def _get_sequence(mol: Molecule, sel):
sel_mask = mol.atomselect(sel)
protein_mask = mol.atomselect("protein")
nucleic_mask = mol.atomselect("nucleic")
if any(protein_mask & sel_mask) and any(nucleic_mask & sel_mask):
raise RuntimeError(
"Your selection contains both protein and nucleic residues. You need to clarify which selection to align."
)
molseg = "protein" if any(protein_mask) else "nucleic"
seqmol, seqidx = mol.sequence(noseg=True, return_idx=True, sel=sel, _logger=False)
seqidx = seqidx[molseg]
seqmol = seqmol[molseg]
segment_type = molseg
return seqmol, seqidx, segment_type
def test_tmscore(self):
from moleculekit.molecule import Molecule
expectedTMscore = np.array([0.21418524, 0.2367377, 0.23433833, 0.21362964, 0.20935164,
0.20279461, 0.27012895, 0.22675238, 0.21230793, 0.2372011])
expectedRMSD = np.array([3.70322128, 3.43637027, 3.188193, 3.84455877, 3.53053882,
3.46781854, 2.93777629, 2.97978692, 2.70792428, 2.63051318])
mol = Molecule(os.path.join(home(dataDir='tmscore'), 'filtered.pdb'))
mol.read(os.path.join(home(dataDir='tmscore'), 'traj.xtc'))
ref = Molecule(os.path.join(home(dataDir='tmscore'), 'ntl9_2hbb.pdb'))
tmscore, rmsd = molTMscore(mol, ref, mol.atomselect('protein'), ref.atomselect('protein'))
self.assertTrue(np.allclose(tmscore, expectedTMscore))
self.assertTrue(np.allclose(rmsd, expectedRMSD))
def _filtSim(i, sims, outFolder, filterSel):
name = _simName(sims[i].trajectory[0])
directory = path.join(outFolder, name)
if not path.exists(directory):
makedirs(directory)
logger.debug("Processing trajectory " + name)
fmolfile = [
path.join(outFolder, "filtered.psf"),
path.join(outFolder, "filtered.pdb"),
]
(traj, outtraj) = _renameSims(sims[i].trajectory, name, outFolder)
if not traj:
ftrajectory = _autoDetectTrajectories(path.join(outFolder, name))
numframes = _getNumFrames(sims[i], ftrajectory)
return Sim(
simid=sims[i].simid,
parent=sims[i],
input=None,
trajectory=ftrajectory,
molfile=fmolfile,
numframes=numframes,
)
try:
from moleculekit.molecule import Molecule
mol = Molecule(sims[i].molfile)
except Exception as e:
logger.warning("Error! Skipping simulation " + name +
" due to error: " + str(e))
return
sel = mol.atomselect(filterSel)
for j in range(0, len(traj)):
try:
mol.read(traj[j])
except IOError as e:
logger.warning("{}, skipping trajectory".format(e))
break
mol.write(outtraj[j], sel)
ftrajectory = _autoDetectTrajectories(path.join(outFolder, name))
numframes = _getNumFrames(sims[i], ftrajectory)
return Sim(
simid=sims[i].simid,
parent=sims[i],
input=None,
trajectory=ftrajectory,
molfile=fmolfile,
numframes=numframes,
)
def write(self, inputdir, outputdir):
""" Writes the production protocol and files into a folder.
Parameters
----------
inputdir : str
Path to a directory containing the files produced by a equilibration process.
outputdir : str
Directory where to write the production setup files.
"""
from moleculekit.molecule import Molecule
# Do version consistency check
if (self._version == 2 and not isinstance(self.acemd, Acemd2)) and \
(self._version == 3 and not isinstance(self.acemd, Acemd)):
raise RuntimeError(
'Acemd object version ({}) inconsistent with protocol version at instantiation '
'({})'.format(type(self.acemd), self._version))
self._findFiles(inputdir)
self._amberFixes()
if self._version == 2:
self.acemd.temperature = str(self.temperature)
self.acemd.langevintemp = str(self.temperature)
elif self._version == 3:
self.acemd.temperature = self.temperature
self.acemd.thermostattemp = self.temperature
from htmd.units import convert
numsteps = convert(self.timeunits,
'timesteps',
self.runtime,
timestep=self.acemd.timestep)
if self._version == 3:
self.acemd.run = str(numsteps)
pdbfile = os.path.join(inputdir, self.acemd.coordinates)
inmol = Molecule(pdbfile)
if np.any(inmol.atomselect('lipids')) and not self.useconstantratio:
logger.warning(
'Lipids detected in input structure. We highly recommend setting useconstantratio=True '
'for membrane simulations.')
if self._version == 2:
if self.restraints:
raise RuntimeWarning(
'restraints are only available on {}(_version=3)'.format(
self.__class__.__name__))
if self.fb_k > 0: # use TCL only for flatbottom
self.acemd.tclforces = 'on'
if isinstance(self.acemd.TCL, tuple):
tcl = list(self.acemd.TCL)
tcl[0] = tcl[0].format(
NUMSTEPS=numsteps,
TEMPERATURE=self.temperature,
KCONST=self.fb_k,
REFINDEX=' '.join(
map(str, inmol.get('index', self.fb_reference))),
SELINDEX=' '.join(
map(str, inmol.get('index', self.fb_selection))),
BOX=' '.join(map(str, self.fb_box)))
self.acemd.TCL = tcl[0] + tcl[1]
else:
logger.warning(
'{} default TCL was already formatted.'.format(
self.__class__.__name__))
else:
self.acemd.TCL = 'set numsteps {NUMSTEPS}\n'.format(
NUMSTEPS=numsteps)
if self.useconstraints:
# Turn on constraints
self.acemd.constraints = 'on'
self.acemd.constraintscaling = '1.0'
else:
if len(self.constraints) != 0:
logger.warning(
'You have setup constraints to {} but constraints are turned off. '
'If you want to use constraints, define '
'useconstraints=True'.format(self.constraints))
elif self._version == 3:
if self.restraints is not None:
logger.info(
'Using user-provided restraints and ignoring constraints and fb_potential'
)
self.acemd.restraints = self.restraints
else:
restraints = list()
if self.fb_k > 0:
logger.warning(
'Converting fb_potential to restraints. This is a convenience '
'functional conversion. We recommend start using restraints with '
'{}(_version=3)'.format(self.__class__.__name__))
restraints += self._fb_potential2restraints(inputdir)
if self.useconstraints:
logger.warning(
'Converting constraints to restraints. This is a convenience '
'functional conversion. We recommend start using restraints with '
'{}(_version=3)'.format(self.__class__.__name__))
restraints += self._constraints2restraints()
else:
if len(self.constraints) != 0:
logger.warning(
'You have setup constraints to {} but constraints are turned off. '
'If you want to use constraints, define '
'useconstraints=True'.format(self.constraints))
if len(restraints) != 0:
self.acemd.restraints = restraints
if self.useconstantratio:
self.acemd.useconstantratio = 'on'
if self.adaptive:
self.acemd.binvelocities = None
self.acemd.setup(inputdir, outputdir, overwrite=True)
if self._version == 2:
# Adding constraints by writing them to the consref file
if self.useconstraints:
inconsreffile = os.path.join(inputdir, self.acemd.consref)
consrefmol = Molecule(inconsreffile)
consrefmol.set('occupancy', 0)
consrefmol.set('beta', 0)
if len(self.constraints) == 0:
raise RuntimeError(
'You have set the production to use constraints (useconstraints=True), but have '
'not defined any constraints (constraints={}).')
else:
for sel in self.constraints:
consrefmol.set('beta', self.constraints[sel], sel)
outconsreffile = os.path.join(outputdir, self.acemd.consref)
consrefmol.write(outconsreffile)
def write(self, inputdir, outputdir):
""" Writes the production protocol and files into a folder.
Parameters
----------
inputdir : str
Path to a directory containing the files produced by a equilibration process.
outputdir : str
Directory where to write the production setup files.
