def check_topology(self):
self.section("topology")
self.log("Number of Atoms: %d" % ilen(self.t.topology.atoms))
self.log("Number of Residues: %d" % ilen(self.t.topology.residues))
self.log("Number of Chains: %d" % ilen(self.t.topology.chains))
self.log("Residue names: %s" % str([r.name for r in self.t.topology.residues]))
self.log("Unique atom names: %s" % np.unique([a.name for a in self.t.topology.atoms]))
def check_topology(self):
self.section('topology')
self.log('Number of Atoms: %d' % ilen(self.t.topology.atoms))
self.log('Number of Residues: %d' % ilen(self.t.topology.residues))
self.log('Number of Chains: %d' % ilen(self.t.topology.chains))
self.log('Residue names: %s' % str([r.name for r in self.t.topology.residues]))
self.log('Unique atom names: %s' % np.unique([a.name for a in self.t.topology.atoms]))
def _topology_from_subset(topology, atom_indices):
"""Create a new topology that only contains the supplied indices
Note
----
This really should be a copy constructor (class method) on Topology,
but I want it to work on either the mdtraj topology OR the OpenMM
topology. An inplace version for the topology object we have here
is also available.
Parameters
----------
topology : topology
The base topology
atom_indices : list([int])
The indices of the atoms to keep
"""
newTopology = Topology()
old_atom_to_new_atom = {}
for chain in topology._chains:
newChain = newTopology.add_chain()
for residue in chain._residues:
resSeq = getattr(residue, 'resSeq', None) or residue.index
newResidue = newTopology.add_residue(residue.name, newChain,
resSeq)
for atom in residue._atoms:
if atom.index in atom_indices:
try: # OpenMM Topology objects don't have serial attributes, so we have to check first.
serial = atom.serial
except AttributeError:
serial = None
newAtom = newTopology.add_atom(atom.name, atom.element,
newResidue, serial=serial)
old_atom_to_new_atom[atom] = newAtom
bondsiter = topology.bonds
if not hasattr(bondsiter, '__iter__'):
bondsiter = bondsiter()
for atom1, atom2 in bondsiter:
try:
newTopology.add_bond(old_atom_to_new_atom[atom1],
old_atom_to_new_atom[atom2])
except KeyError:
pass
# we only put bonds into the new topology if both of their partners
# were indexed and thus HAVE a new atom
# Delete empty residues
for chain in newTopology._chains:
chain._residues = [r for r in chain._residues if len(r._atoms) > 0]
# Delete empty chains
newTopology._chains = [c for c in newTopology._chains
if len(c._residues) > 0]
# Re-set the numAtoms and numResidues
newTopology._numAtoms = ilen(newTopology.atoms)
newTopology._numResidues = ilen(newTopology.residues)
return newTopology
def _topology_from_subset(topology, atom_indices):
"""Create a new topology that only contains the supplied indices
Note
----
This really should be a copy constructor (class method) on Topology,
but I want it to work on either the mdtraj topology OR the OpenMM
topology. An inplace version for the topology object we have here
is also available.
Parameters
----------
topology : topology
The base topology
atom_indices : list([int])
The indices of the atoms to keep
"""
newTopology = Topology()
old_atom_to_new_atom = {}
for chain in topology._chains:
newChain = newTopology.add_chain()
for residue in chain._residues:
resSeq = getattr(residue, 'resSeq', None) or residue.index
newResidue = newTopology.add_residue(residue.name, newChain, resSeq)
for atom in residue._atoms:
if atom.index in atom_indices:
try: # OpenMM Topology objects don't have serial attributes, so we have to check first.
