def test3():
for i in arange(10,31,2.5):
fbase = "{}_{}".format(100.,i)
top = "{}.top".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
gromacs.grompp(f="md3.mdp", o="{}.tpr".format(fbase), c=sol, p=top, n=ndx) #@UndefinedVariable
def make_tpr(box, rrange):
for radius in rrange:
fbase = "{}_{}".format(box,radius)
top = "{}.top".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
gromacs.grompp(f="md2.mdp", o="{}.tpr".format(fbase), c=sol, p=top, n=ndx) #@UndefinedVariable
def test3():
for i in arange(10, 31, 2.5):
fbase = "{}_{}".format(100.0, i)
top = "{}.top".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
gromacs.grompp(f="md3.mdp",
o="{}.tpr".format(fbase),
c=sol,
p=top,
n=ndx) # @UndefinedVariable
def do_energy_min(pdb, itpfn, outpdb="out.pdb"):
# write the topology file
with open("topol_em.top", "w") as fh_t:
fh_t.write(HEAD_TOP_FILE)
fh_t.write("#include " + '"' + itpfn + '"')
fh_t.write(BOTTOM_TOP_FILE)
# TODO function that checks if files are present
gromacs.grompp(f="em.mdp",
c=pdb,
p="topol_em.top",
o="em.tpr",
maxwarn="99")
gromacs.mdrun(s="em.tpr", c=outpdb)
def test2(a,radius,box=100.0,fbase=None):
# l.test2(5.0797,15)
if fbase is None:
fbase = "{}_{}".format(box,radius)
top = "{}.top".format(fbase)
struct = "{}.pdb".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
origin = box/2.
pts=fcc_sphere(a, radius)
w=writer("{}.pdb".format(fbase))
w.CRYST1([box,box,box,90.00,90.00,90.00])
for index,atom in enumerate(pts):
w.ATOM(serial=index+1, name="AU", resName="NP", resSeq=1,
chainID="A", segID="AUNP", element="AU",
x=atom[0]+origin, y=atom[1]+origin, z=atom[2]+origin)
w.close()
#make_index("{}.pdb".format(fbase), "{}.ndx".format(fbase))
with file(top, "w") as t:
t.write(top_src)
t.write("Au {}\n".format(pts.shape[0]))
gromacs.genbox(p=top, cp=struct, cs="spc216.gro", o=sol, vdwd="0.15") #@UndefinedVariable
rc,out,nothing = gromacs.make_ndx(f=sol, n=None, o=ndx, stdout=False, #@UndefinedVariable
input=('', '', 'q'))
gromacs.grompp(f="md2.mdp", o="{}.tpr".format(fbase), c=sol, p=top, n=ndx) #@UndefinedVariable
with file("{}.sh".format(fbase), "w") as f:
f.write("#!/bin/bash\n")
f.write("#PBS -k o\n")
f.write("#PBS -l nodes=1:ppn=12:ccvt,walltime=24:00:00\n")
f.write("#PBS -M [email protected]\n")
f.write("#PBS -m abe\n")
f.write("#PBS -N {}\n".format(fbase))
f.write("#PBS -j oe\n")
f.write("#PBS -q pg\n")
f.write("#PBS -d /N/dc/scratch/somogyie/Au\n")
f.write("mpirun mdrun -deffnm {}".format(fbase))
def make_tpr(box, rrange):
for radius in rrange:
fbase = "{}_{}".format(box, radius)
top = "{}.top".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
gromacs.grompp(f="md2.mdp",
o="{}.tpr".format(fbase),
c=sol,
p=top,
n=ndx) #@UndefinedVariable
def _compile_tpr(self, step_name: str,
geometry: GenPath = None,
trajectory: GenPath = None
) -> pathlib.Path:
"""
Make a tpr file for the chosen step_name and associated mdp file
:param step_name: Key for the mdp file from the dict mdps
:param geometry: Path to the geometry to be used as input. If None,
:attr:`last_geometry` will be used.
:param trajectory: Path to a trajectory file from which to take the
input geometry. This is useful when a full precision geometry is
needed as input and a trr file can be used. If None,
no trajectory will be given to grompp.
