class PDBTrajectoryFile(object):
"""Interface for reading and writing Protein Data Bank (PDB) files
Parameters
----------
filename : str
The filename to open. A path to a file on disk.
mode : {'r', 'w'}
The mode in which to open the file, either 'r' for read or 'w' for write.
force_overwrite : bool
If opened in write mode, and a file by the name of `filename` already
exists on disk, should we overwrite it?
Attributes
----------
positions : np.ndarray, shape=(n_frames, n_atoms, 3)
topology : mdtraj.Topology
closed : bool
Notes
-----
When writing pdb files, mdtraj follows the PDB3.0 standard as closely as
possible. During *reading* however, we try to be more lenient. For instance,
we will parse common nonstandard atom names during reading, and convert them
into the standard names. The replacement table used by mdtraj is at
{mdtraj_source}/formats/pdb/data/pdbNames.xml.
See Also
--------
mdtraj.load_pdb : High-level wrapper that returns a ``md.Trajectory``
"""
distance_unit = 'angstroms'
_residueNameReplacements = {}
_atomNameReplacements = {}
_chain_names = [chr(ord('A') + i) for i in range(26)]
def __init__(self, filename, mode='r', force_overwrite=True):
self._open = False
self._file = None
self._topology = None
self._positions = None
self._mode = mode
self._last_topology = None
if mode == 'r':
PDBTrajectoryFile._loadNameReplacementTables()
if _is_url(filename):
self._file = urlopen(filename)
if filename.lower().endswith('.gz'):
if six.PY3:
self._file = gzip.GzipFile(fileobj=self._file)
else:
self._file = gzip.GzipFile(fileobj=six.StringIO(
self._file.read()))
if six.PY3:
self._file = six.StringIO(self._file.read().decode('utf-8'))
else:
if filename.lower().endswith('.gz'):
self._file = gzip.open(filename, 'r')
self._file = six.StringIO(self._file.read().decode('utf-8'))
else:
self._file = open(filename, 'r')
self._read_models()
elif mode == 'w':
self._header_written = False
self._footer_written = False
if os.path.exists(filename) and not force_overwrite:
raise IOError('"%s" already exists' % filename)
self._file = open(filename, 'w')
else:
raise ValueError("invalid mode: %s" % mode)
self._open = True
def write(self, positions, topology, modelIndex=None, unitcell_lengths=None,
unitcell_angles=None, bfactors=None):
"""Write a PDB file to disk
Parameters
----------
positions : array_like
The list of atomic positions to write.
topology : mdtraj.Topology
The Topology defining the model to write.
modelIndex : {int, None}
If not None, the model will be surrounded by MODEL/ENDMDL records
with this index
unitcell_lengths : {tuple, None}
Lengths of the three unit cell vectors, or None for a non-periodic system
unitcell_angles : {tuple, None}
Angles between the three unit cell vectors, or None for a non-periodic system
bfactors : array_like, default=None, shape=(n_atoms,)
Save bfactors with pdb file. Should contain a single number for
each atom in the topology
"""
if not self._mode == 'w':
raise ValueError('file not opened for writing')
if not self._header_written:
self._write_header(unitcell_lengths, unitcell_angles)
self._header_written = True
if ilen(topology.atoms) != len(positions):
raise ValueError('The number of positions must match the number of atoms')
if np.any(np.isnan(positions)):
raise ValueError('Particle position is NaN')
if np.any(np.isinf(positions)):
raise ValueError('Particle position is infinite')
self._last_topology = topology # Hack to save the topology of the last frame written, allows us to output CONECT entries in write_footer()
if bfactors is None:
