def parse(cls, filename, structure=False):
self = cls(filename)
if structure:
obj = Structure()
for _ in range(self.natom):
# fake
obj.add_atom(Atom(), resname='XXX', resnum=0)
obj.box = self.box
obj.coordinates = self.coordinates[0]
obj._coordinates = self.coordinates
return obj
else:
return self
def parse(cls, filename, structure=False):
self = cls(filename)
if structure:
obj = Structure()
for _ in range(self.natom):
# fake
obj.add_atom(Atom(), resname='XXX', resnum=0)
obj.box = self.box
obj.coordinates = self.coordinates[0]
obj._coordinates = self.coordinates
return obj
else:
return self
def parse(filename):
""" Parses a Gromacs GRO file
Parameters
----------
filename : str or file-like
Name of the file or the GRO file object
Returns
-------
struct : :class:`Structure`
The Structure instance instantiated with *just* residues and atoms
populated (with coordinates)
"""
struct = Structure()
if isinstance(filename, string_types):
fileobj = genopen(filename, 'r')
own_handle = True
else:
fileobj = filename
own_handle = False
try:
# Ignore the title line
fileobj.readline()
try:
natom = int(fileobj.readline().strip())
except ValueError:
raise GromacsError('Could not parse %s as GRO file' % filename)
digits = None
for i, line in enumerate(fileobj):
if i == natom: break
try:
resnum = int(line[:5])
resname = line[5:10].strip()
atomname = line[10:15].strip()
elem = element_by_name(atomname)
atomic_number = AtomicNum[elem]
mass = Mass[elem]
atnum = int(line[15:20])
if atomic_number == 0:
atom = ExtraPoint(name=atomname, number=atnum)
else:
atom = Atom(atomic_number=atomic_number, name=atomname,
number=atnum, mass=mass)
if digits is None:
pdeci = line.index('.', 20)
ndeci = line.index('.', pdeci+1)
digits = ndeci - pdeci
atom.xx, atom.xy, atom.xz = (
float(line[20+i*digits:20+(i+1)*digits])*10
for i in range(3)
)
i = 4
wbeg = (pdeci-4)+(5+ndeci)*(i-1)
wend = (pdeci-4)+(5+ndeci)*i
if line[wbeg:wend].strip():
atom.vx, atom.vy, atom.vz = (
float(line[(pdeci-3)+(6+ndeci)*i:
(pdeci-3)+(6+ndeci)*(i+1)])*10
for i in range(3, 6)
)
except (ValueError, IndexError):
raise GromacsError('Could not parse the atom record of '
'GRO file %s' % filename)
struct.add_atom(atom, resname, resnum)
# Get the box from the last line if it's present
if line.strip():
try:
box = [float(x) for x in line.split()]
except ValueError:
raise GromacsError('Could not understand box line of GRO '
'file %s' % filename)
if len(box) == 3:
struct.box = [box[0]*10, box[1]*10, box[2]*10,
90.0, 90.0, 90.0]
elif len(box) == 9:
# Assume we have vectors
leng, ang = box_vectors_to_lengths_and_angles(
[box[0], box[3], box[4]]*u.nanometers,
[box[5], box[1], box[6]]*u.nanometers,
[box[7], box[8], box[2]]*u.nanometers)
a, b, c = leng.value_in_unit(u.angstroms)
alpha, beta, gamma = ang.value_in_unit(u.degrees)
struct.box = [a, b, c, alpha, beta, gamma]
finally:
if own_handle:
fileobj.close()
return struct
def load_topology(topology, system=None):
"""
Creates a :class:`parmed.structure.Structure` instance from an OpenMM
Topology, optionally filling in parameters from a System
Parameters
----------
topology : :class:`simtk.openmm.app.Topology`
The Topology instance with the list of atoms and bonds for this system
system : :class:`simtk.openmm.System` or str, optional
If provided, parameters from this System will be applied to the
Structure. If a string is given, it will be interpreted as the file name
of an XML-serialized System, and it will be deserialized into a System
before used to supply parameters
Returns
-------
struct : :class:`Structure <parmed.structure.Structure>`
The structure from the provided topology
Raises
------
OpenMMWarning if parameters are found that cannot be interpreted or
processed by ParmEd
TypeError if there are any mismatches between the provided topology and
system (e.g., they have different numbers of atoms)
IOError if system is a string and it is not an existing file
Notes
-----
Due to its flexibility with CustomForces, it is entirely possible that the
functional form of the potential will be unknown to ParmEd. This function
will try to use the energy expression to identify supported potential types
that are implemented as CustomForce objects. In particular, quadratic
improper torsions, when recognized, will be extracted.
