def _get_distance_matrix(self, combo: Chem.Mol, A: Union[Chem.Mol, np.ndarray], B: Union[Chem.Mol, np.ndarray]) -> np.ndarray:
"""
Called by ``_find_closest`` and ``_determine_mergers_novel_ringcore_pair`` in collapse ring (for expansion).
"""
# TODO move to base once made.
# input type
if isinstance(A, Chem.Mol):
mol_A = A
A_idxs = np.arange(mol_A.GetNumAtoms())
else:
mol_A = None
A_idxs = np.array(A)
if isinstance(B, Chem.Mol):
mol_B = B
B_idxs = np.arange(mol_B.GetNumAtoms()) + mol_A.GetNumAtoms()
else:
mol_B = None
B_idxs = np.array(B)
# make matrix
distance_matrix = Chem.Get3DDistanceMatrix(combo)
length = combo.GetNumAtoms()
# nan fill the self values
self._nan_submatrix(distance_matrix, A_idxs)
self._nan_submatrix(distance_matrix, B_idxs)
return distance_matrix
def mol_to_mutliconformer_file(rdkit_mol: Chem.Mol,
file_name: str) -> None:
"""
Write the rdkit molecule to a multi conformer file.
Args:
rdkit_mol:
A complete Chem.Mol instance of a molecule.
file_name:
Name of the file to be created.
"""
file_path = Path(file_name)
# get the file block writer
if file_path.suffix == ".pdb":
writer = Chem.MolToPDBBlock
elif file_path.suffix == ".mol" or file_path.suffix == ".sdf":
writer = Chem.MolToMolBlock
elif file_path.suffix == ".xyz":
writer = Chem.MolToXYZBlock
else:
raise FileTypeError(
f"The file type {file_path.suffix} is not supported please chose from xyz, pdb, mol or sdf."
)
with open(file_name, "w") as out:
for i in range(rdkit_mol.GetNumConformers()):
out.write(writer(rdkit_mol, confId=i))
def conformer_energy(molecule: Chem.Mol, conformer_id: int = 0,
forcefield: str = "UFF") -> float:
""" Get the energy of a conformer in a molecule using the universal forcefield (UFF) forcefield
or the merck molecular forcefield (MMFF) forcefield.
Parameters
----------
molecule : rdkit.Chem.Mol
The molecule which energy will be calculated.
forcefield : {"UFF", "MMFF"}, optional.
The forcefield that will be used to calculate the energy
(default="UFF").
Returns
-------
float
The energy of the molecule.
"""
if molecule.GetNumConformers() == 0:
raise NoConformersError("Molecule must have at least one conformer")
if forcefield == "UFF":
ff = AllChem.UFFGetMoleculeForceField(molecule, confId=conformer_id)
elif forcefield == "MMFF":
props = AllChem.MMFFGetMoleculeProperties(molecule)
ff = AllChem.MMFFGetMoleculeForceField(molecule, props, confId=conformer_id)
return ff.CalcEnergy()
def generate_conformers(rdkit_mol: Chem.Mol,
conformer_no: int) -> List[np.ndarray]:
"""
Generate a set of conformers for the molecule including the input conformer.
Args:
rdkit_mol:
A complete Chem.Mol instance of a molecule.
conformer_no:
The number of conformers made for the molecule
return:
A list of conformer position arrays
"""
AllChem.EmbedMultipleConfs(
rdkit_mol,
numConfs=conformer_no,
randomSeed=1,
clearConfs=False,
useBasicKnowledge=True,
pruneRmsThresh=1,
enforceChirality=True,
)
positions = rdkit_mol.GetConformers()
return [conformer.GetPositions() for conformer in positions]
def guess_origins(self,
mol: Chem.Mol = None,
hits: Optional[List[Chem.Mol]] = None):
"""
Given a positioned mol guess its origins...
