def from_qc_json(cls, qc_json) -> "ReadInput":
"""
Given a QC JSON object, extracts the topology, atoms and coords of the molecule.
#TODO we need to be absle to read mapped smiles for this to work with stereochem and aromaticity
"""
topology = nx.Graph()
atoms = []
for i, atom in enumerate(qc_json.symbols):
atoms.append(
Atom(
atomic_number=Element().number(atom),
atom_index=i,
atom_name=f"{atom}{i}",
))
topology.add_node(i)
for bond in qc_json.connectivity:
topology.add_edge(*bond[:2])
coords = np.array(qc_json.geometry).reshape(
(len(atoms), 3)) * BOHR_TO_ANGS
atoms = atoms or None
return cls(name=None, rdkit_mol=None, coords=coords)
def _read_pdb(self):
"""
Internal pdb reader. Only called when RDKit failed to read the pdb.
Extracts the topology, atoms and coords of the molecule.
"""
coords = []
self.topology = nx.Graph()
atoms = []
atom_count = 0
print('called!')
with open(self.mol_input) as pdb:
for line in pdb:
if 'ATOM' in line or 'HETATM' in line:
print('reading!')
# start collecting the atom class info
atomic_symbol = str(line[76:78])
atomic_symbol = re.sub('[0-9]+', '', atomic_symbol)
atomic_symbol = atomic_symbol.strip()
atom_name = str(line.split()[2])
# If the element column is missing from the pdb, extract the atomic_symbol from the atom name.
if not atomic_symbol:
atomic_symbol = str(line.split()[2])[:-1]
atomic_symbol = re.sub('[0-9]+', '', atomic_symbol)
atomic_number = Element().number(atomic_symbol)
# Now instance the qube atom
qube_atom = Atom(atomic_number, atom_count, atom_name)
atoms.append(qube_atom)
# Also add the atom number as the node in the graph
self.topology.add_node(atom_count)
atom_count += 1
coords.append([
float(line[30:38]),
float(line[38:46]),
float(line[46:54])
])
if 'CONECT' in line:
atom_index = int(line.split()[1]) - 1
# Search the connectivity section and add all edges to the graph corresponding to the bonds.
for i in range(2, len(line.split())):
if int(line.split()[i]) != 0:
bonded_index = int(line.split()[i]) - 1
self.topology.add_edge(atom_index, bonded_index)
atoms[atom_index].add_bond(bonded_index)
atoms[bonded_index].add_bond(atom_index)
# put the object back into the correct place
self.coords = np.array(coords)
self.atoms = atoms or None
def _read_pdb_protein(self):
"""
:return:
"""
with open(self.mol_input, 'r') as pdb:
lines = pdb.readlines()
coords = []
atoms = []
self.topology = nx.Graph()
self.Residues = []
self.pdb_names = []
# atom counter used for graph node generation
atom_count = 0
for line in lines:
if 'ATOM' in line or 'HETATM' in line:
atomic_symbol = str(line[76:78])
atomic_symbol = re.sub('[0-9]+', '', atomic_symbol).strip()
# If the element column is missing from the pdb, extract the atomic_symbol from the atom name.
if not atomic_symbol:
atomic_symbol = str(line.split()[2])
atomic_symbol = re.sub('[0-9]+', '', atomic_symbol)
# now make sure we have a valid element
if atomic_symbol.lower() != 'cl' and atomic_symbol.lower() != 'br':
atomic_symbol = atomic_symbol[0]
atom_name = f'{atomic_symbol}{atom_count}'
qube_atom = Atom(Element().number(atomic_symbol), atom_count, atom_name)
atoms.append(qube_atom)
self.pdb_names.append(str(line.split()[2]))
# also get the residue order from the pdb file so we can rewrite the file
self.Residues.append(str(line.split()[3]))
# Also add the atom number as the node in the graph
self.topology.add_node(atom_count)
atom_count += 1
coords.append([float(line[30:38]), float(line[38:46]), float(line[46:54])])
elif 'CONECT' in line:
conect_terms = line.split()
for atom in conect_terms[2:]:
if int(atom):
self.topology.add_edge(int(conect_terms[1]) - 1, int(atom) - 1)
self.atoms = atoms
self.coords = np.array(coords)
self.residues = [res for res, group in groupby(self.Residues)]
def _read_qc_json(self):
"""
Given a QC JSON object, extracts the topology, atoms and coords of the molecule.
