def __init__(self, fname, seq=None):
super(XyzFile, self).__init__()
if isinstance(fname, string_types):
fxyz = genopen(fname, 'r')
own_handle_xyz = True
else:
fxyz = fname
own_handle_xyz = False
if seq is not None:
seqstruct = load_file(seq)
# Now parse the file
try:
natom = int(fxyz.readline().split()[0])
except (ValueError, IndexError):
raise TinkerError('Bad XYZ file format; first line')
if seq is not None and natom != len(seqstruct.atoms):
raise ValueError('Sequence file %s # of atoms does not match the # '
'of atoms in the XYZ file' % seq)
words = fxyz.readline().split()
if len(words) == 6 and not XyzFile._check_atom_record(words):
self.box = [float(w) for w in words]
words = fxyz.readline().split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1], type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
residue = Residue('SYS')
residue.number = 1
residue._idx = 0
if seq is not None:
residue = seqstruct.residues[0]
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids = [[int(w) for w in words[6:]]]
for i, line in enumerate(fxyz):
words = line.split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1], type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
if seq is not None:
residue = seqstruct.atoms[i+1].residue
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids.append([int(w) for w in words[6:]])
# All of the bonds are stored now -- go ahead and make them now
for atom, bonds in zip(self.atoms, bond_ids):
i = atom.idx + 1
for idx in bonds:
if idx > i:
self.bonds.append(Bond(atom, self.atoms[idx-1]))
if own_handle_xyz:
fxyz.close()
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 __init__(self, fname, seq=None):
super(XyzFile, self).__init__()
if isinstance(fname, string_types):
fxyz = genopen(fname, 'r')
own_handle_xyz = True
else:
fxyz = fname
own_handle_xyz = False
if seq is not None:
seqstruct = load_file(seq)
# Now parse the file
try:
natom = int(fxyz.readline().split()[0])
except (ValueError, IndexError):
raise TinkerError('Bad XYZ file format; first line')
if seq is not None and natom != len(seqstruct.atoms):
raise ValueError(
'Sequence file %s # of atoms does not match the # '
'of atoms in the XYZ file' % seq)
words = fxyz.readline().split()
if len(words) == 6 and not XyzFile._check_atom_record(words):
self.box = [float(w) for w in words]
words = fxyz.readline().split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1],
type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
residue = Residue('SYS')
residue.number = 1
residue._idx = 0
if seq is not None:
residue = seqstruct.residues[0]
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids = [[int(w) for w in words[6:]]]
for i, line in enumerate(fxyz):
words = line.split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1],
type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
if seq is not None:
residue = seqstruct.atoms[i + 1].residue
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids.append([int(w) for w in words[6:]])
# All of the bonds are stored now -- go ahead and make them now
for atom, bonds in zip(self.atoms, bond_ids):
i = atom.idx + 1
for idx in bonds:
if idx > i:
self.bonds.append(Bond(atom, self.atoms[idx - 1]))
if own_handle_xyz:
fxyz.close()
def read(self, line):
""" Reads a line
Parameters
----------
line : str
A line with an atom record from a GRO file
Returns
-------
atom, resname, resnum : Atom, str, int
The Atom instance, residue name, and residue number containing the
atom
"""
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 self._digits is None:
self._pdeci = line.index('.', 20)
self._ndeci = line.index('.', self._pdeci+1)
self._digits = self._ndeci - self._pdeci
atom.xx, atom.xy, atom.xz = (
float(line[20+i*self._digits:20+(i+1)*self._digits])*10
for i in range(3)
)
wbeg = 20 + self._digits * 3
wend = wbeg + self._digits
if line[wbeg:wend].strip():
atom.vx, atom.vy, atom.vz = (
float(line[wbeg+i*self._digits:wend+i*self._digits])*10
for i in range(3)
)
return atom, resname, resnum
def create_random_structure(parametrized, novalence=False):
""" Create a random Structure with random attributes
Parameters
----------
parametrized : bool
If True, add at least two of all kinds of parameters to the
generated random structure. If False, just fill in the atoms and
residues and some random valence terms, but no "types"
novalence : bool, optional
If True, no valence terms will be added. Default is False. This is
set to False if parametrized is True
"""
from parmed.topologyobjects import (
Atom, Bond, AtomType, BondType, AngleType, DihedralType, ImproperType,
CmapType, OutOfPlaneBendType, StretchBendType, TorsionTorsionType,
AmoebaNonbondedExceptionType, Angle, UreyBradley, Dihedral, Improper,
Cmap, TrigonalAngle, OutOfPlaneBend, StretchBend, PiTorsion,
TorsionTorsion, AcceptorDonor, Group, ChiralFrame, MultipoleFrame,
NonbondedException, RBTorsionType)
from parmed import structure
from copy import copy
if parametrized: novalence = False
# Generate random atom and parameter types
atom_types = [
AtomType(''.join(random.sample(uppercase, 3)), i,
random.random() * 16 + 1, random.randint(1, 8))