:param mol:
:return:
"""
if hits is None:
hits = self.hits
mappings = []
for h, hit in enumerate(hits):
hname = hit.GetProp('_Name')
for hi, mi in self.get_positional_mapping(hit, mol).items():
atom = mol.GetAtomWithIdx(mi)
if atom.HasProp('_Novel') and atom.GetBoolProp(
'_Novel') == True:
continue # flagged to avoid.
elif atom.HasProp(
'_Origin') and atom.GetProp('_Origin') != 'none':
origin = json.loads(atom.GetProp('_Origin'))
else:
origin = []
origin.append(f'{hname}.{hi}')
atom.SetProp('_Origin', json.dumps(origin))
def max_from_mol(self, mol: Chem.Mol = None):
if mol is None:
mol = self.positioned_mol
return [
atom.GetDoubleProp('_Max') if atom.HasProp('_Max') else 0
for atom in mol.GetAtoms()
]
def _pre_fragment_pairs(self, scaffold: Chem.Mol, fragmentanda: Chem.Mol, A2B_mapping: Optional = None) \
-> Dict[int, List[Dict]]:
"""
Returns
{4: [{'idx': 5,
'type': rdkit.Chem.rdchem.BondType.SINGLE,
'idx_F': 5,
'idx_S': 1}], ...}
which is slight more than {5: [{'idx': 4, 'type': rdkit.Chem.rdchem.BondType.SINGLE}], ... from categories
idx_F: fragmentanda index
idx_S: scaffold index
required for self.merge, the key is the index of anchoring atom.
Calls get_positional_mapping and _categorise.
:param scaffold: mol to be added to.
:param fragmentanda: mol to be fragmented
:param A2B_mapping: see ``get_positional_mapping``
:return:
"""
# get A2B mapping
if A2B_mapping is None:
A2B_mapping = self.get_positional_mapping(scaffold, fragmentanda)
get_key = lambda d, v: list(d.keys())[list(d.values()).index(v)]
if len(A2B_mapping) == 0:
raise ConnectionError('No overlap!')
# store alternative atom symbols.
for si, fi in A2B_mapping.items():
sa = scaffold.GetAtomWithIdx(si)
sn = sa.GetSymbol()
fn = fragmentanda.GetAtomWithIdx(fi).GetSymbol()
if sn != fn:
sa.SetProp('_AltSymbol', fn)
# prepare.
uniques = set(range(fragmentanda.GetNumAtoms())) - set(
A2B_mapping.values())
categories = self._categorise(fragmentanda, uniques)
pairs = categories['pairs']
for p in pairs: # pairs:Dict[List[Dict]]
for pp in pairs[p]:
pp['idx_F'] = pp['idx'] # less ambiguous: fragmentanda index
pp['idx_S'] = get_key(A2B_mapping, pp['idx']) # scaffold index
return pairs
def convert_to_graph(mol: Chem.Mol, scaffold_ids: t.Tuple[int],
anchors: t.Dict[int, int], hba_ids: t.Tuple[int],
hbd_ids: t.Tuple[int]) -> nx.Graph:
"""
Convert `Chem.Mol` object to `nx.Graph` object
Args:
mol (Chem.Mol):
The molecule object to be converted
scaffold_ids (t.Tuple[int]):
The atom that corresponds to scaffolds
anchors (t.Dict[int, int]):
The mapping from atom in the molecule to atom in scaffold where it
is attached to
hba_ids (t.Tuple[int]):
The atoms corresponding to hydrogen acceptors
hbd_ids (t.Tuple[int]):
The atoms corresponding to hydrogen donnors
Returns:
nx.Graph:
The graph converted
"""
# Initialize graph
graph = nx.Graph()
# Add nodes
nodes = range(mol.GetNumAtoms())
graph.add_nodes_from(nodes)
# Add edges
bond: Chem.Bond
edges = [(bond.GetBeginAtomIdx(), bond.GetEndAtomIdx())
for bond in mol.GetBonds()]
graph.add_edges_from(edges)
# Attach properties to nodes
for node_id in nodes:
atom_i: Chem.Atom = mol.GetAtomWithIdx(node_id)
graph.nodes[node_id]['symbol'] = atom_i.GetSymbol()
