def _build_ensemble(self):
"""Finds simulation files genderated by MDPOW and attempts to build
:class:`MDAnalysis.Universe <MDAnalysis.core.groups.universe.Universe>`
in the lambda directories.
Run if :code:`dirname` argument is given when initializing the class.
First enters FEP directory, then traverses solvent and interaction
directories to search lambda directories for system files."""
fep_dir = os.path.join(self._ensemble_dir, 'FEP')
solv_top_path = None
for solvent in self._solvents: # Ugly set of loops, may have to find way to clean up
if self.top_dict is not None:
solv_top_path = self.top_dict[solvent]
for dirs in self._interactions: # Attribute folder names
int_dir = os.path.join(fep_dir, solvent, dirs)
with in_dir(int_dir,
create=False): # Entering attribute folders
logger.info("Searching %s directory for systems",
os.curdir)
files = os.listdir(os.curdir)
for file in sorted(files): # Traversing lambda directories
if os.path.isdir(file):
with in_dir(file, create=False):
u = self._load_universe_from_dir(
solv_dir=solv_top_path, **self.unv_kwargs)
if u is None:
logger.warning(
f'No system loaded in {file}')
else:
self.add_system((solvent, dirs, file), u)
def _setup_solvate(self, **kwargs):
sol = gromacs.setup.solvate_sol(**kwargs)
with in_dir(self.dirs.solvation, create=False):
u = mda.Universe('solvated.gro')
octanol = u.select_atoms('resname OcOH')
n = octanol.n_residues
with in_dir(self.dirs.topology, create=False):
gromacs.cbook.edit_txt(self.files.topology,
[('OcOH 1', '1', n)])
ionkwargs = kwargs
ionkwargs['struct'] = sol['struct']
params = gromacs.setup.solvate_ion(**ionkwargs)
return params
def data_tofile(data, fid=None, sep="", fmt="%s", dirname="."):
"""
@param data: the data to write to a file. This may be either a string,
a h5py.Dataset, a numpy ndarray, or a MDAnalysis atom group or universe.
@param fid : file or str
An open file object, or a string containing a filename.
This MAY be None if and only if the data is a h5py.Dataset, in shich case, the
output file name is taken from the last part of the dataset name. e.g. if the
dataset name is /foo/bar/fred', 'fred' will be the output file name (but only if
fid is None. If fid is given, this overrides the dataset name.
@param sep : str
Separator between array items for text output. If "" (empty), a binary file is written,
equivalent to file.write(a.tostring()).
@param fmt : str
Format string for text file output. Each entry in the array is formatted to text by
first converting it to the closest Python type, and then using "format" % item.
@param dirname
@return: absolute path of the created file
"""
if data is not None:
if isinstance(data, str) and os.path.isfile(data):
src = os.path.abspath(data)
with utilities.in_dir(dirname):
shutil.copyfile(src, fid)
return os.path.abspath(fid)
else:
with utilities.in_dir(dirname):
if type(data) is h5py.Dataset:
print('name:', data.name)
if fid is None:
fid = os.path.split(data.name)[1]
print('fid:', fid)
# data now becomes an ndarray
data = data[()]
if type(data) is n.ndarray:
print('writing file {} in dir {}'.format(fid, dirname))
data.tofile(fid, sep, fmt)
elif isinstance(
data,
MDAnalysis.core.AtomGroup.AtomGroup) or isinstance(
data, MDAnalysis.core.AtomGroup.Universe):
print("pwd", os.path.curdir)
print("fid", fid)
w = MDAnalysis.Writer(fid, numatoms=len(data.atoms))
w.write(data)
del w
else:
raise TypeError(
"expected either Dataset, ndarray or file path as src")
return os.path.abspath(fid)
def data_tofile(data, fid=None, sep="", fmt="%s", dirname="."):
"""
@param data: the data to write to a file. This may be either a string,
a h5py.Dataset, a numpy ndarray, or a MDAnalysis atom group or universe.
@param fid : file or str
An open file object, or a string containing a filename.
This MAY be None if and only if the data is a h5py.Dataset, in shich case, the
output file name is taken from the last part of the dataset name. e.g. if the
dataset name is /foo/bar/fred', 'fred' will be the output file name (but only if
fid is None. If fid is given, this overrides the dataset name.
@param sep : str
Separator between array items for text output. If "" (empty), a binary file is written,
equivalent to file.write(a.tostring()).
@param fmt : str
Format string for text file output. Each entry in the array is formatted to text by
first converting it to the closest Python type, and then using "format" % item.