"""
self._findFiles(inputdir)
self._amberFixes()
from htmd.units import convert
numsteps = convert(self.timeunits,
'timesteps',
self.runtime,
timestep=self.acemd.timestep)
pdbfile = os.path.join(inputdir, self.acemd.coordinates)
inmol = Molecule(pdbfile)
if np.any(inmol.atomselect('lipids')) and not self.useconstantratio:
logger.warning(
'Lipids detected in input structure. We highly recommend setting useconstantratio=True '
'for membrane simulations.')
if self.fb_k > 0: # use TCL only for flatbottom
self.acemd.tclforces = 'on'
tcl = list(self.acemd.TCL)
tcl[0] = tcl[0].format(
NUMSTEPS=numsteps,
TEMPERATURE=self.temperature,
KCONST=self.fb_k,
REFINDEX=' '.join(
map(str, inmol.get('index', self.fb_reference))),
SELINDEX=' '.join(
map(str, inmol.get('index', self.fb_selection))),
BOX=' '.join(map(str, self.fb_box)))
self.acemd.TCL = tcl[0] + tcl[1]
else:
self.acemd.TCL = 'set numsteps {NUMSTEPS}\n' \
'set temperature {TEMPERATURE}\n'.format(NUMSTEPS=numsteps, TEMPERATURE=self.temperature)
if self.useconstraints:
# Turn on constraints
self.acemd.constraints = 'on'
self.acemd.constraintscaling = '1.0'
else:
if len(self.constraints) != 0:
logger.warning(
'You have setup constraints to {} but constraints are turned off. '
'If you want to use constraints, define useconstraints=True'
.format(self.constraints))
if self.useconstantratio:
self.acemd.useconstantratio = 'on'
if self.adaptive:
self.acemd.binvelocities = None
self.acemd.setup(inputdir, outputdir, overwrite=True)
# Adding constraints by writing them to the consref file
if self.useconstraints:
inconsreffile = os.path.join(inputdir, self.acemd.consref)
consrefmol = Molecule(inconsreffile)
consrefmol.set('occupancy', 0)
consrefmol.set('beta', 0)
if len(self.constraints) == 0:
raise RuntimeError(
'You have set the production to use constraints (useconstraints=True), but have not '
'defined any constraints (constraints={}).')
else:
for sel in self.constraints:
consrefmol.set('beta', self.constraints[sel], sel)
outconsreffile = os.path.join(outputdir, self.acemd.consref)
consrefmol.write(outconsreffile)
class MutualInformation:
def __init__(self, model, mol=None, fstep=0.1, skip=1):
""" Class that calculates the mutual information of protein residues.
Parameters
----------
model : :class:`Model <htmd.model.Model>` object
A Model object with a calculated MSM
mol : :class:`Molecule <moleculekit.molecule.Molecule>` object
A reference molecule from which to obtain structural information. By default model.data.simlist[0].molfile
will be used.
fstep : float
The frame step of the simulations
skip : int
Frame skipping
Examples
--------
>>> from htmd.mutualinformation import MutualInformation
>>> from htmd.ui import *
>>>
>>> sims = simlist(glob('./filtered/*/'), './filtered/filtered.pdb')
>>> mol = Molecule('./filtered/filtered.pdb')
>>>
>>> metr = Metric(sims)
>>> metr.set(MetricDihedral())
>>> datadih = metr.project()
>>> datadih.fstep = 0.1
>>> datadih.dropTraj()
>>>
>>> metr = Metric(datadih.simlist)
>>> metr.set(MetricSelfDistance('protein and name CA', metric='contacts'))
>>> dataco = metr.project()
>>> dataco.fstep = 0.1
>>> dataco.dropTraj()
>>>
>>> tica = TICA(datadih, 20, units='ns')
>>> datatica = tica.project(3)
>>> datatica.cluster(MiniBatchKMeans(n_clusters=1500))
>>> model = Model(datatica)
>>> model.markovModel(12, 4, units='ns')
>>>
>>> mu = MutualInformation(model)
>>> mu.calculate()
>>> mu.saveMI('./mi_matrix.npy')
>>> mu.weightGraph(dataco, 0.005)
>>> mu.save_graphml('./weightgraph.graphml')
"""
self.model = model
self.mol = mol
if mol is None:
self.mol = Molecule(self.model.data.simlist[0].molfile)
self._computeChiDihedrals(fstep=fstep, skip=skip)
self.bindihcat = self._histogram() # Lasts two minutes
self.resids = self.mol.get('resid', 'name CA')
self.residmap = np.ones(self.mol.resid.max() + 1, dtype=int) * -1
self.residmap[self.resids] = np.arange(len(self.resids))
self.mi_matrix = None
self.graph_array = None
self.graph = None
def calculate(self, njobs=1):
"""
Parameters
----------
njobs : int
Number of parallel jobs to spawn for the calculation of MI
"""
from htmd.config import _config
from htmd.parallelprogress import ParallelExecutor
numchi = self.chi.numDimensions
statdist = self.model.msm.stationary_distribution
stconcat = np.concatenate(self.model.data.St)
microcat = self.model.micro_ofcluster[stconcat]
aprun = ParallelExecutor(n_jobs=njobs)
res = aprun(total=numchi, desc='Calculating MI')(delayed(self._parallelAll)(numchi, dih1, 4, self.model.micronum, self.bindihcat, microcat, statdist) for dih1 in range(numchi))
MI_all = np.zeros((len(self.resids), len(self.resids)))
for r in res:
dihcounts = r[0]
pairs = r[1]
for dihc, p in zip(dihcounts, pairs):
dih1, dih2 = p
if dih1 == dih2:
continue
resid1 = self.residmap[self.mol.resid[self.chi.description.atomIndexes[dih1][0]]]
resid2 = self.residmap[self.mol.resid[self.chi.description.atomIndexes[dih2][0]]]
MI_all[resid1][resid2] = self._calcMutualInfo(dihc)
self.mi_matrix = self._cleanautocorrelations(MI_all)
def _computeChiDihedrals(self, fstep=0.1, skip=1):
chis = []
protmol = self.mol.copy()
protmol.filter('protein')
caidx = self.mol.atomselect('protein and name CA')
resids = self.mol.resid[caidx]
resnames = self.mol.resname[caidx]
for residue, resname in zip(resids, resnames):
ch = Dihedral.chi1(protmol, residue)
if ch is not None:
chis.append(ch)
metr = Metric(self.model.data.simlist, skip=skip)
metr.set(MetricDihedral(chis, sincos=False))
data = metr.project()
data.fstep = fstep
self.chi = data
def _histogram(self):
condata = np.concatenate(self.chi.dat)
bins = np.array([-180, -150, -90, -30, 0, 30, 90, 150, 180])
dic = {1: 3, 2: 0, 3: 1, 4: 0, 5: 0, 6: 2, 7: 0, 8: 3, 9:3}
binneddih = np.zeros([condata.shape[0], condata.shape[1]])
for dihedral in range(condata.shape[1]):
binning = np.digitize(condata[:, dihedral], bins)
binneddih[:, dihedral] = [dic[n] if n in dic.keys() else n for n in binning]
return binneddih
def _calcMutualInfo(self, contingency):
# Ripped out of sklearn since it converts floats to integers without warning which breaks our use-case
from math import log
nzx, nzy = np.nonzero(contingency)
nz_val = contingency[nzx, nzy]
contingency_sum = contingency.sum()
pi = np.ravel(contingency.sum(axis=1))
pj = np.ravel(contingency.sum(axis=0))
log_contingency_nm = np.log(nz_val)
contingency_nm = nz_val / contingency_sum
# Don't need to calculate the full outer product, just for non-zeroes
outer = pi.take(nzx) * pj.take(nzy)
log_outer = -np.log(outer) + log(pi.sum()) + log(pj.sum())
mi = (contingency_nm * (log_contingency_nm - log(contingency_sum)) +
contingency_nm * log_outer)
return mi.sum()
def _parallelAll(self, numdih, dih1, numbins, micronum, bindihcat, microcat, stat_dist):
results = []
resultpairs = []
for dih2 in range(dih1, numdih):
dihcounts = np.zeros((numbins, numbins, micronum))
# Find pairs of dihedrals (keys) and which absolute frames they occur in (vals)
df = pd.DataFrame({'a': list(zip(bindihcat[:, dih1], bindihcat[:, dih2]))})
gg = df.groupby(by=df.a).groups
for pair in gg:
microsofpairs = microcat[gg[pair]] # Get the microstates of all frames having given dihedral pair
microsofpairs = np.delete(microsofpairs, np.where(microsofpairs == -1)[0]) # Delete dropped clusters