for i in range(random.randint(8, 20))
]
bond_types = [
BondType(random.random() * 2,
random.random() * 100) for i in range(random.randint(10, 20))
]
angle_types = [
AngleType(random.random() * 50,
random.random() * 120) for i in range(random.randint(10, 20))
]
dihed_types = [
DihedralType(random.random() * 10, random.randint(1, 6),
random.choice([0, 180]))
for i in range(random.randint(10, 20))
]
rb_types = [RBTorsionType(*[random.random() * 10 for i in range(6)])]
imp_types = [
ImproperType(random.random() * 100, random.choice([0, 180]))
for i in range(random.randint(10, 20))
]
cmap_types = [
CmapType(24, [random.random() * 5 for i in range(24 * 24)])
for i in range(random.randint(5, 10))
]
oop_types = [
OutOfPlaneBendType(random.random() * 100)
for i in range(random.randint(10, 20))
]
strbnd_types = [
StretchBendType(random.random() * 10,
random.random() * 10,
random.random() * 2,
random.random() * 2,
random.random() * 120)
for i in range(random.randint(10, 20))
]
ang1, ang2 = list(range(-180, 180, 36)), list(range(-180, 180, 18))
tortor_types = [
TorsionTorsionType((10, 20), ang1[:], ang2[:],
[random.random() * 10 for j in range(200)])
for i in range(random.randint(5, 10))
]
for typ in atom_types:
typ.set_lj_params(random.random() * 2, random.random() * 2)
struct = structure.Structure()
# Add atoms in residues
for res in range(random.randint(20, 30)):
resname = ''.join(random.sample(uppercase, 3))
resid = res + 1
for i in range(random.randint(10, 25)):
name = ''.join(random.sample(uppercase, 4))
if parametrized:
typ = random.choice(atom_types)
type = str(typ)
mass = typ.mass
atomic_number = typ.atomic_number
else:
type = ''.join(random.sample(uppercase, 3))
mass = random.random() * 16 + 1
atomic_number = random.randint(1, 8)
charge = random.random() * 2 - 1
solvent_radius = random.random() * 2
screen = random.random() * 2
atom = Atom(atomic_number=atomic_number,
type=type,
charge=charge,
mass=mass,
solvent_radius=solvent_radius,
screen=screen,
name=name)
if parametrized:
atom.atom_type = typ
struct.add_atom(atom, resname, resid)
if novalence:
return struct
# Possibly add parameter type lists
if parametrized:
struct.bond_types.extend([copy(x) for x in bond_types])
struct.bond_types.claim()
struct.angle_types.extend([copy(x) for x in angle_types])
struct.angle_types.claim()
struct.dihedral_types.extend([copy(x) for x in dihed_types])
struct.dihedral_types.claim()
struct.rb_torsion_types.extend([copy(x) for x in rb_types])
struct.rb_torsion_types.claim()
struct.urey_bradley_types.extend([copy(x) for x in bond_types])
struct.urey_bradley_types.claim()
struct.improper_types.extend([copy(x) for x in imp_types])
struct.improper_types.claim()
struct.cmap_types.extend([copy(x) for x in cmap_types])
struct.cmap_types.claim()
struct.trigonal_angle_types.extend([copy(x) for x in angle_types])
struct.trigonal_angle_types.claim()
struct.out_of_plane_bend_types.extend([copy(x) for x in oop_types])
struct.out_of_plane_bend_types.claim()
struct.pi_torsion_types.extend([copy(x) for x in dihed_types])
struct.pi_torsion_types.claim()
struct.stretch_bend_types.extend([copy(x) for x in strbnd_types])
struct.stretch_bend_types.claim()
struct.torsion_torsion_types.extend([copy(x) for x in tortor_types])
struct.torsion_torsion_types.claim()
struct.adjust_types.extend([
AmoebaNonbondedExceptionType(0.5, 0.5, 0.6, 0.6, 0.7)
for i in range(random.randint(10, 20))
])
struct.adjust_types.claim()
# Add valence terms with optional
for i in range(random.randint(40, 50)):
struct.bonds.append(Bond(*random.sample(struct.atoms, 2)))
if parametrized:
struct.bonds[-1].type = random.choice(struct.bond_types)
for i in range(random.randint(35, 45)):
struct.angles.append(Angle(*random.sample(struct.atoms, 3)))
if parametrized:
struct.angles[-1].type = random.choice(struct.angle_types)
for i in range(random.randint(35, 45)):
struct.urey_bradleys.append(
UreyBradley(*random.sample(struct.atoms, 2)))
if parametrized:
struct.urey_bradleys[-1].type = random.choice(
struct.urey_bradley_types)
for i in range(random.randint(30, 40)):
struct.dihedrals.append(
Dihedral(*random.sample(struct.atoms, 4),
improper=random.choice([True, False])))
if parametrized:
struct.dihedrals[-1].type = random.choice(struct.dihedral_types)
for i in range(random.randint(30, 40)):
struct.rb_torsions.append(Dihedral(*random.sample(struct.atoms, 4)))
if parametrized:
struct.rb_torsions[-1].type = random.choice(
struct.rb_torsion_types)
for i in range(random.randint(10, 20)):
struct.impropers.append(Improper(*random.sample(struct.atoms, 4)))
if parametrized:
struct.impropers[-1].type = random.choice(struct.improper_types)
for i in range(random.randint(25, 35)):
struct.cmaps.append(Cmap(*random.sample(struct.atoms, 5)))
if parametrized:
struct.cmaps[-1].type = random.choice(struct.cmap_types)
for i in range(random.randint(30, 40)):
struct.trigonal_angles.append(
TrigonalAngle(*random.sample(struct.atoms, 4)))
if parametrized:
struct.trigonal_angles[-1].type = random.choice(
struct.trigonal_angle_types)
for i in range(random.randint(30, 40)):
struct.out_of_plane_bends.append(
