def from_rdkit(cls, rd_atom: Chem.Atom) -> "Atom":
"""
Build a QUBEKit atom from an rdkit atom instance.
"""
atomic_number = rd_atom.GetAtomicNum()
index = rd_atom.GetIdx()
formal_charge = rd_atom.GetFormalCharge()
aromatic = rd_atom.GetIsAromatic()
bonds = [a.GetIdx() for a in rd_atom.GetNeighbors()]
# check for names in the normal places pdb, mol2 and mol
if rd_atom.HasProp("_Name"):
name = rd_atom.GetProp("_Name")
elif rd_atom.HasProp("_TriposAtomName"):
name = rd_atom.GetProp("_TriposAtomName")
else:
try:
name = rd_atom.GetMonomerInfo().GetName().strip()
except AttributeError:
name = None
# stereochem
if rd_atom.HasProp("_CIPCode"):
stereo_code = rd_atom.GetProp("_CIPCode")
else:
stereo_code = None
return cls(
atomic_number=atomic_number,
atom_index=index,
atom_name=name,
formal_charge=formal_charge,
aromatic=aromatic,
stereochemistry=stereo_code,
bonds=bonds,
)
def _parse_atom(self, atom: Chem.Atom) -> None:
if atom.GetSymbol() == '*':
neighbor = atom.GetNeighbors()[0]
n_name = self._get_pdb_atomname(neighbor)
self.CONNECT.append([n_name, len(self.CONNECT) + 1, 'CONNECT'])
else:
d = self._get_atom_descriptors(atom)
self.ATOM.append(d)
def _get_square(self, first: Chem.Atom, second: Chem.Atom) -> Union[Tuple[int, int], None]:
for third in [neigh for neigh in second.GetNeighbors() if neigh.GetIdx() != first.GetIdx()]:
fourths = self._get_triangles(first, third)
if fourths and len(fourths) > 1:
fourth = [f for f in fourths if f != second.GetIdx()][0]
return third.GetIdx(), fourth
else:
return None
def assess_atom(atom: Chem.Atom, bt: Chem.BondType) -> Tuple[bool, Chem.BondType]:
if atom.GetAtomicNum() > 8:
return True, bt
elif len(atom.GetNeighbors()) <= 2 and atom.GetIsAromatic():
return True, Chem.BondType.SINGLE
elif len(atom_i.GetNeighbors()) <= 3 and not atom.GetIsAromatic():
return True, bt
else:
return False, bt # too bonded already!
def _is_square(self, first: Chem.Atom, second: Chem.Atom) -> bool:
"""
Get bool of whether two atoms share a common over-neighbor. Ie. joining them would make a square.
Direct bond does not count.
:param first:
:param second:
:return:
"""
for third in [neigh for neigh in second.GetNeighbors() if neigh.GetIdx() != first.GetIdx()]:
if self._is_triangle(first, third) is True:
return True
else:
return False
def _get_unseen_neighbors(self,
atom: Chem.Atom,
seen: List[Chem.Atom],
nondummy: bool = True):
neighbors = atom.GetNeighbors()
if nondummy:
neighbors = [
neighbor for neighbor in neighbors
if neighbor.GetSymbol() != '*'
]
seenIdx = {a.GetIdx() for a in seen}
return [
neighbor for neighbor in neighbors
if neighbor.GetIdx() not in seenIdx
]
def get_substructures_from_atom(
atom: Chem.Atom,
max_size: int,
substructure: Set[int] = None) -> Set[FrozenSet[int]]:
"""
Recursively gets all substructures up to a maximum size starting from an atom in a substructure.
:param atom: The atom to start at.
:param max_size: The maximum size of the substructure to fine.
:param substructure: The current substructure that atom is in.
:return: A set of substructures starting at atom where each substructure is a frozenset of indices.