counts = np.bincount(microsofpairs) # Count number of frames with that pair for each microstate (0, max_micro_seen)
dihcounts[int(pair[0]), int(pair[1]), :len(counts)] += counts # Add the counts
dihcounts /= dihcounts.sum(axis=0).sum(axis=0) # Normalize all slices by total counts in each microstate
dihcounts *= stat_dist # Multiply by stationary distribution of each state
dihcounts = np.sum(dihcounts, axis=2) # Calculate the weighted sum over all states
results.append(dihcounts)
resultpairs.append((dih1, dih2))
return results, resultpairs
def _cleanautocorrelations(self, mi):
# TODO: vectorize this
# np.diag(np.ones(3), k=-1).astype(bool) | np.diag(np.ones(3), k=1).astype(bool) | np.diag(np.ones(4)).astype(bool)
for i in range(mi.shape[0]):
for j in range(mi.shape[1]):
if abs(i - j) < 2:
mi[i][j] = 0
return mi
def saveMI(self, path):
np.save(path, self.mi_matrix)
def loadMI(self, path):
self.mi_matrix = np.load(path)
def weightGraph(self, datacontacts, mi_threshold, time_treshold=0.6):
if len(self.mol.get('resid', 'name CA')) != len(self.resids):
raise Exception('The length of the protein doesn\'t match the Mutual Information data')
contactcat = np.concatenate(datacontacts.dat)
contacts_matrix = np.zeros([len(self.resids), len(self.resids)])
for i in range(contactcat.shape[1]):
counter = np.count_nonzero(contactcat[:, i])
resid1 = self.residmap[self.mol.resid[datacontacts.description.atomIndexes[i][0]]]
resid2 = self.residmap[self.mol.resid[datacontacts.description.atomIndexes[i][1]]]
contacts_matrix[resid1][resid2] = counter
self.graph_array = np.zeros([contacts_matrix.shape[0], contacts_matrix.shape[0]])
mask = (self.mi_matrix > mi_threshold) & (contacts_matrix > (time_treshold * contactcat.shape[0]))
self.graph_array[mask] = self.mi_matrix[mask]
intermed = []
for source in range(self.graph_array.shape[0]):
for target in range(source, self.graph_array.shape[1]):
if self.graph_array[source, target] != 0 and target > source:
intermed.append(
[int(self.resids[source]), int(self.resids[target]), float(self.graph_array[source, target])])
import pandas as pd
import networkx as nx
from sklearn.cluster.spectral import SpectralClustering
pd = pd.DataFrame(intermed, columns=['source', 'target', 'weight'])
pd[['source', 'target']] = pd[['source', 'target']].astype(type('int', (int,), {}))
pd['weight'] = pd['weight'].astype(type('float', (float,), {}))
G = nx.from_pandas_edgelist(pd, 'source', 'target', 'weight')
## setSegment
segids = self.mol.get('segid', 'name CA')
seg_res_dict = {key: value for (key, value) in zip(self.resids, segids) if
np.any(pd.loc[(pd['source'] == key)].index) or np.any(pd.loc[(pd['target'] == key)].index)}
nx.set_node_attributes(G, seg_res_dict, 'Segment')
## set
if not nx.is_connected(G):
G = max(nx.connected_component_subgraphs(G), key=len)
flow_cent = nx.current_flow_betweenness_centrality(G, weight='weight')
nx.set_node_attributes(G, flow_cent, 'flowcent')
Spectre = SpectralClustering(n_clusters=10, affinity='precomputed')
model = Spectre.fit_predict(self.graph_array)
model = model.astype(type('float', (float,), {}))
spectral_dict = {key: value for (key, value) in zip(self.resids, model) if key in G.nodes()}
nx.set_node_attributes(G, spectral_dict, 'spectral')
self.graph = G
def save_graphml(self, path):
import networkx as nx
nx.write_graphml(self.graph, path)
def build(mol,
ff=None,
topo=None,
param=None,
prefix='structure',
outdir='./build',
caps=None,
ionize=True,
saltconc=0,
saltanion=None,
saltcation=None,
disulfide=None,
teleap=None,
teleapimports=None,
execute=True,
atomtypes=None,
offlibraries=None,
gbsa=False,
igb=2):
""" Builds a system for AMBER
Uses tleap to build a system for AMBER. Additionally it allows the user to ionize and add disulfide bridges.
Parameters
----------
mol : :class:`Molecule <moleculekit.molecule.Molecule>` object
The Molecule object containing the system
ff : list of str
A list of leaprc forcefield files.
Use :func:`amber.listFiles <htmd.builder.amber.listFiles>` to get a list of available forcefield files.
Default: :func:`amber.defaultFf <htmd.builder.amber.defaultFf>`
topo : list of str
A list of topology `prepi/prep/in` files.
Use :func:`amber.listFiles <htmd.builder.amber.listFiles>` to get a list of available topology files.
Default: :func:`amber.defaultTopo <htmd.builder.amber.defaultTopo>`
param : list of str
A list of parameter `frcmod` files.
Use :func:`amber.listFiles <htmd.builder.amber.listFiles>` to get a list of available parameter files.
Default: :func:`amber.defaultParam <htmd.builder.amber.defaultParam>`
prefix : str
The prefix for the generated pdb and psf files
outdir : str
The path to the output directory
Default: './build'
caps : dict
A dictionary with keys segids and values lists of strings describing the caps for a particular protein segment.
e.g. caps['P'] = ['ACE', 'NME'] or caps['P'] = ['none', 'none']. Default: will apply ACE and NME caps to every
protein segment.
ionize : bool
Enable or disable ionization
saltconc : float
Salt concentration to add to the system after neutralization.
saltanion : {'Cl-'}
The anion type. Please use only AMBER ion atom names.
saltcation : {'Na+', 'K+', 'Cs+'}
The cation type. Please use only AMBER ion atom names.
disulfide : list of pairs of atomselection strings
If None it will guess disulfide bonds. Otherwise provide a list pairs of atomselection strings for each pair of
residues forming the disulfide bridge.
teleap : str
Path to teLeap executable used to build the system for AMBER
teleapimports : list
A list of paths to pass to teLeap '-I' flag, i.e. directories to be searched
Default: determined from :func:`amber.defaultAmberHome <htmd.builder.amber.defaultAmberHome>` and
:func:`amber.htmdAmberHome <htmd.builder.amber.htmdAmberHome>`
execute : bool
Disable building. Will only write out the input script needed by tleap. Does not include ionization.
atomtypes : list of triplets
Custom atom types defined by the user as ('type', 'element', 'hybrid') triplets
e.g. (('C1', 'C', 'sp2'), ('CI', 'C', 'sp3')). Check `addAtomTypes` in AmberTools docs.
offlibraries : str or list
A path or a list of paths to OFF library files. Check `loadOFF` in AmberTools docs.
gbsa : bool
Modify radii for GBSA implicit water model
igb : int
GB model. Select: 1 for mbondi, 2 and 5 for mbondi2, 7 for bondi and 8 for mbondi3.
Check section 4. The Generalized Born/Surface Area Model of the AMBER manual.
Returns
-------
molbuilt : :class:`Molecule <moleculekit.molecule.Molecule>` object
The built system in a Molecule object
Example
-------
>>> from htmd.ui import * # doctest: +SKIP
>>> mol = Molecule("3PTB")
>>> molbuilt = amber.build(mol, outdir='/tmp/build') # doctest: +SKIP
>>> # More complex example
>>> disu = [['segid P and resid 157', 'segid P and resid 13'], ['segid K and resid 1', 'segid K and resid 25']]
>>> molbuilt = amber.build(mol, outdir='/tmp/build', saltconc=0.15, disulfide=disu) # doctest: +SKIP
"""
# Remove pdb protein bonds as they can be regenerated by tleap. Keep non-protein bonds i.e. for ligands
mol = mol.copy()
_removeProteinBonds(mol)
if teleap is None:
teleap = _findTeLeap()
else:
if shutil.which(teleap) is None:
raise NameError(
'Could not find executable: `{}` in the PATH. Cannot build for AMBER.'