OutOfPlaneBend(*random.sample(struct.atoms, 4)))
if parametrized:
struct.out_of_plane_bends[-1].type = random.choice(
struct.out_of_plane_bend_types)
for i in range(random.randint(30, 40)):
struct.stretch_bends.append(
StretchBend(*random.sample(struct.atoms, 3)))
if parametrized:
struct.stretch_bends[-1].type = random.choice(
struct.stretch_bend_types)
for i in range(random.randint(20, 30)):
struct.pi_torsions.append(PiTorsion(*random.sample(struct.atoms, 6)))
if parametrized:
struct.pi_torsions[-1].type = random.choice(
struct.pi_torsion_types)
for i in range(random.randint(10, 20)):
struct.torsion_torsions.append(
TorsionTorsion(*random.sample(struct.atoms, 5)))
if parametrized:
struct.torsion_torsions[-1].type = random.choice(
struct.torsion_torsion_types)
# Now use some lesser-used features
for i in range(random.randint(5, 10)):
struct.acceptors.append(AcceptorDonor(*random.sample(struct.atoms, 2)))
struct.donors.append(AcceptorDonor(*random.sample(struct.atoms, 2)))
struct.groups.append(Group(random.choice(struct.atoms), 2, 0))
struct.chiral_frames.append(
ChiralFrame(*random.sample(struct.atoms, 2),
chirality=random.choice([-1, 1])))
struct.multipole_frames.append(
MultipoleFrame(random.choice(struct.atoms), 0, 1, 2, 3))
for i in range(random.randint(20, 30)):
struct.adjusts.append(
NonbondedException(*random.sample(struct.atoms, 2)))
if parametrized:
struct.adjusts[-1].type = random.choice(struct.adjust_types)
struct.prune_empty_terms()
struct.unchange()
struct.update_dihedral_exclusions()
return struct
def __init__(self, psf_name=None):
"""
Opens and parses a PSF file, then instantiates a CharmmPsfFile
instance from the data.
"""
global _resre
Structure.__init__(self)
# Bail out if we don't have a filename
if psf_name is None:
return
conv = CharmmPsfFile._convert
# Open the PSF and read the first line. It must start with "PSF"
with closing(genopen(psf_name, 'r')) as psf:
self.name = psf_name
line = psf.readline()
if not line.startswith('PSF'):
raise CharmmError('Unrecognized PSF file. First line is %s' %
line.strip())
# Store the flags
psf_flags = line.split()[1:]
# Now get all of the sections and store them in a dict
psf.readline()
# Now get all of the sections
psfsections = _ZeroDict()
while True:
try:
sec, ptr, data = CharmmPsfFile._parse_psf_section(psf)
except _FileEOF:
break
psfsections[sec] = (ptr, data)
# store the title
self.title = psfsections['NTITLE'][1]
# Next is the number of atoms
natom = conv(psfsections['NATOM'][0], int, 'natom')
# Parse all of the atoms
for i in range(natom):
words = psfsections['NATOM'][1][i].split()
atid = int(words[0])
if atid != i + 1:
raise CharmmError('Nonsequential atoms detected!')
segid = words[1]
rematch = _resre.match(words[2])
if not rematch:
raise CharmmError('Could not interpret residue number %s' % # pragma: no cover
words[2])
resid, inscode = rematch.groups()
resid = conv(resid, int, 'residue number')
resname = words[3]
name = words[4]
attype = words[5]
# Try to convert the atom type to an integer a la CHARMM
try:
attype = int(attype)
except ValueError:
pass
charge = conv(words[6], float, 'partial charge')
mass = conv(words[7], float, 'atomic mass')
props = words[8:]
atom = Atom(name=name, type=attype, charge=charge, mass=mass)
atom.props = props
self.add_atom(atom, resname, resid, chain=segid,
inscode=inscode, segid=segid)
# Now get the number of bonds
nbond = conv(psfsections['NBOND'][0], int, 'number of bonds')
if len(psfsections['NBOND'][1]) != nbond * 2:
raise CharmmError('Got %d indexes for %d bonds' % # pragma: no cover
(len(psfsections['NBOND'][1]), nbond))
it = iter(psfsections['NBOND'][1])
for i, j in zip(it, it):
self.bonds.append(Bond(self.atoms[i-1], self.atoms[j-1]))
# Now get the number of angles and the angle list
ntheta = conv(psfsections['NTHETA'][0], int, 'number of angles')
if len(psfsections['NTHETA'][1]) != ntheta * 3:
raise CharmmError('Got %d indexes for %d angles' % # pragma: no cover
(len(psfsections['NTHETA'][1]), ntheta))
it = iter(psfsections['NTHETA'][1])
for i, j, k in zip(it, it, it):
self.angles.append(
Angle(self.atoms[i-1], self.atoms[j-1], self.atoms[k-1])
)
self.angles[-1].funct = 5 # urey-bradley
# Now get the number of torsions and the torsion list
nphi = conv(psfsections['NPHI'][0], int, 'number of torsions')
if len(psfsections['NPHI'][1]) != nphi * 4:
raise CharmmError('Got %d indexes for %d torsions' % # pragma: no cover
(len(psfsections['NPHI']), nphi))
it = iter(psfsections['NPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.dihedrals.append(
Dihedral(self.atoms[i-1], self.atoms[j-1],
self.atoms[k-1], self.atoms[l-1])
)
self.dihedrals.split = False
# Now get the number of improper torsions
nimphi = conv(psfsections['NIMPHI'][0], int, 'number of impropers')
if len(psfsections['NIMPHI'][1]) != nimphi * 4:
raise CharmmError('Got %d indexes for %d impropers' % # pragma: no cover
(len(psfsections['NIMPHI'][1]), nimphi))
it = iter(psfsections['NIMPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.impropers.append(
Improper(self.atoms[i-1], self.atoms[j-1],
self.atoms[k-1], self.atoms[l-1])
)