bfactors = ['{0:5.2f}'.format(0.0)] * len(positions)
else:
if (np.max(bfactors) >= 100) or (np.min(bfactors) <= -10):
raise ValueError("bfactors must be in (-10, 100)")
bfactors = ['{0:5.2f}'.format(b) for b in bfactors]
atomIndex = 1
posIndex = 0
if modelIndex is not None:
print("MODEL %4d" % modelIndex, file=self._file)
for (chainIndex, chain) in enumerate(topology.chains):
chainName = self._chain_names[chainIndex % len(self._chain_names)]
residues = list(chain.residues)
for (resIndex, res) in enumerate(residues):
if len(res.name) > 3:
resName = res.name[:3]
else:
resName = res.name
for atom in res.atoms:
if len(atom.name) < 4 and atom.name[:1].isalpha() and (atom.element is None or len(atom.element.symbol) < 2):
atomName = ' '+atom.name
elif len(atom.name) > 4:
atomName = atom.name[:4]
else:
atomName = atom.name
coords = positions[posIndex]
if atom.element is not None:
symbol = atom.element.symbol
else:
symbol = ' '
line = "ATOM %5d %-4s %3s %s%4d %s%s%s 1.00 %s %2s " % (
atomIndex % 100000, atomName, resName, chainName,
(res.resSeq) % 10000, _format_83(coords[0]),
_format_83(coords[1]), _format_83(coords[2]),
bfactors[posIndex], symbol)
assert len(line) == 80, 'Fixed width overflow detected'
print(line, file=self._file)
posIndex += 1
atomIndex += 1
if resIndex == len(residues)-1:
print("TER %5d %3s %s%4d" % (atomIndex, resName, chainName, res.resSeq), file=self._file)
atomIndex += 1
if modelIndex is not None:
print("ENDMDL", file=self._file)
def _write_header(self, unitcell_lengths, unitcell_angles, write_metadata=True):
"""Write out the header for a PDB file.
Parameters
----------
unitcell_lengths : {tuple, None}
The lengths of the three unitcell vectors, ``a``, ``b``, ``c``
unitcell_angles : {tuple, None}
The angles between the three unitcell vectors, ``alpha``,
``beta``, ``gamma``
"""
if not self._mode == 'w':
raise ValueError('file not opened for writing')
if unitcell_lengths is None and unitcell_angles is None:
return
if unitcell_lengths is not None and unitcell_angles is not None:
if not len(unitcell_lengths) == 3:
raise ValueError('unitcell_lengths must be length 3')
if not len(unitcell_angles) == 3:
raise ValueError('unitcell_angles must be length 3')
else:
raise ValueError('either unitcell_lengths and unitcell_angles'
'should both be spefied, or neither')
box = list(unitcell_lengths) + list(unitcell_angles)
assert len(box) == 6
if write_metadata:
print("REMARK 1 CREATED WITH MDTraj %s, %s" % (version.version, str(date.today())), file=self._file)
print("CRYST1%9.3f%9.3f%9.3f%7.2f%7.2f%7.2f P 1 1 " % tuple(box), file=self._file)
def _write_footer(self):
if not self._mode == 'w':
raise ValueError('file not opened for writing')
# Identify bonds that should be listed as CONECT records.
standardResidues = ['ALA', 'ASN', 'CYS', 'GLU', 'HIS', 'LEU', 'MET', 'PRO', 'THR', 'TYR',
'ARG', 'ASP', 'GLN', 'GLY', 'ILE', 'LYS', 'PHE', 'SER', 'TRP', 'VAL',
'A', 'G', 'C', 'U', 'I', 'DA', 'DG', 'DC', 'DT', 'DI', 'HOH']
conectBonds = []
if self._last_topology is not None:
for atom1, atom2 in self._last_topology.bonds:
if atom1.residue.name not in standardResidues or atom2.residue.name not in standardResidues:
conectBonds.append((atom1, atom2))
elif atom1.name == 'SG' and atom2.name == 'SG' and atom1.residue.name == 'CYS' and atom2.residue.name == 'CYS':
conectBonds.append((atom1, atom2))
if len(conectBonds) > 0:
# Work out the index used in the PDB file for each atom.
atomIndex = {}
nextAtomIndex = 0
prevChain = None
for chain in self._last_topology.chains:
for atom in chain.atoms:
if atom.residue.chain != prevChain:
