def __init__(self, molid, **kwargs):
"""
Creates a CHARMM writer
Args:
molid (int): VMD molecule ID of system to write
tmp_dir (str): Directory for temporary files. Defaults to "."
lipid_sel (str): Lipid selection string. Defaults to "lipid"
hmr (bool): If hydrogen masses should be repartitioned. Defaults
to False
forcefield (str): Forcefield to use, either "charmm" or "amber"
water_model (str): Water model to use
extra_topos (list of str): Additional topology (.str, .off, .lib) to
include.
extra_params (list of str): Additional parameter sets (.str, .frcmod)
override_defaults (bool): If set, omits default forcefield parameters.
debug_verbose (bool): Prints additional output, like from psfgen.
"""
# Initialize default options
super(CharmmWriter, self).__init__(molid, **kwargs)
# Create a psf generator object
self.psfgen = PsfGen()
# Set forcefield default topologies and parameters
self.forcefield = kwargs.get("forcefield", "charmm")
self.water_model = kwargs.get("water_model", "tip3")
self.topologies = self.get_topologies(self.forcefield,
self.water_model)
self.parameters = self.get_parameters(self.forcefield,
self.water_model)
if "charmm" in self.forcefield:
if self.hmr:
raise DabbleError("HMR not supported with CHARMM ff yet")
# Handle override and extra topologies
if self.override:
self.topologies = []
self.parameters = []
# Now extra topologies (put in self by super __init__)
self.topologies.extend(self.extra_topos)
self.parameters.extend(self.extra_params)
# Once all topologies defined, initialize matcher only if
# using CHARMM topologies (not if we're doing a conversion)
if "charmm" in self.forcefield or "opls" in self.forcefield:
self.matcher = CharmmMatcher(self.topologies)
# Keep track of segment numbers for protein and other
self.segint = 0
def test_formats(tmpdir):
"""
Tests read/write of psf/namdbin files
"""
from psfgen import PsfGen
p = str(tmpdir.mkdir("formats"))
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.add_segment(segid="P0", pdbfile="psf_protein_P0.pdb")
gen.read_coords(segid="P0", filename="psf_protein_P0.pdb")
gen.add_segment(segid="P1", pdbfile="psf_protein_P1.pdb")
gen.read_coords(segid="P1", filename="psf_protein_P1.pdb")
# Write a PSF and a NAMD binary file
gen.write_psf(filename=os.path.join(p, "pdbin.psf"))
gen.write_namdbin(filename=os.path.join(p, "pdbin.bin"))
del gen
# Read in the PSF and NAMD binary file. Topology files should be
# automatically loaded, too. Read in coordinates also as velocities
# to test the velocity read in as well.
gen = PsfGen(output=os.devnull)
os.chdir(p)
gen.read_psf(filename=os.path.join(p, "pdbin.psf"),
namdbinfile=os.path.join(p, "pdbin.bin"),
velnamdbinfile=os.path.join(p, "pdbin.bin"))
assert gen.get_topologies() == ["top_all36_caps.rtf", "top_all36_prot.rtf"]
assert gen.get_segids() == ["P0", "P1"]
assert gen.get_coordinates(segid="P0", resid=1) \
== gen.get_velocities(segid="P0", resid=1)
def test_query():
"""
Tests that query functions work correctly
"""
from psfgen import PsfGen
gen = PsfGen(output=os.devnull)
os.chdir(dir)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.add_segment(segid="P0", pdbfile="psf_protein_P0.pdb")
gen.read_coords(segid="P0", filename="psf_protein_P0.pdb")
gen.add_segment(segid="P1", pdbfile="psf_protein_P1.pdb")
gen.read_coords(segid="P1", filename="psf_protein_P1.pdb")
gen.patch(patchname="DISU", targets=[("P0", "10"), ("P0", "15")])
assert gen.get_topologies() == ["top_all36_caps.rtf", "top_all36_prot.rtf"]
# Check residue names query
resnames = gen.get_residue_types()
assert len(resnames) == 26
assert "CYS" in resnames
assert "TIP3" not in resnames
# Check patches query
patches = gen.get_patches(list_all=True)
assert len(patches) == 25
assert "CYSD" in patches
assert "SEP" not in patches
# Check segids query
assert gen.get_segids() == ["P0", "P1"]
# Check resids query
assert gen.get_resids("P0") == [str(_) for _ in range(1, 26)]
assert gen.get_resids("P1") == [str(_) for _ in range(0, 31)]
# Check resname query, with str or int
assert gen.get_resname(segid="P0", resid="2") == "LEU"
assert gen.get_resname(segid="P1", resid=29) == "SER"
# Check applied patches query
assert gen.get_patches() == [("DISU", "P0", "10"), ("DISU", "P0", "15")]
assert gen.get_first(segid="P0") is None
assert gen.get_last(segid="P1") is None
# Check atom queries
assert gen.get_atom_names(segid="P0", resid="10") \
== ['N', 'HN', 'CA', 'HA', 'CB', 'HB1', 'HB2', 'SG', 'C', 'O']
assert set(gen.get_masses(segid="P0", resid=1)) == {1.008, 12.011, 15.999}
assert gen.get_atom_indices(segid="P1", resid=0) == list(range(1, 7))
assert set(gen.get_charges(segid="P0", resid="10")) \
== {-0.47, 0.07, 0.09, 0.31, -0.1, -0.08, 0.51, -0.51}