.format(teleap))
if not os.path.isdir(outdir):
os.makedirs(outdir)
_cleanOutDir(outdir)
if ff is None:
ff = defaultFf()
if topo is None:
topo = defaultTopo()
if param is None:
param = defaultParam()
if caps is None:
caps = _defaultProteinCaps(mol)
_missingSegID(mol)
_checkMixedSegment(mol)
mol = _charmmLipid2Amber(mol)
_applyProteinCaps(mol, caps)
f = open(os.path.join(outdir, 'tleap.in'), 'w')
f.write('# tleap file generated by amber.build\n')
# Printing out the forcefields
for i, force in enumerate(ensurelist(ff)):
if not os.path.isfile(force):
force = _locateFile(force, 'ff', teleap)
if force is None:
continue
newname = 'ff{}_{}'.format(i, os.path.basename(force))
shutil.copy(force, os.path.join(outdir, newname))
f.write('source {}\n'.format(newname))
f.write('\n')
if gbsa:
gbmodels = {
1: 'mbondi',
2: 'mbondi2',
5: 'mbondi2',
7: 'bondi',
8: 'mbondi3'
}
f.write('set default PBradii {}\n\n'.format(gbmodels[igb]))
# Adding custom atom types
if atomtypes is not None:
atomtypes = ensurelist(tocheck=atomtypes[0], tomod=atomtypes)
f.write('addAtomTypes {\n')
for at in atomtypes:
if len(at) != 3:
raise RuntimeError(
'Atom type definitions have to be triplets. Check the AMBER documentation.'
)
f.write(' {{ "{}" "{}" "{}" }}\n'.format(at[0], at[1], at[2]))
f.write('}\n\n')
# Loading OFF libraries
if offlibraries is not None:
offlibraries = ensurelist(offlibraries)
for i, off in enumerate(offlibraries):
if not os.path.isfile(off):
raise RuntimeError(
'Could not find off-library in location {}'.format(off))
newname = 'offlib{}_{}'.format(i, os.path.basename(off))
shutil.copy(off, os.path.join(outdir, newname))
f.write('loadoff {}\n'.format(newname))
# Loading frcmod parameters
f.write('# Loading parameter files\n')
for i, p in enumerate(param):
if not os.path.isfile(p):
p = _locateFile(p, 'param', teleap)
if p is None:
continue
newname = 'param{}_{}'.format(i, os.path.basename(p))
shutil.copy(p, os.path.join(outdir, newname))
f.write('loadamberparams {}\n'.format(newname))
f.write('\n')
# Loading prepi topologies
f.write('# Loading prepi topologies\n')
for i, t in enumerate(topo):
if not os.path.isfile(t):
t = _locateFile(t, 'topo', teleap)
if t is None:
continue
newname = 'topo{}_{}'.format(i, os.path.basename(t))
shutil.copy(t, os.path.join(outdir, newname))
f.write('loadamberprep {}\n'.format(newname))
f.write('\n')
f.write('# Loading the system\n')
f.write('mol = loadpdb input.pdb\n\n')
if np.sum(mol.atomtype != '') != 0:
logger.debug('Writing mol2 files for input to tleap.')
segs = np.unique(mol.segid[mol.atomtype != ''])
combstr = 'mol = combine {mol'
for s in segs:
name = 'segment{}'.format(s)
mol2name = os.path.join(outdir, '{}.mol2'.format(name))
mol.write(mol2name, (mol.atomtype != '') & (mol.segid == s))
if not os.path.isfile(mol2name):
raise NameError(
'Could not write a mol2 file out of the given Molecule.')
f.write('# Loading the rest of the system\n')
f.write('{} = loadmol2 {}.mol2\n\n'.format(name, name))
combstr += ' {}'.format(name)
combstr += '}\n\n'
f.write(combstr)
# Write patches for disulfide bonds (only after ionizing)
if not ionize:
# TODO: Remove this once we deprecate the class
from htmd.builder.builder import DisulfideBridge
from moleculekit.molecule import UniqueResidueID
if disulfide is not None and len(disulfide) != 0 and isinstance(
disulfide[0], DisulfideBridge):
newdisu = []
for d in disulfide:
r1 = UniqueResidueID.fromMolecule(
mol, 'resid {} and segname {}'.format(d.resid1, d.segid1))
r2 = UniqueResidueID.fromMolecule(
mol, 'resid {} and segname {}'.format(d.resid2, d.segid2))
newdisu.append([r1, r2])
disulfide = newdisu
# TODO: Remove up to here ----------------------
if disulfide is not None and len(disulfide) != 0 and isinstance(
disulfide[0][0], str):
disulfide = convertDisulfide(mol, disulfide)
if disulfide is None:
logger.info('Detecting disulfide bonds.')
disulfide = detectDisulfideBonds(mol)
# Fix structure to match the disulfide patching
if len(disulfide) != 0:
torem = np.zeros(mol.numAtoms, dtype=bool)
f.write('# Adding disulfide bonds\n')
for d in disulfide:
# Rename the residues to CYX if there is a disulfide bond
atoms1 = d[0].selectAtoms(mol, indexes=False)
atoms2 = d[1].selectAtoms(mol, indexes=False)
mol.resname[atoms1] = 'CYX'
mol.resname[atoms2] = 'CYX'
# Remove (eventual) HG hydrogens on these CYS (from proteinPrepare)
torem |= (atoms1 & (mol.name == 'HG')) | (atoms2 &
(mol.name == 'HG'))
# Convert to stupid amber residue numbering
uqseqid = sequenceID(
(mol.resid, mol.insertion, mol.segid)) + mol.resid[0]
uqres1 = int(np.unique(uqseqid[atoms1]))
uqres2 = int(np.unique(uqseqid[atoms2]))
f.write('bond mol.{}.SG mol.{}.SG\n'.format(uqres1, uqres2))
f.write('\n')
mol.remove(torem, _logger=False)
# Calculate the bounding box and store it in the CRD file
f.write('setBox mol "vdw"\n\n')
f.write('# Writing out the results\n')
f.write('saveamberparm mol ' + prefix + '.prmtop ' + prefix + '.crd\n')
f.write('quit')
f.close()
# Printing and loading the PDB file. AMBER can work with a single PDB file if the segments are separate by TER
logger.debug('Writing PDB file for input to tleap.')
pdbname = os.path.join(outdir, 'input.pdb')
# mol2 files have atomtype, here we only write parts not coming from mol2
# We need to write the input.pdb at the end since we modify the resname for disulfide bridges in mol
mol.write(pdbname, mol.atomtype == '')
if not os.path.isfile(pdbname):
raise NameError(
'Could not write a PDB file out of the given Molecule.')
molbuilt = None
if execute:
if not teleapimports:
teleapimports = []
# Source default Amber (i.e. the same paths tleap imports)
amberhome = defaultAmberHome(teleap=teleap)
teleapimports += [
os.path.join(amberhome, s)
for s in _defaultAmberSearchPaths.values()
]
if len(teleapimports) == 0:
raise RuntimeWarning(
'No default Amber force-field found. Check teLeap location: {}'
.format(teleap))
# Source HTMD Amber paths that contain ffs
htmdamberdir = htmdAmberHome()
teleapimports += [
os.path.join(htmdamberdir, os.path.dirname(f)) for f in ff
if os.path.isfile(os.path.join(htmdamberdir, f))
]
if len(teleapimports) == 0:
raise RuntimeError(
'No default Amber force-field imports found. Check '
'`htmd.builder.amber.defaultAmberHome()` and `htmd.builder.amber.htmdAmberHome()`'
)
# Set import flags for teLeap
teleapimportflags = []
for p in teleapimports:
teleapimportflags.append('-I')
teleapimportflags.append(str(p))
logpath = os.path.abspath(os.path.join(outdir, 'log.txt'))
logger.info('Starting the build.')
currdir = os.getcwd()
os.chdir(outdir)
f = open(logpath, 'w')
try:
cmd = [teleap, '-f', './tleap.in']
cmd[1:1] = teleapimportflags
logger.debug(cmd)
call(cmd, stdout=f)
except:
raise NameError('teLeap failed at execution')
f.close()
errors = _logParser(logpath)
os.chdir(currdir)
if errors:
raise BuildError(errors + [
'Check {} for further information on errors in building.'.
format(logpath)
])
logger.info('Finished building.')
if os.path.exists(os.path.join(outdir, 'structure.crd')) and \
os.path.getsize(os.path.join(outdir, 'structure.crd')) != 0 and \
os.path.getsize(os.path.join(outdir, 'structure.prmtop')) != 0:
try:
molbuilt = Molecule(os.path.join(outdir, 'structure.prmtop'))
molbuilt.read(os.path.join(outdir, 'structure.crd'))
except Exception as e:
raise RuntimeError(
'Failed at reading structure.prmtop/structure.crd due to error: {}'
.format(e))
else:
raise BuildError(
'No structure pdb/prmtop file was generated. Check {} for errors in building.'