serial = atom.serial
except AttributeError:
serial = None
newAtom = newTopology.add_atom(atom.name, atom.element, newResidue, serial=serial)
old_atom_to_new_atom[atom] = newAtom
bondsiter = topology.bonds
if not hasattr(bondsiter, '__iter__'):
bondsiter = bondsiter()
for atom1, atom2 in bondsiter:
try:
newTopology.add_bond(old_atom_to_new_atom[atom1],
old_atom_to_new_atom[atom2])
except KeyError:
pass
# we only put bonds into the new topology if both of their partners
# were indexed and thus HAVE a new atom
# Delete empty residues
for chain in newTopology._chains:
chain._residues = [r for r in chain._residues if len(r._atoms) > 0]
# Delete empty chains
newTopology._chains = [c for c in newTopology._chains if len(c._residues) > 0]
# Re-set the numAtoms and numResidues
newTopology._numAtoms = ilen(newTopology.atoms)
newTopology._numResidues = ilen(newTopology.residues)
return newTopology
def check_topology(self):
self.section("topology")
self.log("Number of Atoms: %d" % ilen(self.topology.atoms))
self.log("Number of Residues: %d" % ilen(self.topology.residues))
self.log("Number of Chains: %d" % ilen(self.topology.chains))
self.log(
"Residues: %s" % ", ".join(["%s (%d atoms)" % (r, ilen(r.atoms)) for r in self.topology.residues])
)
self.log("Unique atom names: %s" % ", ".join(np.unique([a.name for a in self.topology.atoms])))
def check_topology(self):
self.section('topology')
self.log('Number of Atoms: %d' % ilen(self.topology.atoms))
self.log('Number of Residues: %d' % ilen(self.topology.residues))
self.log('Number of Chains: %d' % ilen(self.topology.chains))
self.log('Residues: %s' % ', '.join([
'%s (%d atoms)' % (r, ilen(r.atoms))
for r in self.topology.residues
]))
self.log('Unique atom names: %s' %
', '.join(np.unique([a.name for a in self.topology.atoms])))
def test_load_multiframe():
with open(get_fn('multiframe.pdb')) as f:
pdb = PdbStructure(f)
yield lambda: eq(len(pdb.models), 2)
yield lambda: eq(len(pdb.models[0].chains), 1)
yield lambda: eq(len(pdb.models[0].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[0].iter_atoms()), 22)
yield lambda: eq(len(pdb.models[1].chains), 1)
yield lambda: eq(len(pdb.models[1].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[1].iter_atoms()), 22)
t = load(get_fn('multiframe.pdb'))
yield lambda: eq(t.n_frames, 2)
yield lambda: eq(t.n_atoms, 22)
yield lambda: eq(t.xyz[0], t.xyz[1])
def test_load_multiframe():
with open(get_fn('multiframe.pdb')) as f:
pdb = PdbStructure(f)
yield lambda: eq(len(pdb.models), 2)
yield lambda: eq(len(pdb.models[0].chains), 1)
yield lambda: eq(len(pdb.models[0].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[0].iter_atoms()), 22)
yield lambda: eq(len(pdb.models[1].chains), 1)
yield lambda: eq(len(pdb.models[1].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[1].iter_atoms()), 22)
t = load(get_fn('multiframe.pdb'))
yield lambda: eq(t.n_frames, 2)
yield lambda: eq(t.n_atoms, 22)
yield lambda: eq(t.xyz[0], t.xyz[1])
def test_legacy_hdf1():
t0 = load(fn, chunk=1)
t1 = load(fn, chunk=10)
t2 = load(fn, chunk=100)
yield lambda: eq(t0.xyz, t1.xyz)
yield lambda: eq(t0.xyz, t2.xyz)
yield lambda: t0.topology == load(nat).topology
yield lambda: eq(ilen(t0.topology.bonds), 14)
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 test_2EQQ_0():
# this is an nmr structure with 20 models
t = load(get_fn('2EQQ.pdb'))
yield lambda: eq(t.n_frames, 20)
yield lambda: eq(t.n_atoms, 423)
yield lambda: eq(ilen(t.top.residues), 28)
def test_2EQQ_0():
# this is an nmr structure with 20 models
t = load(get_fn('2EQQ.pdb'))
yield lambda: eq(t.n_frames, 20)
yield lambda: eq(t.n_atoms, 423)
yield lambda: eq(ilen(t.top.residues), 28)
def extract(item, atom_indices='all', copy_if_all=True, check=True):
if check:
digest_item(item, 'mdtraj.Topology')
atom_indices = digest_atom_indices(atom_indices)
if atom_indices is 'all':
if copy_if_all:
from copy import deepcopy
tmp_item = deepcopy(item)
else:
tmp_item = item
else:
from mdtraj.core.topology import Topology
from mdtraj.utils import ilen
atom_indices_to_be_kept = set(atom_indices)
newTopology = Topology()
old_atom_to_new_atom = {}
for chain in item._chains:
newChain = newTopology.add_chain()
for group in chain._groups:
resSeq = getattr(group, 'resSeq', None) or group.index
newResidue = newTopology.add_group(group.name, newChain,
resSeq, group.segment_id)
for atom in group._atoms:
if atom.index in atom_indices_to_be_kept:
try: # OpenMM Topology objects don't have serial attributes, so we have to check first.