@param dirname
@return: absolute path of the created file
"""
if data is not None:
if isinstance(data, str) and os.path.isfile(data):
src = os.path.abspath(data)
with utilities.in_dir(dirname):
shutil.copyfile(src, fid)
return os.path.abspath(fid)
else:
with utilities.in_dir(dirname):
if type(data) is h5py.Dataset:
print('name:',data.name)
if fid is None:
fid = os.path.split(data.name)[1]
print('fid:',fid)
# data now becomes an ndarray
data = data[()]
if type(data) is n.ndarray:
print('writing file {} in dir {}'.format(fid,dirname))
data.tofile(fid,sep,fmt)
elif isinstance(data, MDAnalysis.core.AtomGroup.AtomGroup) or isinstance(data, MDAnalysis.core.AtomGroup.Universe):
print("pwd", os.path.curdir)
print("fid", fid)
w = MDAnalysis.Writer(fid,numatoms=len(data.atoms))
w.write(data)
del w
else:
raise TypeError("expected either Dataset, ndarray or file path as src")
return os.path.abspath(fid)
def _setup(self, **kwargs):
mdp = config.get_template("bar_opls.mdp")
with in_dir(self.tmpdir.name, create=False):
self.Gsolv = fep.Gsolv(simulation=self.S,
molecule='BNZ',
mdp=mdp,
**kwargs)
self.Gsolv.setup(maxwarn=1)
def test_basic_run(self):
with in_dir(self.tmpdir.name, create=False):
try:
self._run_equil('water', 'benzene/')
self._new_structures()
except Exception as err:
raise AssertionError(
'Equilibration simulations failed with exception:\n{0}'.
format(str(err)))
def test_default_run(self):
with in_dir(self.tmpdir.name, create=False):
try:
self._run_fep('water', 'benzene/')
except:
raise AssertionError('FEP simulations failed with exception:\n{0}'.format(str(err)))
assert os.path.exists(os.path.join(self.tmpdir.name,
'benzene', 'FEP', 'water', 'VDW', '0000', 'md.edr'))
def solvation(setup, solvent, ff='OPLS-AA'):
itp = test_file[ff]
with in_dir(setup, create=False):
try:
S = sims[solvent](molecule='BNZ', forcefield=ff)
S.topology(itp=itp)
S.solvate(struct='benzene.pdb')
except Exception:
raise AssertionError('Solvation failed.')
def _test_solvation(self, solvent):
with in_dir(self.tmpdir.name, create=False):
try:
if isinstance(solvent, list):
for sol in solvent:
S = self.sims[sol](molecule='BNZ')
S.topology(itp='benzene.itp')
S.solvate(struct='benzene.pdb')
else:
S = self.sims[solvent](molecule='BNZ')
S.topology(itp='benzene.itp')
S.solvate(struct='benzene.pdb')
except Exception:
raise AssertionError('Solvation failed.')
def test_add_remove_systems(self):
with in_dir(self.tmpdir.name, create=False):
bnz = Ensemble()
l_dir = os.path.join(os.curdir, 'FEP', 'water', 'Coulomb', '0000')
top_dir = os.path.join(l_dir, 'md.gro')
trj_dir = os.path.join(l_dir, 'md_red.xtc')
U = mda.Universe(top_dir, trj_dir)
bnz.add_system(('water', 'Coulomb', '0000'), U)
assert bnz.keys() == [('water', 'Coulomb', '0000')]
assert bnz._num_systems == 1
assert bnz.__repr__() == "<Ensemble Containing 1 System>"
assert len(bnz) == 1
bnz.pop(('water', 'Coulomb', '0000'))
assert bnz._num_systems == 0
assert len(bnz) == 0
def topology(struct=None, protein='protein',
top='system.top', dirname='top',
posres="posres.itp", **pdb2gmx_args):
"""Build Gromacs topology files from pdb.
:Keywords:
*struct*
input structure (**required**)
*protein*
name of the output files
*top*
name of the topology file
*dirname*
directory in which the new topology will be stored
*pdb2gmxargs*
arguments for ``pdb2gmx`` such as ``ff``, ``water``, ...
.. note::
At the moment this function simply runs ``pdb2gmx`` and uses
the resulting topology file directly. If you want to create
more complicated topologies and maybe also use additional itp
files or make a protein itp file then you will have to do this
manually.
"""
structure = realpath(struct)
new_struct = protein + '.gro'
if posres is None:
posres = protein + '_posres.itp'
pdb2gmx_args.update({'f': structure, 'o': new_struct, 'p': top, 'i': posres})
with in_dir(dirname):
logger.info("[%(dirname)s] Building topology %(top)r from struct = %(struct)r" % vars())
# perhaps parse output from pdb2gmx 4.5.x to get the names of the chain itp files?
gromacs.pdb2gmx(**pdb2gmx_args)
return { \
'top': realpath(dirname, top), \
'struct': realpath(dirname, new_struct), \
'posres' : realpath(dirname, posres) }
def topology(self, itp='drug.itp', **kwargs):
"""Generate a topology for compound *molecule*.
:Keywords:
*itp*
Gromacs itp file; will be copied to topology dir and
included in topology
*dirname*
name of the topology directory ["top"]
*kwargs*
see source for *top_template*, *topol*
"""
self.journal.start('topology')
dirname = kwargs.pop('dirname', self.BASEDIR('top'))
self.dirs.topology = realpath(dirname)
top_template = config.get_template(kwargs.pop('top_template', 'system.top'))
topol = kwargs.pop('topol', os.path.basename(top_template))
itp = os.path.realpath(itp)
_itp = os.path.basename(itp)
with in_dir(dirname):
shutil.copy(itp, _itp)
gromacs.cbook.edit_txt(top_template,
[('#include +"compound\.itp"', 'compound\.itp', _itp),
('#include +"oplsaa\.ff/tip4p\.itp"', 'tip4p\.itp', self.solvent.itp),
('Compound', 'solvent', self.solvent_type),
('Compound', 'DRUG', self.molecule),
('DRUG\s*1', 'DRUG', self.molecule),
],
newname=topol)
logger.info('[%(dirname)s] Created topology %(topol)r that includes %(_itp)r', vars())
# update known files and dirs
self.files.topology = realpath(dirname, topol)
if not self.dirs.topology in self.dirs.includes:
self.dirs.includes.append(self.dirs.topology)
self.journal.completed('topology')
return {'dirname': dirname, 'topol': topol}
def topology(self, itp='drug.itp', **kwargs):
"""Generate a topology for compound *molecule*.