Other CustomForces, including the CustomNonbondedForce used to implement
NBFIX (off-diagonal L-J modifications) and the 12-6-4 potential, will not be
processed and will result in an unknown functional form
"""
struct = Structure()
atommap = dict()
for c in topology.chains():
chain = c.id
for r in c.residues():
residue = r.name
resid = r.index
for a in r.atoms():
if a.element is None:
atom = ExtraPoint(name=a.name)
else:
atom = Atom(atomic_number=a.element.atomic_number,
name=a.name, mass=a.element.mass)
struct.add_atom(atom, residue, resid, chain)
atommap[a] = atom
for a1, a2 in topology.bonds():
struct.bonds.append(Bond(atommap[a1], atommap[a2]))
vectors = topology.getPeriodicBoxVectors()
if vectors is not None:
leng, ang = box_vectors_to_lengths_and_angles(*vectors)
leng = leng.value_in_unit(u.angstroms)
ang = ang.value_in_unit(u.degrees)
struct.box = [leng[0], leng[1], leng[2], ang[0], ang[1], ang[2]]
if struct.box is not None:
struct.box = create_array(struct.box)
if system is None:
return struct
if isinstance(system, string_types):
with open(system, 'r') as f:
system = mm.XmlSerializer.deserialize(f.read())
# We have a system, try to extract parameters from it
if len(struct.atoms) != system.getNumParticles():
raise TypeError('Topology and System have different numbers of atoms '
'(%d vs. %d)' % (len(struct.atoms), system.getNumParticles()))
processed_forces = set()
ignored_forces = (mm.CMMotionRemover, mm.AndersenThermostat,
mm.MonteCarloBarostat, mm.MonteCarloAnisotropicBarostat,
mm.MonteCarloMembraneBarostat, mm.CustomExternalForce,
mm.GBSAOBCForce, mm.CustomGBForce)
if system.usesPeriodicBoundaryConditions():
vectors = system.getDefaultPeriodicBoxVectors()
leng, ang = box_vectors_to_lengths_and_angles(*vectors)
leng = leng.value_in_unit(u.angstroms)
ang = ang.value_in_unit(u.degrees)
struct.box = create_array(
[leng[0], leng[1], leng[2], ang[0], ang[1], ang[2]]
)
else:
struct.box = None
for force in system.getForces():
if isinstance(force, mm.HarmonicBondForce):
if mm.HarmonicBondForce in processed_forces:
# Try to process this HarmonicBondForce as a Urey-Bradley term
_process_urey_bradley(struct, force)
else:
_process_bond(struct, force)
elif isinstance(force, mm.HarmonicAngleForce):
_process_angle(struct, force)
elif isinstance(force, mm.PeriodicTorsionForce):
_process_dihedral(struct, force)
elif isinstance(force, mm.RBTorsionForce):
_process_rbtorsion(struct, force)
elif isinstance(force, mm.CustomTorsionForce):
if not _process_improper(struct, force):
struct.unknown_functional = True
warnings.warn('Unknown functional form of CustomTorsionForce',
OpenMMWarning)
elif isinstance(force, mm.CMAPTorsionForce):
_process_cmap(struct, force)
elif isinstance(force, mm.NonbondedForce):
_process_nonbonded(struct, force)
elif isinstance(force, ignored_forces):
continue
else:
struct.unknown_functional = True
warnings.warn('Unsupported Force type %s' % type(force).__name__,
OpenMMWarning)
processed_forces.add(type(force))
return struct
def parse(filename, skip_bonds=False):
""" Parses a Gromacs GRO file
Parameters
----------
filename : str or file-like
Name of the file or the GRO file object
skip_bonds : bool, optional
If True, skip trying to assign bonds. This can save substantial time
when parsing large files with non-standard residue names. However,
no bonds are assigned. This is OK if, for instance, the GRO file is
being parsed simply for its coordinates. This will also reduce the
accuracy of assigned atomic numbers for typical ions. Default is
False.
Returns
-------
struct : :class:`Structure`
The Structure instance instantiated with *just* residues and atoms
populated (with coordinates)
"""
struct = Structure()
if isinstance(filename, string_types):
fileobj = genopen(filename, 'r')
own_handle = True
else:
fileobj = filename
own_handle = False
try:
# Ignore the title line
fileobj.readline()
try:
natom = int(fileobj.readline().strip())
except ValueError:
raise GromacsError('Could not parse %s as GRO file' % filename)
line_parser = _AtomLineParser()
for i, line in enumerate(fileobj):
if i == natom: break
try:
atom, resname, resnum = line_parser.read(line)
except (ValueError, IndexError):
raise GromacsError('Could not parse the atom record of '
'GRO file %s' % filename)
struct.add_atom(atom, resname, resnum)
else:
# If no box exists, the break did not hit, so line still
# contains the last atom (which cannot be interpreted as a box).
# This wipes out line (IFF fileobj reached the line)
line = fileobj.readline()
if i+1 != natom:
raise GromacsError('Truncated GRO file. Found %d of %d '
'atoms' % (i+1, natom))
# Get the box from the last line if it's present
if line.strip():
try:
box = [float(x) for x in line.split()]
except ValueError:
raise GromacsError('Could not understand box line of GRO '
'file %s' % filename)
if len(box) == 3:
struct.box = [box[0]*10, box[1]*10, box[2]*10,
90.0, 90.0, 90.0]
elif len(box) == 9:
# Assume we have vectors
leng, ang = box_vectors_to_lengths_and_angles(
[box[0], box[3], box[4]]*u.nanometers,
[box[5], box[1], box[6]]*u.nanometers,
[box[7], box[8], box[2]]*u.nanometers)
a, b, c = leng.value_in_unit(u.angstroms)
alpha, beta, gamma = ang.value_in_unit(u.degrees)
struct.box = [a, b, c, alpha, beta, gamma]
finally:
if own_handle:
fileobj.close()
# Assign bonds (and improved element guesses)
if not skip_bonds:
struct.assign_bonds()
return struct
def parse(filename, structure=False):
""" Parses a mol2 file (or mol3) file
Parameters
----------
filename : str or file-like
Name of the file to parse or file-like object to parse from
structure : bool, optional
If True, the return value is a :class:`Structure` instance. If
False, it is either a :class:`ResidueTemplate` or
:class:`ResidueTemplateContainter` instance, depending on whether
there is one or more than one residue defined in it. Default is
False
Returns
-------
molecule : :class:`Structure`, :class:`ResidueTemplate`, or
:class:`ResidueTemplateContainer`
The molecule defined by this mol2 file
Raises
------
Mol2Error
If the file format is not recognized or non-numeric values are
present where integers or floating point numbers are expected. Also
raises Mol2Error if you try to parse a mol2 file that has multiple
@<MOLECULE> entries with ``structure=True``.