nextAtomIndex += 1
prevChain = atom.residue.chain
atomIndex[atom] = nextAtomIndex
nextAtomIndex += 1
# Record which other atoms each atom is bonded to.
atomBonds = {}
for atom1, atom2 in conectBonds:
index1 = atomIndex[atom1]
index2 = atomIndex[atom2]
if index1 not in atomBonds:
atomBonds[index1] = []
if index2 not in atomBonds:
atomBonds[index2] = []
atomBonds[index1].append(index2)
atomBonds[index2].append(index1)
# Write the CONECT records.
for index1 in sorted(atomBonds):
bonded = atomBonds[index1]
while len(bonded) > 4:
print("CONECT%5d%5d%5d%5d" % (index1, bonded[0], bonded[1], bonded[2]), file=self._file)
del bonded[:4]
line = "CONECT%5d" % index1
for index2 in bonded:
line = "%s%5d" % (line, index2)
print(line, file=self._file)
print("END", file=self._file)
self._footer_written = True
@classmethod
def set_chain_names(cls, values):
"""Set the cycle of chain names used when writing PDB files
When writing PDB files, PDBTrajectoryFile translates each chain's
index into a name -- the name is what's written in the file. By
default, chains are named with the letters A-Z.
Parameters
----------
values : list
A list of chacters (strings of length 1) that the PDB writer will
cycle through to choose chain names.
"""
for item in values:
if not isinstance(item, six.string_types) and len(item) == 1:
raise TypeError('Names must be a single character string')
cls._chain_names = values
@property
def positions(self):
"""The cartesian coordinates of all of the atoms in each frame. Available when a file is opened in mode='r'
"""
return self._positions
@property
def topology(self):
"""The topology from this PDB file. Available when a file is opened in mode='r'
"""
return self._topology
@property
def unitcell_lengths(self):
"The unitcell lengths (3-tuple) in this PDB file. May be None"
return self._unitcell_lengths
@property
def unitcell_angles(self):
"The unitcell angles (3-tuple) in this PDB file. May be None"
return self._unitcell_angles
@property
def closed(self):
"Whether the file is closed"
return not self._open
def close(self):
"Close the PDB file"
if self._mode == 'w' and not self._footer_written:
self._write_footer()
if self._open:
self._file.close()
self._open = False
def _read_models(self):
if not self._mode == 'r':
raise ValueError('file not opened for reading')
self._topology = Topology()
pdb = PdbStructure(self._file, load_all_models=True)
atomByNumber = {}
for chain in pdb.iter_chains():
c = self._topology.add_chain()
for residue in chain.iter_residues():
resName = residue.get_name()
if resName in PDBTrajectoryFile._residueNameReplacements:
resName = PDBTrajectoryFile._residueNameReplacements[resName]
r = self._topology.add_residue(resName, c, residue.number)
if resName in PDBTrajectoryFile._atomNameReplacements:
atomReplacements = PDBTrajectoryFile._atomNameReplacements[resName]
else:
atomReplacements = {}
for atom in residue.atoms:
atomName = atom.get_name()
if atomName in atomReplacements:
atomName = atomReplacements[atomName]
atomName = atomName.strip()
element = atom.element
if element is None:
element = self._guess_element(atomName, residue)
newAtom = self._topology.add_atom(atomName, element, r, serial=atom.serial_number)
atomByNumber[atom.serial_number] = newAtom
# load all of the positions (from every model)
_positions = []
for model in pdb.iter_models(use_all_models=True):
coords = []
for chain in model.iter_chains():
for residue in chain.iter_residues():
for atom in residue.atoms:
coords.append(atom.get_position())
_positions.append(coords)
if not all(len(f) == len(_positions[0]) for f in _positions):
raise ValueError('PDB Error: All MODELs must contain the same number of ATOMs')
self._positions = np.array(_positions)
## The atom positions read from the PDB file
self._unitcell_lengths = pdb.get_unit_cell_lengths()
self._unitcell_angles = pdb.get_unit_cell_angles()
self._topology.create_standard_bonds()
self._topology.create_disulfide_bonds(self.positions[0])
# Add bonds based on CONECT records.
connectBonds = []
for connect in pdb.models[0].connects:
i = connect[0]
for j in connect[1:]:
if i in atomByNumber and j in atomByNumber:
connectBonds.append((atomByNumber[i], atomByNumber[j]))
if len(connectBonds) > 0:
# Only add bonds that don't already exist.
existingBonds = set(self._topology.bonds)
for bond in connectBonds:
if bond not in existingBonds and (bond[1], bond[0]) not in existingBonds:
self._topology.add_bond(bond[0], bond[1])
existingBonds.add(bond)
@staticmethod
def _loadNameReplacementTables():
"""Load the list of atom and residue name replacements."""