:Keywords:
*itp*
Gromacs itp file; will be copied to topology dir and
included in topology
*dirname*
name of the topology directory ["top"]
*kwargs*
see source for *top_template*, *topol*
"""
self.journal.start('topology')
dirname = kwargs.pop('dirname', self.BASEDIR('top'))
self.dirs.topology = realpath(dirname)
top_template = config.get_template(kwargs.pop('top_template', 'system.top'))
topol = kwargs.pop('topol', os.path.basename(top_template))
itp = os.path.realpath(itp)
_itp = os.path.basename(itp)
with in_dir(dirname):
shutil.copy(itp, _itp)
gromacs.cbook.edit_txt(top_template,
[('#include +"compound\.itp"', 'compound\.itp', _itp),
('#include +"oplsaa\.ff/tip4p\.itp"', 'tip4p\.itp', self.solvent.itp),
('Compound', 'solvent', self.solvent_type),
('Compound', 'DRUG', self.molecule),
('DRUG\s*1', 'DRUG', self.molecule),
],
newname=topol)
logger.info('[%(dirname)s] Created topology %(topol)r that includes %(_itp)r', vars())
# update known files and dirs
self.files.topology = realpath(dirname, topol)
if not self.dirs.topology in self.dirs.includes:
self.dirs.includes.append(self.dirs.topology)
self.journal.completed('topology')
return {'dirname': dirname, 'topol': topol}
def _setup_MD(
dirname,
deffnm="md",
mdp=config.templates["md_OPLSAA.mdp"],
struct=None,
top="top/system.top",
ndx=None,
mainselection='"Protein"',
qscript=config.qscript_template,
qname=None,
startdir=None,
mdrun_opts="",
budget=None,
walltime=1 / 3.0,
dt=0.002,
runtime=1e3,
**mdp_kwargs
):
"""Generic function to set up a ``mdrun`` MD simulation.
See the user functions for usage.
"""
if struct is None:
raise ValueError("struct must be set to a input structure")
structure = realpath(struct)
topology = realpath(top)
try:
index = realpath(ndx)
except AttributeError: # (that's what realpath(None) throws...)
index = None # None is handled fine below
qname = mdp_kwargs.pop("sgename", qname) # compatibility for old scripts
qscript = mdp_kwargs.pop("sge", qscript) # compatibility for old scripts
qscript_template = config.get_template(qscript)
mdp_template = config.get_template(mdp)
nsteps = int(float(runtime) / float(dt))
mdp = deffnm + ".mdp"
tpr = deffnm + ".tpr"
mainindex = deffnm + ".ndx"
final_structure = deffnm + ".gro" # guess... really depends on templates,could also be DEFFNM.pdb
# write the processed topology to the default output
mdp_parameters = {"nsteps": nsteps, "dt": dt, "pp": "processed.top"}
mdp_parameters.update(mdp_kwargs)
add_mdp_includes(topology, mdp_parameters)
logger.info("[%(dirname)s] input mdp = %(mdp_template)r", vars())
with in_dir(dirname):
if not (mdp_parameters.get("Tcoupl", "").lower() == "no" or mainselection is None):
logger.info("[%(dirname)s] Automatic adjustment of T-coupling groups" % vars())
# make index file in almost all cases; with mainselection == None the user
# takes FULL control and also has to provide the template or index
groups = make_main_index(structure, selection=mainselection, oldndx=index, ndx=mainindex)
natoms = dict([(g["name"], float(g["natoms"])) for g in groups])
tc_group_names = ("__main__", "__environment__") # defined in make_main_index()
try:
x = natoms["__main__"] / natoms["__environment__"]
except KeyError:
x = 0 # force using SYSTEM in code below
wmsg = (
"Missing __main__ and/or __environment__ index group.\n"
"This probably means that you have an atypical system. You can "
"set mainselection=None and provide your own mdp and index files "
"in order to set up temperature coupling.\n"
"If no T-coupling is required then set Tcoupl='no'.\n"
"For now we will just couple everything to 'System'."
)
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
if x < 0.1:
# couple everything together
tau_t = firstof(mdp_parameters.pop("tau_t", 0.1))
ref_t = firstof(mdp_parameters.pop("ref_t", 300))
# combine all in one T-coupling group
mdp_parameters["tc-grps"] = "System"
mdp_parameters["tau_t"] = tau_t # this overrides the commandline!
mdp_parameters["ref_t"] = ref_t # this overrides the commandline!