:return: The Path to the tpr file
"""
geometry = self.last_geometry if geometry is None else geometry
tpr = '{}-{}.tpr'.format(self.name, step_name)
p_tpr = self._fp(tpr)
self.tprs[step_name] = p_tpr
rc, output, junk = gromacs.grompp(c=geometry,
p=self.top,
f=self.mdps[step_name],
o=tpr,
t=trajectory,
stdout=False)
# Doesn't capture output if failed?
self.outputs['compile_{}'.format(step_name)] = output
return p_tpr
def grompp(f, c, p, prefix="topol"):
s = prefix + '.tpr'
po = prefix + '.mdp'
rc, output, junk = gromacs.grompp(f=f, p=p, c=c, o=s, po=po, stdout=False, stderr=False)
#print(rc, output, junk)
assert rc == 0, \
"grompp -o {0} -po {1} -f {2} -c {3} -p {4} failed".format(s, po, f, c, p)
return s
def test_grompp(self, tmpdir):
"""Check if grompp can be run successfully at all"""
f = self.mdp
c = self.conf
p = self.processed
with tmpdir.as_cwd():
o = 'topol.tpr'
po = 'mdout.mdp'
rc, output, junk = gromacs.grompp(f=f, p=p, c=c, o=o, po=po, stdout=False, stderr=False)
assert rc == 0, "grompp -f {0} -o {1} ... failed to run".format(f, o)
def grompp(f, c, p, prefix="topol"):
s = prefix + '.tpr'
po = prefix + '.mdp'
rc, output, junk = gromacs.grompp(f=f,
p=p,
c=c,
o=s,
po=po,
stdout=False,
stderr=False)
#print(rc, output, junk)
assert rc == 0, \
"grompp -o {0} -po {1} -f {2} -c {3} -p {4} failed".format(s, po, f, c, p)
return s
def test_grompp(self, tmpdir):
"""Check if grompp can be run successfully at all"""
f = self.mdp
c = self.conf
p = self.processed
with tmpdir.as_cwd():
o = 'topol.tpr'
po = 'mdout.mdp'
rc, output, junk = gromacs.grompp(f=f,
p=p,
c=c,
o=o,
po=po,
stdout=False,
stderr=False)
assert rc == 0, "grompp -f {0} -o {1} ... failed to run".format(
f, o)
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, **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)
cbook.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".format(**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 = {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 = 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 = 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 energy_minimize(dirname='em', mdp=config.templates['em.mdp'],
struct='solvate/ionized.gro', 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.gro]
*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` instance; by default we
just try :func:`gromacs.mdrun_d` and :func:`gromacs.mdrun` but a
MDrunner instance 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} ...".format(**vars()))
cbook.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.".format(**vars())
logger.warn(wmsg)
warnings.warn(wmsg, category=BadParameterWarning)
with in_dir(dirname):
unprocessed = 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:
mdrun = run.get_double_or_single_prec_mdrun()
mdrun(**mdrun_args)
else:
if type(mdrunner) is type:
# class
# 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()
else:
# anything with a run() method that takes mdrun arguments...
try:
mdrunner.run(mdrunargs=mdrun_args)
except AttributeError:
logger.error("mdrunner: Provide a gromacs.run.MDrunner class or instance or a callback with a run() method")
raise TypeError("mdrunner: Provide a gromacs.run.MDrunner class or instance or a callback with a run() method")
# 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}".format(**vars()))
return {'struct': final_struct,
'top': topology,
'mainselection': mainselection,
}
def test_harmonic_potential():
import gmx
import os
import myplugin
cwd = os.path.dirname(__file__)
water = os.path.join(cwd, 'data', 'water.gro')
import shutil
shutil.copy(water, './')
try:
# use GromacsWrapper if available
import gromacs
import gromacs.formats
from gromacs.tools import Solvate as solvate
solvate(o='water.gro', box=[5, 5, 5])
mdpparams = [('integrator', 'md'), ('nsteps', 1000), ('nstxout', 100),
('nstvout', 100), ('nstfout', 100),
('tcoupl', 'v-rescale'), ('tc-grps', 'System'),
('tau-t', 1), ('ref-t', 298)]
mdp = gromacs.formats.MDP()
for param, value in mdpparams:
mdp[param] = value
mdp.write('water.mdp')
with open('input.top', 'w') as fh:
fh.write("""#include "gromos43a1.ff/forcefield.itp"
#include "gromos43a1.ff/spc.itp"
[ system ]
; Name
spc
[ molecules ]
; Compound #mols
SOL 4055
""")
gromacs.grompp(f='water.mdp',
c='water.gro',
po='water.mdp',
pp='water.top',
o='water.tpr',
p='input.top')
tpr_filename = os.path.abspath('water.tpr')
except:
from gmx.data import tpr_filename
print("Testing plugin potential with input file {}".format(
os.path.abspath(tpr_filename)))
md = gmx.workflow.from_tpr(tpr_filename, append_output=False)
context = gmx.context.ParallelArrayContext(md)
with context as session:
session.run()
# Create a WorkElement for the potential
#potential = gmx.core.TestModule()
params = {'sites': [1, 4], 'R0': 2.0, 'k': 10000.0}
potential_element = gmx.workflow.WorkElement(namespace="myplugin",
operation="create_restraint",
params=params)