"""
if isinstance(filename, string_types):
f = genopen(filename, 'r')
own_handle = True
else:
f = filename
own_handle = False
rescont = ResidueTemplateContainer()
struct = Structure()
restemp = ResidueTemplate()
mol_info = []
multires_structure = False
try:
section = None
last_residue = None
headtail = 'head'
molecule_number = 0
for line in f:
if line.startswith('#'): continue
if not line.strip() and section is None: continue
if line.startswith('@<TRIPOS>'):
section = line[9:].strip()
if section == 'MOLECULE' and (restemp.atoms or rescont):
if structure:
raise Mol2Error('Cannot convert MOL2 with multiple '
'@<MOLECULE>s to a Structure')
# Set the residue name from the MOL2 title if the
# molecule had only 1 residue and it was given a name in
# the title
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
multires_structure = False
rescont.append(restemp)
restemp = ResidueTemplate()
struct = Structure()
last_residue = None
molecule_number += 1
mol_info = []
continue
if section is None:
raise Mol2Error('Bad mol2 file format')
if section == 'MOLECULE':
# Section formatted as follows:
# mol_name
# num_atoms [num_bonds [num_substr [num_feat [num_sets]]]]
# mol_type
# charge_type
# [status_bits]
# [mol_comment]
# TODO: Do something with the name.
if len(mol_info) == 0:
mol_info.append(line.strip())
elif len(mol_info) == 1:
mol_info.append([int(x) for x in line.split()])
elif len(mol_info) == 2:
mol_info.append(line.strip())
elif len(mol_info) == 3:
mol_info.append(line.strip())
# Ignore the rest
continue
if section == 'ATOM':
# Section formatted as follows:
# atom_id -- serial number of atom
# atom_name -- name of the atom
# x -- X-coordinate of the atom
# y -- Y-coordinate of the atom
# z -- Z-coordinate of the atom
# atom_type -- type of the atom
# subst_id -- Residue serial number
# subst_name -- Residue name
# charge -- partial atomic charge
# status_bit -- ignored
words = line.split()
id = int(words[0])
name = words[1]
x = float(words[2])
y = float(words[3])
z = float(words[4])
typ = words[5]
try:
resid = int(words[6])
except IndexError:
resid = 0
try:
resname = words[7]
except IndexError:
resname = 'UNK'
if 'NO_CHARGES' not in mol_info:
try:
charge = float(words[8])
except IndexError:
charge = 0
else:
charge = 0
if last_residue is None:
last_residue = (resid, resname)
restemp.name = resname
atom = Atom(name=name, type=typ, number=id, charge=charge)
atom.xx, atom.xy, atom.xz = x, y, z
struct.add_atom(atom, resname, resid)
if last_residue != (resid, resname):
rescont.append(restemp)
restemp = ResidueTemplate()
restemp.name = resname
last_residue = (resid, resname)
multires_structure = True
restemp.add_atom(copy.copy(atom))
continue
if section == 'BOND':
# Section formatted as follows:
# bond_id -- serial number of bond (ignored)
# origin_atom_id -- serial number of first atom in bond
# target_atom_id -- serial number of other atom in bond
# bond_type -- string describing bond type (ignored)
# status_bits -- ignored
words = line.split()
int(words[0]) # Bond serial number... redundant and ignored
a1 = int(words[1])
a2 = int(words[2])
atom1 = struct.atoms.find_original_index(a1)
atom2 = struct.atoms.find_original_index(a2)
struct.bonds.append(Bond(atom1, atom2))
# Now add it to our residue container
# See if it's a head/tail connection
if atom1.residue is not atom2.residue:
if atom1.residue.idx == len(rescont):
res1 = restemp
elif atom1.residue.idx < len(rescont):
res1 = rescont[atom1.residue.idx]
else:
raise Mol2Error('Bad bonding pattern detected')
if atom2.residue.idx == len(rescont):
res2 = restemp
elif atom1.residue.idx < len(rescont):
res2 = rescont[atom2.residue.idx]
else:
raise Mol2Error('Bad bonding pattern detected')
assert res1 is not res2, 'BAD identical residues'
idx1 = atom1.idx - atom1.residue[0].idx
idx2 = atom2.idx - atom2.residue[0].idx
if atom1.residue.idx < atom2.residue.idx:
res1.tail = res1[idx1]
res2.head = res2[idx2]
else:
res1.head = res1[idx1]
res2.tail = res2[idx2]
elif not multires_structure:
restemp.add_bond(a1-1, a2-1)
else:
# Same residue, add the bond
offset = atom1.residue[0].idx
if atom1.residue.idx == len(rescont):
res = restemp
else:
res = rescont[atom1.residue.idx]
res.add_bond(atom1.idx-offset, atom2.idx-offset)