mdp_parameters["gen-temp"] = mdp_parameters.pop("gen_temp", ref_t)
wmsg = (
"Size of __main__ is only %.1f%% of __environment__ so "
"we use 'System' for T-coupling and ref_t = %g K and "
"tau_t = %g 1/ps (can be changed in mdp_parameters).\n" % (x * 100, ref_t, tau_t)
)
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
else:
# couple protein and bath separately
n_tc_groups = len(tc_group_names)
tau_t = asiterable(mdp_parameters.pop("tau_t", 0.1))
ref_t = asiterable(mdp_parameters.pop("ref_t", 300))
if len(tau_t) != n_tc_groups:
tau_t = n_tc_groups * [tau_t[0]]
wmsg = (
"%d coupling constants should have been supplied for tau_t. "
"Using %f 1/ps for all of them." % (n_tc_groups, tau_t[0])
)
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
if len(ref_t) != n_tc_groups:
ref_t = n_tc_groups * [ref_t[0]]
wmsg = "%d temperatures should have been supplied for ref_t. " "Using %g K for all of them." % (
n_tc_groups,
ref_t[0],
)
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
mdp_parameters["tc-grps"] = tc_group_names
mdp_parameters["tau_t"] = tau_t
mdp_parameters["ref_t"] = ref_t
mdp_parameters["gen-temp"] = mdp_parameters.pop("gen_temp", ref_t[0])
index = realpath(mainindex)
if mdp_parameters.get("Tcoupl", "").lower() == "no":
logger.info("Tcoupl == no: disabling all temperature coupling mdp options")
mdp_parameters["tc-grps"] = ""
mdp_parameters["tau_t"] = ""
mdp_parameters["ref_t"] = ""
mdp_parameters["gen-temp"] = ""
if mdp_parameters.get("Pcoupl", "").lower() == "no":
logger.info("Pcoupl == no: disabling all pressure coupling mdp options")
mdp_parameters["tau_p"] = ""
mdp_parameters["ref_p"] = ""
mdp_parameters["compressibility"] = ""
unprocessed = gromacs.cbook.edit_mdp(mdp_template, new_mdp=mdp, **mdp_parameters)
check_mdpargs(unprocessed)
gromacs.grompp(f=mdp, p=topology, c=structure, n=index, o=tpr, **unprocessed)
runscripts = gromacs.qsub.generate_submit_scripts(
qscript_template,
deffnm=deffnm,
jobname=qname,
budget=budget,
startdir=startdir,
mdrun_opts=mdrun_opts,
walltime=walltime,
)
logger.info("[%(dirname)s] output mdp = %(mdp)r", vars())
logger.info("[%(dirname)s] output ndx = %(ndx)r", vars())
logger.info("[%(dirname)s] output tpr = %(tpr)r", vars())
logger.info("[%(dirname)s] output runscripts = %(runscripts)r", vars())
logger.info(
"[%(dirname)s] All files set up for a run time of %(runtime)g ps " "(dt=%(dt)g, nsteps=%(nsteps)g)" % vars()
)
kwargs = {
"struct": realpath(os.path.join(dirname, final_structure)), # guess
"top": topology,
"ndx": index, # possibly mainindex
"qscript": runscripts,
"mainselection": mainselection,
"deffnm": deffnm, # return deffnm (tpr = deffnm.tpr!)
}
kwargs.update(mdp_kwargs) # return extra mdp args so that one can use them for prod run
return kwargs
def _setup_MD(dirname,
deffnm='md', mdp=config.templates['md_OPLSAA.mdp'],
struct=None,
top='top/system.top', ndx=None,
mainselection='"Protein"',
qscript=config.qscript_template, qname=None, startdir=None, mdrun_opts="", budget=None, walltime=1/3.,
dt=0.002, runtime=1e3, multi=1, **mdp_kwargs):
"""Generic function to set up a ``mdrun`` MD simulation.
See the user functions for usage.
@param qname: name of the queing system, may be None.
@param multi: setup multiple concurrent simulations. These are based upon deffnm being set,
and a set of mdp / tpr are created named [deffnm]0.tpr. [deffnm]1.tpr, ...
"""
if struct is None:
raise ValueError('struct must be set to a input structure')
structure = realpath(struct)
topology = realpath(top)
try:
index = realpath(ndx)
except AttributeError: # (that's what realpath(None) throws...)
index = None # None is handled fine below
qname = mdp_kwargs.pop('sgename', qname) # compatibility for old scripts
qscript = mdp_kwargs.pop('sge', qscript) # compatibility for old scripts
qscript_template = config.get_template(qscript)
mdp_template = config.get_template(mdp)
nsteps = int(float(runtime)/float(dt))
mainindex = deffnm + '.ndx'
final_structure = deffnm + '.pdb' # guess... really depends on templates,could also be DEFFNM.pdb
# write the processed topology to the default output
mdp_parameters = {'nsteps':nsteps, 'dt':dt}
mdp_parameters.update(mdp_kwargs)
add_mdp_includes(topology, mdp_parameters)
# the basic result dictionary
# depending on options, various bits might be added to this.
result = {'struct': realpath(os.path.join(dirname, final_structure)), # guess
'top': topology,
'ndx': index, # possibly mainindex
'mainselection': mainselection,
'deffnm': deffnm, # return deffnm (tpr = deffnm.tpr!)