serial = atom.serial
except AttributeError:
serial = None
newAtom = newTopology.add_atom(atom.name,
atom.element,
newResidue,
serial=serial)
old_atom_to_new_atom[atom] = newAtom
bondsiter = item.bonds
if not hasattr(bondsiter, '__iter__'):
bondsiter = bondsiter()
for bond in bondsiter:
try:
atom1, atom2 = bond
newTopology.add_bond(old_atom_to_new_atom[atom1],
old_atom_to_new_atom[atom2],
type=bond.type,
order=bond.order)
except KeyError:
pass
# we only put bonds into the new topology if both of their partners
# were indexed and thus HAVE a new atom
# Delete empty groups
newTopology._groups = [
r for r in newTopology._groups if len(r._atoms) > 0
]
for chain in newTopology._chains:
chain._groups = [r for r in chain._groups if len(r._atoms) > 0]
# Delete empty chains
newTopology._chains = [
c for c in newTopology._chains if len(c._groups) > 0
]
# Re-set the numAtoms and numResidues
newTopology._numAtoms = ilen(newTopology.atoms)
newTopology._numResidues = ilen(newTopology.groups)
tmp_item = newTopology
return tmp_item
def write(self, positions, topology, frame_ind=None, pcv_ind=None, unitcell_lengths=None,
unitcell_angles=None):
"""Write a PDB file to disk using plumed2 PCV format
Parameters
----------
positions : array_like
The list of atomic positions to write.
topology : mdtraj.Topology
The Topology defining the model to write.
frame_ind : {int, None}
If not None, the index of frames will be surrounded by REMARK X=? and END
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
"""
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()
posIndex = 0
if frame_ind is not None:
print("REMARK X=%d" % frame_ind, 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 %1s%4d %s%s%s %5.2f %5.2f %-4s%-2s " % (
pcv_ind.atomInd[posIndex] % 100000, atomName, resName, chainName,
(res.resSeq) % 10000, _format_83(coords[0]),
_format_83(coords[1]), _format_83(coords[2]),
pcv_ind.alignPLU[posIndex], pcv_ind.rmsPLU[posIndex], atom.segment_id[:4], symbol[-2:])
assert len(line) == 80, 'Fixed width overflow detected'
print(line, file=self._file)
posIndex += 1
if resIndex == len(residues) - 1:
print("END", file=self._file)
def n_bonds(self):
"""Number of bonds in which the atom participates."""
# TODO: this info could be cached.
return ilen(bond for bond in self.residue.chain.topology.bonds
if self in bond)
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 n_bonds(self):
"""Number of bonds in which the atom participates."""
# TODO: this info could be cached.
return ilen(bond for bond in self.residue.chain.topology.bonds
if self in bond)