# Note that we could flexibly capture accessor methods as workflow elements, too. Maybe we can
# hide the extra Python bindings by letting myplugin.HarmonicRestraint automatically convert
# to a WorkElement when add_dependency is called on it.
potential_element.name = "harmonic_restraint"
before = md.workspec.elements[md.name]
md.add_dependency(potential_element)
assert potential_element.name in md.workspec.elements
assert potential_element.workspec is md.workspec
after = md.workspec.elements[md.name]
assert not before is after
# Context will need to do these in __enter__
# potential = myplugin.HarmonicRestraint()
# potential.set_params(1, 4, 2.0, 10000.0)
context = gmx.context.ParallelArrayContext(md)
with context as session:
session.run()
def add_box_mutation(self):
"""
Function to add the solvent and run a minimization of the local amino acid with the waters included
Output:
system.gro -- file containing the complete system minimized after the mutation
"""
# Read the solvent
os.system("grep ATOM {path}/solvent/solvent_{itera}.pdb | grep -v ENDMDL > {path}/solvent.pdb".format(path=self.path,itera=self.last_iteration))
# Concatenate complex and system
os.system("cat {path}/complex.pdb {path}/solvent.pdb > {path}/system.pdb".format(path=self.path))
rc,sout,serr=gromacs.editconf(f=self.path+"/system.pdb", o=self.path+"/system_mod.pdb", stdout=False)
os.system("mv {}/system_mod.pdb {}/system.pdb".format(self.path,self.path))
# Make an index of the system
rc,sout,serr=gromacs.make_ndx(f=self.path+"/system.pdb", o=self.path+"/index.ndx", stdout=False, input=('chain {} \n q'.format(self.pep_chain)))
# Copy the topol files of the target chains, which are the same always
if self.target=="protein":
for ch in self.chain_join:
os.system("cp {}/topol_Protein_chain_{}.itp {}/system_Protein_chain_{}.itp".format(self.path,ch,self.path,ch))
if self.target=="drug":
for ch in self.chain_join:
os.system("cp {}/topol_Drug_chain_{}.itp {}/system_Drug_chain_{}.itp".format(self.path,ch,self.path,ch))
# Copy the topol.top to system.top
os.system("cp {}/topol.top {}/system.top".format(self.path,self.path))
os.system("cp {}/complex_Protein_chain_{}.itp {}/system_Protein_chain_{}.itp".format(self.path,self.pep_chain,self.path,self.pep_chain))
os.system("sed -i 's/topol_/system_/g' {}/system.top".format(self.path))
# Select water and ions within 0.2 distance of the residue
rc,sout,serr=gromacs.make_ndx(f=self.path+"/system.pdb", o=self.path+"/index.ndx", stdout=False, input=('chain {}'.format(self.pep_chain),'q'))
rc,sout,serr=gromacs.select(f=self.path+"/system.pdb", n=self.path+"/index.ndx", s=self.path+"/system.pdb", on=self.path+"/index_sol.ndx", stdout=False, select="group Water_and_ions and same residue as within 0.2 of (group ch{} and resnr {})".format(self.pep_chain,self.pep_position))
# Solve the issue with atoms overlapped with the selected residue
values=[x.strip() for x in open(self.path+"/index_sol.ndx")]
atomsSOL=[]
for i,v in enumerate(values):
if i!=0:
info=v.split()
atomsSOL=atomsSOL+info
# List of the atoms that will be deleted
atomsDelete=[]
# Check the overlapped atoms
for a in atomsSOL:
# Obtain the list of atoms from the index file
bash = "awk '$2 == '{}' {{print $6','$7','$8}}' {}/system.pdb".format(a,self.path)
coordinates = subprocess.check_output(['bash','-c', bash])
comp=coordinates.strip().split()
comparison=[]
for c in comp: comparison.append(float(c))
ndComp=np.array(comparison)
distancesSOL=[]
# Read the structure in biopython
parser = PDBParser()