# Check coordinates and velocities
assert len(gen.get_coordinates(segid="P1", resid=25)) == 17
assert set(gen.get_velocities(segid="P1", resid=1)) == {(0., 0., 0.)}
def test_single_chain(tmpdir):
"""
Tests simple realistic system building
"""
from psfgen import PsfGen
p = str(tmpdir.mkdir("single_chain"))
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.read_topology("top_water_ions.rtf")
# Read protein
gen.add_segment(segid="P0", pdbfile="psf_protein_P0.pdb")
gen.read_coords(segid="P0", filename="psf_protein_P0.pdb")
gen.add_segment(segid="P1", pdbfile="psf_protein_P1.pdb")
gen.read_coords(segid="P1", filename="psf_protein_P1.pdb")
# Read waters, with 10k atoms per file to avoid PDB limitations
gen.add_segment(segid="W0", pdbfile="psf_wat_0.pdb")
gen.read_coords(segid="W0", filename="psf_wat_0.pdb")
gen.add_segment(segid="W1", pdbfile="psf_wat_1.pdb")
gen.read_coords(segid="W1", filename="psf_wat_1.pdb")
# Read ions
gen.add_segment(segid="I", pdbfile="psf_ions.pdb")
gen.read_coords(segid="I", filename="psf_ions.pdb")
# Add disulfides
gen.patch(patchname="DISU", targets=[("P0", "10"), ("P0", "15")])
gen.patch(patchname="DISU", targets=[("P0", "24"), ("P1", "23")])
gen.patch(patchname="DISU", targets=[("P0", "11"), ("P1", "11")])
# Regenerate
gen.regenerate_angles()
gen.regenerate_dihedrals()
# Write
os.chdir(p)
gen.write_psf(filename="output.psf")
gen.write_pdb(filename="output.pdb")
# Load as a molecule with vmd-python and check it's correct
m = molecule.load("psf", "output.psf", "pdb", "output.pdb")
check_correctness(m)
molecule.delete(m)
def test_case_sensitivity():
"""
Tests setting case sensitivity. Do this with 2 objects because you can't
change the setting after reading in topology files.
"""
from psfgen import PsfGen
os.chdir(dir)
gen = PsfGen(case_sensitive=True, output=os.devnull)
gen.read_topology("top_casesensitive.rtf")
assert gen.get_residue_types() == ["ACE", "Ace"]
# Can't change case sensitivity after topologies have been read
with pytest.raises(ValueError):
gen.case_sensitive = False
del gen
gen = PsfGen(case_sensitive=True, output=os.devnull)
gen.case_sensitive = False
gen.read_topology("top_casesensitive.rtf")
assert gen.get_residue_types() == ["ACE"]
def test_delete():
"""
Tests removing atoms
"""
from psfgen import PsfGen
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.add_segment(segid="P0", pdbfile="psf_protein_P0.pdb")
gen.read_coords(segid="P0", filename="psf_protein_P0.pdb")
# Delete a specific atom
assert "CAY" in gen.get_atom_names(segid="P0", resid=1)
gen.delete_atoms(segid="P0", resid=1, atomname="CAY")
assert "CAY" not in gen.get_atom_names(segid="P0", resid=1)
# Try deleting a capping group
assert gen.get_resids("P0") == [str(_) for _ in range(1, 26)]
gen.delete_atoms(segid="P0", resid=1)
assert gen.get_resids("P0") == [str(_) for _ in range(2, 26)]
# Add and then delete a segment
gen.add_segment(segid="DELETE", pdbfile="psf_protein_P1.pdb")
assert gen.get_segids() == ["P0", "DELETE"]
gen.delete_atoms(segid="DELETE")
assert gen.get_segids() == ["P0"]
def test_mutation(tmpdir):
"""
Tests mutation of L2A in chain 0. Also as a result tests guessing
coordinates
"""
from psfgen import PsfGen
p = str(tmpdir.mkdir("mutation"))
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.add_segment(segid="P0",
pdbfile="psf_protein_P0.pdb",
mutate=[("2", "ALA")])
gen.read_coords(segid="P0", filename="psf_protein_P0.pdb")
gen.patch(patchname="DISU", targets=[("P0", "10"), ("P0", "15")])
# Guess coordinates for ALA mutation
gen.guess_coords()
# Set one specific coordinate
gen.set_position(segid="P0",
resid="2",
atomname="HB1",
position=(1.0, 2.0, 3.0))
# Regenerate
gen.regenerate_angles()
gen.regenerate_dihedrals()
# Write
os.chdir(p)
gen.write_psf(filename="output.psf")
gen.write_pdb(filename="output.pdb")
# Check results with vmd-python
m = molecule.load("psf", "output.psf", "pdb", "output.pdb")
assert len(set(atomsel("protein").fragment)) == 1
assert len(set(atomsel("resname ACE NMA NME").residue)) == 2
# Test mutation happened and resid 2 is ALA not LEU
assert set(atomsel("resid 2").resname) == set(["ALA"])
# Check coordinate guessing happened and HB3 has a nonzero position
assert atomsel("resid 2 and name HB3").x != [0.0]
assert atomsel("resid 2 and name HB3").y != [0.0]
assert atomsel("resid 2 and name HB3").z != [0.0]
# Check manual coordinate setting happened
assert atomsel("resid 2 and name HB1").x == [1.0]
assert atomsel("resid 2 and name HB1").y == [2.0]
assert atomsel("resid 2 and name HB1").z == [3.0]
molecule.delete(m)
def test_ends(tmpdir):
"""
Tests adding patches to the beginning and end, as well as adding
residues in the segment
"""
from psfgen import PsfGen
p = str(tmpdir.mkdir("mutation"))
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_prot.rtf")