.format(logpath))
if ionize:
shutil.move(os.path.join(outdir, 'structure.crd'),
os.path.join(outdir, 'structure.noions.crd'))
shutil.move(os.path.join(outdir, 'structure.prmtop'),
os.path.join(outdir, 'structure.noions.prmtop'))
totalcharge = np.sum(molbuilt.charge)
nwater = np.sum(molbuilt.atomselect('water and noh'))
anion, cation, anionatom, cationatom, nanion, ncation = ionizef(
totalcharge,
nwater,
saltconc=saltconc,
anion=saltanion,
cation=saltcation)
newmol = ionizePlace(mol, anion, cation, anionatom, cationatom,
nanion, ncation)
# Redo the whole build but now with ions included
return build(newmol,
ff=ff,
topo=topo,
param=param,
prefix=prefix,
outdir=outdir,
caps={},
ionize=False,
execute=execute,
saltconc=saltconc,
disulfide=disulfide,
teleap=teleap,
atomtypes=atomtypes,
offlibraries=offlibraries)
tmpbonds = molbuilt.bonds
molbuilt.bonds = [] # Removing the bonds to speed up writing
molbuilt.write(os.path.join(outdir, 'structure.pdb'))
molbuilt.bonds = tmpbonds # Restoring the bonds
detectCisPeptideBonds(molbuilt) # Warn in case of cis bonds
return molbuilt
def write(self, inputdir, outputdir):
"""Write the equilibration protocol
Writes the equilibration protocol and files into a folder for execution
using files inside the inputdir directory
Parameters
----------
inputdir : str
Path to a directory containing the files produced by a build process.
outputdir : str
Directory where to write the equilibration setup files.
Examples
--------
>>> md = Equilibration()
>>> md.write('./build','./equil')
"""
from moleculekit.molecule import Molecule
self._findFiles(inputdir)
self._amberFixes()
from htmd.units import convert
numsteps = convert(self.timeunits,
"timesteps",
self.runtime,
timestep=self.acemd.timestep)
self.acemd.temperature = self.temperature
self.acemd.thermostattemperature = self.temperature
self.acemd.run = str(numsteps)
pdbfile = os.path.join(inputdir, self.acemd.coordinates)
inmol = Molecule(pdbfile)
from htmd.builder.builder import detectCisPeptideBonds
detectCisPeptideBonds(inmol)
if np.any(inmol.atomselect("lipids")) and not self.useconstantratio:
logger.warning(
"Lipids detected in input structure. We highly recommend setting useconstantratio=True "
"for membrane simulations.")
if self.restraintsteps is None:
constrsteps = int(numsteps / 2)
else:
constrsteps = int(self.restraintsteps)
if self.restraints is not None:
self.acemd.restraints = self.restraints
else:
self.acemd.restraints = self.defaultEquilRestraints(constrsteps,
mol=inmol)
if self.acemd.boxsize is None and self.acemd.extendedsystem is None:
coords = inmol.get("coords", sel="water")
if coords.size == 0: # It's a vacuum simulation
coords = inmol.get("coords", sel="all")
dim = np.max(coords, axis=0) - np.min(coords, axis=0)
dim += 12.0
else:
dim = np.max(coords, axis=0) - np.min(coords, axis=0)
self.acemd.boxsize = "{} {} {}".format(dim[0], dim[1], dim[2])
if self.useconstantratio:
self.acemd.useconstantratio = "on"
self.acemd.setup(inputdir, outputdir, overwrite=True)
def write(self, inputdir, outputdir, cisbondcheck=True):
"""Writes the production protocol and files into a folder.
Parameters
----------
inputdir : str
Path to a directory containing the files produced by a equilibration process.
outputdir : str
Directory where to write the production setup files.
cisbondcheck : bool
Set to False to disable the cis peptidic bond checks
"""
from moleculekit.molecule import Molecule
from htmd.units import convert
from htmd.builder.builder import detectCisPeptideBonds
self._findFiles(inputdir)
self._amberFixes()
self.acemd.temperature = self.temperature
self.acemd.thermostattemperature = self.temperature
numsteps = convert(self.timeunits,
"timesteps",
self.runtime,
timestep=self.acemd.timestep)
self.acemd.run = str(numsteps)
structfile = os.path.join(inputdir, self.acemd.structure)
pdbfile = os.path.join(inputdir, self.acemd.coordinates)
inmol = Molecule([structfile, pdbfile])
if cisbondcheck:
detectCisPeptideBonds(inmol, respect_bonds=True)
if np.any(inmol.atomselect("lipids")) and not self.useconstantratio:
logger.warning(
"Lipids detected in input structure. We highly recommend setting useconstantratio=True "
"for membrane simulations.")
if self.restraints is not None:
logger.info(
"Using user-provided restraints and ignoring constraints and fb_potential"
)
self.acemd.restraints = self.restraints
else:
restraints = list()
if self.fb_k > 0:
logger.warning(
"Converting fb_potential to restraints. This is a convenience "
"functional conversion. We recommend to start using restraints."
)
restraints += self._fb_potential2restraints(inputdir)
if self.useconstraints:
logger.warning(
"Converting constraints to restraints. This is a convenience "
"functional conversion. We recommend to start using restraints"
)
restraints += self._constraints2restraints()
else:
if len(self.constraints) != 0:
logger.warning(
"You have setup constraints to {} but constraints are turned off. "
"If you want to use constraints, define "
"useconstraints=True".format(self.constraints))
if len(restraints) != 0:
self.acemd.restraints = restraints
if self.useconstantratio:
self.acemd.barostatconstratio = "on"
if self.adaptive:
self.acemd.binvelocities = None
self.acemd.setup(inputdir, outputdir, overwrite=True)
def build(mol, topo=None, param=None, stream=None, prefix='structure', outdir='./build', caps=None, ionize=True, saltconc=0,
saltanion=None, saltcation=None, disulfide=None, patches=None, noregen=None, aliasresidues=None, psfgen=None, execute=True, _clean=True):
""" Builds a system for CHARMM
Uses VMD and psfgen to build a system for CHARMM. Additionally it allows for ionization and adding of disulfide bridges.
Parameters
----------
mol : :class:`Molecule <moleculekit.molecule.Molecule>` object
The Molecule object containing the system
topo : list of str
A list of topology `rtf` files.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available topology files.
Default: ['top/top_all36_prot.rtf', 'top/top_all36_lipid.rtf', 'top/top_water_ions.rtf']
param : list of str
A list of parameter `prm` files.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available parameter files.
Default: ['par/par_all36_prot_mod.prm', 'par/par_all36_lipid.prm', 'par/par_water_ions.prm']
stream : list of str
A list of stream `str` files containing topologies and parameters.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available stream files.
Default: ['str/prot/toppar_all36_prot_arg0.str']
prefix : str
The prefix for the generated pdb and psf files
outdir : str
The path to the output directory
Default: './build'
caps : dict
A dictionary with keys segids and values lists of strings describing the caps of that segment.
e.g. caps['P'] = ['first ACE', 'last CT3'] or caps['P'] = ['first none', 'last none'].
Default: will apply ACE and CT3 caps to proteins and none caps to the rest.
ionize : bool
Enable or disable ionization
saltconc : float
Salt concentration (in Molar) to add to the system after neutralization.
saltanion : {'CLA'}
The anion type. Please use only CHARMM ion atom names.
saltcation : {'SOD', 'MG', 'POT', 'CES', 'CAL', 'ZN2'}
The cation type. Please use only CHARMM ion atom names.
disulfide : list of pairs of atomselection strings
If None it will guess disulfide bonds. Otherwise provide a list pairs of atomselection strings for each pair of
residues forming the disulfide bridge.
patches : list of str
Any further patches the user wants to apply
noregen : list of str
A list of patches that must not be regenerated (angles and dihedrals)
Default: ['FHEM', 'PHEM', 'PLOH', 'PLO2', 'PLIG', 'PSUL']
aliasresidues : dict of aliases
A dictionary of key: value pairs of residue names we want to alias
psfgen : str
Path to psfgen executable used to build for CHARMM
execute : bool
Disable building. Will only write out the input script needed by psfgen. Does not include ionization.
Returns
-------
molbuilt : :class:`Molecule <moleculekit.molecule.Molecule>` object
The built system in a Molecule object
Example
-------
>>> from htmd.ui import *
>>> mol = Molecule("3PTB")
>>> mol.filter("not resname BEN")
>>> molbuilt = charmm.build(mol, outdir='/tmp/build', ionize=False) # doctest: +ELLIPSIS
Bond between A: [serial 185 resid 42 resname CYS chain A segid 0]
B: [serial 298 resid 58 resname CYS chain A segid 0]...