"""
assert max_size >= 1
if substructure is None:
substructure = {atom.GetIdx()}
substructures = {frozenset(substructure)}
if len(substructure) == max_size:
return substructures
# Get neighbors which are not already in the substructure
new_neighbors = [
neighbor for neighbor in atom.GetNeighbors()
if neighbor.GetIdx() not in substructure
]
for neighbor in new_neighbors:
# Define new substructure with neighbor
new_substructure = deepcopy(substructure)
new_substructure.add(neighbor.GetIdx())
# Skip if new substructure has already been considered
if frozenset(new_substructure) in substructures:
continue
# Recursively get substructures including this substructure plus neighbor
new_substructures = get_substructures_from_atom(
neighbor, max_size, new_substructure)
# Add those substructures to current set of substructures
substructures |= new_substructures
return substructures
def _parse_atom(self, atom: Chem.Atom) -> None:
if atom.GetSymbol() == '*':
neighbor = atom.GetNeighbors()[0]
n_name = self._get_PDBInfo_atomname(neighbor)
if self.is_aminoacid() and neighbor.GetSymbol() == 'N':
# atom_name, index, connect_type, connect_name
self.CONNECT.append([n_name, 1, 'LOWER_CONNECT', 'LOWER'])
elif self.is_aminoacid() and neighbor.GetSymbol() == 'C':
# atom_name, index, connect_type, connect_name
self.CONNECT.append([n_name, 2, 'UPPER_CONNECT', 'UPPER'])
elif self.is_aminoacid():
i = max(3, len(self.CONNECT) + 1)
self.CONNECT.append([n_name, i, 'CONNECT'])
else:
self.CONNECT.append([n_name, len(self.CONNECT) + 1, 'CONNECT'])
else:
d = self._get_atom_descriptors(atom)
self.ATOM.append(d)
def assess_atom(atom: Chem.Atom, bt: Chem.BondType) -> Tuple[bool, Chem.BondType]:
"""
True means add, False means delete
:param atom:
:param bt:
:return:
"""
n_neigh = sum([self._is_count_valid(neigh) for neigh in atom.GetNeighbors()])
if atom.GetAtomicNum() > 8:
return True, bt
elif atom.HasProp('DELETE'): # if it is to be deleted it should be fine.
return True, bt
elif n_neigh <= 2 and atom.GetIsAromatic():
return True, Chem.BondType.SINGLE
elif n_neigh <= 3 and not atom.GetIsAromatic():
return True, bt
else:
return False, bt # too bonded already!
def _assess_atom_for_possible_bonding(self, atom: Chem.Atom,
bt: Chem.BondType) -> bool:
"""
Method for add_bond_if_possible
True means add, False means delete
:param atom:
:param bt:
:return:
"""
n_neigh = sum(
[self._is_count_valid(neigh) for neigh in atom.GetNeighbors()])
if atom.GetAtomicNum() > 8:
return True
elif atom.HasProp(
'DELETE'): # if it is to be deleted it should be fine.
return True
elif n_neigh <= 2 and atom.GetIsAromatic():
return True, Chem.BondType.SINGLE
elif n_neigh <= 3 and not atom.GetIsAromatic():
return True
else:
return False # too bonded already!
def _parse_atom(self, atom: Chem.Atom) -> None:
self.log.debug(
f'Parsing {atom.GetSymbol()} at position {atom.GetIdx()}')
if atom.GetSymbol() == '*':
neighbor = atom.GetNeighbors()[0]
n_name = self._get_PDBInfo_atomname(neighbor)
if self.is_aminoacid() and neighbor.GetSymbol() == 'N':
# atom_name, index, connect_type, connect_name
self.CONNECT.append([n_name, 1, 'LOWER_CONNECT', 'LOWER'])
elif self.is_aminoacid() and neighbor.GetSymbol() == 'C':
# atom_name, index, connect_type, connect_name
self.CONNECT.append([n_name, 2, 'UPPER_CONNECT', 'UPPER'])
elif self.is_aminoacid():
i = max(3, len(self.CONNECT) + 1)
self.CONNECT.append([n_name, i, 'CONNECT'])
else:
self.CONNECT.append([n_name, len(self.CONNECT) + 1, 'CONNECT'])
else:
d = self._get_atom_descriptors(
atom) # dict of 'name', 'rtype': 'mtype', 'partial'
self.ATOM.append(d)
formal = atom.GetFormalCharge()
if formal != 0:
self.CHARGE.append([d['name'], formal])
def is_hydrogen_bonding(self, atom: Atom):
return any([
nbr.GetSymbol() in ["N", "O", "S"] for nbr in atom.GetNeighbors()
])