# Add neutral N-terminus
# Add an alanine then a protonated glutamate at the C-terminus.
gen.add_segment(segid="P",
pdbfile="protein_nocaps.pdb",
first="NTER",
last="GLUP",
residues=[("25", "ALA"), ("26", "GLU")])
# Set coordinates and regenerate angles and dihedrals
gen.read_coords(segid="P", filename="protein_nocaps.pdb")
gen.guess_coords()
# Check internal state
assert gen.get_resids("P") == [str(_) for _ in range(2, 27)]
assert gen.get_resname(segid="P", resid=25) == "ALA"
assert gen.get_patches(list_defaults=True) == [('GLUP', 'P', '26'),
('NTER', 'P', '2')]
assert gen.get_first(segid="P") == "NTER"
assert gen.get_last(segid="P") == "GLUP"
# Output
os.chdir(p)
gen.write_psf(filename="output.psf")
gen.write_pdb(filename="output.pdb")
# Check all resids are present and that 2 extra ones were added
m = molecule.load("psf", "output.psf", "pdb", "output.pdb")
assert list(set(atomsel("all").resid)) == list(range(2, 27))
assert len(atomsel("all")) == 382
assert set(atomsel("resid 25").resname) == set(["ALA"])
# Check patches were applied correctly
assert "HT1" in atomsel("resid 2").name
assert "HN" not in atomsel("resid 2").name
assert "HE2" in atomsel("resid 26").name
# Check all coordinates are set
assert 0.0 not in atomsel("all").x
assert 0.0 not in atomsel("all").y
assert 0.0 not in atomsel("all").z
molecule.delete(m)
def test_alias():
"""
Tests atom and residue aliases, either at the topology or the PDB level
"""
from psfgen import PsfGen
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.alias_residue(top_resname="LEU", pdb_resname="LEX")
gen.alias_residue(top_resname="ARG", pdb_resname="AAA")
gen.alias_atom(top_atomname="N", pdb_atomname="NOOO", resname="PHE")
gen.add_segment(segid="P", pdbfile="protein_newnames.pdb")
gen.read_coords(segid="P", filename="psf_protein_P0.pdb")
assert gen.get_resname(segid="P", resid=2) == "LEU"
assert gen.get_resname(segid="P", resid=5) == "ALA"
assert "N" in gen.get_atom_names(segid="P", resid=23)
def test_set():
"""
Tests that setters work correctly
"""
from psfgen import PsfGen
os.chdir(dir)
gen = PsfGen(output=os.devnull)
gen.read_topology("top_all36_caps.rtf")
gen.read_topology("top_all36_prot.rtf")
gen.add_segment(segid="P", pdbfile="psf_protein_P1.pdb")
assert gen.get_segids() == ["P"]
# Set segid
gen.set_segid(segid="P", new_segid="P1")
assert gen.get_segids() == ["P1"]
gen.read_coords(segid="P1", filename="psf_protein_P1.pdb")
# Set resname
assert gen.get_resname(segid="P1", resid="1") == "ASP"
gen.set_resname(segid="P1", resid="1", new_resname="ASH")
assert gen.get_resname(segid="P1", resid="1") == "ASH"
# Set charge
gen.set_charge(segid="P1", resid="1", atomname="O", charge=-1.)
assert -1.0 in gen.get_charges(segid="P1", resid="1")
# Set atom name
gen.set_atom_name(segid="P1", resid="1", atomname="N", new_atomname="NO")
assert "N" not in gen.get_atom_names(segid="P1", resid="1")
assert "NO" in gen.get_atom_names(segid="P1", resid="1")
# Set coord
gen.set_position(segid="P1",
resid="1",
atomname="HN",
position=(0., 0., -1.))
assert (0., 0., -1.) in gen.get_coordinates(segid="P1", resid="1")
# Set velocity
gen.set_velocity(segid="P1",
resid="1",
atomname="NO",
velocity=(
5.,
5.,
3.,
))
assert (
5.,
5.,
3.,
) in gen.get_velocities(segid="P1", resid="1")
class CharmmWriter(MoleculeWriter):
"""
An object that handles all the conversions to a psf file
by interfacing with psfgen.