# Now handle the donors (what is this used for??)
ndon = conv(psfsections['NDON'][0], int, 'number of donors')
if len(psfsections['NDON'][1]) != ndon * 2:
raise CharmmError('Got %d indexes for %d donors' % # pragma: no cover
(len(psfsections['NDON'][1]), ndon))
it = iter(psfsections['NDON'][1])
for i, j in zip(it, it):
self.donors.append(
AcceptorDonor(self.atoms[i-1], self.atoms[j-1])
)
# Now handle the acceptors (what is this used for??)
nacc = conv(psfsections['NACC'][0], int, 'number of acceptors')
if len(psfsections['NACC'][1]) != nacc * 2:
raise CharmmError('Got %d indexes for %d acceptors' % # pragma: no cover
(len(psfsections['NACC'][1]), nacc))
it = iter(psfsections['NACC'][1])
for i, j in zip(it, it):
self.acceptors.append(
AcceptorDonor(self.atoms[i-1], self.atoms[j-1])
)
# Now get the group sections
try:
ngrp, nst2 = psfsections['NGRP NST2'][0]
except ValueError: # pragma: no cover
raise CharmmError('Could not unpack GROUP pointers') # pragma: no cover
tmp = psfsections['NGRP NST2'][1]
self.groups.nst2 = nst2
# Now handle the groups
if len(psfsections['NGRP NST2'][1]) != ngrp * 3:
raise CharmmError('Got %d indexes for %d groups' % # pragma: no cover
(len(tmp), ngrp))
it = iter(psfsections['NGRP NST2'][1])
for i, j, k in zip(it, it, it):
self.groups.append(Group(self.atoms[i], j, k))
# Assign all of the atoms to molecules recursively
tmp = psfsections['MOLNT'][1]
set_molecules(self.atoms)
molecule_list = [a.marked for a in self.atoms]
if len(tmp) == len(self.atoms):
if molecule_list != tmp:
warnings.warn('Detected PSF molecule section that is WRONG. '
'Resetting molecularity.', CharmmWarning)
# We have a CHARMM PSF file; now do NUMLP/NUMLPH sections
numlp, numlph = psfsections['NUMLP NUMLPH'][0]
if numlp != 0 or numlph != 0:
raise NotImplementedError('Cannot currently handle PSFs with '
'lone pairs defined in the NUMLP/'
'NUMLPH section.')
# Now do the CMAPs
ncrterm = conv(psfsections['NCRTERM'][0], int, 'Number of cross-terms')
if len(psfsections['NCRTERM'][1]) != ncrterm * 8:
raise CharmmError('Got %d CMAP indexes for %d cmap terms' % # pragma: no cover
(len(psfsections['NCRTERM']), ncrterm))
it = iter(psfsections['NCRTERM'][1])
for i, j, k, l, m, n, o, p in zip(it, it, it, it, it, it, it, it):
self.cmaps.append(
Cmap.extended(self.atoms[i-1], self.atoms[j-1],
self.atoms[k-1], self.atoms[l-1],
self.atoms[m-1], self.atoms[n-1],
self.atoms[o-1], self.atoms[p-1])
)
self.unchange()
self.flags = psf_flags
def create_random_structure(parametrized, novalence=False):
""" Create a random Structure with random attributes
Parameters
----------
parametrized : bool
If True, add at least two of all kinds of parameters to the
generated random structure. If False, just fill in the atoms and
residues and some random valence terms, but no "types"
novalence : bool, optional
If True, no valence terms will be added. Default is False. This is
set to False if parametrized is True
"""
from parmed.topologyobjects import (Atom, Bond, AtomType, BondType,
AngleType, DihedralType, ImproperType, CmapType, OutOfPlaneBendType,
StretchBendType, TorsionTorsionType, AmoebaNonbondedExceptionType,
Angle, UreyBradley, Dihedral, Improper, Cmap, TrigonalAngle,
OutOfPlaneBend, StretchBend, PiTorsion, TorsionTorsion,
AcceptorDonor, Group, ChiralFrame, MultipoleFrame,
NonbondedException, RBTorsionType)
from parmed import structure
from copy import copy
if parametrized: novalence = False
# Generate random atom and parameter types
atom_types = [AtomType(''.join(random.sample(uppercase, 3)),
i, random.random()*16+1, random.randint(1, 8))
for i in range(random.randint(8, 20))]
bond_types = [BondType(random.random()*2, random.random()*100)
for i in range(random.randint(10, 20))]
angle_types = [AngleType(random.random()*50, random.random()*120)
for i in range(random.randint(10, 20))]
dihed_types = [DihedralType(random.random()*10, random.randint(1, 6),
random.choice([0, 180]))
for i in range(random.randint(10, 20))]
rb_types = [RBTorsionType(*[random.random()*10 for i in range(6)])]
imp_types = [ImproperType(random.random()*100, random.choice([0, 180]))
for i in range(random.randint(10, 20))]
cmap_types = [CmapType(24, [random.random()*5 for i in range(24*24)])
for i in range(random.randint(5, 10))]
oop_types = [OutOfPlaneBendType(random.random()*100)
for i in range(random.randint(10, 20))]
strbnd_types = [StretchBendType(random.random()*10, random.random()*10,
random.random()*2, random.random()*2,
random.random()*120)
for i in range(random.randint(10, 20))]
ang1, ang2 = list(range(-180,180,36)), list(range(-180,180,18))
tortor_types = [TorsionTorsionType((10, 20), ang1[:], ang2[:],
[random.random()*10 for j in range(200)])
for i in range(random.randint(5, 10))]
for typ in atom_types:
typ.set_lj_params(random.random()*2, random.random()*2)
struct = structure.Structure()
# Add atoms in residues
for res in range(random.randint(20, 30)):
resname = ''.join(random.sample(uppercase, 3))
resid = res + 1
for i in range(random.randint(10, 25)):
name = ''.join(random.sample(uppercase, 4))
if parametrized:
typ = random.choice(atom_types)
type = str(typ)
mass = typ.mass