if len(PDBTrajectoryFile._residueNameReplacements) == 0:
tree = etree.parse(os.path.join(os.path.dirname(__file__), 'data', 'pdbNames.xml'))
allResidues = {}
proteinResidues = {}
nucleicAcidResidues = {}
for residue in tree.getroot().findall('Residue'):
name = residue.attrib['name']
if name == 'All':
PDBTrajectoryFile._parseResidueAtoms(residue, allResidues)
elif name == 'Protein':
PDBTrajectoryFile._parseResidueAtoms(residue, proteinResidues)
elif name == 'Nucleic':
PDBTrajectoryFile._parseResidueAtoms(residue, nucleicAcidResidues)
for atom in allResidues:
proteinResidues[atom] = allResidues[atom]
nucleicAcidResidues[atom] = allResidues[atom]
for residue in tree.getroot().findall('Residue'):
name = residue.attrib['name']
for id in residue.attrib:
if id == 'name' or id.startswith('alt'):
PDBTrajectoryFile._residueNameReplacements[residue.attrib[id]] = name
if 'type' not in residue.attrib:
atoms = copy(allResidues)
elif residue.attrib['type'] == 'Protein':
atoms = copy(proteinResidues)
elif residue.attrib['type'] == 'Nucleic':
atoms = copy(nucleicAcidResidues)
else:
atoms = copy(allResidues)
PDBTrajectoryFile._parseResidueAtoms(residue, atoms)
PDBTrajectoryFile._atomNameReplacements[name] = atoms
def _guess_element(self, atom_name, residue):
"Try to guess the element name"
upper = atom_name.upper()
if upper.startswith('CL'):
element = elem.chlorine
elif upper.startswith('NA'):
element = elem.sodium
elif upper.startswith('MG'):
element = elem.magnesium
elif upper.startswith('BE'):
element = elem.beryllium
elif upper.startswith('LI'):
element = elem.lithium
elif upper.startswith('K'):
element = elem.potassium
elif upper.startswith('ZN'):
element = elem.zinc
elif len(residue) == 1 and upper.startswith('CA'):
element = elem.calcium
# TJL has edited this. There are a few issues here. First,
# parsing for the element is non-trivial, so I do my best
# below. Second, there is additional parsing code in
# pdbstructure.py, and I am unsure why it doesn't get used
# here...
elif len(residue) > 1 and upper.startswith('CE'):
element = elem.carbon # (probably) not Celenium...
elif len(residue) > 1 and upper.startswith('CD'):
element = elem.carbon # (probably) not Cadmium...
elif residue.name in ['TRP', 'ARG', 'GLN', 'HIS'] and upper.startswith('NE'):
element = elem.nitrogen # (probably) not Neon...
elif residue.name in ['ASN'] and upper.startswith('ND'):
element = elem.nitrogen # (probably) not ND...
elif residue.name == 'CYS' and upper.startswith('SG'):
element = elem.sulfur # (probably) not SG...
else:
try:
element = elem.get_by_symbol(atom_name[0])
except KeyError:
try:
symbol = atom_name[0:2].strip().rstrip("AB0123456789").lstrip("0123456789")
element = elem.get_by_symbol(symbol)
except KeyError:
element = None
return element
@staticmethod
def _parseResidueAtoms(residue, map):
for atom in residue.findall('Atom'):
name = atom.attrib['name']
for id in atom.attrib:
map[atom.attrib[id]] = name
def __del__(self):
self.close()
def __enter__(self):
return self
def __exit__(self, *exc_info):
self.close()
def __len__(self):
"Number of frames in the file"
if str(self._mode) != 'r':
raise NotImplementedError('len() only available in mode="r" currently')
if not self._open:
raise ValueError('I/O operation on closed file')
return len(self._positions)
class PDBTrajectoryFile(object):
"""Interface for reading and writing Protein Data Bank (PDB) files
Parameters
----------
filename : str
The filename to open. A path to a file on disk.
mode : {'r', 'w'}
The mode in which to open the file, either 'r' for read or 'w' for write.
force_overwrite : bool
If opened in write mode, and a file by the name of `filename` already
exists on disk, should we overwrite it?