continue
if section == 'CRYSIN':
# Section formatted as follows:
# a -- length of first unit cell vector
# b -- length of second unit cell vector
# c -- length of third unit cell vector
# alpha -- angle b/w b and c
# beta -- angle b/w a and c
# gamma -- angle b/w a and b
# space group -- number of space group (ignored)
# space group setting -- ignored
words = line.split()
box = [float(x) for x in words[:6]]
if len(box) != 6:
raise ValueError('%d box dimensions found; needed 6' %
len(box))
struct.box = copy.copy(box)
rescont.box = copy.copy(box)
continue
if section == 'SUBSTRUCTURE':
# Section formatted as follows:
# subst_id -- residue number
# subst_name -- residue name
# root_atom -- first atom of residue
# subst_type -- ignored (usually 'RESIDUE')
# dict_type -- type of substructure (ignored)
# chain -- chain ID of residue
# sub_type -- type of the chain
# inter_bonds -- # of inter-substructure bonds
# status -- ignored
# comment -- ignored
words = line.split()
if not words: continue
id = int(words[0])
resname = words[1]
try:
chain = words[5]
except IndexError:
chain = ''
# Set the chain ID
for res in struct.residues:
if res.number == id and res.name == resname:
res.chain = chain
continue
# MOL3 sections
if section == 'HEADTAIL':
atname, residx = line.split()
residx = int(residx)
if residx in (0, 1) or residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in head/tail')
for atom in res:
if atom.name == atname:
if headtail == 'head':
res.head = atom
headtail = 'tail'
else:
res.tail = atom
headtail = 'head'
break
else:
if headtail == 'head':
headtail = 'tail'
else:
headtail = 'head'
continue
if section == 'RESIDUECONNECT':
words = line.split()
residx = int(words[0])
if residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in '
'residueconnect')
for a in words[3:]:
if a == '0': continue
for atom in res:
if atom.name == a:
atom.connections.append(atom)
break
else:
raise Mol2Error('Residue connection atom %s not '
'found in residue %d' % (a, residx))
if structure:
return struct
elif len(rescont) > 0:
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
rescont.append(restemp)
return rescont
else:
return restemp
except ValueError as e:
raise Mol2Error('String conversion trouble: %s' % e)
finally:
if own_handle: f.close()
def parse(filename, structure=False):
""" Parses a mol2 file (or mol3) file
Parameters
----------
filename : str or file-like
Name of the file to parse or file-like object to parse from
structure : bool, optional
If True, the return value is a :class:`Structure` instance. If
False, it is either a :class:`ResidueTemplate` or
:class:`ResidueTemplateContainter` instance, depending on whether
there is one or more than one residue defined in it. Default is
False
Returns
-------
molecule : :class:`Structure`, :class:`ResidueTemplate`, or
:class:`ResidueTemplateContainer`
The molecule defined by this mol2 file
Raises
------
Mol2Error
If the file format is not recognized or non-numeric values are
present where integers or floating point numbers are expected. Also
raises Mol2Error if you try to parse a mol2 file that has multiple
@<MOLECULE> entries with ``structure=True``.
"""
if isinstance(filename, string_types):
f = genopen(filename, 'r')
own_handle = True
else:
f = filename
own_handle = False
rescont = ResidueTemplateContainer()
struct = Structure()
restemp = ResidueTemplate()
mol_info = []
multires_structure = False
try:
section = None
last_residue = None
headtail = 'head'
molecule_number = 0
for line in f:
if line.startswith('#'): continue
if not line.strip() and section is None: continue
if line.startswith('@<TRIPOS>'):
section = line[9:].strip()
if section == 'MOLECULE' and (restemp.atoms or rescont):
if structure:
raise Mol2Error('Cannot convert MOL2 with multiple '
'@<MOLECULE>s to a Structure')
# Set the residue name from the MOL2 title if the
# molecule had only 1 residue and it was given a name in
# the title
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
multires_structure = False
rescont.append(restemp)
restemp = ResidueTemplate()
struct = Structure()
last_residue = None
molecule_number += 1
mol_info = []
continue
if section is None:
raise Mol2Error('Bad mol2 file format')
if section == 'MOLECULE':