for node_id in anchors:
graph.nodes[node_id]['anchor'] = anchors[node_id]
for node_id in hba_ids:
graph.nodes[node_id]['is_hba'] = True
for node_id in hbd_ids:
graph.nodes[node_id]['is_hbd'] = True
for node_id in scaffold_ids:
graph.nodes[node_id]['is_scaffold'] = True
return graph
def merge(self, scaffold: Chem.Mol, fragmentanda: Chem.Mol,
anchor_index: int, attachment_details: List[Dict]) -> Chem.Mol:
for detail in attachment_details:
attachment_index = detail['idx_F'] # fragmentanda attachment_index
scaffold_attachment_index = detail['idx_S']
bond_type = detail['type']
f = Chem.FragmentOnBonds(fragmentanda, [
fragmentanda.GetBondBetweenAtoms(anchor_index,
attachment_index).GetIdx()
],
addDummies=False)
frag_split = []
fragmols = Chem.GetMolFrags(f,
asMols=True,
fragsMolAtomMapping=frag_split,
sanitizeFrags=False)
if self._debug_draw:
print(frag_split)
# Get the fragment of interest.
ii = 0
for mol_N, indices in enumerate(frag_split):
if anchor_index in indices:
break
ii += len(indices)
else:
raise Exception
frag = fragmols[mol_N]
frag_anchor_index = indices.index(anchor_index)
if self._debug_draw:
self.draw_nicely(frag)
combo = Chem.RWMol(rdmolops.CombineMols(scaffold, frag))
scaffold_anchor_index = frag_anchor_index + scaffold.GetNumAtoms()
if self._debug_draw:
print(scaffold_anchor_index, scaffold_attachment_index,
anchor_index, scaffold.GetNumAtoms())
self.draw_nicely(combo)
combo.AddBond(scaffold_anchor_index, scaffold_attachment_index,
bond_type)
Chem.SanitizeMol(
combo,
sanitizeOps=Chem.rdmolops.SanitizeFlags.SANITIZE_ADJUSTHS +
Chem.rdmolops.SanitizeFlags.SANITIZE_SETAROMATICITY,
catchErrors=True)
if self._debug_draw:
self.draw_nicely(combo)
scaffold = combo
return scaffold
def _prevent_weird_rings(self, mol: Chem.Mol):
if not isinstance(mol, Chem.RWMol):
mol = Chem.RWMol(mol)
ringatoms = self._get_ring_info(mol) #GetRingInfo().AtomRings()
for ring_A, ring_B in itertools.combinations(ringatoms, r=2):
shared = set(ring_A).intersection(set(ring_B))
if len(shared) == 0:
log.debug('This molecule has some separate rings')
pass # separate rings
elif len(shared) == 1:
log.debug('This molecule has a spiro bicycle')
pass # spiro ring.
elif len(shared) == 2:
log.debug('This molecule has a fused ring')
if mol.GetBondBetweenAtoms(*shared) is not None:
pass # indole/naphtalene
small, big = sorted([ring_A, ring_B], key=lambda ring: len(ring))
if len(small) == 4:
log.warning('This molecule has a benzo-azetine–kind-of-thing: expanding to indole')
# Chem.MolFromSmiles('C12CCCCC1CC2')
# benzo-azetine is likely an error: add and extra atom
a, b = set(small).difference(big)
self._place_between(mol, a, b)
elif len(small) == 3:
log.warning('This molecule has a benzo-cyclopropane–kind-of-thing: expanding to indole')
# Chem.MolFromSmiles('C12CCCCC1C2')
# benzo-cyclopronane is actually impossible at this stage.
a = list(set(small).difference(big))[0]
for b in shared:
self._place_between(mol, a, b)
else:
pass # indole and nathalene
elif (len(ring_A), len(ring_B)) == (6, 6):
raise Exception('This is utterly impossible')
else:
print(f'mysterious ring system {len(ring_A)} + {len(ring_B)}')
pass # ????