>>> # More complex example
>>> topos = ['top/top_all36_prot.rtf', './benzamidine.rtf', 'top/top_water_ions.rtf']
>>> params = ['par/par_all36_prot_mod.prm', './benzamidine.prm', 'par/par_water_ions.prm']
>>> disu = [['segid P and resid 157', 'segid P and resid 13'], ['segid K and resid 1', 'segid K and resid 25']]
>>> ar = {'SAPI24': 'SP24'} # Alias large resnames to a short-hand version
>>> molbuilt = charmm.build(mol, topo=topos, param=params, outdir='/tmp/build', saltconc=0.15, disulfide=disu, aliasresidues=ar) # doctest: +SKIP
"""
mol = mol.copy()
_missingSegID(mol)
_checkMixedSegment(mol)
_checkLongResnames(mol, aliasresidues)
if psfgen is None:
psfgen = shutil.which('psfgen', mode=os.X_OK)
if not psfgen:
raise FileNotFoundError('Could not find psfgen executable, or no execute permissions are given. '
'Run `conda install psfgen`.')
if not os.path.isdir(outdir):
os.makedirs(outdir)
if _clean:
_cleanOutDir(outdir)
if topo is None:
topo = defaultTopo()
if param is None:
param = defaultParam()
if stream is None:
stream = defaultStream()
if caps is None:
caps = _defaultCaps(mol)
# patches that must _not_ be regenerated
if noregen is None:
noregen = ['FHEM', 'PHEM', 'PLOH', 'PLO2', 'PLIG', 'PSUL']
alltopo = topo.copy()
allparam = param.copy()
# Splitting the stream files and adding them to the list of parameter and topology files
charmmdir = path.join(home(), 'builder', 'charmmfiles')
for s in stream:
if s[0] != '.' and path.isfile(path.join(charmmdir, s)):
s = path.join(charmmdir, s)
outrtf, outprm = _prepareStream(s)
alltopo.append(outrtf)
allparam.append(outprm)
#_missingChain(mol)
#_checkProteinGaps(mol)
if patches is None:
patches = []
if isinstance(patches, str):
patches = [patches]
allpatches = []
allpatches += patches
# Find protonated residues and add patches for them
allpatches += _protonationPatches(mol)
f = open(path.join(outdir, 'build.vmd'), 'w')
f.write('# psfgen file generated by charmm.build\n')
f.write('package require psfgen;\n')
f.write('psfcontext reset;\n\n')
# Copying and printing out the topologies
if not path.exists(path.join(outdir, 'topologies')):
os.makedirs(path.join(outdir, 'topologies'))
for i in range(len(alltopo)):
if alltopo[i][0] != '.' and path.isfile(path.join(charmmdir, alltopo[i])):
alltopo[i] = path.join(charmmdir, alltopo[i])
localname = '{}.'.format(i) + path.basename(alltopo[i])
shutil.copy(alltopo[i], path.join(outdir, 'topologies', localname))
f.write('topology ' + path.join('topologies', localname) + '\n')
f.write('\n')
_printAliases(f)
if aliasresidues is not None: # User defined aliases
for key, val in aliasresidues.items():
mol.resname[mol.resname == key] = val
f.write(' pdbalias residue {} {}\n'.format(val, key))
# Printing out segments
if not path.exists(path.join(outdir, 'segments')):
os.makedirs(path.join(outdir, 'segments'))
logger.info('Writing out segments.')
segments = _getSegments(mol)
wateratoms = mol.atomselect('water')
for seg in segments:
pdbname = 'segment' + seg + '.pdb'
segatoms = mol.segid == seg
mol.write(path.join(outdir, 'segments', pdbname), sel=segatoms)
segwater = wateratoms & segatoms
f.write('segment ' + seg + ' {\n')
if np.all(segatoms == segwater): # If segment only contains waters, set: auto none
f.write('\tauto none\n')
f.write('\tpdb ' + path.join('segments', pdbname) + '\n')
if caps is not None and seg in caps:
for c in caps[seg]:
f.write('\t' + c + '\n')
f.write('}\n')
f.write('coordpdb ' + path.join('segments', pdbname) + ' ' + seg + '\n\n')
# Printing out patches for the disulfide bridges
# TODO: Remove this once we deprecate the class
from htmd.builder.builder import DisulfideBridge
from moleculekit.molecule import UniqueResidueID
if disulfide is not None and len(disulfide) != 0 and isinstance(disulfide[0], DisulfideBridge):
newdisu = []
for d in disulfide:
r1 = UniqueResidueID.fromMolecule(mol, 'resid {} and segname {}'.format(d.resid1, d.segid1))
r2 = UniqueResidueID.fromMolecule(mol, 'resid {} and segname {}'.format(d.resid2, d.segid2))
newdisu.append([r1, r2])
disulfide = newdisu
# TODO: Remove up to here ----------------------
if disulfide is not None and len(disulfide) != 0 and isinstance(disulfide[0][0], str):
disulfide = convertDisulfide(mol, disulfide)
if disulfide is None:
disulfide = detectDisulfideBonds(mol)
if len(disulfide) != 0:
for d in disulfide:
str0 = '{}:{}{}'.format(d[0].segid, d[0].resid, d[0].insertion)
str1 = '{}:{}{}'.format(d[1].segid, d[1].resid, d[1].insertion)
f.write('patch DISU {} {}\n'.format(str0, str1))
f.write('\n')
noregenpatches = [p for p in allpatches if p.split()[1] in noregen]
regenpatches = [p for p in allpatches if p.split()[1] not in noregen]
# Printing regenerable patches
if len(regenpatches) != 0:
for p in regenpatches:
f.write(p + '\n')
f.write('\n')
# Regenerate angles and dihedrals
f.write('regenerate angles dihedrals\n')
f.write('\n')
# Printing non-regenerable patches
if len(noregenpatches) != 0:
for p in noregenpatches:
f.write(p + '\n')
f.write('\n')
f.write('guesscoord\n')
f.write('writepsf ' + prefix + '.psf\n')
f.write('writepdb ' + prefix + '.pdb\n')
#f.write('quit\n')
f.close()
if allparam is not None:
combine(allparam, path.join(outdir, 'parameters'))
molbuilt = None
if execute:
logpath = os.path.abspath('{}/log.txt'.format(outdir))
logger.info('Starting the build.')
currdir = os.getcwd()
os.chdir(outdir)
f = open(logpath, 'w')
#call([vmd, '-dispdev', 'text', '-e', './build.vmd'], stdout=f)
my_env = os.environ.copy()
my_env['LC_ALL'] = 'C'
call([psfgen, './build.vmd'], stdout=f, stderr=f, env=my_env)
f.close()
errors = _logParser(logpath)
os.chdir(currdir)
if errors:
raise BuildError(errors + ['Check {} for further information on errors in building.'.format(logpath)])
logger.info('Finished building.')
if path.isfile(path.join(outdir, 'structure.pdb')) and path.isfile(path.join(outdir, 'structure.psf')):
molbuilt = Molecule(path.join(outdir, 'structure.pdb'), validateElements=False)
molbuilt.read(path.join(outdir, 'structure.psf'), validateElements=False)
else:
raise BuildError('No structure pdb/psf file was generated. Check {} for errors in building.'.format(logpath))
if ionize:
os.makedirs(path.join(outdir, 'pre-ionize'))
data = glob(path.join(outdir, '*'))
for f in data:
shutil.move(f, path.join(outdir, 'pre-ionize'))
totalcharge = np.sum(molbuilt.charge)
nwater = np.sum(molbuilt.atomselect('water and noh'))
anion, cation, anionatom, cationatom, nanion, ncation = ionizef(totalcharge, nwater, saltconc=saltconc,
anion=saltanion, cation=saltcation)
newmol = ionizePlace(mol, anion, cation, anionatom, cationatom, nanion, ncation)
# Redo the whole build but now with ions included
return build(newmol, topo=alltopo, param=allparam, stream=[], prefix=prefix, outdir=outdir, ionize=False,
caps=caps, execute=execute, saltconc=saltconc, disulfide=disulfide, patches=patches,
noregen=noregen, aliasresidues=aliasresidues, psfgen=psfgen, _clean=False)
_checkFailedAtoms(molbuilt)
_recoverProtonations(molbuilt)
return molbuilt
def write(self, inputdir, outputdir):
""" Write the equilibration protocol
Writes the equilibration protocol and files into a folder for execution
using files inside the inputdir directory
Parameters
----------
inputdir : str
Path to a directory containing the files produced by a build process.
outputdir : str
Directory where to write the equilibration setup files.
