_ligands["E"][1] == 'no_ligand'
sc = LigandScorer(mdl, trg.Select("cname=A"), None, None)
- assert sc.unassigned_model_ligands["L_2"][1] == ('no_ligand', 'No ligand in the target')
+ assert sc.unassigned_model_ligands["L_2"][1] == 'no_ligand'
sc = LigandScorer(mdl.Select("cname=A"), trg, None, None,
unassigned=True, rmsd_assignment=True)
- assert sc.unassigned_target_ligands["E"][1] == ('no_ligand', 'No ligand in the model')
+ assert sc.unassigned_target_ligands["E"][1] == 'no_ligand'
sc = LigandScorer(mdl, trg.Select("cname=A"), None, None,
unassigned=True, rmsd_assignment=True)
- assert sc.unassigned_model_ligands["L_2"][1] == ('no_ligand', 'No ligand in the target')
+ assert sc.unassigned_model_ligands["L_2"][1] == 'no_ligand'
# However not everything must be missing
with self.assertRaises(ValueError):