Attributes
----------
positions : np.ndarray, shape=(n_frames, n_atoms, 3)
topology : mdtraj.Topology
closed : bool
Notes
-----
When writing pdb files, mdtraj follows the PDB3.0 standard as closely as
possible. During *reading* however, we try to be more lenient. For instance,
we will parse common nonstandard atom names during reading, and convert them
into the standard names. The replacement table used by mdtraj is at
{mdtraj_source}/formats/pdb/data/pdbNames.xml.
See Also
--------
mdtraj.load_pdb : High-level wrapper that returns a ``md.Trajectory``
"""
distance_unit = 'angstroms'
_residueNameReplacements = {}
_atomNameReplacements = {}
_chain_names = [chr(ord('A') + i) for i in range(26)]
def __init__(self, filename, mode='r', force_overwrite=True):
self._open = False
self._file = None
self._topology = None
self._positions = None
self._mode = mode
self._last_topology = None
if mode == 'r':
PDBTrajectoryFile._loadNameReplacementTables()
if _is_url(filename):
self._file = urlopen(filename)
if filename.lower().endswith('.gz'):
if six.PY3:
self._file = gzip.GzipFile(fileobj=self._file)
else:
self._file = gzip.GzipFile(
fileobj=six.StringIO(self._file.read()))
if six.PY3:
self._file = six.StringIO(
self._file.read().decode('utf-8'))
else:
if filename.lower().endswith('.gz'):
self._file = gzip.open(filename, 'r')
self._file = six.StringIO(
self._file.read().decode('utf-8'))
else:
self._file = open(filename, 'r')
self._read_models()
elif mode == 'w':
self._header_written = False
self._footer_written = False
if os.path.exists(filename) and not force_overwrite:
raise IOError('"%s" already exists' % filename)
self._file = open(filename, 'w')
else:
raise ValueError("invalid mode: %s" % mode)
self._open = True
def write(self,
positions,
topology,
modelIndex=None,
unitcell_lengths=None,
unitcell_angles=None,
bfactors=None):
"""Write a PDB file to disk
Parameters
----------
positions : array_like
The list of atomic positions to write.
topology : mdtraj.Topology
The Topology defining the model to write.
modelIndex : {int, None}
If not None, the model will be surrounded by MODEL/ENDMDL records
with this index
unitcell_lengths : {tuple, None}
Lengths of the three unit cell vectors, or None for a non-periodic system
unitcell_angles : {tuple, None}
Angles between the three unit cell vectors, or None for a non-periodic system
bfactors : array_like, default=None, shape=(n_atoms,)
Save bfactors with pdb file. Should contain a single number for
each atom in the topology
"""
if not self._mode == 'w':
raise ValueError('file not opened for writing')
if not self._header_written:
self._write_header(unitcell_lengths, unitcell_angles)
self._header_written = True
if ilen(topology.atoms) != len(positions):
raise ValueError(
'The number of positions must match the number of atoms')
if np.any(np.isnan(positions)):
raise ValueError('Particle position is NaN')
if np.any(np.isinf(positions)):
raise ValueError('Particle position is infinite')
self._last_topology = topology # Hack to save the topology of the last frame written, allows us to output CONECT entries in write_footer()
if bfactors is None:
bfactors = ['{0:5.2f}'.format(0.0)] * len(positions)
else:
if (np.max(bfactors) >= 100) or (np.min(bfactors) <= -10):
raise ValueError("bfactors must be in (-10, 100)")
bfactors = ['{0:5.2f}'.format(b) for b in bfactors]
atomIndex = 1