elif len(shared) < self.atoms_in_bridge_cutoff:
#adamantene/norbornane/tropinone kind of thing
log.warning('This molecule has a bridge: leaving')
pass # ideally check if planar...
else:
log.warning('This molecule has a bridge that will be removed')
mol = self._prevent_bridge_ring(mol, ring_A)
# start from scratch.
return self._prevent_weird_rings(mol)
return mol.GetMol()
def get_pharmacophoric_point(ligand: Chem.Mol, feat_name: str,
atom_indices: Sequence, conformer_index: int,
radius: float,
directionality: bool) -> PharmacophoricPoint:
""" Obtain the coordinates and if specified the direction vector and return a pharmacophoric point.
Parameters
----------
ligand : rdkit.Chem.Mol
A ligand
conformer_index : int
The conformer whose coordinates will be used to obtain the pharmacophoric
points.
radius : float
Lenght of the radius in angstroms of the parmacohporic point.
directionality : bool
Whether to compute the direction vectgor of that point.
Returns
-------
openpharmacophore.PharmacophoricPoint
A pharmacophoric point.
"""
if len(atom_indices) > 1:
# Find the centroid
# Aromatic, hydrophobic, positive or negative feature
coords = PharmacophoricPointExtractor._feature_centroid(
ligand, atom_indices, conformer_index)
# Find direction vector
if directionality:
direction = PharmacophoricPointExtractor._aromatic_direction_vector(
ligand, atom_indices, conformer_index)
else:
direction = None
else:
# Find the centroid
# Donor or acceptor feature
position = ligand.GetConformer(conformer_index).GetAtomPosition(
atom_indices[0])
coords = np.zeros((3, ))
coords[0] = position.x
coords[1] = position.y
coords[2] = position.z
# Find direction vector
if directionality:
direction = PharmacophoricPointExtractor._donor_acceptor_direction_vector(
ligand, atom_indices[0], coords, conformer_index)
else:
direction = None
return PharmacophoricPoint(feat_type=feat_name,
center=puw.quantity(coords, "angstroms"),
radius=puw.quantity(radius, "angstroms"),
direction=direction,
atom_indices=atom_indices)
def CalculateChi0(mol: Chem.Mol) -> float:
"""Calculate molecular connectivity chi index for path order 0."""
deltas = [x.GetDegree() for x in mol.GetAtoms()]
while 0 in deltas:
deltas.remove(0)
deltas = numpy.array(deltas, 'd')
res = sum(numpy.sqrt(1. / deltas))
return res
def CalculateMeanWeiner(mol: Chem.Mol) -> float:
"""Get Mean Weiner index of a molecule.
Or AW.
"""
N = mol.GetNumAtoms()
WeinerNumber = CalculateWeiner(mol)
return 2.0 * WeinerNumber / (N * (N - 1))
def CalculateQuadratic(mol: Chem.Mol) -> float:
"""Get Quadratic index.
Or Qindex.
"""
M = CalculateZagreb1(mol)
N = mol.GetNumAtoms()
return 3 - 2 * N + M / 2.0
def toxicity(self, molecule: Chem.Mol) -> bool:
"""
Checks if a given molecule fails the structural filters.