Writes a pdb/psf file pair from the current molecule using the
CHARMM36 topology and atom names/types. Interfaces with psfgen by
dynamically generating the .tcl file that psfgen takes as input.
Prompts the user for additional topology files and helps with
matching atom names that cannot be automatically translated to the
charmm naming conventions.
"""
#+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# CONSTANTS #
#+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
WATER_NAMES = {
"tip3": "TIP3",
"tip4e": "TP4E",
"spce": "SPCE",
}
WATER_O_NAME = "OH2"
WATER_H_NAMES = ["H1", "H2"]
#==========================================================================
def __init__(self, molid, **kwargs):
"""
Creates a CHARMM writer
Args:
molid (int): VMD molecule ID of system to write
tmp_dir (str): Directory for temporary files. Defaults to "."
lipid_sel (str): Lipid selection string. Defaults to "lipid"
hmr (bool): If hydrogen masses should be repartitioned. Defaults
to False
forcefield (str): Forcefield to use, either "charmm" or "amber"
water_model (str): Water model to use
extra_topos (list of str): Additional topology (.str, .off, .lib) to
include.
extra_params (list of str): Additional parameter sets (.str, .frcmod)
override_defaults (bool): If set, omits default forcefield parameters.
debug_verbose (bool): Prints additional output, like from psfgen.
"""
# Initialize default options
super(CharmmWriter, self).__init__(molid, **kwargs)
# Create a psf generator object
self.psfgen = PsfGen()
# Set forcefield default topologies and parameters
self.forcefield = kwargs.get("forcefield", "charmm")
self.water_model = kwargs.get("water_model", "tip3")
self.topologies = self.get_topologies(self.forcefield,
self.water_model)
self.parameters = self.get_parameters(self.forcefield,
self.water_model)
if "charmm" in self.forcefield:
if self.hmr:
raise DabbleError("HMR not supported with CHARMM ff yet")
# Handle override and extra topologies
if self.override:
self.topologies = []
self.parameters = []
# Now extra topologies (put in self by super __init__)
self.topologies.extend(self.extra_topos)
self.parameters.extend(self.extra_params)
# Once all topologies defined, initialize matcher only if
# using CHARMM topologies (not if we're doing a conversion)
if "charmm" in self.forcefield or "opls" in self.forcefield:
self.matcher = CharmmMatcher(self.topologies)
# Keep track of segment numbers for protein and other
self.segint = 0
#=========================================================================
def write(self, filename):
"""
Writes the parameter and topology files
Args:
filename (str): File name to write. File type suffix will be added.
"""
self.outprefix = filename
# Put our molecule on top
old_top = molecule.get_top()
molecule.set_top(self.molid)
# Amber forcefield done with AmberWriter then conversion
if "amber" in self.forcefield:
# Avoid circular import by doing it here
from dabble.param import AmberWriter
prmtopgen = AmberWriter(molid=self.molid,
tmp_dir=self.tmp_dir,
forcefield=self.forcefield,
water_model=self.water_model,
hmr=self.hmr,
lipid_sel=self.lipid_sel,
extra_topos=self.extra_topos,
extra_params=self.extra_params,
override_defaults=self.override,
debug_verbose=self.debug)
prmtopgen.write(self.outprefix)
self._prmtop_to_charmm()
# Charmm forcefield
elif "charmm" in self.forcefield:
self._run_psfgen()
# OPLS forcefield. Same as charmm but list separately for readability
elif "opls" in self.forcefield:
self._run_psfgen()
else:
raise DabbleError("Unsupported forcefield '%s' for CharmmWriter" %
self.forcefield)
# Check output and finish up
self._check_psf_output()
# Reset top molecule
molecule.set_top(old_top)
#=========================================================================
# Static methods #
#=========================================================================
@classmethod
def get_topologies(cls, forcefield, water_model):
if forcefield == "charmm":
topos = [
"top_all36_caps.rtf", "top_all36_cgenff.rtf",
"top_all36_prot.rtf", "top_all36_lipid.rtf",
"top_all36_carb.rtf", "top_all36_na.rtf",
"toppar_all36_prot_na_combined.str",
"toppar_all36_prot_fluoro_alkanes.str"
]
if water_model == "tip3":
topos.append("toppar_water_ions.str")
elif water_model == "tip4e":
topos.append("toppar_water_ions_tip4p_ew.str")
elif water_model == "spce":
topos.append("toppar_water_ions_spc_e.str")
elif forcefield == "opls":
topos = ["opls_aam.rtf", "opls_aam_caps.rtf"]
if water_model != "tip3":
raise DabbleError("Only TIP3 water model supported for OPLS")
elif forcefield == "amber":
from dabble.param import AmberWriter # avoid circular dependency
return AmberWriter.get_topologies(forcefield, water_model)
else:
raise ValueError("Invalid forcefield: '%s'" % forcefield)
return [cls._get_forcefield_path(top) for top in topos]
#=========================================================================