posIndex = 0
if modelIndex is not None:
print("MODEL %4d" % modelIndex, file=self._file)
for (chainIndex, chain) in enumerate(topology.chains):
chainName = self._chain_names[chainIndex % len(self._chain_names)]
residues = list(chain.residues)
for (resIndex, res) in enumerate(residues):
if len(res.name) > 3:
resName = res.name[:3]
else:
resName = res.name
for atom in res.atoms:
if len(atom.name) < 4 and atom.name[:1].isalpha() and (
atom.element is None
or len(atom.element.symbol) < 2):
atomName = ' ' + atom.name
elif len(atom.name) > 4:
atomName = atom.name[:4]
else:
atomName = atom.name
coords = positions[posIndex]
if atom.element is not None:
symbol = atom.element.symbol
else:
symbol = ' '
line = "ATOM %5d %-4s %3s %s%4d %s%s%s 1.00 %s %2s " % (
atomIndex % 100000, atomName, resName, chainName,
(res.resSeq) % 10000, _format_83(coords[0]),
_format_83(coords[1]), _format_83(
coords[2]), bfactors[posIndex], symbol)
assert len(line) == 80, 'Fixed width overflow detected'
print(line, file=self._file)
posIndex += 1
atomIndex += 1
if resIndex == len(residues) - 1:
print("TER %5d %3s %s%4d" %
(atomIndex, resName, chainName, res.resSeq),
file=self._file)
atomIndex += 1
if modelIndex is not None:
print("ENDMDL", file=self._file)
def _write_header(self,
unitcell_lengths,
unitcell_angles,
write_metadata=True):
"""Write out the header for a PDB file.
Parameters
----------
unitcell_lengths : {tuple, None}
The lengths of the three unitcell vectors, ``a``, ``b``, ``c``
unitcell_angles : {tuple, None}
The angles between the three unitcell vectors, ``alpha``,
``beta``, ``gamma``
"""
if not self._mode == 'w':
raise ValueError('file not opened for writing')
if unitcell_lengths is None and unitcell_angles is None:
return
if unitcell_lengths is not None and unitcell_angles is not None:
if not len(unitcell_lengths) == 3:
raise ValueError('unitcell_lengths must be length 3')
if not len(unitcell_angles) == 3:
raise ValueError('unitcell_angles must be length 3')
else:
raise ValueError('either unitcell_lengths and unitcell_angles'
'should both be spefied, or neither')
box = list(unitcell_lengths) + list(unitcell_angles)
assert len(box) == 6
if write_metadata:
print("REMARK 1 CREATED WITH MDTraj %s, %s" %
(version.version, str(date.today())),
file=self._file)
print("CRYST1%9.3f%9.3f%9.3f%7.2f%7.2f%7.2f P 1 1 " %
tuple(box),
file=self._file)
def _write_footer(self):
if not self._mode == 'w':
raise ValueError('file not opened for writing')
# Identify bonds that should be listed as CONECT records.
standardResidues = [
'ALA', 'ASN', 'CYS', 'GLU', 'HIS', 'LEU', 'MET', 'PRO', 'THR',
'TYR', 'ARG', 'ASP', 'GLN', 'GLY', 'ILE', 'LYS', 'PHE', 'SER',
'TRP', 'VAL', 'A', 'G', 'C', 'U', 'I', 'DA', 'DG', 'DC', 'DT',
'DI', 'HOH'
]
conectBonds = []
if self._last_topology is not None:
for atom1, atom2 in self._last_topology.bonds:
if atom1.residue.name not in standardResidues or atom2.residue.name not in standardResidues:
conectBonds.append((atom1, atom2))
elif atom1.name == 'SG' and atom2.name == 'SG' and atom1.residue.name == 'CYS' and atom2.residue.name == 'CYS':
conectBonds.append((atom1, atom2))
if len(conectBonds) > 0:
# Work out the index used in the PDB file for each atom.
atomIndex = {}
nextAtomIndex = 0
prevChain = None
for chain in self._last_topology.chains:
for atom in chain.atoms:
if atom.residue.chain != prevChain:
nextAtomIndex += 1
prevChain = atom.residue.chain
atomIndex[atom] = nextAtomIndex
nextAtomIndex += 1
# Record which other atoms each atom is bonded to.
atomBonds = {}
for atom1, atom2 in conectBonds:
index1 = atomIndex[atom1]
index2 = atomIndex[atom2]
if index1 not in atomBonds:
atomBonds[index1] = []
if index2 not in atomBonds:
atomBonds[index2] = []
atomBonds[index1].append(index2)
atomBonds[index2].append(index1)
# Write the CONECT records.
for index1 in sorted(atomBonds):
bonded = atomBonds[index1]
while len(bonded) > 4:
print("CONECT%5d%5d%5d%5d" %
(index1, bonded[0], bonded[1], bonded[2]),
file=self._file)
del bonded[:4]
line = "CONECT%5d" % index1
for index2 in bonded:
line = "%s%5d" % (line, index2)
print(line, file=self._file)
print("END", file=self._file)
self._footer_written = True
@classmethod
def set_chain_names(cls, values):
"""Set the cycle of chain names used when writing PDB files
When writing PDB files, PDBTrajectoryFile translates each chain's
index into a name -- the name is what's written in the file. By
default, chains are named with the letters A-Z.
Parameters
----------
values : list
A list of chacters (strings of length 1) that the PDB writer will
cycle through to choose chain names.
"""
for item in values:
if not isinstance(item, six.string_types) and len(item) == 1:
raise TypeError('Names must be a single character string')
cls._chain_names = values
@property
def positions(self):
"""The cartesian coordinates of all of the atoms in each frame. Available when a file is opened in mode='r'
"""
return self._positions
@property
def topology(self):
"""The topology from this PDB file. Available when a file is opened in mode='r'
"""
return self._topology
@property
def unitcell_lengths(self):
"The unitcell lengths (3-tuple) in this PDB file. May be None"
return self._unitcell_lengths
@property
def unitcell_angles(self):
"The unitcell angles (3-tuple) in this PDB file. May be None"
return self._unitcell_angles
@property
def closed(self):
"Whether the file is closed"
return not self._open
def close(self):
"Close the PDB file"
if self._mode == 'w' and not self._footer_written:
self._write_footer()
if self._open:
self._file.close()
self._open = False
def _read_models(self):
if not self._mode == 'r':
raise ValueError('file not opened for reading')
self._topology = Topology()
pdb = PdbStructure(self._file, load_all_models=True)
atomByNumber = {}
for chain in pdb.iter_chains():
c = self._topology.add_chain()
for residue in chain.iter_residues():
resName = residue.get_name()
if resName in PDBTrajectoryFile._residueNameReplacements:
resName = PDBTrajectoryFile._residueNameReplacements[
resName]
r = self._topology.add_residue(resName, c, residue.number)
if resName in PDBTrajectoryFile._atomNameReplacements:
atomReplacements = PDBTrajectoryFile._atomNameReplacements[
resName]
else:
atomReplacements = {}
for atom in residue.atoms:
atomName = atom.get_name()
if atomName in atomReplacements:
atomName = atomReplacements[atomName]
atomName = atomName.strip()
element = atom.element
if element is None:
element = self._guess_element(atomName, residue)
newAtom = self._topology.add_atom(
atomName, element, r, serial=atom.serial_number)
atomByNumber[atom.serial_number] = newAtom
# load all of the positions (from every model)
_positions = []
for model in pdb.iter_models(use_all_models=True):
coords = []
for chain in model.iter_chains():
for residue in chain.iter_residues():
for atom in residue.atoms:
coords.append(atom.get_position())
_positions.append(coords)
if not all(len(f) == len(_positions[0]) for f in _positions):
raise ValueError(
'PDB Error: All MODELs must contain the same number of ATOMs')
self._positions = np.array(_positions)
## The atom positions read from the PDB file
self._unitcell_lengths = pdb.get_unit_cell_lengths()
self._unitcell_angles = pdb.get_unit_cell_angles()
self._topology.create_standard_bonds()
self._topology.create_disulfide_bonds(self.positions[0])
# Add bonds based on CONECT records.
connectBonds = []
for connect in pdb.models[0].connects:
i = connect[0]
for j in connect[1:]:
if i in atomByNumber and j in atomByNumber:
connectBonds.append((atomByNumber[i], atomByNumber[j]))
if len(connectBonds) > 0:
# Only add bonds that don't already exist.
existingBonds = set(self._topology.bonds)
for bond in connectBonds:
if bond not in existingBonds and (
bond[1], bond[0]) not in existingBonds:
self._topology.add_bond(bond[0], bond[1])
existingBonds.add(bond)
@staticmethod
def _loadNameReplacementTables():
"""Load the list of atom and residue name replacements."""