atomic_number = typ.atomic_number
else:
type = ''.join(random.sample(uppercase, 3))
mass = random.random() * 16 + 1
atomic_number = random.randint(1, 8)
charge = random.random() * 2 - 1
solvent_radius = random.random() * 2
screen = random.random() * 2
atom = Atom(atomic_number=atomic_number, type=type, charge=charge,
mass=mass, solvent_radius=solvent_radius,
screen=screen, name=name)
if parametrized:
atom.atom_type = typ
struct.add_atom(atom, resname, resid)
if novalence:
return struct
# Possibly add parameter type lists
if parametrized:
struct.bond_types.extend([copy(x) for x in bond_types])
struct.bond_types.claim()
struct.angle_types.extend([copy(x) for x in angle_types])
struct.angle_types.claim()
struct.dihedral_types.extend([copy(x) for x in dihed_types])
struct.dihedral_types.claim()
struct.rb_torsion_types.extend([copy(x) for x in rb_types])
struct.rb_torsion_types.claim()
struct.urey_bradley_types.extend([copy(x) for x in bond_types])
struct.urey_bradley_types.claim()
struct.improper_types.extend([copy(x) for x in imp_types])
struct.improper_types.claim()
struct.cmap_types.extend([copy(x) for x in cmap_types])
struct.cmap_types.claim()
struct.trigonal_angle_types.extend([copy(x) for x in angle_types])
struct.trigonal_angle_types.claim()
struct.out_of_plane_bend_types.extend([copy(x) for x in oop_types])
struct.out_of_plane_bend_types.claim()
struct.pi_torsion_types.extend([copy(x) for x in dihed_types])
struct.pi_torsion_types.claim()
struct.stretch_bend_types.extend([copy(x) for x in strbnd_types])
struct.stretch_bend_types.claim()
struct.torsion_torsion_types.extend([copy(x) for x in tortor_types])
struct.torsion_torsion_types.claim()
struct.adjust_types.extend([AmoebaNonbondedExceptionType(0.5, 0.5, 0.6, 0.6, 0.7)
for i in range(random.randint(10, 20))])
struct.adjust_types.claim()
# Add valence terms with optional
for i in range(random.randint(40, 50)):
struct.bonds.append(Bond(*random.sample(struct.atoms, 2)))
if parametrized:
struct.bonds[-1].type = random.choice(struct.bond_types)
for i in range(random.randint(35, 45)):
struct.angles.append(Angle(*random.sample(struct.atoms, 3)))
if parametrized:
struct.angles[-1].type = random.choice(struct.angle_types)
for i in range(random.randint(35, 45)):
struct.urey_bradleys.append(UreyBradley(*random.sample(struct.atoms, 2)))
if parametrized:
struct.urey_bradleys[-1].type = random.choice(struct.urey_bradley_types)
for i in range(random.randint(30, 40)):
struct.dihedrals.append(Dihedral(*random.sample(struct.atoms, 4),
improper=random.choice([True, False])))
if parametrized:
struct.dihedrals[-1].type = random.choice(struct.dihedral_types)
for i in range(random.randint(30, 40)):
struct.rb_torsions.append(Dihedral(*random.sample(struct.atoms, 4)))
if parametrized:
struct.rb_torsions[-1].type = random.choice(struct.rb_torsion_types)
for i in range(random.randint(10, 20)):
struct.impropers.append(Improper(*random.sample(struct.atoms, 4)))
if parametrized:
struct.impropers[-1].type = random.choice(struct.improper_types)
for i in range(random.randint(25, 35)):
struct.cmaps.append(Cmap(*random.sample(struct.atoms, 5)))
if parametrized:
struct.cmaps[-1].type = random.choice(struct.cmap_types)
for i in range(random.randint(30, 40)):
struct.trigonal_angles.append(TrigonalAngle(*random.sample(struct.atoms, 4)))
if parametrized:
struct.trigonal_angles[-1].type = random.choice(struct.trigonal_angle_types)
for i in range(random.randint(30, 40)):
struct.out_of_plane_bends.append(OutOfPlaneBend(*random.sample(struct.atoms, 4)))
if parametrized:
struct.out_of_plane_bends[-1].type = random.choice(struct.out_of_plane_bend_types)
for i in range(random.randint(30, 40)):
struct.stretch_bends.append(StretchBend(*random.sample(struct.atoms, 3)))
if parametrized:
struct.stretch_bends[-1].type = random.choice(struct.stretch_bend_types)
for i in range(random.randint(20, 30)):
struct.pi_torsions.append(PiTorsion(*random.sample(struct.atoms, 6)))
if parametrized:
struct.pi_torsions[-1].type = random.choice(struct.pi_torsion_types)
for i in range(random.randint(10, 20)):
struct.torsion_torsions.append(TorsionTorsion(*random.sample(struct.atoms, 5)))
if parametrized:
struct.torsion_torsions[-1].type = random.choice(struct.torsion_torsion_types)
# Now use some lesser-used features
for i in range(random.randint(5, 10)):
struct.acceptors.append(AcceptorDonor(*random.sample(struct.atoms, 2)))
struct.donors.append(AcceptorDonor(*random.sample(struct.atoms, 2)))
struct.groups.append(Group(random.choice(struct.atoms), 2, 0))
struct.chiral_frames.append(ChiralFrame(*random.sample(struct.atoms, 2),
chirality=random.choice([-1, 1])))
struct.multipole_frames.append(MultipoleFrame(random.choice(struct.atoms),
0, 1, 2, 3))
for i in range(random.randint(20, 30)):
struct.adjusts.append(NonbondedException(*random.sample(struct.atoms, 2)))
if parametrized:
struct.adjusts[-1].type = random.choice(struct.adjust_types)
struct.prune_empty_terms()
struct.unchange()
struct.update_dihedral_exclusions()
return struct
def __init__(self, psf_name=None):
"""
Opens and parses a PSF file, then instantiates a CharmmPsfFile
instance from the data.