}
with in_dir(dirname):
if not (mdp_parameters.get('Tcoupl','').lower() == 'no' or mainselection is None):
logger.info("[%(dirname)s] Automatic adjustment of T-coupling groups" % vars())
# make index file in almost all cases; with mainselection == None the user
# takes FULL control and also has to provide the template or index
groups = make_main_index(structure, selection=mainselection,
oldndx=index, ndx=mainindex)
natoms = dict([(g['name'], float(g['natoms'])) for g in groups])
tc_group_names = ('__main__', '__environment__') # defined in make_main_index()
try:
x = natoms['__main__']/natoms['__environment__']
except KeyError:
x = 0 # force using SYSTEM in code below
wmsg = "Missing __main__ and/or __environment__ index group.\n" \
"This probably means that you have an atypical system. You can " \
"set mainselection=None and provide your own mdp and index files " \
"in order to set up temperature coupling.\n" \
"If no T-coupling is required then set Tcoupl='no'.\n" \
"For now we will just couple everything to 'System'."
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
if x < 0.1:
# couple everything together
tau_t = firstof(mdp_parameters.pop('tau_t', 0.1))
ref_t = firstof(mdp_parameters.pop('ref_t', 300))
# combine all in one T-coupling group
mdp_parameters['tc-grps'] = 'System'
mdp_parameters['tau_t'] = tau_t # this overrides the commandline!
mdp_parameters['ref_t'] = ref_t # this overrides the commandline!
mdp_parameters['gen-temp'] = mdp_parameters.pop('gen_temp', ref_t)
wmsg = "Size of __main__ is only %.1f%% of __environment__ so " \
"we use 'System' for T-coupling and ref_t = %g K and " \
"tau_t = %g 1/ps (can be changed in mdp_parameters).\n" \
% (x * 100, ref_t, tau_t)
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
else:
# couple protein and bath separately
n_tc_groups = len(tc_group_names)
tau_t = asiterable(mdp_parameters.pop('tau_t', 0.1))
ref_t = asiterable(mdp_parameters.pop('ref_t', 300))
if len(tau_t) != n_tc_groups:
tau_t = n_tc_groups * [tau_t[0]]
wmsg = "%d coupling constants should have been supplied for tau_t. "\
"Using %f 1/ps for all of them." % (n_tc_groups, tau_t[0])
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
if len(ref_t) != n_tc_groups:
ref_t = n_tc_groups * [ref_t[0]]
wmsg = "%d temperatures should have been supplied for ref_t. "\
"Using %g K for all of them." % (n_tc_groups, ref_t[0])
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
mdp_parameters['tc-grps'] = tc_group_names
mdp_parameters['tau_t'] = tau_t
mdp_parameters['ref_t'] = ref_t
mdp_parameters['gen-temp'] = mdp_parameters.pop('gen_temp', ref_t[0])
index = realpath(mainindex)
if mdp_parameters.get('Tcoupl','').lower() == 'no':
logger.info("Tcoupl == no: disabling all temperature coupling mdp options")
mdp_parameters['tc-grps'] = ""
mdp_parameters['tau_t'] = ""
mdp_parameters['ref_t'] = ""
mdp_parameters['gen-temp'] = ""
if mdp_parameters.get('Pcoupl','').lower() == 'no':
logger.info("Pcoupl == no: disabling all pressure coupling mdp options")
mdp_parameters['tau_p'] = ""
mdp_parameters['ref_p'] = ""
mdp_parameters['compressibility'] = ""
# do multiple concurrent simulations - ensemble sampling
if multi > 1:
for i in range(multi):
new_mdp = deffnm + str(i) + ".mdp"
mdout = deffnm + "out" + str(i) + ".mdp"
pp = "processed" + str(i) + ".top"
tpr = deffnm + str(i) + ".tpr"
# doing ensemble sampling, so give differnt seeds for each one
# if we are using 32 bit gromacs, make seeds are are 32 bit even on
# 64 bit machine
mdp_parameters["andersen_seed"] = random.randint(0,2**31)
mdp_parameters["gen_seed"] = random.randint(0,2**31)
mdp_parameters["ld_seed"] = random.randint(0,2**31)
unprocessed = gromacs.cbook.edit_mdp(mdp_template, new_mdp=new_mdp, **mdp_parameters)
check_mdpargs(unprocessed)
gromacs.grompp(f=new_mdp, p=topology, c=structure, n=index, o=tpr,
po=mdout, pp=pp, **unprocessed)
# only add multi to result if we really are doing multiple runs
result["multi"] = multi
else:
new_mdp = deffnm + '.mdp'
tpr = deffnm + '.tpr'
unprocessed = gromacs.cbook.edit_mdp(mdp_template, new_mdp=new_mdp, **mdp_parameters)
check_mdpargs(unprocessed)
gromacs.grompp(f=new_mdp, p=topology, c=structure, n=index, o=tpr,
po="mdout.mdp", pp="processed.top", **unprocessed)
# generate scripts for queing system if requested
if qname is not None:
runscripts = gromacs.qsub.generate_submit_scripts(
qscript_template, deffnm=deffnm, jobname=qname, budget=budget,
startdir=startdir, mdrun_opts=mdrun_opts, walltime=walltime)
result["qscript"] =runscripts
logger.info("[%(dirname)s] All files set up for a run time of %(runtime)g ps "
"(dt=%(dt)g, nsteps=%(nsteps)g)" % vars())
result.update(mdp_kwargs) # return extra mdp args so that one can use them for prod run
result.pop('define', None) # but make sure that -DPOSRES does not stay...
return result
def energy_minimize(dirname='em', mdp=config.templates['em.mdp'],
struct='solvate/ionized.pdb', top='top/system.top',
output='em.pdb', deffnm="em",
mdrunner=None, **kwargs):
"""Energy minimize the system.