if len(PDBTrajectoryFile._residueNameReplacements) == 0:
tree = etree.parse(
os.path.join(os.path.dirname(__file__), 'data',
'pdbNames.xml'))
allResidues = {}
proteinResidues = {}
nucleicAcidResidues = {}
for residue in tree.getroot().findall('Residue'):
name = residue.attrib['name']
if name == 'All':
PDBTrajectoryFile._parseResidueAtoms(residue, allResidues)
elif name == 'Protein':
PDBTrajectoryFile._parseResidueAtoms(
residue, proteinResidues)
elif name == 'Nucleic':
PDBTrajectoryFile._parseResidueAtoms(
residue, nucleicAcidResidues)
for atom in allResidues:
proteinResidues[atom] = allResidues[atom]
nucleicAcidResidues[atom] = allResidues[atom]
for residue in tree.getroot().findall('Residue'):
name = residue.attrib['name']
for id in residue.attrib:
if id == 'name' or id.startswith('alt'):
PDBTrajectoryFile._residueNameReplacements[
residue.attrib[id]] = name
if 'type' not in residue.attrib:
atoms = copy(allResidues)
elif residue.attrib['type'] == 'Protein':
atoms = copy(proteinResidues)
elif residue.attrib['type'] == 'Nucleic':
atoms = copy(nucleicAcidResidues)
else:
atoms = copy(allResidues)
PDBTrajectoryFile._parseResidueAtoms(residue, atoms)
PDBTrajectoryFile._atomNameReplacements[name] = atoms
def _guess_element(self, atom_name, residue):
"Try to guess the element name"
upper = atom_name.upper()
if upper.startswith('CL'):
element = elem.chlorine
elif upper.startswith('NA'):
element = elem.sodium
elif upper.startswith('MG'):
element = elem.magnesium
elif upper.startswith('BE'):
element = elem.beryllium
elif upper.startswith('LI'):
element = elem.lithium
elif upper.startswith('K'):
element = elem.potassium
elif upper.startswith('ZN'):
element = elem.zinc
elif len(residue) == 1 and upper.startswith('CA'):
element = elem.calcium
# TJL has edited this. There are a few issues here. First,
# parsing for the element is non-trivial, so I do my best
# below. Second, there is additional parsing code in
# pdbstructure.py, and I am unsure why it doesn't get used
# here...
elif len(residue) > 1 and upper.startswith('CE'):
element = elem.carbon # (probably) not Celenium...
elif len(residue) > 1 and upper.startswith('CD'):
element = elem.carbon # (probably) not Cadmium...
elif residue.name in ['TRP', 'ARG', 'GLN', 'HIS'
] and upper.startswith('NE'):
element = elem.nitrogen # (probably) not Neon...
elif residue.name in ['ASN'] and upper.startswith('ND'):
element = elem.nitrogen # (probably) not ND...
elif residue.name == 'CYS' and upper.startswith('SG'):
element = elem.sulfur # (probably) not SG...
else:
try:
element = elem.get_by_symbol(atom_name[0])
except KeyError:
try:
symbol = atom_name[0:2].strip().rstrip(
"AB0123456789").lstrip("0123456789")
element = elem.get_by_symbol(symbol)
except KeyError:
element = None
return element
@staticmethod
def _parseResidueAtoms(residue, map):
for atom in residue.findall('Atom'):
name = atom.attrib['name']
for id in atom.attrib:
map[atom.attrib[id]] = name
def __del__(self):
self.close()
def __enter__(self):
return self
def __exit__(self, *exc_info):
self.close()
def __len__(self):
"Number of frames in the file"
if str(self._mode) != 'r':
raise NotImplementedError(
'len() only available in mode="r" currently')
if not self._open:
raise ValueError('I/O operation on closed file')
return len(self._positions)