"""
for (pattern, tolerance) in zip(self.pattern_list, self.tolerance_list):
if len(molecule.GetSubstructMatches(pattern)) > tolerance:
return True
return False
def store(self, combined: Chem.Mol, combined_map: Dict[int, int],
disregarded: List[Chem.Mol]):
combined.SetProp('parts',
json.dumps([m.GetProp('_Name') for m in disregarded]))
self.c_map_options.append(combined_map)
self.c_options.append(combined)
self.c_disregarded_options.append(disregarded)
return None
def _get_ori_i(self, mol: Chem.Mol, include_collapsed=True):
indices = [atom.GetIntProp('_ori_i') for atom in mol.GetAtoms()]
if include_collapsed:
for atom in self._get_collapsed_atoms(mol):
indices.extend(json.loads(atom.GetProp('_ori_is')))
else:
pass
return indices
def find_closest_to_ligand(cls, pdb: Chem.Mol, ligand_resn: str) -> Tuple[Chem.Atom, Chem.Atom]:
"""
Find the closest atom to the ligand
:param pdb: a rdkit Chem object
:param ligand_resn: 3 letter code
:return: tuple of non-ligand atom and ligand atom
"""
ligand = [atom.GetIdx() for atom in pdb.GetAtoms() if atom.GetPDBResidueInfo().GetResidueName() == ligand_resn]
dm = Chem.Get3DDistanceMatrix(pdb)
mini = np.take(dm, ligand, 0)
mini[mini == 0] = np.nan
mini[:, ligand] = np.nan
a, b = np.where(mini == np.nanmin(mini))
lig_atom = pdb.GetAtomWithIdx(ligand[int(a[0])])
nonlig_atom = pdb.GetAtomWithIdx(int(b[0]))
return (nonlig_atom, lig_atom)
def _get_mol_sender_receivers(mol: Chem.Mol) -> Tuple[np.ndarray, np.ndarray]:
"""Get connectivity (messages) info for a data_dict."""
senders, receivers = [], []
for bond in mol.GetBonds():
id1 = bond.GetBeginAtom().GetIdx()
id2 = bond.GetEndAtom().GetIdx()
senders.extend([id1, id2])
receivers.extend([id2, id1])
return np.array(senders), np.array(receivers)
def store_positions(self, mol: Chem.Mol) -> Chem.Mol:
"""
Saves positional data as _x, _y, _z and majorly ``_ori_i``, the original index.
The latter gets used by ``_get_new_index``.
:param mol:
:return:
"""
conf = mol.GetConformer()
name = mol.GetProp('_Name')
for i, atom in enumerate(mol.GetAtoms()):
pos = conf.GetAtomPosition(i)
atom.SetIntProp('_ori_i', i)
atom.SetProp('_ori_name', name)
atom.SetDoubleProp('_x', pos.x)
atom.SetDoubleProp('_y', pos.y)
atom.SetDoubleProp('_z', pos.z)
return mol
def process(mol: Mol, device: torch.device, **kwargs):
n = mol.GetNumAtoms() + 1
graph = DGLGraph()
graph.add_nodes(n)
graph.add_edges(graph.nodes(), graph.nodes())
graph.add_edges(range(1, n), 0)
# graph.add_edges(0, range(1, n))
for e in mol.GetBonds():
u, v = e.GetBeginAtomIdx(), e.GetEndAtomIdx()
graph.add_edge(u + 1, v + 1)
graph.add_edge(v + 1, u + 1)
adj = graph.adjacency_matrix(transpose=False).to_dense()
v, m = feature.mol_feature(mol)
vec = torch.cat([torch.zeros((1, m)), v]).to(device)
return ChebNetData(n, adj, vec)
def apply_to_mol(self, mol: Chem.Mol):
results_dict = asdict(self)
results_dict.update({
"dG_bind": self.dG_bind,
"dG_bind_err": self.dG_bind_err,
})
for field, val in results_dict.items():
field_name = self._convert_field_to_sdf_field(field)
mol.SetProp(field_name, str(val))
def get_conn(cls, mol: Chem.Mol) -> Chem.Atom:
"""
Get connecting atom of mol.