This sets up the system (creates run input files) and also runs
``mdrun_d``. Thus it can take a while.
Additional itp files should be in the same directory as the top file.
Many of the keyword arguments below already have sensible values.
:Keywords:
*dirname*
set up under directory dirname [em]
*struct*
input structure (gro, pdb, ...) [solvate/ionized.pdb]
*output*
output structure (will be put under dirname) [em.pdb]
*deffnm*
default name for mdrun-related files [em]
*top*
topology file [top/system.top]
*mdp*
mdp file (or use the template) [templates/em.mdp]
*includes*
additional directories to search for itp files
*mdrunner*
:class:`gromacs.run.MDrunner` class; by defauly we
just try :func:`gromacs.mdrun_d` and :func:`gromacs.mdrun` but a
MDrunner class gives the user the ability to run mpi jobs
etc. [None]
*kwargs*
remaining key/value pairs that should be changed in the
template mdp file, eg ``nstxtcout=250, nstfout=250``.
.. note:: If :func:`~gromacs.mdrun_d` is not found, the function
falls back to :func:`~gromacs.mdrun` instead.
"""
structure = realpath(struct)
topology = realpath(top)
mdp_template = config.get_template(mdp)
deffnm = deffnm.strip()
# write the processed topology to the default output
kwargs.setdefault('pp', 'processed.top')
# filter some kwargs that might come through when feeding output
# from previous stages such as solvate(); necessary because *all*
# **kwargs must be *either* substitutions in the mdp file *or* valid
# command line parameters for ``grompp``.
kwargs.pop('ndx', None)
# mainselection is not used but only passed through; right now we
# set it to the default that is being used in all argument lists
# but that is not pretty. TODO.
mainselection = kwargs.pop('mainselection', '"Protein"')
# only interesting when passed from solvate()
qtot = kwargs.pop('qtot', 0)
mdp = deffnm+'.mdp'
tpr = deffnm+'.tpr'
logger.info("[%(dirname)s] Energy minimization of struct=%(struct)r, top=%(top)r, mdp=%(mdp)r ..." % vars())
add_mdp_includes(topology, kwargs)
if qtot != 0:
# At the moment this is purely user-reported and really only here because
# it might get fed into the function when using the keyword-expansion pipeline
# usage paradigm.
wmsg = "Total charge was reported as qtot = %(qtot)g <> 0; probably a problem." % vars()
logger.warn(wmsg)
warnings.warn(wmsg, category=BadParameterWarning)
with in_dir(dirname):
unprocessed = gromacs.cbook.edit_mdp(mdp_template, new_mdp=mdp, **kwargs)
check_mdpargs(unprocessed)
gromacs.grompp(f=mdp, o=tpr, c=structure, p=topology, **unprocessed)
mdrun_args = dict(v=True, stepout=10, deffnm=deffnm, c=output)
if mdrunner is None:
try:
gromacs.mdrun_d(**mdrun_args)
except (AttributeError, OSError):
# fall back to mdrun if no double precision binary
wmsg = "No 'mdrun_d' binary found so trying 'mdrun' instead.\n"\
"(Note that energy minimization runs better with mdrun_d.)"
logger.warn(wmsg)
warnings.warn(wmsg, category=AutoCorrectionWarning)
gromacs.mdrun(**mdrun_args)
else:
# user wants full control and provides simulation.MDrunner **class**
# NO CHECKING --- in principle user can supply any callback they like
mdrun = mdrunner(**mdrun_args)
mdrun.run()
# em.gro --> gives 'Bad box in file em.gro' warning --- why??
# --> use em.pdb instead.
if not os.path.exists(output):
errmsg = "Energy minimized system NOT produced."
logger.error(errmsg)
raise GromacsError(errmsg)
final_struct = realpath(output)
logger.info("[%(dirname)s] energy minimized structure %(final_struct)r" % vars())
return {'struct': final_struct,
'top': topology,
'mainselection': mainselection,
}
def solvate(struct='top/protein.pdb', top='top/system.top',
distance=0.9, boxtype='dodecahedron',
concentration=0, cation='NA', anion='CL',
water='spc', solvent_name='SOL', with_membrane=False,
ndx = 'main.ndx', mainselection = '"Protein"',
dirname='solvate',
**kwargs):
"""Put protein into box, add water, add counter-ions.
Currently this really only supports solutes in water. If you need
to embedd a protein in a membrane then you will require more
sophisticated approaches.
However, you *can* supply a protein already inserted in a
bilayer. In this case you will probably want to set *distance* =
``None`` and also enable *with_membrane* = ``True`` (using extra
big vdw radii for typical lipids).
.. Note:: The defaults are suitable for solvating a globular
protein in a fairly tight (increase *distance*!) dodecahedral
box.
:Arguments:
*struct* : filename
pdb or gro input structure
*top* : filename
Gromacs topology
*distance* : float
When solvating with water, make the box big enough so that
at least *distance* nm water are between the solute *struct*
and the box boundary.
Set *boxtype* to ``None`` in order to use a box size in the input
file (gro or pdb).
*boxtype* or *bt*: string
Any of the box types supported by :class:`~gromacs.tools.Editconf`
(triclinic, cubic, dodecahedron, octahedron). Set the box dimensions
either with *distance* or the *box* and *angle* keywords.