@classmethod
def get_parameters(cls, forcefield, water_model):
if forcefield == "charmm":
prms = [
"par_all36m_prot.prm", "par_all36_cgenff.prm",
"par_all36_lipid.prm", "par_all36_carb.prm",
"par_all36_na.prm", "toppar_all36_prot_na_combined.str"
]
if water_model == "tip3":
prms.append("toppar_water_ions.str")
elif water_model == "tip4e":
prms.append("toppar_water_ions_tip4p_ew.str")
elif water_model == "spce":
prms.append("toppar_water_ions_spc_e.str")
elif forcefield == "amber":
from dabble.param import AmberWriter # avoid circular dependency
return AmberWriter.get_parameters(forcefield, water_model)
elif forcefield == "opls":
prms = ["opls_aam.prm"]
if water_model != "tip3":
raise DabbleError("Only TIP3 water model supported for OPLS")
else:
raise ValueError("Invalid forcefield: '%s'" % forcefield)
return [cls._get_forcefield_path(par) for par in prms]
#=========================================================================
# Private methods #
#=========================================================================
def _write_water_blocks(self):
"""
Writes a lot of temporary files with 10000 waters each, to bypass
psfgen being stupid with files containing more than 10000 of a residue.
"""
# Set water names and write them to PDB file(s)
self._set_water_names()
pdbs = self._write_water_pdbs()
for i, pdb in enumerate(pdbs):
self.psfgen.add_segment(segid="W%d" % i, pdbfile=pdb)
self.psfgen.read_coords(segid="W%d" % i, filename=pdb)
# If water model includes dummy atoms, guess the coordinates
# This is safe as only waters have been added to the psfgen state
# so far, so actually broken atoms won't be fixed on accident.
if self.water_model != "tip3":
self.psfgen.guess_coords()
self.psfgen.regenerate_angles()
self.psfgen.regenerate_dihedrals()
#==========================================================================
def _write_lipid_blocks(self):
"""
Writes a temporary PDB file containing the lipids for later use by
psfgen. Renumbers the lipid residues because some can have **** instead
of an integer for resid in large systems, which will crash psfgen. Also
sets atom names for some common lipids (currently POPC)
Raises:
NotImplementedError if more than 10,000 lipids are present since it
doesn't support feeding multiple lipid blocks to psfgen currently
NotImplementedError if lipid other than POPC,POPE,POPG is found
"""
# Put current molecule on top to simplify atom selection
old_top = molecule.get_top()
molecule.set_top(self.molid)
# Collect lipid residues up
alll = atomsel('(%s) and user 1.0' % self.lipid_sel)
residues = list(set(alll.residue))
# Lipids not compatible with AMBER parameters, CHARMM format
if alll and ("amber" in self.forcefield or "opls" in self.forcefield):
raise ValueError(
"AMBER or OPLS parameters not supported for lipids"
" in CHARMM output format")
# Sanity check for < 10k lipids
if len(residues) >= 10000:
raise NotImplementedError("More than 10k lipids found")
# Loop through all residues and renumber and correctly name them
lipress = []
for resname in set(alll.resname):
lipress.extend(self._rename_by_resname(resname, renumber=True))
# Write temporary lipid pdb
_, temp = tempfile.mkstemp(suffix='.pdb',
prefix='psf_lipid_',
dir=self.tmp_dir)
os.close(_)
saved_lips = atomsel("residue %s" % ' '.join(str(_) for _ in lipress))
saved_lips.user = 0.0
saved_lips.write('pdb', temp)
# Generate lipid segment
self.psfgen.add_segment(segid="L", pdbfile=temp)
self.psfgen.read_coords(segid="L", filename=temp)
# Put old top back
molecule.set_top(old_top)
#==========================================================================
def _write_ion_blocks(self):
"""
Writes a PDB file containing correctly named ions for use by
psfgen, and instructs psfgen to use it in TCL code.
"""
# Put our molecule on top to simplify atom selection language
old_top = molecule.get_top()
molecule.set_top(self.molid)
# Select all ions
allions = []
for resname in set(atomsel("numbonds 0").resname):
allions.extend(self._rename_by_resname(resname, renumber=True))
# Stop if no ions were found
if not allions:
return
# Save ions as pdb
allsel = atomsel("residue %s" % " ".join(str(_) for _ in allions))
allsel.resid = range(len(allsel))
allsel.user = 0.0
_, temp = tempfile.mkstemp(suffix=".pdb",
prefix="psf_ions_",
dir=self.tmp_dir)
os.close(_)
allsel.write("pdb", temp)
self.psfgen.add_segment(segid="I", pdbfile=temp)
self.psfgen.read_coords(segid="I", filename=temp)
molecule.set_top(old_top)
#==========================================================================
def _find_single_residue_names(self, resname, molid):
"""
Uses graph matcher and available topologies to match up
ligand names automatically. Tries to use graphs, and if there's an
uneven number of atoms tries to match manually to suggest which atoms
are most likely missing.