Examples
--------
>>> md = Equilibration()
>>> md.write('./build','./equil')
"""
from moleculekit.molecule import Molecule
# Do version consistency check
if (self._version == 2 and not isinstance(self.acemd, Acemd2)) and \
(self._version == 3 and not isinstance(self.acemd, Acemd)):
raise RuntimeError(
'Acemd object version ({}) inconsistent with protocol version at instantiation '
'({})'.format(type(self.acemd), self._version))
self._findFiles(inputdir)
self._amberFixes()
from htmd.units import convert
numsteps = convert(self.timeunits,
'timesteps',
self.runtime,
timestep=self.acemd.timestep)
if self._version == 3:
self.acemd.temperature = self.temperature
self.acemd.thermostattemperature = self.temperature
self.acemd.run = str(numsteps)
pdbfile = os.path.join(inputdir, self.acemd.coordinates)
inmol = Molecule(pdbfile)
from htmd.builder.builder import detectCisPeptideBonds
detectCisPeptideBonds(inmol)
if np.any(inmol.atomselect('lipids')) and not self.useconstantratio:
logger.warning(
'Lipids detected in input structure. We highly recommend setting useconstantratio=True '
'for membrane simulations.')
if self.constraintsteps is None:
constrsteps = int(numsteps / 2)
else:
constrsteps = int(self.constraintsteps)
if self._version == 2:
if self.restraints:
raise RuntimeWarning(
'restraints are only available on {}(_version=3)'.format(
self.__class__.__name__))
if isinstance(self.acemd.TCL, tuple):
tcl = list(self.acemd.TCL)
tcl[0] = tcl[0].format(
NUMSTEPS=numsteps,
KCONST=self.fb_k,
REFINDEX=' '.join(
map(str, inmol.get('index', self.fb_reference))),
SELINDEX=' '.join(
map(str, inmol.get('index', self.fb_selection))),
BOX=' '.join(map(str, self.fb_box)),
NVTSTEPS=self.nvtsteps,
CONSTRAINTSTEPS=constrsteps,
TEMPERATURE=self.temperature)
self.acemd.TCL = tcl[0] + tcl[1]
else:
logger.warning('{} default TCL was already formatted.'.format(
self.__class__.__name__))
elif self._version == 3:
if self.restraints is not None:
logger.info(
'Using user-provided restraints and ignoring constraints and fb_potential'
)
self.acemd.restraints = self.restraints
else:
logger.warning(
'Converting constraints and fb_potential to restraints. This is a convenience '
'functional conversion. We recommend start using restraints with '
'{}(_version=3)'.format(self.__class__.__name__))
restraints = list()
# convert constraints to restraints and add them
if self.constraints is not None:
restraints += self._constraints2restraints(constrsteps)
# convert fb_potential to restraints and add them
if self.fb_k > 0:
restraints += self._fb_potential2restraints(inputdir)
self.acemd.restraints = restraints
if ((self._version == 2) and self.acemd.celldimension is None and self.acemd.extendedsystem is None) or \
((self._version == 3) and self.acemd.boxsize is None and self.acemd.extendedsystem is None):
coords = inmol.get('coords', sel='water')
if coords.size == 0: # It's a vacuum simulation
coords = inmol.get('coords', sel='all')
dim = np.max(coords, axis=0) - np.min(coords, axis=0)
dim += 12.
else:
dim = np.max(coords, axis=0) - np.min(coords, axis=0)
if self._version == 2:
self.acemd.celldimension = '{} {} {}'.format(
dim[0], dim[1], dim[2])
else:
self.acemd.boxsize = '{} {} {}'.format(dim[0], dim[1], dim[2])
if self.useconstantratio:
self.acemd.useconstantratio = 'on'
self.acemd.setup(inputdir, outputdir, overwrite=True)
if self._version == 2:
# Adding constraints by writing them to the consref file
inconsreffile = os.path.join(inputdir, self.acemd.consref)
consrefmol = Molecule(inconsreffile)
consrefmol.set('occupancy', 0)
consrefmol.set('beta', 0)
if len(self.constraints) == 0:
raise RuntimeError(
'You have not defined any constraints for the {} ('
'constraints={{}}).'.format(self.__class__.__name__))
else:
for sel in self.constraints:
consrefmol.set('beta', self.constraints[sel], sel)
outconsreffile = os.path.join(outputdir, self.acemd.consref)
consrefmol.write(outconsreffile)
# Write and test your answer here
np.unique(m.chain)
"""## How many residues per each chain?"""
for ch in np.unique(m.chain):
tmp = m.copy()
tmp.filter(f"protein and chain {ch} and name CA")
nres = tmp.numAtoms
print(f"Chain {ch} has {nres} residues (CA atoms)")
# Alternative (non-destructive) implementation
for ch in np.unique(m.chain):
asel=m.atomselect(f"protein and chain {ch} and name CA")
nres = np.sum(asel)
print(f"Chain {ch} has {nres} residues (CA atoms)")
"""## How many atoms?"""
for ch in np.unique(m.chain):
tmp = m.copy()
tmp.filter(f"protein and chain {ch}")
nat = tmp.numAtoms
print(f"Chain {ch} has {nat} atoms")
"""## Is your structure solvated? If so, remove solvent."""
nwat = np.sum(m.atomselect("water"))
print(nwat)
def build(
mol,
topo=None,
param=None,
stream=None,
prefix="structure",
outdir="./build",
caps=None,
ionize=True,
saltconc=0,
saltanion=None,
saltcation=None,
disulfide=None,
regenerate=["angles", "dihedrals"],
patches=None,
noregen=None,
aliasresidues=None,
psfgen=None,
execute=True,
_clean=True,
):
"""Builds a system for CHARMM
Uses VMD and psfgen to build a system for CHARMM. Additionally it allows for ionization and adding of disulfide bridges.
Parameters
----------
mol : :class:`Molecule <moleculekit.molecule.Molecule>` object
The Molecule object containing the system
topo : list of str
A list of topology `rtf` files.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available topology files.
Default: ['top/top_all36_prot.rtf', 'top/top_all36_lipid.rtf', 'top/top_water_ions.rtf']
param : list of str
A list of parameter `prm` files.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available parameter files.
Default: ['par/par_all36_prot.prm', 'par/par_all36_lipid.prm', 'par/par_water_ions.prm']
stream : list of str
A list of stream `str` files containing topologies and parameters.
Use :func:`charmm.listFiles <htmd.builder.charmm.listFiles>` to get a list of available stream files.
Default: ['str/prot/toppar_all36_prot_arg0.str']
prefix : str
The prefix for the generated pdb and psf files
outdir : str
The path to the output directory
Default: './build'
caps : dict
A dictionary with keys segids and values lists of strings describing the caps of that segment.
e.g. caps['P'] = ['first ACE', 'last CT3'] or caps['P'] = ['first none', 'last none'].
Default: will apply ACE and CT3 caps to proteins and none caps to the rest.
ionize : bool
Enable or disable ionization
saltconc : float
Salt concentration (in Molar) to add to the system after neutralization.
saltanion : {'CLA'}
The anion type. Please use only CHARMM ion atom names.
saltcation : {'SOD', 'MG', 'POT', 'CES', 'CAL', 'ZN2'}
The cation type. Please use only CHARMM ion atom names.
disulfide : list of pairs of atomselection strings
If None it will guess disulfide bonds. Otherwise provide a list pairs of atomselection strings for each pair of
residues forming the disulfide bridge.
regenerate : None or list of strings of: ['angles', 'dihedrals']
Disable angle/dihedral regeneration with `regenerate=None`, or enable it with `regenerate=['angles', 'diheldrals']`
or just one of the two options with `regenerate=['angles']` or `regenerate=['diheldrals']`.
patches : list of str
Any further patches the user wants to apply
noregen : list of str
A list of patches that must not be regenerated (angles and dihedrals)
Default: ['FHEM', 'PHEM', 'PLOH', 'PLO2', 'PLIG', 'PSUL']
aliasresidues : dict of aliases
A dictionary of key: value pairs of residue names we want to alias
psfgen : str
Path to psfgen executable used to build for CHARMM
execute : bool
Disable building. Will only write out the input script needed by psfgen. Does not include ionization.