"""
self.topology = nx.Graph()
atoms = []
for i, atom in enumerate(self.mol_input.symbols):
atoms.append(Atom(atomic_number=Element().number(atom), atom_index=i, atom_name=f'{atom}{i}'))
self.topology.add_node(i)
for bond in self.mol_input.connectivity:
self.topology.add_edge(*bond[:2])
self.coords = np.array(self.mol_input.geometry).reshape((len(atoms), 3)) * constants.BOHR_TO_ANGS
self.atoms = atoms or None
def from_pdb(cls, file_name: str, name: Optional[str] = None):
"""
Read the protein input pdb file.
:return:
"""
with open(file_name, "r") as pdb:
lines = pdb.readlines()
coords = []
atoms = []
bonds = []
Residues = []
pdb_names = []
# atom counter used for graph node generation
atom_count = 0
for line in lines:
if "ATOM" in line or "HETATM" in line:
atomic_symbol = str(line[76:78])
atomic_symbol = re.sub("[0-9]+", "", atomic_symbol).strip()
# If the element column is missing from the pdb, extract the atomic_symbol from the atom name.
if not atomic_symbol:
atomic_symbol = str(line.split()[2])
atomic_symbol = re.sub("[0-9]+", "", atomic_symbol)
# now make sure we have a valid element
if atomic_symbol.lower() != "cl" and atomic_symbol.lower(
) != "br":
atomic_symbol = atomic_symbol[0]
atom_name = f"{atomic_symbol}{atom_count}"
# TODO should we use a protein pdb package for this?
qube_atom = Atom(
atomic_number=Element().number(atomic_symbol),
atom_index=atom_count,
atom_name=atom_name,
formal_charge=0,
aromatic=False,
)
atoms.append(qube_atom)
pdb_names.append(str(line.split()[2]))
# also get the residue order from the pdb file so we can rewrite the file
Residues.append(str(line.split()[3]))
atom_count += 1
coords.append([
float(line[30:38]),
float(line[38:46]),
float(line[46:54])
])
elif "CONECT" in line:
conect_terms = line.split()
for atom in conect_terms[2:]:
if int(atom):
bond = Bond(
atom1_index=int(conect_terms[1]) - 1,
atom2_index=int(atom) - 1,
bond_order=1,
aromatic=False,
)
bonds.append(bond)
coords = np.array(coords)
residues = [res for res, group in groupby(Residues)]
if name is None:
name = Path(file_name).stem
return cls(
atoms=atoms,
bonds=bonds,
coords=coords,
pdb_names=pdb_names,
residues=residues,
name=name,
)
def _read_mol2(self):
"""
Internal mol2 reader. Only called when RDKit failed to read the mol2.
Extracts the topology, atoms and coords of the molecule.
"""
coords = []
self.topology = nx.Graph()
atoms = []
atom_count = 0
with open(self.mol_input, 'r') as mol2:
atom_flag = False
bond_flag = False
for line in mol2:
if '@<TRIPOS>ATOM' in line:
atom_flag = True
continue
elif '@<TRIPOS>BOND' in line:
atom_flag = False
bond_flag = True
continue
elif '@<TRIPOS>SUBSTRUCTURE' in line:
bond_flag = False
continue
if atom_flag:
# Add the molecule information
atomic_symbol = line.split()[1][:2]
atomic_symbol = re.sub('[0-9]+', '', atomic_symbol)
atomic_symbol = atomic_symbol.strip().title()
atomic_number = Element().number(atomic_symbol)
coords.append([float(line.split()[2]), float(line.split()[3]), float(line.split()[4])])
# Collect the atom names
atom_name = str(line.split()[1])
# Add the nodes to the topology object
self.topology.add_node(atom_count)
atom_count += 1
# Get the atom types
atom_type = line.split()[5]
atom_type = atom_type.replace(".", "")
# Make the qube_atom
qube_atom = Atom(atomic_number, atom_count, atom_name)
qube_atom.atom_type = atom_type
atoms.append(qube_atom)
if bond_flag:
# Add edges to the topology network
atom_index, bonded_index = int(line.split()[1]) - 1, int(line.split()[2]) - 1
self.topology.add_edge(atom_index, bonded_index)
atoms[atom_index].add_bond(bonded_index)
atoms[bonded_index].add_bond(atom_index)
# put the object back into the correct place
self.coords = np.array(coords)
self.atoms = atoms or None