def load_traj(filepath: PathLike, **kwargs) -> mdtraj.Trajectory:
"""Load a trajectory, if a PDB fails with zero box volume disable volume check and try again."""
filepath = pathlib.Path(filepath)
if filepath.suffix.lower() == ".pdb":
# PDBs have a couple of things that can go wrong - we handle these here...
kwargs.pop("top", None) # Can't specify a topology for `load_pdb`
try:
return mdtraj.load_pdb(str(filepath), **kwargs)
except FloatingPointError:
# PDB file loading checks density using the simulation box
# This can fail if the box volume is zero
trajectory = mdtraj.load_pdb(str(filepath),
no_boxchk=True,
**kwargs)
logger.warning(
"Unitcell has zero volume - periodic boundaries will not be accounted for. "
"If the molecule is split by a periodic boundary, results will be incorrect."
)
return trajectory
return mdtraj.load(str(filepath), **kwargs)
def from_pdb(self, filename, selection=None, probe_radius=1.4, **kwargs):
r"""Calculate solvent accessible surface area (SASA) from a PDB file
If the PBD contains more than one structure, calculation is performed
only for the first one.
SASA units are Angstroms squared
Parameters
----------
filename: str
Path to the PDB file
selection: str
Atomic selection for calculating SASA. All atoms considered if
default None is passed. See the
`selections page <https://www.mdanalysis.org/docs/documentation_pages/selections.html>`_
for atom selection syntax.
probe_radius: float
The radius of the probe, in Angstroms
kwargs: dict
Optional arguments for the underlying mdtraj.shrake_rupley
algorithm doing the actual SaSa calculation
Returns
-------
self: :class:`~idpflex.properties.SaSa`
Instantiated SaSa property object
""" # noqa: E501
self.selection = selection
a_traj = mdtraj.load_pdb(filename)
if selection is not None:
selection = a_traj.top.select(selection) # atomic indices
a_traj = mdtraj.load_pdb(filename, atom_indices=selection)
return self.from_mdtraj(a_traj, probe_radius=probe_radius, **kwargs)
def verify_transform(pdb1, pdb2, lenVector):
"""Verifie que c'est la même protéine dans les deux fichiers pdb puis revoie leur coordonnées
sous forme de deux listes
-Args:
_pdb1: fichier au format pdb
_pdb2: fichier au format pdb
"""
pdb_1 = md.load_pdb(pdb1) # lecture du fichier pdb par le module mdtraj
pdb_2 = md.load_pdb(pdb2)
if pdb_1.n_atoms != pdb_2.n_atoms or pdb_1.n_residues != pdb_2.n_residues:
print(
"Pas le même nombre d'atomes/residues entre les deux structures !!"
)
print("Nombre d'atomes : {} vs {}".format(pdb_1.n_atoms,
pdb_2.n_atoms))
print("Nombre de résidues : {} vs {}".format(pdb_1.n_residues,
pdb_2.n_residues))
sys.exit()
if pdb_1.n_atoms * 3 != lenVector:
print(
"Nombre de coordonnées du système: {} vs nombre de coordonnées normales{}"
.format(pdb_1.n_atoms * 3, lenVector))
print(
"Pas le même nombre de coordonnées normales et de coordonnées atomiques !!"