If set to ``None`` it will ignore *distance* and use the box
inside the *struct* file.
*bt* overrides the value of *boxtype*.
*box*
List of three box lengths [A,B,C] that are used by :class:`~gromacs.tools.Editconf`
in combination with *boxtype* (``bt`` in :program:`editconf`) and *angles*.
Setting *box* overrides *distance*.
*angles*
List of three angles (only necessary for triclinic boxes).
*concentration* : float
Concentration of the free ions in mol/l. Note that counter
ions are added in excess of this concentration.
*cation* and *anion* : string
Molecule names of the ions. This depends on the chosen force field.
*water* : string
Name of the water model; one of "spc", "spce", "tip3p",
"tip4p". This should be appropriate for the chosen force
field. If an alternative solvent is required, simply supply the path to a box
with solvent molecules (used by :func:`~gromacs.genbox`'s *cs* argument)
and also supply the molecule name via *solvent_name*.
*solvent_name*
Name of the molecules that make up the solvent (as set in the itp/top).
Typically needs to be changed when using non-standard/non-water solvents.
["SOL"]
*with_membrane* : bool
``True``: use special ``vdwradii.dat`` with 0.1 nm-increased radii on
lipids. Default is ``False``.
*ndx* : filename
How to name the index file that is produced by this function.
*mainselection* : string
A string that is fed to :class:`~gromacs.tools.Make_ndx` and
which should select the solute.
*dirname* : directory name
Name of the directory in which all files for the solvation stage are stored.
*includes*
List of additional directories to add to the mdp include path
*kwargs*
Additional arguments are passed on to
:class:`~gromacs.tools.Editconf` or are interpreted as parameters to be
changed in the mdp file.
"""
structure = realpath(struct)
topology = realpath(top)
# arguments for editconf that we honour
editconf_keywords = ["box", "bt", "angles", "c", "center", "aligncenter",
"align", "translate", "rotate", "princ"]
editconf_kwargs = dict((k,kwargs.pop(k,None)) for k in editconf_keywords)
editconf_boxtypes = ["triclinic", "cubic", "dodecahedron", "octahedron", None]
# needed for topology scrubbing
scrubber_kwargs = {'marker': kwargs.pop('marker',None)}
# sanity checks and argument dependencies
bt = editconf_kwargs.pop('bt')
boxtype = bt if bt else boxtype # bt takes precedence over boxtype
if not boxtype in editconf_boxtypes:
msg = "Unsupported boxtype %(boxtype)r: Only %(boxtypes)r are possible." % vars()
logger.error(msg)
raise ValueError(msg)
if editconf_kwargs['box']:
distance = None # if box is set then user knows what she is doing...
# handle additional include directories (kwargs are also modified!)
mdp_kwargs = add_mdp_includes(topology, kwargs)
if water.lower() in ('spc', 'spce'):
water = 'spc216'
elif water.lower() == 'tip3p':
water = 'spc216'
logger.warning("TIP3P water model selected: using SPC equilibrated box "
"for initial solvation because it is a reasonable starting point "
"for any 3-point model. EQUILIBRATE THOROUGHLY!")
# By default, grompp should not choke on a few warnings because at
# this stage the user cannot do much about it (can be set to any
# value but is kept undocumented...)
grompp_maxwarn = kwargs.pop('maxwarn',10)
# clean topology (if user added the marker; the default marker is
# ; Gromacs auto-generated entries follow:
n_removed = gromacs.cbook.remove_molecules_from_topology(topology, **scrubber_kwargs)
with in_dir(dirname):
logger.info("[%(dirname)s] Solvating with water %(water)r..." % vars())
if boxtype is None:
hasBox = False
ext = os.path.splitext(structure)[1]
if ext == '.gro':
hasBox = True
elif ext == '.pdb':
with open(structure) as struct:
for line in struct:
if line.startswith('CRYST'):
hasBox = True
break
if not hasBox:
msg = "No box data in the input structure %(structure)r and boxtype is set to None" % vars()
logger.exception(msg)
raise MissingDataError(msg)
distance = boxtype = None # ensures that editconf just converts
editconf_kwargs.update({'f': structure, 'o': 'boxed.gro',
'bt': boxtype, 'd': distance})
gromacs.editconf(**editconf_kwargs)
if with_membrane:
vdwradii_dat = get_lipid_vdwradii() # need to clean up afterwards
logger.info("Using special vdW radii for lipids %r" % vdw_lipid_resnames)
try:
gromacs.genbox(p=topology, cp='boxed.gro', cs=water, o='solvated.gro')
except:
if with_membrane:
# remove so that it's not picked up accidentally
gromacs.utilities.unlink_f(vdwradii_dat)
raise
logger.info("Solvated system with %s", water)
with open('none.mdp','w') as mdp:
mdp.write('; empty mdp file\ninclude = %(include)s\nrcoulomb = 1\nrvdw = 1\nrlist = 1\n' % mdp_kwargs)
qtotgmx = gromacs.cbook.grompp_qtot(f='none.mdp', o='topol.tpr', c='solvated.gro',
p=topology, stdout=False, maxwarn=grompp_maxwarn)
qtot = round(qtotgmx)
logger.info("[%(dirname)s] After solvation: total charge qtot = %(qtotgmx)r = %(qtot)r" % vars())
if concentration != 0:
logger.info("[%(dirname)s] Adding ions for c = %(concentration)f M..." % vars())
# target concentration of free ions c ==>
# N = N_water * c/c_water
# add ions for concentration to the counter ions (counter ions are less free)
#
# get number of waters (count OW ... works for SPC*, TIP*P water models)
rc,output,junk = gromacs.make_ndx(f='topol.tpr', o='ow.ndx',
input=('keep 0', 'del 0', 'a OW*', 'name 0 OW', '', 'q'),
stdout=False)
groups = gromacs.cbook.parse_ndxlist(output)
gdict = dict([(g['name'], g) for g in groups]) # overkill...