structure = parser.get_structure('PEP', self.path+"/system.pdb")
model = structure[0]
# Check the distances with all the atoms from the selected residue
for residue in model[self.pep_chain]:
resC=residue.get_resname()
resNumber=residue.get_full_id()[3][1]
if resNumber==self.pep_position:
for atom in residue:
idAtom = atom.get_id()
if idAtom[0].isdigit() == False:
if resC=="ILE" and idAtom=="CD": idAtom="CD1"
diff = atom.coord - ndComp
diffValue=np.sqrt(np.sum(diff * diff))
distancesSOL.append(float(diffValue))
# Threshold to determine which atoms can be overlapped
if min(distancesSOL)<1.0:
if a not in atomsDelete: atomsDelete.append(a)
# Selection of the final atoms that will be included in the index
final_index=[]
for element in atomsSOL:
flag=0
for delete in atomsDelete:
if abs(int(element)-int(delete))<=2: flag=1
if flag==0:
final_index.append(element)
# Update of the index sol file
new_index=open(self.path+"/index_sol2.ndx","w")
new_index.write("{}\n".format(values[0]))
group=[]
counter=1
for ele in final_index:
if counter <15:
group.append(ele)
counter+=1
else:
group.append(ele)
new_index.write(" ".join(group)+" \n")
counter=1
group=[]
new_index.write(" ".join(group)+" ")
new_index.close()
# Update the file
os.system("mv {}/index_sol2.ndx {}/index_sol.ndx".format(self.path,self.path))
ref_ndx = NDX()
ref_ndx.read(self.path+"/index.ndx")
bash="grep '\[' {}/index.ndx | wc -l".format(self.path)
number_index = subprocess.check_output(['bash','-c', bash])
index_ref=int(number_index)-1
#index_ref=len(ref_ndx)-1
# Create the side chain index in a template file
os.system("echo 'name 0 overlap' > %s/template" %self.path)
os.system("echo '\"SideChain\" & \"ch{}\" & r {}' >> {}/template".format(self.pep_chain,str(self.pep_position),self.path))
os.system("echo '\"overlap\" | \"SideChain_&_ch{}_&_r_{}\"' >> {}/template".format(self.pep_chain,str(self.pep_position),self.path))
os.system("echo '\"System\" &! \"overlap_SideChain_&_ch{}_&_r_{}\"' >> {}/template".format(self.pep_chain,str(self.pep_position),self.path))
os.system("echo 'q' >> {}/template".format(self.path))
# Create an index joining both created before
os.system("gmx -quiet make_ndx -f {path}/system.pdb -n {path}/index_sol.ndx {path}/index.ndx -o {path}/total_index.ndx < {path}/template".format(path=self.path))
os.system("sed -i 's/System_&_\!overlap_SideChain_&_ch{}_&_r_{}/to_block/g' {}/total_index.ndx".format(self.pep_chain,str(self.pep_position),self.path))
# Generate the gro file
rc,sout,serr=gromacs.editconf(f=self.path+"/system.pdb", o=self.path+"/system.gro", stdout=False)
# Prepare the files for the minimization and run
rc,sout,serr=gromacs.grompp(f=self.path+"/mdp/minim_overlap.mdp", o=self.path+"/systemNEW.tpr", p=self.path+"/system.top", n=self.path+"/total_index.ndx", c=self.path+"/system.gro", stdout=False)
gromacs.utilities.unlink_gmx("mdout.mdp")
print("Running second minimization ...")
rc,sout,serr=gromacs.mdrun(deffnm=self.path+"/systemNEW", stdout=False)
# Copy the system.gro file that will be used to run the last minimization
os.system("cp {}/systemNEW.gro {}/system.gro".format(self.path,self.path))
os.system("sed -i '$ d' {}/system.gro".format(self.path))
os.system("tail -n1 {path}/npt-pbc.gro >> {path}/system.gro".format(path=self.path))
# Delete temporal files
os.system("rm {path}/complex.pdb {path}/solvent.pdb {path}/systemNEW* {path}/template {path}/index.ndx {path}/index_sol.ndx {path}/total_index.ndx *.itp".format(path=self.path))
def test_ensemble_potential_nompi():
"""Test ensemble potential without an ensemble.