"""
global _resre
Structure.__init__(self)
# Bail out if we don't have a filename
if psf_name is None:
return
conv = CharmmPsfFile._convert
# Open the PSF and read the first line. It must start with "PSF"
with closing(genopen(psf_name, 'r')) as psf:
self.name = psf_name
line = psf.readline()
if not line.startswith('PSF'):
raise CharmmError('Unrecognized PSF file. First line is %s' %
line.strip())
# Store the flags
psf_flags = line.split()[1:]
# Now get all of the sections and store them in a dict
psf.readline()
# Now get all of the sections
psfsections = _ZeroDict()
while True:
try:
sec, ptr, data = CharmmPsfFile._parse_psf_section(psf)
except _FileEOF:
break
psfsections[sec] = (ptr, data)
# store the title
self.title = psfsections['NTITLE'][1]
# Next is the number of atoms
natom = conv(psfsections['NATOM'][0], int, 'natom')
# Parse all of the atoms
for i in range(natom):
words = psfsections['NATOM'][1][i].split()
atid = int(words[0])
if atid != i + 1:
raise CharmmError('Nonsequential atoms detected!')
segid = words[1]
rematch = _resre.match(words[2])
if not rematch:
raise CharmmError('Could not interpret residue number %s'
% # pragma: no cover
words[2])
resid, inscode = rematch.groups()
resid = conv(resid, int, 'residue number')
resname = words[3]
name = words[4]
attype = words[5]
# Try to convert the atom type to an integer a la CHARMM
try:
attype = int(attype)
except ValueError:
pass
charge = conv(words[6], float, 'partial charge')
mass = conv(words[7], float, 'atomic mass')
props = words[8:]
atom = Atom(name=name, type=attype, charge=charge, mass=mass)
atom.props = props
self.add_atom(atom,
resname,
resid,
chain=segid,
inscode=inscode,
segid=segid)
# Now get the number of bonds
nbond = conv(psfsections['NBOND'][0], int, 'number of bonds')
if len(psfsections['NBOND'][1]) != nbond * 2:
raise CharmmError(
'Got %d indexes for %d bonds' % # pragma: no cover
(len(psfsections['NBOND'][1]), nbond))
it = iter(psfsections['NBOND'][1])
for i, j in zip(it, it):
self.bonds.append(Bond(self.atoms[i - 1], self.atoms[j - 1]))
# Now get the number of angles and the angle list
ntheta = conv(psfsections['NTHETA'][0], int, 'number of angles')
if len(psfsections['NTHETA'][1]) != ntheta * 3:
raise CharmmError(
'Got %d indexes for %d angles' % # pragma: no cover
(len(psfsections['NTHETA'][1]), ntheta))
it = iter(psfsections['NTHETA'][1])
for i, j, k in zip(it, it, it):
self.angles.append(
Angle(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1]))
self.angles[-1].funct = 5 # urey-bradley
# Now get the number of torsions and the torsion list
nphi = conv(psfsections['NPHI'][0], int, 'number of torsions')
if len(psfsections['NPHI'][1]) != nphi * 4:
raise CharmmError(
'Got %d indexes for %d torsions' % # pragma: no cover
(len(psfsections['NPHI']), nphi))
it = iter(psfsections['NPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.dihedrals.append(
Dihedral(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1]))
self.dihedrals.split = False
# Now get the number of improper torsions
nimphi = conv(psfsections['NIMPHI'][0], int, 'number of impropers')
if len(psfsections['NIMPHI'][1]) != nimphi * 4:
raise CharmmError(
'Got %d indexes for %d impropers' % # pragma: no cover
(len(psfsections['NIMPHI'][1]), nimphi))
it = iter(psfsections['NIMPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.impropers.append(
Improper(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1]))