Args:
resname (str): Residue name of the ligand that will be written.
All ligands will be checked separately against the graphs.
molid (int): VMD molecule ID to consider
Returns:
(list of ints): Residue numbers (not resid) of all input ligands
that were successfully matched. Need to do it this way since
residue names can be changed in here to different things.
Raises:
ValueError if number of resids does not match number of residues as
interpreted by VMD
NotImplementedError if a residue could not be matched to a graph.
"""
# Put our molecule on top
old_top = molecule.get_top()
molecule.set_top(molid)
# Sanity check that there is no discrepancy between defined resids and
# residues as interpreted by VMD.
residues = set(atomsel("user 1.0 and resname '%s'" % resname).residue)
for chain in set(atomsel("user 1.0 and resname '%s'" % resname).chain):
tempres = set(
atomsel("user 1.0 and resname '%s' and chain %s" %
(resname, chain)).residue)
resids = set(
atomsel("user 1.0 and resname '%s' and chain %s" %
(resname, chain)).resid)
if len(tempres) != len(resids):
raise DabbleError("VMD found %d residues for resname '%s', "
"but there are %d resids in chain %s! "
"Check input." %
(len(tempres), resname, len(resids), chain))
for residue in residues:
sel = atomsel("residue %s and resname '%s' and user 1.0" %
(residue, resname))
newname, atomnames = self.matcher.get_names(sel,
print_warning=True)
if not newname:
resname, patch, atomnames = self.matcher.get_patches(sel)
if not newname:
print(
"ERROR: Could not find a residue definition for %s:%s"
% (resname, residue))
raise NotImplementedError(
"No residue definition for %s:%s" % (resname, residue))
print("\tApplying patch %s to ligand %s" % (patch, newname))
# Do the renaming
self._apply_naming_dictionary(atomnames=atomnames,
resnames=newname,
verbose=True)
molecule.set_top(old_top)
return list(residues)
#==========================================================================
def _write_generic_block(self, residues):
"""
Matches ligands to available topology file, renames atoms, and then
writes temporary files for the ligands
Args:
residues (list of int): Residue numbers to be written. Will all
be written to one segment.
Returns:
True if successful
"""
# Put our molecule on top to simplify atom selection language
old_top = molecule.get_top()
molecule.set_top(self.molid)
alig = atomsel('user 1.0 and residue %s' %
" ".join([str(x) for x in residues]))
# Write temporary file containg the residues and update tcl commands
_, temp = tempfile.mkstemp(suffix='.pdb',
prefix='psf_block_',
dir=self.tmp_dir)
os.close(_)
alig.write('pdb', temp)
alig.user = 0.0
# Get next available segment name
segname = "B%d" % self.segint
self.segint += 1
self.psfgen.add_segment(segid=segname, pdbfile=temp)
self.psfgen.read_coords(segid=segname, filename=temp)
if old_top != -1:
molecule.set_top(old_top)
return True
#==========================================================================
def _write_protein_blocks(self, molid, frag):
"""
Writes a protein fragment to a pdb file for input to psfgen
Automatically assigns amino acid names
Args:
molid (int): VMD molecule ID of renumbered protein
frag (str): Fragment to write
Returns:
(list of Patches): Patches to add to psfgen input files
"""
print("Setting protein atom names")
# Put our molecule on top to simplify atom selection language
old_top = molecule.get_top()
molecule.set_top(molid)
patches = set()
extpatches = set()
# Get a unique and reliabe segment name
seg = self.matcher.get_protein_segname(molid, frag)
fragsel = atomsel("fragment '%s'" % frag)
residues = list(set(fragsel.residue))
for residue in residues:
sel = atomsel('residue %s' % residue)
resid = sel.resid[0]
# Only try to match single amino acid if there are 1 or 2 bonds
if len(self.matcher.get_extraresidue_atoms(sel)) < 3:
(newname, atomnames) = self.matcher.get_names(sel, False)
# See if it's a disulfide bond participant
else:
(newname, patch, atomnames) = \
self.matcher.get_disulfide("residue %d" % residue,
molid)
if newname:
extpatches.add(patch)
# Couldn't find a match. See if it's a patched residue
if not newname:
(newname, patchname, atomnames) = self.matcher.get_patches(sel)
if newname:
# This returns patch name only, not a Patch object
patches.add(
Patch(name=patchname, segids=[seg], resids=[resid]))
# Fall through to error condition
if not newname:
raise DabbleError("Couldn't find a patch for %s:%s" %
(sel.resname[0], resid))
# Do the renaming
self._apply_naming_dictionary(atomnames=atomnames,
resnames=newname)
# Save protein chain in the correct order
filename = self.tmp_dir + '/psf_protein_%s.pdb' % seg
_write_ordered_pdb(filename, "fragment '%s'" % frag, molid)
print("\tWrote %d atoms to the protein segment %s" %