Still requires ParallelArrayContext.
"""
import gmx
import os
import myplugin
cwd = os.path.dirname(__file__)
water = os.path.join(cwd, 'data', 'water.gro')
import shutil
shutil.copy(water, './')
# assert False
try:
# use GromacsWrapper if available
import gromacs
import gromacs.formats
from gromacs.tools import Solvate as solvate
solvate(o='water.gro', box=[5, 5, 5])
mdpparams = [('integrator', 'md'), ('nsteps', 1000), ('nstxout', 100),
('nstvout', 100), ('nstfout', 100),
('tcoupl', 'v-rescale'), ('tc-grps', 'System'),
('tau-t', 1), ('ref-t', 298)]
mdp = gromacs.formats.MDP()
for param, value in mdpparams:
mdp[param] = value
mdp.write('water.mdp')
with open('input.top', 'w') as fh:
fh.write("""#include "gromos43a1.ff/forcefield.itp"
#include "gromos43a1.ff/spc.itp"
[ system ]
; Name
spc
[ molecules ]
; Compound #mols
SOL 4055
""")
gromacs.grompp(f='water.mdp',
c='water.gro',
po='water.mdp',
pp='water.top',
o='water.tpr',
p='input.top')
tpr_filename = os.path.abspath('water.tpr')
except:
from gmx.data import tpr_filename
print("Testing plugin potential with input file {}".format(
os.path.abspath(tpr_filename)))
assert gmx.version.api_is_at_least(0, 0, 5)
md = gmx.workflow.from_tpr([tpr_filename], append_output=False)
# Create a WorkElement for the potential
#potential = gmx.core.TestModule()
params = {
'sites': [1, 4],
'nbins': 10,
'binWidth': 0.1,
'min_dist': 0.,
'max_dist': 10.,
'experimental': [1.] * 10,
'nsamples': 1,
'sample_period': 0.001,
'nwindows': 4,
'k': 10000.,
'sigma': 1.
}
potential = gmx.workflow.WorkElement(namespace="myplugin",
operation="ensemble_restraint",
params=params)
# Note that we could flexibly capture accessor methods as workflow elements, too. Maybe we can
# hide the extra Python bindings by letting myplugin.HarmonicRestraint automatically convert
# to a WorkElement when add_dependency is called on it.
potential.name = "ensemble_restraint"
md.add_dependency(potential)
context = gmx.context.ParallelArrayContext(md)
with context as session:
session.run()
def test_failure_raises():
# unknown option
with pytest.raises(gromacs.GromacsError):
gromacs.grompp(y=True)
def energy_minimize(dirname='em',
mdp=config.templates['em.mdp'],
struct='solvate/ionized.gro',
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.gro]
*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` instance; by default we
just try :func:`gromacs.mdrun_d` and :func:`gromacs.mdrun` but a
MDrunner instance 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} ..."
.format(**vars()))
cbook.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.".format(
**vars())
logger.warn(wmsg)
warnings.warn(wmsg, category=BadParameterWarning)
with in_dir(dirname):
unprocessed = 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:
mdrun = run.get_double_or_single_prec_mdrun()
mdrun(**mdrun_args)
else:
if type(mdrunner) is type:
# class
# 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()
else:
# anything with a run() method that takes mdrun arguments...
try:
mdrunner.run(mdrunargs=mdrun_args)
except AttributeError:
logger.error(
"mdrunner: Provide a gromacs.run.MDrunner class or instance or a callback with a run() method"
)
raise TypeError(
"mdrunner: Provide a gromacs.run.MDrunner class or instance or a callback with a run() method"
)
# 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}".format(
**vars()))
return {
'struct': final_struct,
'top': topology,
'mainselection': mainselection,
}
def test2(a, radius, box=100.0, fbase=None):
# l.test2(5.0797,15)
if fbase is None:
fbase = "{}_{}".format(box, radius)
top = "{}.top".format(fbase)
struct = "{}.pdb".format(fbase)
sol = "{}.sol.pdb".format(fbase)
ndx = "{}.ndx".format(fbase)
origin = box / 2.