"""
for atom in mol.GetAtoms():
if atom.GetSymbol() == '*':
return atom.GetNeighbors()[0]
else:
raise ValueError('Dummy atom not found')
def mol_to_nx(mol: Chem.Mol):
G = nx.Graph()
for atom in mol.GetAtoms():
G.add_node(
atom.GetIdx(),
atomic_num=atom.GetAtomicNum(),
formal_charge=atom.GetFormalCharge(),
chiral_tag=atom.GetChiralTag(),
hybridization=atom.GetHybridization(),
num_explicit_hs=atom.GetNumExplicitHs(),
is_aromatic=atom.GetIsAromatic(),
)
for bond in mol.GetBonds():
G.add_edge(
bond.GetBeginAtomIdx(), bond.GetEndAtomIdx(), bond_type=bond.GetBondType()
)
return G
def map_query(mol: Chem.Mol, query: Chem.Mol) -> t.Tuple[int]:
"""
Get the set of indices of all atoms in molecule `mol` matching the query
`query`
"""
match = set()
for match_i in mol.GetSubstructMatches(query):
match = match | set(match_i)
match = tuple(match)
return match
def CalculateKappa2(mol: Chem.Mol) -> float:
"""Calculate molecular shape index for two bonded fragments."""
P2 = len(Chem.FindAllPathsOfLengthN(mol, 2))
A = mol.GetNumHeavyAtoms()
denom = P2 + 0.0
if denom:
kappa = (A - 1) * (A - 2)**2 / denom**2
else:
kappa = 0.0
return round(kappa, 3)
def merge_pair(self,
scaffold: Chem.Mol,
fragmentanda: Chem.Mol,
mapping: Optional = None) -> Chem.Mol:
"""
To specify attachments use ``.merge``.
To understand what is going on see ``.categorise``
:param scaffold: mol to be added to.
:param fragmentanda: mol to be fragmented
:param mapping: see ``get_positional_mapping``. Optional in _pre_fragment_pairs
:return:
"""
done_already = []
fp = self._pre_fragment_pairs(scaffold, fragmentanda, mapping)
# confusingly these are hit indexed.
for anchor_index, attachment_details in fp.items():
# anchor index is the fragment-to-added's internal atom that attaches
if anchor_index in done_already:
continue
# fix rings.
uniques = {
atom.GetIdx()
for atom in fragmentanda.GetAtoms()
if 'overlapping' not in atom.GetProp('_Category')
}
team = self._recruit_team(fragmentanda, anchor_index, uniques)
other_attachments = list((team & set(fp.keys())) - {anchor_index})
other_attachment_details = []
for other in other_attachments:
other_attachment_details.append(fp[other])
done_already.append(other)
scaffold = self._merge_part(
scaffold,
fragmentanda,
anchor_index=anchor_index,
attachment_details=attachment_details,
other_attachments=other_attachments,
other_attachment_details=other_attachment_details)
new_name = self._get_combined_name(scaffold, fragmentanda)
scaffold.SetProp('_Name', new_name)
self.keep_copy(scaffold, 'pair_merged')
return scaffold
def CalculateChi1(mol: Chem.Mol):
"""Calculate molecular connectivity chi index for path order 1."""
cc = [x.GetBeginAtom().GetDegree() * x.GetEndAtom().GetDegree() for x in mol.GetBonds()]
if len(cc) == 0:
return 0.0
while 0 in cc:
cc.remove(0)
cc = numpy.array(cc, 'd')
res = sum(numpy.sqrt(1. / cc))
return res
def _CalculateBondNumber(mol: Chem.Mol, bondtype: str = 'SINGLE') -> float:
"""Calculate number of bond of specified type.
:param bondtype: can be SINGLE, DOUBLE, TRIPLE or AROMATIC.
"""
i = 0
for bond in mol.GetBonds():
if bond.GetBondType().name == bondtype:
i += 1
return i
def CalculateMeanRandic(mol: Chem.Mol) -> float:
"""Calculate mean chi1 (Randic) connectivity index."""
cc = [x.GetBeginAtom().GetDegree() * x.GetEndAtom().GetDegree() for x in mol.GetBonds()]
if len(cc) == 0:
return 0.0
while 0 in cc:
cc.remove(0)
cc = numpy.array(cc, 'd')
res = numpy.mean(numpy.sqrt(1. / cc))
return res