)
sys.exit()
xyz1 = [
c * 10 for l in pdb_1[0].xyz[0] for c in l
] # conversion des coordonnées dans un autre format et passage des nm aux angström
xyz2 = [c * 10 for l in pdb_2[0].xyz[0] for c in l]
return (xyz1, xyz2)
def test_pdb_from_url():
# load pdb from URL
t1 = load_pdb('http://www.rcsb.org/pdb/files/4K6Q.pdb.gz')
t2 = load_pdb('http://www.rcsb.org/pdb/files/4K6Q.pdb')
eq(t1.n_frames, 1)
eq(t2.n_frames, 1)
eq(t1.n_atoms, 2208)
eq(t2.n_atoms, 2208)
def test_pdb_from_url():
# load pdb from URL
t1 = load_pdb('http://www.rcsb.org/pdb/files/4K6Q.pdb.gz')
t2 = load_pdb('http://www.rcsb.org/pdb/files/4K6Q.pdb')
eq(t1.n_frames, 1)
eq(t2.n_frames, 1)
eq(t1.n_atoms, 2208)
eq(t2.n_atoms, 2208)
def test_load_pdb_input_top(get_fn):
pdb = get_fn('native.pdb')
p_1 = load_pdb(pdb)
p_2 = load_pdb(pdb, top=p_1.topology)
eq(p_1.xyz, p_2.xyz)
eq(p_1.topology, p_2.topology)
def __init__(self, FG_fname, CG_fname):
self.FG_fname = FG_fname
self.CG_fname = CG_fname
self.FG_trj = md.load_pdb(filename=self.FG_fname).center_coordinates()
self.CG_trj = md.load_pdb(filename=self.CG_fname).center_coordinates()
self._parse()
self._create_mapping()
self._align()
def loader(file):
try:
md.load_pdb(file)
except:
return None
print(file)
return None
def loader(file):
try:
md.load_pdb(file)
except:
return None
print(file)
return None
def __init__(self, CG_pdb_f_name, AA_pdb_f_name):
self.CG_pdb_f_name = CG_pdb_f_name
self.AA_pdb_f_name = AA_pdb_f_name
self.CG_trj = md.load_pdb(filename=self.CG_pdb_f_name).remove_solvent()
self.CG_top = self.CG_trj.top
self.AA_trj = md.load_pdb(filename=self.AA_pdb_f_name).remove_solvent()
self.AA_top = self.AA_trj.top
self.CG_beads = parse_CG_pdb(self.CG_pdb_f_name)
self.AA_beads = parse_AA_pdb(self.AA_pdb_f_name)
def _construct_traj(self):
logger.debug('Loading Trajectory object for model {0} ({1}/{2})'.format(self.df.templateid.iloc[0], 0, len(self.df.model_filepath)))
traj = mdtraj.load_pdb(self.df.model_filepath[0])
remove_disulfide_bonds_from_topology(traj.topology)
self.traj = traj
for m, model_filepath in enumerate(self.df.model_filepath[1:]):
logger.debug('Loading Trajectory object for model {0} ({1}/{2})'.format(self.df.templateid.iloc[m+1], m+1, len(self.df.model_filepath)))
traj = mdtraj.load_pdb(model_filepath)
remove_disulfide_bonds_from_topology(traj.topology)
self.traj += traj
def makeDatabase(self, q_values, num_iter, lmax, save_path, beta = 0.0, alpha = 1.0, zeta = 0.0): self.alpha = alpha self.beta = beta self.zeta = zeta self.q_values = q_values self.lmax = lmax self.num_iter = num_iter traj_full = mdtraj.load_pdb( self.traj_full_path ) traj_guess = mdtraj.load_pdb( self.traj_guess_path ) if type(q_values) == str and q_values == 'all': try: assert( np.isclose( self.model_guess.q_values, self.model_full.q_values ).all() ) except AssertionError: print( "q_values of the model_full and model_guess do not match. \ Cannot use all q_values." ) return self.q_values = self.model_full.q_values # do nothing and do not slice the models else: self.model_full.slice_by_qvalues( q_values=self.q_values, inplace=True ) self.model_guess.slice_by_qvalues( q_values=self.q_values, inplace=True) try: assert( np.isclose( self.model_guess.q_values, self.model_full.q_values ).all() ) except AssertionError: print( "q_values of the model_full and model_guess do not match. \ Failed!!!" ) return # retriever the phases self._retrieve_phases() # compare guess to full model self._compare_full_guess() # compute RMSD try: self.rmsd = mdtraj.rmsd( traj_full, traj_guess, frame = 0) except ValueError: self.rmsd = 0 # save self._saveDB( save_path )
def test_segment_id(get_fn):
pdb = load_pdb(get_fn('ala_ala_ala.pdb'))
pdb.save_pdb(temp)
pdb2 = load_pdb(temp)
correct_segment_id = 'AAL'
# check that all segment ids are set correctly
for ridx,r in enumerate(pdb.top.residues):