# Now handle the donors (what is this used for??)
ndon = conv(psfsections['NDON'][0], int, 'number of donors')
if len(psfsections['NDON'][1]) != ndon * 2:
raise CharmmError(
'Got %d indexes for %d donors' % # pragma: no cover
(len(psfsections['NDON'][1]), ndon))
it = iter(psfsections['NDON'][1])
for i, j in zip(it, it):
self.donors.append(
AcceptorDonor(self.atoms[i - 1], self.atoms[j - 1]))
# Now handle the acceptors (what is this used for??)
nacc = conv(psfsections['NACC'][0], int, 'number of acceptors')
if len(psfsections['NACC'][1]) != nacc * 2:
raise CharmmError(
'Got %d indexes for %d acceptors' % # pragma: no cover
(len(psfsections['NACC'][1]), nacc))
it = iter(psfsections['NACC'][1])
for i, j in zip(it, it):
self.acceptors.append(
AcceptorDonor(self.atoms[i - 1], self.atoms[j - 1]))
# Now get the group sections
try:
ngrp, nst2 = psfsections['NGRP NST2'][0]
except ValueError: # pragma: no cover
raise CharmmError(
'Could not unpack GROUP pointers') # pragma: no cover
tmp = psfsections['NGRP NST2'][1]
self.groups.nst2 = nst2
# Now handle the groups
if len(psfsections['NGRP NST2'][1]) != ngrp * 3:
raise CharmmError(
'Got %d indexes for %d groups' % # pragma: no cover
(len(tmp), ngrp))
it = iter(psfsections['NGRP NST2'][1])
for i, j, k in zip(it, it, it):
self.groups.append(Group(self.atoms[i], j, k))
# Assign all of the atoms to molecules recursively
tmp = psfsections['MOLNT'][1]
set_molecules(self.atoms)
molecule_list = [a.marked for a in self.atoms]
if len(tmp) == len(self.atoms):
if molecule_list != tmp:
warnings.warn(
'Detected PSF molecule section that is WRONG. '
'Resetting molecularity.', CharmmWarning)
# We have a CHARMM PSF file; now do NUMLP/NUMLPH sections
numlp, numlph = psfsections['NUMLP NUMLPH'][0]
if numlp != 0 or numlph != 0:
raise NotImplementedError(
'Cannot currently handle PSFs with '
'lone pairs defined in the NUMLP/'
'NUMLPH section.')
# Now do the CMAPs
ncrterm = conv(psfsections['NCRTERM'][0], int,
'Number of cross-terms')
if len(psfsections['NCRTERM'][1]) != ncrterm * 8:
raise CharmmError('Got %d CMAP indexes for %d cmap terms'
% # pragma: no cover
(len(psfsections['NCRTERM']), ncrterm))
it = iter(psfsections['NCRTERM'][1])
for i, j, k, l, m, n, o, p in zip(it, it, it, it, it, it, it, it):
self.cmaps.append(
Cmap.extended(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1],
self.atoms[m - 1], self.atoms[n - 1],
self.atoms[o - 1], self.atoms[p - 1]))
self.unchange()
self.flags = psf_flags
def load(pose):
"""
Load a :class:`Pose` object and return a populated :class:`Structure`
instance
Parameters
----------
pose : :class:`Pose`
PyRosetta :class:`Pose` object to convert
"""
if not Pose or not AtomID:
raise ImportError('Could not load the PyRosetta module.')
if not isinstance(pose, Pose):
raise TypeError('Object is not a PyRosetta Pose object.')
struct = Structure()
atnum = 1
conf = pose.conformation()
for resid in range(1, pose.total_residue() + 1):
res = pose.residue(resid)
resname = res.name3().strip()
chain = chr(res.chain() + ord('A') - 1)
for atno, at in enumerate(res.atoms(), start=1):
try:
atinfo = res.atom_type(atno)
atname = res.atom_name(atno).strip()
if atinfo.is_virtual():
atsym = 'EP'
else:
atsym = atinfo.element()
rmin = atinfo.lj_radius()
epsilon = atinfo.lj_wdepth()
atomic_number = AtomicNum[atsym]
mass = Mass[atsym]
except KeyError:
raise RosettaError('Could not recognize element: %s.' %
atsym)
params = dict(atomic_number=atomic_number,
name=atname,
charge=0.0,
mass=mass,
occupancy=0.0,
bfactor=0.0,
altloc='',
number=atnum,
rmin=rmin,
epsilon=epsilon)
if atinfo.is_virtual():
atom = ExtraPoint(**params)
else:
atom = Atom(**params)
atom.xx, atom.xy, atom.xz = tuple(at.xyz())
struct.add_atom(atom, resname, resid, chain, '')
atnum += 1
try:
for nbr in conf.bonded_neighbor_all_res(AtomID(
atno, resid)):
if nbr.rsd() < resid or (nbr.rsd() == resid
and nbr.atomno() < atno):
struct.bonds.append(
Bond(struct.atoms[_n_prior(pose, nbr)], atom))
except:
raise RosettaError('Could not add bonds.')
struct.unchange()
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 apply_patch(self, patch, precision=4):
"""
Apply the specified PatchTemplate to the ResidueTemplate.
This only handles patches that affect a single residue.
An exception is thrown if patch is incompatible because
* The patch specifies that an atom is to be deleted that doesn't exist in the residue
* A bond specified as being added in the patch does not have both atom names present after adding/deleting atoms from the patch
* The new net charge is not integral to the specified precision
* The residue is not modified in any way (no atoms or bonds added/changed/deleted)
Parameters
----------
patch : PatchTemplate
The patch to apply to this residue
precision : int, optional
Each valid patch should be produce a net charge that is integral to
this many decimal places.