# Section formatted as follows:
# mol_name
# num_atoms [num_bonds [num_substr [num_feat [num_sets]]]]
# mol_type
# charge_type
# [status_bits]
# [mol_comment]
# TODO: Do something with the name.
if len(mol_info) == 0:
mol_info.append(line.strip())
elif len(mol_info) == 1:
mol_info.append([int(x) for x in line.split()])
elif len(mol_info) == 2:
mol_info.append(line.strip())
elif len(mol_info) == 3:
mol_info.append(line.strip())
# Ignore the rest
continue
if section == 'ATOM':
# Section formatted as follows:
# atom_id -- serial number of atom
# atom_name -- name of the atom
# x -- X-coordinate of the atom
# y -- Y-coordinate of the atom
# z -- Z-coordinate of the atom
# atom_type -- type of the atom
# subst_id -- Residue serial number
# subst_name -- Residue name
# charge -- partial atomic charge
# status_bit -- ignored
words = line.split()
id = int(words[0])
name = words[1]
x = float(words[2])
y = float(words[3])
z = float(words[4])
typ = words[5]
try:
resid = int(words[6])
except IndexError:
resid = 0
try:
resname = words[7]
except IndexError:
resname = 'UNK'
if 'NO_CHARGES' not in mol_info:
try:
charge = float(words[8])
except IndexError:
charge = 0
else:
charge = 0
if last_residue is None:
last_residue = (resid, resname)
restemp.name = resname
atom = Atom(name=name, type=typ, number=id, charge=charge)
atom.xx, atom.xy, atom.xz = x, y, z
struct.add_atom(atom, resname, resid)
if last_residue != (resid, resname):
rescont.append(restemp)
restemp = ResidueTemplate()
restemp.name = resname
last_residue = (resid, resname)
multires_structure = True
try:
restemp.add_atom(copy.copy(atom))
except ValueError:
# Allow mol2 files being parsed as a Structure to have
# duplicate atom names
if not structure:
raise
continue
if section == 'BOND':
# Section formatted as follows:
# bond_id -- serial number of bond (ignored)
# origin_atom_id -- serial number of first atom in bond
# target_atom_id -- serial number of other atom in bond
# bond_type -- string describing bond type (ignored)
# status_bits -- ignored
words = line.split()
int(words[0]) # Bond serial number... redundant and ignored
a1 = int(words[1])
a2 = int(words[2])
atom1 = struct.atoms.find_original_index(a1)
atom2 = struct.atoms.find_original_index(a2)
struct.bonds.append(Bond(atom1, atom2))
# Now add it to our residue container
# See if it's a head/tail connection
if atom1.residue is not atom2.residue:
if atom1.residue.idx == len(rescont):
res1 = restemp
elif atom1.residue.idx < len(rescont):
res1 = rescont[atom1.residue.idx]
assert atom.residue.idx <= len(rescont), 'Bad bond!'
if atom2.residue.idx == len(rescont):
res2 = restemp
elif atom2.residue.idx < len(rescont):
res2 = rescont[atom2.residue.idx]
assert atom.residue.idx <= len(rescont), 'Bad bond!'
assert res1 is not res2, 'BAD identical residues'
idx1 = atom1.idx - atom1.residue[0].idx
idx2 = atom2.idx - atom2.residue[0].idx
if atom1.residue.idx < atom2.residue.idx:
res1.tail = res1[idx1]
res2.head = res2[idx2]
else:
res1.head = res1[idx1]
res2.tail = res2[idx2]
elif not multires_structure:
if not structure:
restemp.add_bond(a1-1, a2-1)
else:
# Same residue, add the bond
offset = atom1.residue[0].idx
if atom1.residue.idx == len(rescont):
res = restemp
else:
res = rescont[atom1.residue.idx]
res.add_bond(atom1.idx-offset, atom2.idx-offset)
continue
if section == 'CRYSIN':
# Section formatted as follows:
# a -- length of first unit cell vector
# b -- length of second unit cell vector
# c -- length of third unit cell vector
# alpha -- angle b/w b and c
# beta -- angle b/w a and c
# gamma -- angle b/w a and b
# space group -- number of space group (ignored)
# space group setting -- ignored
words = line.split()
box = [float(w) for w in words[:6]]
if len(box) != 6:
raise ValueError('%d box dimensions found; needed 6' %
len(box))
struct.box = copy.copy(box)
rescont.box = copy.copy(box)
continue
if section == 'SUBSTRUCTURE':
# Section formatted as follows:
# subst_id -- residue number
# subst_name -- residue name
# root_atom -- first atom of residue
# subst_type -- ignored (usually 'RESIDUE')
# dict_type -- type of substructure (ignored)
# chain -- chain ID of residue
# sub_type -- type of the chain
# inter_bonds -- # of inter-substructure bonds
# status -- ignored
# comment -- ignored
words = line.split()
if not words: continue
id = int(words[0])
resname = words[1]
try:
chain = words[5]
except IndexError:
chain = ''
# Set the chain ID
for res in struct.residues:
if res.number == id and res.name == resname:
res.chain = chain
continue
# MOL3 sections
if section == 'HEADTAIL':
atname, residx = line.split()
residx = int(residx)
if residx in (0, 1) or residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in head/tail')
for atom in res:
if atom.name == atname:
if headtail == 'head':
res.head = atom
headtail = 'tail'
else:
res.tail = atom
headtail = 'head'
break
else:
if headtail == 'head':
headtail = 'tail'
else:
headtail = 'head'
continue
if section == 'RESIDUECONNECT':
words = line.split()
residx = int(words[0])
if residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in '
'residueconnect')
for a in words[3:]:
if a == '0': continue
for atom in res:
if atom.name == a:
res.connections.append(atom)
break
else:
raise Mol2Error('Residue connection atom %s not '
'found in residue %d' % (a, residx))
if structure:
return struct
elif len(rescont) > 0:
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
rescont.append(restemp)
return rescont
else:
return restemp
except ValueError as e:
raise Mol2Error('String conversion trouble: %s' % e)
finally:
if own_handle: f.close()
def parse(filename):
""" Read a PQR file and return a populated `Structure` class
Parameters
----------
filename : str or file-like
Name of the PQR file to read, or a file-like object that can iterate
over the lines of a PQR. Compressed file names can be specified and
are determined by file-name extension (e.g., file.pqr.gz,
file.pqr.bz2)
Returns
-------
structure : :class:`Structure`
The Structure object initialized with all of the information from
the PDB file. No bonds or other topological features are added by
default.