assert r.segment_id == correct_segment_id, "residue %i (0-indexed) does not have segment_id set correctly from ala_ala_ala.pdb"%(ridx)
# check that all segment ids are set correctly after a new pdb file is written
for ridx,(r1,r2) in enumerate(zip(pdb.top.residues,pdb2.top.residues)):
assert r1.segment_id == r2.segment_id, "segment_id of residue %i (0-indexed) in ala_ala_ala.pdb does not agree with value in after being written out to a new pdb file"%(ridx)
def test_1vii_url_and_gz():
t1 = load_pdb('http://www.rcsb.org/pdb/files/1vii.pdb.gz')
t2 = load_pdb('http://www.rcsb.org/pdb/files/1vii.pdb')
t3 = load_pdb(get_fn('1vii.pdb.gz'))
t4 = load_pdb(get_fn('1vii.pdb'))
eq(t1.n_frames, 1)
eq(t1.n_frames, t2.n_frames)
eq(t1.n_frames, t3.n_frames)
eq(t1.n_frames, t4.n_frames)
eq(t1.n_atoms, t2.n_atoms)
eq(t1.n_atoms, t3.n_atoms)
eq(t1.n_atoms, t4.n_atoms)
def test_segment_id(get_fn):
pdb = load_pdb(get_fn('ala_ala_ala.pdb'))
pdb.save_pdb(temp)
pdb2 = load_pdb(temp)
correct_segment_id = 'AAL'
# check that all segment ids are set correctly
for ridx,r in enumerate(pdb.top.residues):
assert r.segment_id == correct_segment_id, "residue %i (0-indexed) does not have segment_id set correctly from ala_ala_ala.pdb"%(ridx)
# check that all segment ids are set correctly after a new pdb file is written
for ridx,(r1,r2) in enumerate(zip(pdb.top.residues,pdb2.top.residues)):
assert r1.segment_id == r2.segment_id, "segment_id of residue %i (0-indexed) in ala_ala_ala.pdb does not agree with value in after being written out to a new pdb file"%(ridx)
def test_1vii_url_and_gz():
t1 = load_pdb('http://www.rcsb.org/pdb/files/1vii.pdb.gz')
t2 = load_pdb('http://www.rcsb.org/pdb/files/1vii.pdb')
t3 = load_pdb(get_fn('1vii.pdb.gz'))
t4 = load_pdb(get_fn('1vii.pdb'))
eq(t1.n_frames, 1)
eq(t1.n_frames, t2.n_frames)
eq(t1.n_frames, t3.n_frames)
eq(t1.n_frames, t4.n_frames)
eq(t1.n_atoms, t2.n_atoms)
eq(t1.n_atoms, t3.n_atoms)
eq(t1.n_atoms, t4.n_atoms)
def _get_models(self):
self.model = {}
root, dirnames, filenames = next(os.walk(self.models_target_dir))
for dirname in dirnames:
if 'implicit' in self.model and 'explicit' in self.model:
break
if 'implicit' not in self.model:
implicit_model_filename = os.path.join(self.models_target_dir, dirname, 'implicit-refined.pdb.gz')
if os.path.exists(implicit_model_filename):
self.model['implicit'] = mdtraj.load_pdb(implicit_model_filename)
if 'explicit' not in self.model:
explicit_model_filename = os.path.join(self.models_target_dir, dirname, 'explicit-refined.pdb.gz')
if os.path.exists(explicit_model_filename):
self.model['explicit'] = mdtraj.load_pdb(explicit_model_filename)
def func(arg):
xtcf, topf, reffs, p, sel = arg
toppdb = md.load_pdb(topf)
atm0 = toppdb.top.select(sel)
xtc = md.load_xtc(xtcf, topf)
xtc = xtc.atom_slice(atm0)
rmsd = []
for reff in reffs:
refpdb = md.load_pdb(reff)
atm1 = refpdb.top.select(sel)
ref = refpdb.atom_slice(atm1)
rmsd.append(md.rmsd(xtc, ref))
rmsd.append([p] * len(xtc))
return np.array(rmsd).T
def get(self, pdbid):
pdbid = pdbid.lower()
if len(pdbid) != 4:
raise Exception("pdb id must be four characters long")
filename = os.path.join(self.temp_dir, '%s.pdb.gz' % pdbid)
with open(filename, 'wb') as filehandler:
self.conn.retrbinary(
'RETR pub/pdb/data/structures/divided/pdb/%s/pdb%s.ent.gz' %
(pdbid[1:3], pdbid), filehandler.write)
if self.just_files:
return filename
else:
# then we actually want to load it as a pdb file object and return an mdtraj trajectory
with gzip.open(filename) as filehandler:
text = filehandler.read()
with open(filename[:-3], 'w') as filehandler:
filehandler.write(text)
pdb = mdtraj.load_pdb(filename[:-3])
return pdb
def clean_pdb(name, route=None, chain_num=None):
""" Cleans the structure to only leave the important part.