(len(atomsel("fragment %s" % frag)), seg))
# Now invoke psfgen for the protein segments
self.psfgen.add_segment(segid=seg, pdbfile=filename)
print("Applying the following single-residue patches to P%s:\n" % frag)
print("\t%s" % "\t".join(str(_) for _ in patches))
for p in patches:
self.psfgen.patch(patchname=p.name, targets=p.targets())
self.psfgen.read_coords(segid=seg, filename=filename)
# Fix coordinates that are out of bounds, ie 5 characters
badidxs = atomsel(
"fragment '%s' and (abs(x) >= 100 or abs(y) >= 100 "
"or abs(z) >= 100)" % frag, molid).index
for idx in badidxs:
atom = atomsel("index %d" % idx, molid)
self.psfgen.set_position(segid=seg,
resid=atom.resid[0],
atomname=atom.name[0],
position=(atom.x[0], atom.y[0],
atom.z[0]))
if old_top != -1:
molecule.set_top(old_top)
fragsel.user = 0.0
return extpatches
#==========================================================================
def _check_psf_output(self):
"""
Scans the output psf from psfgen for atoms where the coordinate
could not be set, indicating an unmatched atom. This check is necessary
because sometimes psfgen will run with no errors or warnings but will
have unmatched atoms that are all at (0,0,0).
"""
# Check file was written at all
if not os.path.isfile('%s.pdb' % self.outprefix):
raise DabbleError("\nERROR: psf file failed to write.\n"
" Please see log above.\n")
# Open the pdb file in VMD and check for atoms with no occupancy
fileh = molecule.load('pdb', '%s.pdb' % self.outprefix)
errors = atomsel("occupancy=-1", molid=fileh)
# Print out error messages
if errors:
errstr = "\nERROR: Couldn't find the following atoms.\n"
for i in range(len(errors)):
errstr += "\t%s%s:%s\n" % (errors.resname[i], errors.resid[i],
errors.name[i])
errstr += "Check if they are present in the original structure.\n"
raise DabbleError(errstr)
print("\nChecked output pdb/psf has all atoms present "
"and correct.\n")
#==========================================================================
def _find_residue_in_rtf(self, resname, molid):
"""
Scans the input topology files to find a name match for the given
residue name, then pulls out the atoms involved and checks that they
are all present in the input coordinates, prompting the user to correct
the names of atoms that could not be matched.
Residue ID is used because there can be multiple copies of a residue
with the same name, but only one has missing or extra atoms.
Args:
resname (str): Residue name to check
molid (int): VMD molecule ID
Returns:
True if all matching was successful
False if the residue name cannot be found
"""
print("Finding residue name '%s'" % resname)
for top in self.topologies:
topfile = open(top, 'r')
topo_atoms = _get_atoms_from_rtf(text=topfile.readlines(),
resname=resname)
# Use first definition found of this residue
if topo_atoms:
break
topfile.close()
if not topo_atoms:
return False
print("Successfully found residue %s in input topologies" % resname)
# Match up atoms with python sets
pdb_atoms = set(
atomsel("resname '%s' and user 1.0" % resname, molid=molid).name)
pdb_only = pdb_atoms - topo_atoms
topo_only = topo_atoms - pdb_atoms
# If uneven number of atoms, there are missing or additional atoms
if len(pdb_atoms) > len(topo_atoms):
raise DabbleError(
"\nERROR: Cannot process modified residue %s.\n"
"There are %d extra atoms in the input structure "
"that are undefined in the topology file. The "
"following atoms could not be matched and may "
"either be misnamed, or additional atoms:\n"
"[ %s ]\n" % (resname, len(pdb_atoms) - len(topo_atoms),
" ".join(pdb_only)))
if len(topo_atoms) > len(pdb_atoms):
raise DabbleError(
"\nERROR: Cannot process modified residue %s.\n"
"There are %d missing atoms in the input structure "
"that are defined in the topology file. The "
"following atoms could not be matched and may "
"either be misnamed or deleted atoms:\n"
"[ %s ]\n" % (resname, len(topo_atoms) - len(pdb_atoms),
" ".join(topo_only)))
# Offer to rename atoms that couldn't be matched to the topology
if pdb_only:
print("\nWARNING: Having some trouble with modified residue %s.\n"
" The following atom names cannot be matched up "
" to the input topologies. They are probably "
" misnamed.\n" % resname)
print(" To help you, here are the atom names that "
" should be present according to the topology "
" but were not found:\n")
print(" [ %s ]\n" % ' '.join([str(t) for t in topo_only]))
print(" Please enter a valid name for each atom as "
"it appears or CTRL+D to quit..\n")
for unmatched in pdb_only:
print("Unmatched topology names: [ %s ]" % ' '.join(topo_only))
newname = input(" %s -> " % unmatched)
while newname not in topo_only:
print("'%s' is not an available name in the topology."