pts = fcc_sphere(a, radius)
w = writer("{}.pdb".format(fbase))
w.CRYST1([box, box, box, 90.00, 90.00, 90.00])
for index, atom in enumerate(pts):
w.ATOM(serial=index + 1,
name="AU",
resName="NP",
resSeq=1,
chainID="A",
segID="AUNP",
element="AU",
x=atom[0] + origin,
y=atom[1] + origin,
z=atom[2] + origin)
w.close()
#make_index("{}.pdb".format(fbase), "{}.ndx".format(fbase))
with file(top, "w") as t:
t.write(top_src)
t.write("Au {}\n".format(pts.shape[0]))
gromacs.genbox(p=top, cp=struct, cs="spc216.gro", o=sol,
vdwd="0.15") #@UndefinedVariable
rc, out, nothing = gromacs.make_ndx(
f=sol,
n=None,
o=ndx,
stdout=False, #@UndefinedVariable
input=('', '', 'q'))
gromacs.grompp(f="md2.mdp", o="{}.tpr".format(fbase), c=sol, p=top,
n=ndx) #@UndefinedVariable
with file("{}.sh".format(fbase), "w") as f:
f.write("#!/bin/bash\n")
f.write("#PBS -k o\n")
f.write("#PBS -l nodes=1:ppn=12:ccvt,walltime=24:00:00\n")
f.write("#PBS -M [email protected]\n")
f.write("#PBS -m abe\n")
f.write("#PBS -N {}\n".format(fbase))
f.write("#PBS -j oe\n")
f.write("#PBS -q pg\n")
f.write("#PBS -d /N/dc/scratch/somogyie/Au\n")
f.write("mpirun mdrun -deffnm {}".format(fbase))
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 _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 test_ensemble_potential_withmpi():
import gmx
import os
import shutil
import myplugin
from mpi4py import MPI
rank = MPI.COMM_WORLD.Get_rank()
tests_dir = os.path.dirname(__file__)
water = os.path.join(tests_dir, 'data', 'water.gro')
rank_dir = os.path.join(os.getcwd(), str(rank))
os.mkdir(rank_dir)
shutil.copy(water, rank_dir)
# In MPI, this never makes it to grompp. We should get rid of this...
try:
# use GromacsWrapper if available
import gromacs
import gromacs.formats
from gromacs.tools import Solvate as solvate
solvate(o=os.path.join(rank_dir, 'water.gro'), box=[5, 5, 5])
mdpparams = [('integrator', 'md'), ('nsteps', 1000), ('nstxout', 100),
('nstvout', 100), ('nstfout', 100),
('tcoupl', 'v-rescale'), ('tc-grps', 'System'),
('tau-t', 1), ('ref-t', 298)]
mdp = gromacs.formats.MDP()
for param, value in mdpparams:
mdp[param] = value
mdp.write(os.path.join(rank_dir, 'water.mdp'))
with open(os.path.join(rank_dir, 'input.top'), 'w') as fh:
fh.write("""#include "gromos43a1.ff/forcefield.itp"
#include "gromos43a1.ff/spc.itp"
[ system ]
; Name
spc
[ molecules ]
; Compound #mols
SOL 4055
""")
gromacs.grompp(f=os.path.join(rank_dir, 'water.mdp'),
c=os.path.join(rank_dir, 'water.gro'),
po=os.path.join(rank_dir, 'water.mdp'),
pp=os.path.join(rank_dir, 'water.top'),
o=os.path.join(rank_dir, 'water.tpr'),
p=os.path.join(rank_dir, 'input.top'))
tpr_filename = os.path.join(rank_dir, 'water.tpr')
except:
from gmx.data import tpr_filename
logger.info("Testing plugin potential with input file {}".format(
os.path.abspath(tpr_filename)))
assert gmx.version.api_is_at_least(0, 0, 5)
md = gmx.workflow.from_tpr([tpr_filename, tpr_filename],
append_output=False)
# Create a WorkElement for the potential
#potential = gmx.core.TestModule()
params = {
'sites': [1, 4],
'nbins': 10,
'binWidth': 0.1,
'min_dist': 0.,
'max_dist': 10.,
'experimental': [0.5] * 10,
'nsamples': 1,
'sample_period': 0.001,
'nwindows': 4,
'k': 10000.,
'sigma': 1.