N_water = gdict['OW']['natoms'] # ... but dict lookup is nice
N_ions = int(N_water * concentration/CONC_WATER) # number of monovalents
else:
N_ions = 0
# neutralize (or try -neutral switch of genion???)
n_cation = n_anion = 0
if qtot > 0:
n_anion = int(abs(qtot))
elif qtot < 0:
n_cation = int(abs(qtot))
n_cation += N_ions
n_anion += N_ions
if n_cation != 0 or n_anion != 0:
# sanity check:
assert qtot + n_cation - n_anion < 1e-6
logger.info("[%(dirname)s] Adding n_cation = %(n_cation)d and n_anion = %(n_anion)d ions..." % vars())
gromacs.genion(s='topol.tpr', o='ionized.gro', p=topology,
pname=cation, nname=anion, np=n_cation, nn=n_anion,
input=solvent_name)
else:
# fake ionized file ... makes it easier to continue without too much fuzz
try:
os.unlink('ionized.gro')
except OSError, err:
if err.errno != errno.ENOENT:
raise
os.symlink('solvated.gro', 'ionized.gro')
qtot = gromacs.cbook.grompp_qtot(f='none.mdp', o='ionized.tpr', c='ionized.gro',
p=topology, stdout=False, maxwarn=grompp_maxwarn)
if abs(qtot) > 1e-4:
wmsg = "System has non-zero total charge qtot = %(qtot)g e." % vars()
warnings.warn(wmsg, category=BadParameterWarning)
logger.warn(wmsg)
# make main index
try:
make_main_index('ionized.tpr', selection=mainselection, ndx=ndx)
except GromacsError, err:
# or should I rather fail here?
wmsg = "Failed to make main index file %r ... maybe set mainselection='...'.\n"\
"The error message was:\n%s\n" % (ndx, str(err))
logger.warn(wmsg)
warnings.warn(wmsg, category=GromacsFailureWarning)
def _setup(self, **kwargs):
with in_dir(self.tmpdir.name, create=False):
self.Gsolv = mdpow.fep.Gsolv(simulation=self.S,
molecule='BNZ',
**kwargs)
self.Gsolv.setup(maxwarn=1)
def topology(self, itp='drug.itp', prm=None, **kwargs):
"""Generate a topology for compound *molecule*.
:Keywords:
*itp*
Gromacs itp file; will be copied to topology dir and
included in topology
*prm*
Gromacs prm file; if given, will be copied to topology
dir and included in topology
*dirname*
name of the topology directory ["top"]
*kwargs*
see source for *top_template*, *topol*
"""
self.journal.start('topology')
dirname = kwargs.pop('dirname', self.BASEDIR('top'))
self.dirs.topology = realpath(dirname)
setting = forcefields.get_ff_paths(self.forcefield)
template = forcefields.get_top_template(self.solvent_type)
top_template = config.get_template(kwargs.pop('top_template',
template))
topol = kwargs.pop('topol', os.path.basename(top_template))
self.top_template = top_template
itp = os.path.realpath(itp)
_itp = os.path.basename(itp)
if prm is None:
prm_kw = ''
else:
prm = os.path.realpath(prm)
_prm = os.path.basename(prm)
prm_kw = '#include "{}"'.format(_prm)
with in_dir(dirname):
shutil.copy(itp, _itp)
if prm is not None:
shutil.copy(prm, _prm)
gromacs.cbook.edit_txt(top_template, [
(r'#include +"oplsaa\.ff/forcefield\.itp"', r'oplsaa\.ff/',
setting[0]),
(r'#include +"compound\.itp"', r'compound\.itp', _itp),
(r'#include +"oplsaa\.ff/tip4p\.itp"',
r'oplsaa\.ff/tip4p\.itp', setting[0] + self.solvent.itp),
(r'#include +"oplsaa\.ff/ions_opls\.itp"',
r'oplsaa\.ff/ions_opls\.itp', setting[1]),
(r'#include +"compound\.prm"', r'#include +"compound\.prm"',
prm_kw),
(r'#include +"water\.itp"', r'water\.itp', setting[2]),
(r'Compound', 'solvent', self.solvent_type),
(r'Compound', 'DRUG', self.molecule),
(r'DRUG\s*1', 'DRUG', self.molecule),
],
newname=topol)
logger.info(
'[%(dirname)s] Created topology %(topol)r that includes %(_itp)r',
vars())
# update known files and dirs
self.files.topology = realpath(dirname, topol)
if not self.dirs.topology in self.dirs.includes:
self.dirs.includes.append(self.dirs.topology)
self.journal.completed('topology')
return {'dirname': dirname, 'topol': topol}