SCHWED-2440: use "peptide" flag to filter chains in mol.alg.Superpose.

This changes behaviors in MatchResidueByLocalAln and MatchResidueByGlobalAln (
or Superpose with match = local-aln or global-aln) which previously failed for
non-"protein" chains (i.e. anything w/o peptide bonds). Now, amino acid chains
without peptide bonds (e.g. CA-only) are acceptable and other chains (e.g.
ligands / water) are skipped correctly.

modules/mol/alg/pymod/superpose.py

@@ -193,8 +193,8 @@ def _MatchResidueByAln(ent_a, ent_b, atoms, alnmethod):
     ## fetch chains (peptide-linking residues only)
     chain_a = ent_a.chains[i]
     chain_b = ent_b.chains[i]
-    chain_view_a = chain_a.Select('protein=true')
-    chain_view_b = chain_b.Select('protein=true')
+    chain_view_a = chain_a.Select('peptide=true')
+    chain_view_b = chain_b.Select('peptide=true')
     if chain_view_a.chain_count == 0 or chain_view_b.chain_count == 0:
       # skip empty chains
       continue
@@ -240,12 +240,12 @@ def MatchResidueByLocalAln(ent_a, ent_b, atoms='all'):
   """
   Match residues by local alignment. Takes **ent_a** and **ent_b**, extracts the
   sequences chain-wise and aligns them in Smith/Waterman manner using the
-  BLOSUM62 matrix for scoring. Only residues which are marked as "protein" (see
-  :attr:`~ost.mol.ResidueHandle.is_protein`) are considered for alignment. The
-  residues of the entities are then matched based on this alignment. Only atoms
-  present in both residues are included in the views. Chains are processed in
-  order of appearance. If **ent_a** and **ent_b** contain a different number of
-  chains, processing stops with the lower count.
+  BLOSUM62 matrix for scoring. Only residues which are marked as :attr:`peptide
+  linking <ost.mol.ResidueHandle.peptide_linking>` are considered for alignment.
+  The residues of the entities are then matched based on this alignment. Only
+  atoms present in both residues are included in the views. Chains are processed
+  in order of appearance. If **ent_a** and **ent_b** contain a different number
+  of chains, processing stops with the lower count.
 
   :param ent_a: The first entity
   :type ent_a: :class:`~ost.mol.EntityView` or :class:`~ost.mol.EntityHandle`

modules/mol/alg/tests/test_convenient_superpose.py

@@ -176,7 +176,7 @@ class TestConvenientSuperpose(unittest.TestCase):
     view1, view2 = mol.alg.MatchResidueByLocalAln(ent_view_wrong, ent_view)
     self.assertEqualAtomOrder(view1, view2)
 
-  def testWrongResidueOrder(self): 
+  def testWrongResidueOrder(self):
     ent_view = io.LoadEntity(os.path.join("testfiles","1aho.pdb")).Select("")
     ent_view_wrong = ent_view.CreateEmptyView()
     for c in ent_view.chains:
@@ -193,6 +193,37 @@ class TestConvenientSuperpose(unittest.TestCase):
     view1, view2 = mol.alg.MatchResidueByNum(ent_view_wrong, ent_view)
     self.assertEqualAtomOrder(view1, view2)
 
+  def testPeptideFilter(self):
+    # check that CA only chains are ok and ligand chains are skipped
+    ent_1_full = io.LoadPDB(os.path.join("testfiles", "5tglA.pdb"))
+    ent_2_full = io.LoadPDB(os.path.join("testfiles", "5tglA_modified.pdb"))
+    exp_atom_count = 253
+    # check CA-only chain
+    ent_1 = ent_1_full.Select("cname=A")
+    ent_2 = ent_2_full.Select("cname=A")
+    view1, view2 = mol.alg.MatchResidueByLocalAln(ent_1, ent_2)
+    self.assertEqual(view1.atom_count, exp_atom_count)
+    self.assertEqual(view2.atom_count, exp_atom_count)
+    view1, view2 = mol.alg.MatchResidueByGlobalAln(ent_1, ent_2)
+    self.assertEqual(view1.atom_count, exp_atom_count)
+    self.assertEqual(view2.atom_count, exp_atom_count)
+    # check ligand chain
+    ent_1_ligand = ent_1_full.Select("cname='_'")
+    ent_2_ligand = ent_2_full.Select("cname='_'")
+    view1, view2 = mol.alg.MatchResidueByLocalAln(ent_1_ligand, ent_2_ligand)
+    self.assertEqual(view1.atom_count, 0)
+    self.assertEqual(view2.atom_count, 0)
+    view1, view2 = mol.alg.MatchResidueByGlobalAln(ent_1_ligand, ent_2_ligand)
+    self.assertEqual(view1.atom_count, 0)
+    self.assertEqual(view2.atom_count, 0)
+    # check both together
+    view1, view2 = mol.alg.MatchResidueByLocalAln(ent_1_full, ent_2_full)
+    self.assertEqual(view1.atom_count, exp_atom_count)
+    self.assertEqual(view2.atom_count, exp_atom_count)
+    view1, view2 = mol.alg.MatchResidueByGlobalAln(ent_1_full, ent_2_full)
+    self.assertEqual(view1.atom_count, exp_atom_count)
+    self.assertEqual(view2.atom_count, exp_atom_count)
+
 if __name__ == "__main__":
   from ost import testutils
   testutils.RunTests()