Inputs:
* name: str. route of the input .pdb file
* route: str. route of the output. will overwrite input if not provided
* chain_num: int. index of chain to select (1-indexed as pdb files)
Output: route of destin file.
"""
destin = route if route is not None else name
# read input
raw_prot = mdtraj.load_pdb(name)
# iterate over prot and select the specified chains
idxs = []
for chain in raw_prot.topology.chains:
# if arg passed, only select that chain
if chain_num is not None:
if chain_num != chain.index:
continue
# select indexes of chain
chain_idxs = raw_prot.topology.select(f"chainid == {str(chain.index)}")
idxs.extend( chain_idxs.tolist() )
# sort: topology and xyz selection are ordered
idxs = sorted(idxs)
# get new trajectory from the sleected subset of indexes and save
prot = mdtraj.Trajectory(xyz=raw_prot.xyz[:, idxs],
topology=raw_prot.topology.subset(idxs))
prot.save(destin)
return destin
def _load_traj(ag):
h, tmp = tempfile.mkstemp(dir='.', suffix='.pdb')
os.close(h)
prody.writePDB(tmp, ag)
traj = md.load_pdb(tmp, frame=0)
os.remove(tmp)
return traj
def test_4():
t = md.load_pdb(get_fn('1am7_protein.pdb'))
a = md.compute_dssp(t, simplified=True)
b = md.compute_dssp(t, simplified=False)
assert len(a) == len(b)
assert len(a[0]) == len(b[0])
assert list(np.unique(a[0])) == ['C', 'E', 'H']
def custom2pdb(coords, proteinnet_id, route):
""" Takes a custom representation and turns into a .pdb file.
Inputs:
* coords: array/tensor of shape (3 x N) or (N x 3). in Angstroms.
same order as in the proteinnnet is assumed (same as raw pdb file)
* proteinnet_id: str. proteinnet id format (<class>#<pdb_id>_<chain_number>_<chain_id>)
see: https://github.com/aqlaboratory/proteinnet/
* route: str. destin route.
Output: tuple of routes: (original, generated) for the structures.