"""
if isinstance(filename, string_types):
own_handle = True
fileobj = genopen(filename, "r")
else:
own_handle = False
fileobj = filename
struct = Structure()
# Add metadata fields
modelno = 1 # For PDB files with multiple MODELs
atomno = 0
coordinates = []
all_coordinates = []
# Support hexadecimal numbering like that printed by VMD
try:
for line in fileobj:
words = line.split()
if words[0] in ("ATOM", "HETATM"):
atomno += 1
if len(words) == 10:
_, num, nam, res, resn, x, y, z, chg, rad = words
chn = ""
elif len(words) >= 11:
_, num, nam, res, chn, resn, x, y, z, chg, rad = (words[i] for i in range(11))
# If the radius is not a float (but rather a letter,
# like the element or something), then the chain might
# be missing. In this case, shift all tokens "back" one
# and empty the chn string
try:
float(rad)
except ValueError:
resn, x, y, z, chg, rad = chn, resn, x, y, z, chg
else:
raise ValueError("Illegal PQR record format: expected " "10 or 11 tokens on the atom line")
x, y, z = float(x), float(y), float(z)
chg, rad = float(chg), float(rad)
resn, num = int(resn), int(num)
elem = element_by_name(nam) # Yuck
atomic_number = AtomicNum[elem]
mass = Mass[elem]
if nam in ("EP", "LP"): # lone pair
atom = ExtraPoint(
atomic_number=atomic_number, name=nam, charge=chg, mass=mass, number=num, solvent_radius=rad
)
else:
atom = Atom(
atomic_number=atomic_number, name=nam, charge=chg, mass=mass, number=num, solvent_radius=rad
)
atom.xx, atom.xy, atom.xz = float(x), float(y), float(z)
if modelno == 1:
struct.add_atom(atom, res, resn, chn)
else:
try:
orig_atom = struct.atoms[atomno - 1]
except IndexError:
raise PDBError("Extra atom in MODEL %d" % modelno)
if orig_atom.residue.name != res.strip() or orig_atom.name != nam.strip():
raise PDBError(
"Atom %d differs in MODEL %d [%s %s "
"vs. %s %s]" % (atomno, modelno, orig_atom.residue.name, orig_atom.name, res, nam)
)
coordinates.extend([atom.xx, atom.xy, atom.xz])
elif words[0] == "ENDMDL":
# End the current model
if len(struct.atoms) == 0:
raise PDBError("MODEL ended before any atoms read in")
modelno += 1
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError("Inconsistent atom numbers in some PDB models")
all_coordinates.append(coordinates)
atomno = 0
coordinates = []
elif words[0] == "MODEL":
if modelno == 1 and len(struct.atoms) == 0:
continue
if len(coordinates) > 0:
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError("Inconsistent atom numbers in " "some PDB models")
warnings.warn("MODEL not explicitly ended", PDBWarning)
all_coordinates.append(coordinates)
coordinates = []
modelno += 1
atomno = 0
elif words[0] == "CRYST1":
a, b, c = (float(w) for w in words[1:4])
try:
A, B, C = (float(w) for w in words[4:7])
except ValueError:
A = B = C = 90.0
struct.box = [a, b, c, A, B, C]
finally:
if own_handle:
fileobj.close()
struct.unchange()
if coordinates:
if len(coordinates) != 3 * len(struct.atoms):
raise PDBError("bad number of atoms in some PQR models")
all_coordinates.append(coordinates)
struct._coordinates = np.array(all_coordinates).reshape((-1, len(struct.atoms), 3))
return struct
def parse(filename):
""" Read a PQR file and return a populated `Structure` class
Parameters
----------
filename : str or file-like
Name of the PQR file to read, or a file-like object that can iterate
over the lines of a PQR. Compressed file names can be specified and
are determined by file-name extension (e.g., file.pqr.gz,
file.pqr.bz2)
Returns
-------
structure : :class:`Structure`
The Structure object initialized with all of the information from
the PDB file. No bonds or other topological features are added by
default.