}
potential = gmx.workflow.WorkElement(namespace="myplugin",
operation="ensemble_restraint",
params=params)
# Note that we could flexibly capture accessor methods as workflow elements, too. Maybe we can
# hide the extra Python bindings by letting myplugin.HarmonicRestraint automatically convert
# to a WorkElement when add_dependency is called on it.
potential.name = "ensemble_restraint"
before = md.workspec.elements[md.name]
md.add_dependency(potential)
context = gmx.context.ParallelArrayContext(md)
with context as session:
session.run()
#print(np.asarray(substract).shape)
ID = substract.index(min(substract))
index.append(ID)
dist.append(data[ID])
print(index)
print(dist)
# Generate the frames for umbrella sampling
u = mda.Universe(sys.argv[2], sys.argv[3])
for ts in u.trajectory[np.array(index)[:]]:
print("Frame: {0:7d}, timestep: {1:12.4f}".format(ts.frame, ts.time))
groname = "conf" + str(ts.frame) + ".gro"
u.atoms.write(groname)
# creating the dictionary for pairing the elements in index list and other in distance list as key:value
dic = dict(zip(index, dist))
for i, key in enumerate(list(dic.keys())):
if i == 2:
break
print(i, key, dic[key])
shutil.copy(sys.argv[5], "umbrella_cp.mdp")
with fileinput.FileInput("umbrella_cp.mdp", inplace=True,
backup='.bak') as file:
for line in file:
print(line.replace("CHANGE", str(dic[key] / 10)), end='')
gromacs.grompp(f="umbrella_cp.mdp",
c="conf" + str(key) + ".gro",
p="topols.top",
n="index.ndx",
o="umbrella" + str(key) + ".tpr")
def run_minim_complex(self,run_minim=False):
"""
Function to run a local minimization on the side chain that was mutated
Arguments:
run_minim -- boolean flag that will control if the minimization is run or not
Output:
complex.pdb -- new complex pdb with the minimization and the new itp files
"""
# Get the chain with the peptide to generate a novel itp file
os.system("python3 {}/src/scores/get_chains.py {}/complex.pdb {}".format(self.path_scores,self.path,self.path))
rc,sout,serr=gromacs.pdb2gmx(f=self.path+"/complex_"+self.pep_chain+".pdb", p=self.path+"/binder.top", o=self.path+"/complex_"+self.pep_chain+".gro", stdout=False, input=('6','6'))
os.system("sed -i '/forcefield/d' {}/binder.top".format(self.path))
os.system("sed -i '/\[ system \]/,$d' {}/binder.top".format(self.path))
os.system("mv {}/binder.top {}/complex_Protein_chain_{}.itp".format(self.path,self.path,self.pep_chain))
rc,sout,serr=gromacs.editconf(f=self.path+"/complex_"+self.pep_chain+".gro", o=self.path+"/complex_"+self.pep_chain+".pdb", stdout=False)
# Fix the amino acid nomenclature
os.system("for i in ASP ARG HIS HIE HID HIP LYS GLU SER THR ASN GLN CYS CYX GLY PRO ALA VAL ILE LEU MET PHE TYR TRP; do sed -i s/\"$i \"/\"$i {}\"/g {}/complex_{}.pdb; done".format(self.pep_chain,self.path,self.pep_chain))
for i,ch in enumerate(self.chain_join):
if i==0:
os.system("grep ATOM {}/complex_{}.pdb > {}/complex.pdb".format(self.path,ch,self.path))
os.system("echo 'TER' >> {}/complex.pdb".format(self.path))
else:
os.system("grep ATOM {}/complex_{}.pdb >> {}/complex.pdb".format(self.path,ch,self.path))
os.system("echo 'TER' >> {}/complex.pdb".format(self.path))
# Get the new complex.pdb and the peptide chain itp file and delete temporal files
os.system("grep ATOM {}/complex_{}.pdb >> {}/complex.pdb".format(self.path,self.pep_chain,self.path))
os.system("echo 'TER' >> {}/complex.pdb".format(self.path))
os.system("rm {}/complex_*.pdb".format(self.path))
os.system("rm {}/complex_*.gro".format(self.path))
os.system("rm {}/chains.seq".format(self.path))
os.system("head -n -18 {}/complex_Protein_chain_{}.itp > {}/temp; mv {}/temp {}/complex_Protein_chain_{}.itp".format(self.path,self.pep_chain,self.path,self.path,self.path,self.pep_chain))
# Copy the topol files of the target chains, which are the same always
# Copy the topol files of the target chains, which are the same always
if self.target=="protein":
for ch in self.chain_join:
os.system("cp {}/topol_Protein_chain_{}.itp {}/complex_Protein_chain_{}.itp".format(self.path,ch,self.path,ch))
if self.target=="drug":
for ch in self.chain_join:
os.system("cp {}/topol_Drug_chain_{}.itp {}/complex_Drug_chain_{}.itp".format(self.path,ch,self.path,ch))
# Copy the topol.top to complex.top and delete all the additional atoms
os.system("cp {}/topol.top {}/complex.top".format(self.path,self.path))
os.system("sed -i '/Ion/d' {}/complex.top".format(self.path))
os.system("sed -i '/SOL/d' {}/complex.top".format(self.path))
os.system("sed -i '/NA/d' {}/complex.top".format(self.path))
os.system("sed -i '/CL/d' {}/complex.top".format(self.path))
os.system("sed -i '/solvent/d' {}/complex.top".format(self.path))
os.system("sed -i 's/topol_/complex_/g' {}/complex.top".format(self.path))
# Get a pdb of the complex where an index will be created
rc,sout,serr=gromacs.make_ndx(f=self.path+"/complex.pdb", o=self.path+"/reference.ndx", stdout=False, input=('q'))
ref_ndx = NDX()
ref_ndx.read(self.path+"/reference.ndx")
#index_ref=len(ref_ndx)-1
bash="grep '\[' {}/reference.ndx | wc -l".format(self.path)
number_index = subprocess.check_output(['bash','-c', bash])
index_ref=int(number_index)-1
gromacs.utilities.unlink_gmx(self.path+"/reference.ndx")
# Create the side chain index
input_for_ndx=()
counter=index_ref
input_for_ndx+=('chain {}'.format(self.pep_chain),); counter+=1
input_for_ndx+=('name {} binder'.format(counter),);
input_for_ndx+=('"SideChain" & "binder"'+' & r {}'.format(self.pep_position),); counter+=1
input_for_ndx+=('"System" &! {}'.format(counter),); counter+=1
input_for_ndx+=('name {} scmut'.format(counter),);
sentence=""
for i,ch in enumerate(self.chain_join):
if i==0: sentence=sentence+"chain {}".format(ch)
else: sentence=sentence+" | chain {}".format(ch)
input_for_ndx+=(sentence,); counter+=1
input_for_ndx+=('name {} target'.format(counter),)
input_for_ndx+=('\"target\" | \"binder\"',); counter+=1
input_for_ndx+=('name {} complex'.format(counter),)
input_for_ndx+=('q',)
# Generate the index file
rc,sout,serr=gromacs.make_ndx(f=self.path+"/complex.pdb", o=self.path+"/scmut.ndx", stdout=False, input=input_for_ndx)
# Generate the gro file
rc,sout,serr=gromacs.editconf(f=self.path+"/complex.pdb", o=self.path+"/complex.gro", stdout=False)
# Add a small box for the residues
os.system("sed -i '$ d' {}/complex.gro".format(self.path))
os.system('echo " 20.0 20.0 20.0" >> {path}/complex.gro'.format(path=self.path))
# Prepare the files for the minimization
rc,sout,serr=gromacs.grompp(f=self.path+"/mdp/minim_scmut.mdp", o=self.path+"/complex.tpr", p=self.path+"/complex.top", n=self.path+"/scmut.ndx", c=self.path+"/complex.gro", stdout=False)
gromacs.utilities.unlink_gmx("mdout.mdp")
# Run the minimization of the side chain alone and the residues around it
if run_minim:
# Run the minimization
print("Running first minimization ...")
rc,sout,serr=gromacs.mdrun(deffnm=self.path+"/complex", stdout=False)
# Get the complex pdb file
rc,sout,serr=gromacs.trjconv(f=self.path+"/complex.gro",s=self.path+"/complex.tpr", n=self.path+"/scmut.ndx", o=self.path+"/min_complex.pdb",stdout=False,input=("complex"))
os.system("rm posre.itp {path}/complex.tpr {path}/complex.top; grep -v ENDMDL {path}/min_complex.pdb | grep -v MODEL > {path}/complex.pdb; rm {path}/min_complex.pdb {path}/complex.log {path}/complex.trr {path}/complex.edr {path}/scmut.ndx".format(path=self.path))
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,
**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)
cbook.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".
format(**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 = {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 = 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 = 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 test_failure_raises():
# unknown option
with pytest.raises(gromacs.GromacsError):
gromacs.grompp(y=True)