MOL2 parser populates elements attribute

package/CHANGELOG

@@ -117,6 +117,7 @@ Fixes
   * new `Results` class now can be pickled/unpickled. (PR #3309) 
 
 Enhancements
+  * MOL2 parser populates elements attribute from SYMBL atom types (Issue #3062)
   * Added guessers for aromaticity and Gasteiger partial charges (Issue #2468,
     PR #2926)
   * Added `Results` class for storing analysis results (#3115, PR #3233)

package/MDAnalysis/topology/MOL2Parser.py

@@ -53,13 +53,17 @@
     Atomtypes,
     Bonds,
     Charges,
+    Elements,
     Masses,
     Resids,
     Resnums,
     Resnames,
     Segids,
 )
 from ..core.topology import Topology
+from .tables import SYBYL2SYMB
+
+import warnings
 
 
 class MOL2Parser(TopologyReaderBase):
@@ -70,13 +74,18 @@ class MOL2Parser(TopologyReaderBase):
      - Atomnames
      - Atomtypes
      - Charges
-     - Resids,
+     - Resids
      - Resnames
      - Bonds
+     - Elements
 
     Guesses the following:
      - masses
 
+    Notes
+    -----
+    Elements are obtained directly from the SYBYL atom types.
+
     .. versionchanged:: 0.9
        Now subclasses TopologyReaderBase
     .. versionchanged:: 0.20.0
@@ -93,6 +102,10 @@ def parse(self, **kwargs):
         Returns
         -------
         A MDAnalysis Topology object
+
+
+        .. versionchanged: 2.0.0
+           Parse elements from atom types.
         """
         blocks = []
 
@@ -148,7 +161,22 @@ def parse(self, **kwargs):
 
         n_atoms = len(ids)
 
-        masses = guessers.guess_masses(types)
+        validated_elements = []
+        invalid_elements = set()
+        for at in types:
+            try:
+                validated_elements.append(SYBYL2SYMB[at])
+            except KeyError:
+                invalid_elements.add(at)
+                validated_elements.append('')
+
+        # Print single warning for all unknown elements, if any
+        if invalid_elements:
+            warnings.warn("Unknown elements found for some "
+                          f"atoms: {invalid_elements}. "
+                          "These have been given an empty element record.")
+
+        masses = guessers.guess_masses(validated_elements)
 
         attrs = []
         attrs.append(Atomids(np.array(ids, dtype=np.int32)))
@@ -157,6 +185,9 @@ def parse(self, **kwargs):
         attrs.append(Charges(np.array(charges, dtype=np.float32)))
         attrs.append(Masses(masses, guessed=True))
 
+        if not np.all(validated_elements == ''):
+            attrs.append(Elements(np.array(validated_elements, dtype="U3")))
+
         resids = np.array(resids, dtype=np.int32)
         resnames = np.array(resnames, dtype=object)
 

package/MDAnalysis/topology/tables.py

@@ -394,3 +394,37 @@ def kv2dict(s, convertor=str):
           113: 'Nh', 114: 'Fl', 115: 'Mc', 116: 'Lv', 117: 'Ts', 118: 'Og'}
 
 SYMB2Z = {v:k for k, v in Z2SYMB.items()}
+
+# Conversion between SYBYL atom types and corresponding elements
+# Tripos MOL2 file format: https://zhanglab.ccmb.med.umich.edu/DockRMSD/mol2.pdf
+SYBYL2SYMB = {
+    "H": "H", "H.spc": "H", "H.t3p": "H",
+    "C.3": "C", "C.2": "C", "C.1": "C", "C.ar": "C", "C.cat": "C",
+    "N.3": "N", "N.2": "N", "N.1": "N", "N.ar": "N",
+    "N.am": "N", "N.pl3": "N", "N.4": "N",
+    "O.3": "O", "O.2": "O", "O.co2": "O", "O.spc": "O", "O.t3p": "O",
+    "S.3": "S", "S.2": "S", "S.O": "S", "S.O2": "S",
+    "S.o": "S", "S.o2": "S",  # Non-standard but often found in the wild...
+    "P.3": "P",
+    "F": "F",
+    "Li": "Li",
+    "Na": "Na",
+    "Mg": "Mg",
+    "Al": "Al",
+    "Si": "Si",
+    "K": "K",
+    "Ca": "Ca",
+    "Cr.th": "Cr",
+    "Cr.oh": "Cr",
+    "Mn": "Mn",
+    "Fe": "Fe",
+    "Co.oh": "Co",
+    "Cu": "Cu",
+    "Cl": "Cl",
+    "Br": "Br",
+    "I": "I",
+    "Zn": "Zn",
+    "Se": "Se",
+    "Mo": "Mo",
+    "Sn": "Sn",
+}

testsuite/MDAnalysisTests/converters/test_rdkit.py

@@ -229,6 +229,10 @@ def test_identical_topology(self, rdmol):
 
     def test_raise_requires_elements(self):
         u = mda.Universe(mol2_molecule)
+
+        # Delete topology attribute (PR #3069)
+        u.del_TopologyAttr('elements')
+
         with pytest.raises(
             AttributeError,
             match="`elements` attribute is required for the RDKitConverter"

testsuite/MDAnalysisTests/topology/test_mol2.py

@@ -30,13 +30,15 @@
     mol2_molecule,
     mol2_molecules,
 )
-from numpy.testing import assert_equal
 