"Please try again.\n" % newname)
newname = input(" %s -> " % unmatched)
atomsel("resname '%s' and user 1.0 and name '%s'" %
(resname, unmatched)).name = newname
pdb_atoms = set(
atomsel("resname '%s' and user 1.0" % resname).name)
topo_only = topo_atoms - pdb_atoms
resname = newname
# Recurse to check that everything is assigned correctly
self._find_residue_in_rtf(resname, molid)
print("Matched up all atom names for resname '%s'\n" % resname)
return True
#==========================================================================
def _get_patch(self, seg, resid):
"""
Prompts the user for a patch to apply for the given residue.
Gathers available patches from topology files
Args:
seg (str): Segment to apply the patch to
resid (int): Residue ID to apply the patch to
Returns:
(str) patch line to put in the psfgen input file
"""
avail_patches = self._get_avail_patches()
print("What is the patch name I should apply?")
print("Type NONE for no patch, if your residue is completely "
"defined in a str file")
print("Or type HELP for a list of all patches I know about")
patchname = input("> ")
if patchname == "HELP":
print(" PATCH COMMENT")
print(" ----- -------")
for patch in avail_patches:
print("%7s %s" % (patch, avail_patches[patch]))
patchname = input("> ")
while (patchname not in avail_patches) and (patchname != "NONE"):
print("I don't know about patch %s" % patchname)
patchname = input("Try again > ")
if patchname == "NONE":
return ""
return "patch %s %s:%d\n" % (patchname, seg, resid)
#==========================================================================
def _get_avail_patches(self):
"""
Gathers the patches defined in all topology files.
Returns:
(dict str -> str): Patch names as keys, comment as value
"""
avail_patches = {}
for top in self.topologies:
topfile = open(top, 'r')
for line in topfile:
tokens = line.split()
if not tokens:
continue
if tokens[0] == "PRES":
comment = ' '.join(tokens[tokens.index("!") + 1:])
avail_patches[tokens[1]] = comment
return avail_patches
#==========================================================================
def _run_psfgen(self):
# Read topology files in to psfgen
print("Using the following topologies:")
for top in self.topologies:
print(" - %s" % os.path.split(top)[1])
self.psfgen.read_topology(top)
# Mark all atoms as unsaved with the user field
atomsel('all', molid=self.molid).user = 1.0
check_atom_names(molid=self.molid)
# Save water 10k molecules at a time
if atomsel('water', molid=self.molid):
self._write_water_blocks()
# Now ions if present, changing the atom names
if atomsel('ions', molid=self.molid):
self._write_ion_blocks()
# Now lipid
if atomsel(self.lipid_sel):
self._write_lipid_blocks()
# Now handle the protein
# Save and reload the protein so residue looping is correct
if atomsel("resname %s" % PATCHABLE_ACIDS, molid=self.molid):
extpatches = set()
for frag in sorted(
set(
atomsel("resname %s" % PATCHABLE_ACIDS,
molid=self.molid).fragment)):
extpatches.update(self._write_protein_blocks(self.molid, frag))
# List all patches applied to the protein
print("Applying the following patches:\n")
print("\t%s" % "\n\t".join(str(_) for _ in extpatches))
# Apply all multi segment patches to the protein
for p in extpatches:
self.psfgen.patch(p.name, p.targets())
else:
print("\n\tDidn't find any protein. Continuing...\n")
# Regenerate angles and dihedrals after applying patches
# Angles must be regenerated FIRST!
# See http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/4137.html
self.psfgen.regenerate_angles()
self.psfgen.regenerate_dihedrals()
# Check if there is anything else and let the user know about it
leftovers = atomsel('user 1.0', molid=self.molid)
for lig in set(leftovers.resname):
residues = self._find_single_residue_names(resname=lig,
molid=self.molid)
self._write_generic_block(residues)
# Write the output files and run
self.psfgen.write_psf(filename="%s.psf" % self.outprefix,
type="x-plor")
self.psfgen.write_pdb(filename="%s.pdb" % self.outprefix)
#==========================================================================
def _prmtop_to_charmm(self):
"""
Converts an AMBER prmtop with AMBER parameters to a psf file,
using ParmEd.
"""
# Save PSF topology and parameter file
parmstruct = load_file(self.outprefix + ".prmtop",
xyz=self.outprefix + ".inpcrd",
structure=True)
parmstruct.save(self.outprefix + ".psf", format="psf")
# Save PDB file with coordinates
m = molecule.load("parm7", self.outprefix + ".prmtop", "rst7",
self.outprefix + ".inpcrd")
atomsel("all", m).write("pdb", self.outprefix + ".pdb")
molecule.delete(m)