"""
if isinstance(filename, string_types):
own_handle = True
fileobj = genopen(filename, 'r')
else:
own_handle = False
fileobj = filename
struct = Structure()
# Add metadata fields
modelno = 1 # For PDB files with multiple MODELs
atomno = 0
coordinates = []
all_coordinates = []
# Support hexadecimal numbering like that printed by VMD
try:
for line in fileobj:
words = line.split()
if words[0] in ('ATOM', 'HETATM'):
atomno += 1
if len(words) == 10:
_, num, nam, res, resn, x, y, z, chg, rad = words
chn = ''
elif len(words) >= 11:
_, num, nam, res, chn, resn, x, y, z, chg, rad = (
words[i] for i in range(11))
# If the radius is not a float (but rather a letter,
# like the element or something), then the chain might
# be missing. In this case, shift all tokens "back" one
# and empty the chn string
try:
float(rad)
except ValueError:
resn, x, y, z, chg, rad = chn, resn, x, y, z, chg
else:
raise ValueError('Illegal PQR record format: expected '
'10 or 11 tokens on the atom line')
x, y, z = float(x), float(y), float(z)
chg, rad = float(chg), float(rad)
resn, num = int(resn), int(num)
elem = element_by_name(nam) # Yuck
atomic_number = AtomicNum[elem]
mass = Mass[elem]
if nam in ('EP', 'LP'): # lone pair
atom = ExtraPoint(atomic_number=atomic_number,
name=nam,
charge=chg,
mass=mass,
number=num,
solvent_radius=rad)
else:
atom = Atom(atomic_number=atomic_number,
name=nam,
charge=chg,
mass=mass,
number=num,
solvent_radius=rad)
atom.xx, atom.xy, atom.xz = float(x), float(y), float(z)
if modelno == 1:
struct.add_atom(atom, res, resn, chn)
else:
try:
orig_atom = struct.atoms[atomno - 1]
except IndexError:
raise PDBError('Extra atom in MODEL %d' % modelno)
if (orig_atom.residue.name != res.strip()
or orig_atom.name != nam.strip()):
raise PDBError(
'Atom %d differs in MODEL %d [%s %s '
'vs. %s %s]' %
(atomno, modelno, orig_atom.residue.name,
orig_atom.name, res, nam))
coordinates.extend([atom.xx, atom.xy, atom.xz])
elif words[0] == 'ENDMDL':
# End the current model
if len(struct.atoms) == 0:
raise PDBError('MODEL ended before any atoms read in')
modelno += 1
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError(
'Inconsistent atom numbers in some PDB models')
all_coordinates.append(coordinates)
atomno = 0
coordinates = []
elif words[0] == 'MODEL':
if modelno == 1 and len(struct.atoms) == 0: continue
if len(coordinates) > 0:
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError('Inconsistent atom numbers in '
'some PDB models')
warnings.warn('MODEL not explicitly ended', PDBWarning)
all_coordinates.append(coordinates)
coordinates = []
modelno += 1
atomno = 0
elif words[0] == 'CRYST1':
a, b, c = (float(w) for w in words[1:4])
try:
A, B, C = (float(w) for w in words[4:7])
except ValueError:
A = B = C = 90.0
struct.box = [a, b, c, A, B, C]
finally:
if own_handle: fileobj.close()
struct.unchange()
if coordinates:
if len(coordinates) != 3 * len(struct.atoms):
raise PDBError('bad number of atoms in some PQR models')
all_coordinates.append(coordinates)
struct._coordinates = np.array(all_coordinates).reshape(
(-1, len(struct.atoms), 3))
return struct
def load_topology(topology, system=None, xyz=None, box=None):
"""
Creates a :class:`parmed.structure.Structure` instance from an OpenMM
Topology, optionally filling in parameters from a System
Parameters
----------
topology : :class:`simtk.openmm.app.Topology`
The Topology instance with the list of atoms and bonds for this system
system : :class:`simtk.openmm.System` or str, optional
If provided, parameters from this System will be applied to the
Structure. If a string is given, it will be interpreted as the file name
of an XML-serialized System, and it will be deserialized into a System
before used to supply parameters
xyz : str or array of float
Name of a file containing coordinate information or an array of
coordinates. If file has unit cell information, it also uses that
information unless ``box`` (below) is also specified
box : array of 6 floats
Unit cell dimensions
Returns
-------
struct : :class:`Structure <parmed.structure.Structure>`
The structure from the provided topology
Raises
------
OpenMMWarning if parameters are found that cannot be interpreted or
processed by ParmEd
TypeError if there are any mismatches between the provided topology and
system (e.g., they have different numbers of atoms)
IOError if system is a string and it is not an existing file
Notes
-----
Due to its flexibility with CustomForces, it is entirely possible that the
functional form of the potential will be unknown to ParmEd. This function
will try to use the energy expression to identify supported potential types
that are implemented as CustomForce objects. In particular, quadratic
improper torsions, when recognized, will be extracted.