+import numpy as np
+from io import StringIO
 
 class TestMOL2Base(ParserBase):
     parser = mda.topology.MOL2Parser.MOL2Parser
     expected_attrs = [
-        'ids', 'names', 'types', 'charges', 'resids', 'resnames', 'bonds'
+        'ids', 'names', 'types', 'charges', 'resids', 'resnames', 'bonds',
+        'elements',
     ]
     guessed_attrs = ['masses']
     expected_n_atoms = 49
@@ -50,6 +52,7 @@ def test_attr_size(self, top):
         assert len(top.charges) == top.n_atoms
         assert len(top.resids) == top.n_residues
         assert len(top.resnames) == top.n_residues
+        assert len(top.elements) == top.n_atoms
 
     def test_bonds(self, top):
         assert len(top.bonds) == 49  # bonds for 49 atoms
@@ -68,3 +71,99 @@ def test_bond_orders():
     u = mda.Universe(mol2_molecule)
     orders = [bond.order for bond in u.atoms.bonds]
     assert_equal(orders, ref_orders)
+
+
+def test_elements():
+    u = mda.Universe(mol2_molecule)
+
+    assert_equal(
+        u.atoms.elements[:5],
+        np.array(["N", "S", "N", "N", "O"], dtype="U3")
+    )
+
+
+# Test for #2927
+def test_elements_selection():
+    u = mda.Universe(mol2_molecule)
+    ag = u.select_atoms("element S")
+
+    assert_equal(
+        ag.elements,
+        np.array(["S", "S"], dtype="U3")
+    )
+
+
+mol2_wrong_element = """\
+@<TRIPOS>MOLECULE
+FXA101_1
+49 51 1 0 0
+SMALL
+USER_CHARGES
+
+
+@<TRIPOS>ATOM
+  1 N1       6.8420     9.9900    22.7430 N.am  1 Q101  -0.8960
+  2 S1       8.1400     9.2310    23.3330 X.o2  1 Q101   1.3220
+  3 N2       4.4000     9.1300    20.4710 XX.am  1 Q101  -0.3970
+"""
+
+
+def test_wrong_elements_warnings():
+    with pytest.warns(UserWarning, match='Unknown elements found') as record:
+        u = mda.Universe(StringIO(mol2_wrong_element), format='MOL2')
+
+    # One warning from invalid elements, one from invalid masses
+    assert len(record) == 2
+    print(record[0].message)
+
+    expected = np.array(['N', '', ''], dtype=object)
+    assert_equal(u.atoms.elements, expected)
+
+
+mol2_fake = """\
+@<TRIPOS>MOLECULE
+FXA101_1
+49 0 0 0 0
+SMALL
+USER_CHARGES
+
+
+@<TRIPOS>ATOM
+  1 H1       0.0000     1.0000    10.0000 H.spc  1 XXXX   5.0000
+  2 H2       0.0000     1.0000    10.0000 H.t3p  1 XXXX   5.0000
+  3 H3       0.0000     1.0000    10.0000 H.xyz  1 XXXX   5.0000
+  4 C1       0.0000     1.0000    10.0000 C.1  1 XXXX   5.0000
+  5 C2       0.0000     1.0000    10.0000 C.2  1 XXXX   5.0000
+  5 C3       0.0000     1.0000    10.0000 C.3  1 XXXX   5.0000
+  6 C4       0.0000     1.0000    10.0000 C.ar  1 XXXX   5.0000
+  7 C5       0.0000     1.0000    10.0000 C.cat  1 XXXX   5.0000
+  8 C6       0.0000     1.0000    10.0000 C.xyz  1 XXXX   5.0000
+  9 N1       0.0000     1.0000    10.0000 N.1  1 XXXX   5.0000
+  10 N2       0.0000     1.0000    10.0000 N.2  1 XXXX   5.0000
+  11 N3       0.0000     1.0000    10.0000 N.3  1 XXXX   5.0000
+  12 N4       0.0000     1.0000    10.0000 N.ar  1 XXXX   5.0000
+  13 O1       0.0000     1.0000    10.0000 O.2  1 XXXX   5.0000
+  14 O2       0.0000     1.0000    10.0000 O.3  1 XXXX   5.0000
+  15 O3       0.0000     1.0000    10.0000 O.co2  1 XXXX   5.0000
+  16 O4       0.0000     1.0000    10.0000 O.spc  1 XXXX   5.0000
+  16 O5       0.0000     1.0000    10.0000 O.t3p  1 XXXX   5.0000
+  17 S1       0.0000     1.0000    10.0000 S.3  1 XXXX   5.0000
+  18 S2       0.0000     1.0000    10.0000 S.2  1 XXXX   5.0000
+  19 S3       0.0000     1.0000    10.0000 S.O  1 XXXX   5.0000
+  20 S4       0.0000     1.0000    10.0000 S.O2  1 XXXX   5.0000
+  21 S5       0.0000     1.0000    10.0000 S.o  1 XXXX   5.0000
+  22 S6       0.0000     1.0000    10.0000 S.o2  1 XXXX   5.0000
+  23 P1       0.0000     1.0000    10.0000 P.3  1 XXXX   5.0000
+  24 Cr1       0.0000     1.0000    10.0000 Cr.th  1 XXXX   5.0000
+  25 Cr2       0.0000     1.0000    10.0000 Cr.oh  1 XXXX   5.0000
+  26 Co1       0.0000     1.0000    10.0000 Co.oh  1 XXXX   5.0000
+"""
+
+def test_all_elements():
+    with pytest.warns(UserWarning, match='Unknown elements found') as record:
+        u = mda.Universe(StringIO(mol2_fake), format='MOL2')
+
+    expected = ["H"] * 2 + [""] + ["C"] * 5 + [""] + ["N"] * 4 + ["O"] * 5 + \
+        ["S"] * 6 + ["P"] + ["Cr"] * 2 + ["Co"]
+    expected = np.array(expected, dtype=object)
+    assert_equal(u.atoms.elements, expected)