Returns
-------
molbuilt : :class:`Molecule <moleculekit.molecule.Molecule>` object
The built system in a Molecule object
Example
-------
>>> from htmd.ui import *
>>> mol = Molecule("3PTB")
>>> mol.filter("not resname BEN")
>>> molbuilt = charmm.build(mol, outdir='/tmp/build', ionize=False) # doctest: +ELLIPSIS
Bond between A: [serial 185 resid 42 resname CYS chain A segid 0]
B: [serial 298 resid 58 resname CYS chain A segid 0]...
>>> # More complex example
>>> topos = ['top/top_all36_prot.rtf', './BEN.rtf', 'top/top_water_ions.rtf']
>>> params = ['par/par_all36_prot.prm', './BEN.prm', 'par/par_water_ions.prm']
>>> disu = [['segid P and resid 157', 'segid P and resid 13'], ['segid K and resid 1', 'segid K and resid 25']]
>>> ar = {'SAPI24': 'SP24'} # Alias large resnames to a short-hand version
>>> molbuilt = charmm.build(mol, topo=topos, param=params, outdir='/tmp/build', saltconc=0.15, disulfide=disu, aliasresidues=ar) # doctest: +SKIP
"""
mol = mol.copy()
_missingSegID(mol)
_checkMixedSegment(mol)
_checkLongResnames(mol, aliasresidues)
if psfgen is None:
psfgen = shutil.which("psfgen", mode=os.X_OK)
if not psfgen:
raise FileNotFoundError(
"Could not find psfgen executable, or no execute permissions are given. "
"Run `conda install psfgen -c acellera`.")
if not os.path.isdir(outdir):
os.makedirs(outdir)
if _clean:
_cleanOutDir(outdir)
if topo is None:
topo = defaultTopo()
if param is None:
param = defaultParam()
if stream is None:
stream = defaultStream()
if caps is None:
caps = _defaultCaps(mol)
# patches that must _not_ be regenerated
if noregen is None:
noregen = ["FHEM", "PHEM", "PLOH", "PLO2", "PLIG", "PSUL"]
alltopo = topo.copy()
allparam = param.copy()
# Splitting the stream files and adding them to the list of parameter and topology files
charmmdir = htmdCharmmHome()
for s in stream:
if s[0] != "." and path.isfile(path.join(charmmdir, s)):
s = path.join(charmmdir, s)
outrtf, outprm = _prepareStream(s)
alltopo.append(outrtf)
allparam.append(outprm)
# _missingChain(mol)
# _checkProteinGaps(mol)
if patches is None:
patches = []
if isinstance(patches, str):
patches = [patches]
allpatches = []
allpatches += patches
# Find protonated residues and add patches for them
allpatches += _protonationPatches(mol)
f = open(path.join(outdir, "build.vmd"), "w")
f.write("# psfgen file generated by charmm.build\n")
f.write("package require psfgen;\n")
f.write("psfcontext reset;\n\n")
# Copying and printing out the topologies
if not path.exists(path.join(outdir, "topologies")):
os.makedirs(path.join(outdir, "topologies"))
for i in range(len(alltopo)):
if alltopo[i][0] != "." and path.isfile(
path.join(charmmdir, alltopo[i])):
alltopo[i] = path.join(charmmdir, alltopo[i])
localname = "{}.".format(i) + path.basename(alltopo[i])
shutil.copy(alltopo[i], path.join(outdir, "topologies", localname))
f.write("topology " + path.join("topologies", localname) + "\n")
f.write("\n")
_printAliases(f)
if aliasresidues is not None: # User defined aliases
for key, val in aliasresidues.items():
mol.resname[mol.resname == key] = val
f.write(" pdbalias residue {} {}\n".format(val, key))
# Printing out segments
if not path.exists(path.join(outdir, "segments")):
os.makedirs(path.join(outdir, "segments"))
logger.info("Writing out segments.")
segments = _getSegments(mol)
wateratoms = mol.atomselect("water")
for seg in segments:
pdbname = "segment" + seg + ".pdb"
segatoms = mol.segid == seg
mol.write(path.join(outdir, "segments", pdbname), sel=segatoms)
segwater = wateratoms & segatoms
f.write("segment " + seg + " {\n")
if np.all(segatoms ==
segwater): # If segment only contains waters, set: auto none
f.write("\tauto none\n")
f.write("\tpdb " + path.join("segments", pdbname) + "\n")
if caps is not None and seg in caps:
for c in caps[seg]:
f.write("\t" + c + "\n")
f.write("}\n")
f.write("coordpdb " + path.join("segments", pdbname) + " " + seg +
"\n\n")
if (disulfide is not None and len(disulfide) != 0
and isinstance(disulfide[0][0], str)):
disulfide = convertDisulfide(mol, disulfide)
if disulfide is None:
disulfide = detectDisulfideBonds(mol)
if len(disulfide) != 0:
for d in sorted(disulfide, key=lambda x: x[0].segid):
str0 = f"{d[0].segid}:{d[0].resid}{d[0].insertion}"
str1 = f"{d[1].segid}:{d[1].resid}{d[1].insertion}"
f.write(f"patch DISU {str0} {str1}\n")
f.write("\n")
noregenpatches = [p for p in allpatches if p.split()[1] in noregen]
regenpatches = [p for p in allpatches if p.split()[1] not in noregen]
# Printing regenerable patches
if len(regenpatches) != 0:
for p in regenpatches:
f.write(p + "\n")
f.write("\n")
# Regenerate angles and dihedrals
if regenerate is not None:
f.write("regenerate {}\n".format(" ".join(regenerate)))
f.write("\n")
# Printing non-regenerable patches
if len(noregenpatches) != 0:
for p in noregenpatches:
f.write(p + "\n")
f.write("\n")
f.write("guesscoord\n")
f.write("writepsf " + prefix + ".psf\n")
f.write("writepdb " + prefix + ".pdb\n")
# f.write('quit\n')
f.close()
if allparam is not None:
combine(allparam, path.join(outdir, "parameters"))
molbuilt = None
if execute:
logpath = os.path.abspath("{}/log.txt".format(outdir))
logger.info("Starting the build.")
currdir = os.getcwd()
os.chdir(outdir)
f = open(logpath, "w")
# call([vmd, '-dispdev', 'text', '-e', './build.vmd'], stdout=f)
my_env = os.environ.copy()
my_env["LC_ALL"] = "C"
call([psfgen, "./build.vmd"], stdout=f, stderr=f, env=my_env)
f.close()
errors = _logParser(logpath)
os.chdir(currdir)
if errors:
raise BuildError(errors + [
"Check {} for further information on errors in building.".
format(logpath)
])
logger.info("Finished building.")
if path.isfile(path.join(outdir, "structure.pdb")) and path.isfile(
path.join(outdir, "structure.psf")):
molbuilt = Molecule(path.join(outdir, "structure.pdb"))
molbuilt.read(path.join(outdir, "structure.psf"))
else:
raise BuildError(
"No structure pdb/psf file was generated. Check {} for errors in building."
.format(logpath))
if ionize:
os.makedirs(path.join(outdir, "pre-ionize"))
data = glob(path.join(outdir, "*"))
for f in data:
shutil.move(f, path.join(outdir, "pre-ionize"))
totalcharge = np.sum(molbuilt.charge)
nwater = np.sum(molbuilt.atomselect("water and noh"))
anion, cation, anionatom, cationatom, nanion, ncation = ionizef(
molbuilt,
totalcharge,
nwater,
saltconc=saltconc,
anion=saltanion,
cation=saltcation,
)
newmol = ionizePlace(mol, anion, cation, anionatom, cationatom,
nanion, ncation)
# Redo the whole build but now with ions included
return build(
newmol,
topo=alltopo,
param=allparam,
stream=[],
prefix=prefix,
outdir=outdir,
ionize=False,
caps=caps,
execute=execute,
saltconc=saltconc,
disulfide=disulfide,
regenerate=regenerate,
patches=patches,
noregen=noregen,
aliasresidues=aliasresidues,
psfgen=psfgen,
_clean=False,
)
_checkFailedAtoms(molbuilt)
_recoverProtonations(molbuilt)
detectCisPeptideBonds(molbuilt,
respect_bonds=True) # Warn in case of cis bonds
return molbuilt