Other CustomForces, including the CustomNonbondedForce used to implement
NBFIX (off-diagonal L-J modifications) and the 12-6-4 potential, will not be
processed and will result in an unknown functional form
"""
import simtk.openmm as mm
struct = Structure()
atommap = dict()
for c in topology.chains():
chain = c.id
for r in c.residues():
residue = r.name
resid = r.index
for a in r.atoms():
if a.element is None:
atom = ExtraPoint(name=a.name)
else:
atom = Atom(atomic_number=a.element.atomic_number,
name=a.name,
mass=a.element.mass)
struct.add_atom(atom, residue, resid, chain)
atommap[a] = atom
for a1, a2 in topology.bonds():
struct.bonds.append(Bond(atommap[a1], atommap[a2]))
vectors = topology.getPeriodicBoxVectors()
if vectors is not None:
leng, ang = box_vectors_to_lengths_and_angles(*vectors)
leng = leng.value_in_unit(u.angstroms)
ang = ang.value_in_unit(u.degrees)
struct.box = [leng[0], leng[1], leng[2], ang[0], ang[1], ang[2]]
loaded_box = False
if xyz is not None:
if isinstance(xyz, string_types):
xyz = load_file(xyz, skip_bonds=True)
struct.coordinates = xyz.coordinates
if struct.box is not None:
if xyz.box is not None:
loaded_box = True
struct.box = xyz.box
else:
struct.coordinates = xyz
if box is not None:
loaded_box = True
struct.box = box
if struct.box is not None:
struct.box = np.asarray(struct.box)
if system is None:
return struct
if isinstance(system, string_types):
system = load_file(system)
if not isinstance(system, mm.System):
raise TypeError('system must be an OpenMM System object or serialized '
'XML of an OpenMM System object')
# We have a system, try to extract parameters from it
if len(struct.atoms) != system.getNumParticles():
raise TypeError('Topology and System have different numbers of atoms '
'(%d vs. %d)' %
(len(struct.atoms), system.getNumParticles()))
processed_forces = set()
ignored_forces = (mm.CMMotionRemover, mm.AndersenThermostat,
mm.MonteCarloBarostat, mm.MonteCarloAnisotropicBarostat,
mm.MonteCarloMembraneBarostat, mm.CustomExternalForce,
mm.GBSAOBCForce, mm.CustomGBForce)
if system.usesPeriodicBoundaryConditions():
if not loaded_box:
vectors = system.getDefaultPeriodicBoxVectors()
leng, ang = box_vectors_to_lengths_and_angles(*vectors)
leng = leng.value_in_unit(u.angstroms)
ang = ang.value_in_unit(u.degrees)
struct.box = np.asarray(
[leng[0], leng[1], leng[2], ang[0], ang[1], ang[2]])
else:
struct.box = None
for force in system.getForces():
if isinstance(force, mm.HarmonicBondForce):
if mm.HarmonicBondForce in processed_forces:
# Try to process this HarmonicBondForce as a Urey-Bradley term
_process_urey_bradley(struct, force)
else:
_process_bond(struct, force)
elif isinstance(force, mm.HarmonicAngleForce):
_process_angle(struct, force)
elif isinstance(force, mm.PeriodicTorsionForce):
_process_dihedral(struct, force)
elif isinstance(force, mm.RBTorsionForce):
_process_rbtorsion(struct, force)
elif isinstance(force, mm.CustomTorsionForce):
if not _process_improper(struct, force):
struct.unknown_functional = True
warnings.warn('Unknown functional form of CustomTorsionForce',
OpenMMWarning)
elif isinstance(force, mm.CMAPTorsionForce):
_process_cmap(struct, force)
elif isinstance(force, mm.NonbondedForce):
_process_nonbonded(struct, force)
elif isinstance(force, ignored_forces):
continue
else:
struct.unknown_functional = True
warnings.warn('Unsupported Force type %s' % type(force).__name__,
OpenMMWarning)
processed_forces.add(type(force))
return struct
def parse(filename):
""" Parses a Gromacs GRO file
Parameters
----------
filename : str or file-like
Name of the file or the GRO file object
Returns
-------
struct : :class:`Structure`
The Structure instance instantiated with *just* residues and atoms
populated (with coordinates)
"""
struct = Structure()
if isinstance(filename, string_types):
fileobj = genopen(filename, 'r')
own_handle = True
else:
fileobj = filename
own_handle = False
try:
# Ignore the title line
fileobj.readline()
try:
natom = int(fileobj.readline().strip())
except ValueError:
raise GromacsError('Could not parse %s as GRO file' % filename)
digits = None
line_parser = _AtomLineParser()
for i, line in enumerate(fileobj):
if i == natom: break
try:
atom, resname, resnum = line_parser.read(line)
except (ValueError, IndexError):
raise GromacsError('Could not parse the atom record of '
'GRO file %s' % filename)
struct.add_atom(atom, resname, resnum)
else:
# If no box exists, the break did not hit, so line still
# contains the last atom (which cannot be interpreted as a box).
# This wipes out line (IFF fileobj reached the line)
line = fileobj.readline()
if i+1 != natom:
raise GromacsError('Truncated GRO file. Found %d of %d '
'atoms' % (i+1, natom))
# Get the box from the last line if it's present
if line.strip():
try:
box = [float(x) for x in line.split()]
except ValueError:
raise GromacsError('Could not understand box line of GRO '
'file %s' % filename)
if len(box) == 3:
struct.box = [box[0]*10, box[1]*10, box[2]*10,
90.0, 90.0, 90.0]
elif len(box) == 9:
# Assume we have vectors
leng, ang = box_vectors_to_lengths_and_angles(
[box[0], box[3], box[4]]*u.nanometers,
[box[5], box[1], box[6]]*u.nanometers,
[box[7], box[8], box[2]]*u.nanometers)
a, b, c = leng.value_in_unit(u.angstroms)
alpha, beta, gamma = ang.value_in_unit(u.degrees)
struct.box = [a, b, c, alpha, beta, gamma]
finally:
if own_handle:
fileobj.close()
return struct