Default is 4
Returns
-------
residue : ResidueTemplate
A new ResidueTemplate corresponding to the patched residue is returned.
The original remains unmodified.
"""
# Create a copy
# TODO: Once ResidueTemplate.from_residue() actually copies all info, use that instead?
residue = _copy.copy(self)
# Record whether we've actually modified the residue.
modifications_made = False
# Delete atoms
for atom_name in patch.delete_atoms:
try:
residue.delete_atom(atom_name)
modifications_made = True
except (KeyError, MoleculeError) as e:
if atom_name.startswith('D') and atom_name[
1:] in self and atom_name[1:] in patch.delete_atoms:
# This is a Drude particle. We're also deleting its parent atom, so don't report an error.
pass
else:
raise IncompatiblePatchError(
'Atom %s could not be deleted from the patched residue: atoms are %s (exception: %s)'
% (atom_name, list(residue._map.keys()), str(e)))
# Add or replace atoms
for atom in patch.atoms:
if atom.name in residue:
# Overwrite type and charge
residue[atom.name].type = atom.type
residue[atom.name].charge = atom.charge
else:
residue.add_atom(
Atom(name=atom.name, type=atom.type, charge=atom.charge))
modifications_made = True
# Add bonds
for (atom1_name, atom2_name, order) in patch.add_bonds:
try:
# Remove dangling bonds
for name in [atom1_name, atom2_name]:
if residue.head and (name == residue.head.name):
residue.head = None
if residue.tail and (name == residue.tail.name):
residue.tail = None
# Add bond
residue.add_bond(atom1_name, atom2_name, order)
modifications_made = True
except (IndexError, MoleculeError) as e:
raise IncompatiblePatchError(
'Bond %s-%s could not be added to patched residue: atoms are %s (exception: %s)'
% (atom1_name, atom2_name, list(
residue._map.keys()), str(e)))
# Delete impropers
for impr in patch.delete_impropers:
try:
residue._impr.remove(impr)
# removal of impropers doesn't do anything as far as OpenMM is concerned, so don't note this as a modification having been made
except ValueError as e:
raise IncompatiblePatchError(
'Improper %s was not found in residue to be patched.' %
impr)
# Check that the net charge is integral.
net_charge = residue.net_charge
is_integral = (round(net_charge, precision) - round(net_charge)) == 0.0
if not is_integral:
raise IncompatiblePatchError(
'Patch is not compatible with residue due to non-integral charge (charge was %f).'
% net_charge)
# Ensure residue is connected
import networkx as nx
G = residue.to_networkx(False)
if not nx.is_connected(G):
components = [c for c in nx.connected_components(G)]
raise IncompatiblePatchError(
'Patched residue bond graph is not a connected graph: %s' %
str(components))
# Make sure the patch has actually modified the residue
if not modifications_made:
raise IncompatiblePatchError('Patch did not modify residue.')
return residue
def apply_patch(self, patch, precision=4):
"""
Apply the specified PatchTemplate to the ResidueTemplate.
This only handles patches that affect a single residue.
An exception is thrown if patch is incompatible because
* The patch specifies that an atom is to be deleted that doesn't exist in the residue
* A bond specified as being added in the patch does not have both atom names present after adding/deleting atoms from the patch
* The new net charge is not integral to the specified precision
Parameters
----------
patch : PatchTemplate
The patch to apply to this residue
precision : int, optional
Each valid patch should be produce a net charge that is integral to
this many decimal places.
Default is 4
Returns
-------
residue : ResidueTemplate
A new ResidueTemplate corresponding to the patched residue is returned.
The original remains unmodified.
"""
# Create a copy
# TODO: Once ResidueTemplate.from_residue() actually copies all info, use that instead?
residue = _copy.copy(self)
# Delete atoms
for atom_name in patch.delete_atoms:
try:
residue.delete_atom(atom_name)
except KeyError as e:
raise IncompatiblePatchError(str(e))
# Add or replace atoms
for atom in patch.atoms:
if atom.name in residue:
# Overwrite type and charge
residue[atom.name].type = atom.type
residue[atom.name].charge = atom.charge
else:
residue.add_atom(
Atom(name=atom.name, type=atom.type, charge=atom.charge))
# Add bonds
for (atom1_name, atom2_name, order) in patch.add_bonds:
try:
# Remove dangling bonds
for name in [atom1_name, atom2_name]:
if name == residue.head:
residue.head = None
if name == residue.tail:
residue.tail = None
# Add bond
residue.add_bond(atom1_name, atom2_name, order)
except IndexError as e:
raise IncompatiblePatchError(
'Bond %s-%s could not be added to patched residue: atoms are %s'
% (atom1_name, atom2_name, list(residue._map.keys())))
# Delete impropers
for impr in patch.delete_impropers:
try:
residue._impr.remove(impr)
except ValueError as e:
raise IncompatiblePatchError(
'Improper %s was not found in residue to be patched.' %
impr)
# Check that the net charge is integral.
net_charge = residue.net_charge
is_integral = (round(net_charge, precision) - round(net_charge)) == 0.0
if not is_integral:
raise IncompatiblePatchError(
'Patch is not compatible with residue due to non-integral charge (charge was %f).'
% net_charge)
# Ensure residue is connected
import networkx as nx
G = residue.to_networkx()
if not nx.is_connected(G):
components = [c for c in nx.connected_components(G)]
raise IncompatiblePatchError(
'Patched residue bond graph is not a connected graph: %s' %
str(components))
return residue
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