"""
# convert to numpy
if isinstance(coords, torch.Tensor):
coords = coords.detach().cpu().numpy()
# ensure (1, N, 3)
if coords.shape[1] == 3:
coords = coords.T
coords = np.newaxis(coords, axis=0)
# get pdb id and chain num
pdb_name, chain_num = proteinnet_id.split("#")[-1].split("_")[:-1]
pdb_destin = "/".join(route.split("/")[:-1])+"/"+pdb_name+".pdb"
# download pdb file and select appropiate
download_pdb(pdb_name, pdb_destin)
clean_pdb(pdb_destin, chain_num=chain_num)
# load trajectory scaffold and replace coordinates - assumes same order
scaffold = mdtraj.load_pdb(pdb_destin)
scaffold.xyz = coords
scaffold.save(route)
return pdb_destin, route
def main(args):
traj = md.load_pdb('http://www.rcsb.org/pdb/files/2EQQ.pdb')
#strip sidechains of residues 2 and 27
idxs = traj.topology.select(
'name CB CA N C or not (resSeq 2 or resSeq 27)')
traj = traj.atom_slice(idxs)
bins = np.linspace(0.0, 80.0, 41)
bin_centers = bins[:-1] + (bins[1] - bins[0]) / 2.0
#get the index of donor and acceptor attachment atoms
donorAttIdx = traj.topology.select('name CB and resSeq 2')[0]
acceptorAttIdx = traj.topology.select('name CB and resSeq 27')[0]
#print header
print('Frame#\t' + 'A\t'.join('{:.1f}'.format(e) for e in bin_centers))
for frIdx in range(traj.n_frames):
gridDonor = genAV(traj, frIdx, donorAttIdx, 20.0, 2.0, 3.5, 0.9, 0.3,
3.5 + 3.0)
gridAcceptor = genAV(traj, frIdx, acceptorAttIdx, 22.0, 2.0, 3.5, 0.9,
0.4, 3.5 + 3.0)
donorVolSize = np.count_nonzero(np.array(gridDonor.grid) > 1.0)
accVolSize = np.count_nonzero(np.array(gridAcceptor.grid) > 1.0)
if donorVolSize == 0 or accVolSize == 0:
continue
distances = ll.sampleDistanceDistInv(gridDonor, gridAcceptor, 1000000)
freq, _ = np.histogram(distances, bins=bins)
print('{}\t'.format(frIdx) + '\t'.join(str(e) for e in freq))
def _construct_traj(self):
logger.debug(
'Loading Trajectory object for model {0} ({1}/{2})'.format(
self.df.templateid.iloc[0], 0, len(self.df.model_filepath)))
traj = mdtraj.load_pdb(self.df.model_filepath[0])
remove_disulfide_bonds_from_topology(traj.topology)
self.traj = traj
for m, model_filepath in enumerate(self.df.model_filepath[1:]):
logger.debug(
'Loading Trajectory object for model {0} ({1}/{2})'.format(
self.df.templateid.iloc[m + 1], m + 1,
len(self.df.model_filepath)))
traj = mdtraj.load_pdb(model_filepath)
remove_disulfide_bonds_from_topology(traj.topology)
self.traj += traj
def load_pdb(fn):
t = mt.load_pdb(fn)
top = t.topology
protein_indices = top.select("protein")
t = t.atom_slice(protein_indices)
return t[int(t.n_frames * 0.8): ]
def metal_searcher(file):
metal_searcher_results = {}
metal_searcher_results['ok_file_count'] = 0
metal_searcher_results['error_file_count'] = 0
metal_searcher_results['metals_in_file'] = {}
metal_searcher_results['any_metal_in_file'] = 0
for i in metals_list:
metal_searcher_results['metals_in_file'][i] = 0
print(file)
try:
traj = md.load_pdb(file)
metal_searcher_results['ok_file_count'] = 1
except:
metal_searcher_results['error_file_count'] = 1
return metal_searcher_results
topo = traj.topology
one_atom_residue_atoms = [atom.name for atom in topo.atoms if atom.name == atom.residue.name and atom.name in metals_list]
if one_atom_residue_atoms:
metal_searcher_results['any_metal_in_file'] = 1
for atom_name in one_atom_residue_atoms:
metal_searcher_results['metals_in_file'][atom_name] += 1
print(metal_searcher_results)
return metal_searcher_results
def get(self, pdbid):
pdbid = pdbid.lower()
if len(pdbid) != 4:
raise Exception("pdb id must be four characters long")
filename = os.path.join(self.temp_dir, '%s.pdb.gz' % pdbid)
with open(filename, 'wb') as filehandler:
self.conn.retrbinary('RETR pub/pdb/data/structures/divided/pdb/%s/pdb%s.ent.gz' % (pdbid[1:3], pdbid),
filehandler.write)
if self.just_files:
return filename
else:
# then we actually want to load it as a pdb file object and return an mdtraj trajectory
with gzip.open(filename) as filehandler:
text = filehandler.read()
with open(filename[:-3], 'w') as filehandler:
filehandler.write(text)
pdb = mdtraj.load_pdb(filename[:-3])
return pdb
def test_3(tmpdir):
# 1COY gives a small error, due to a broken chain.
pdbids = ['1GAI', '6gsv', '2AAC']
for pdbid in pdbids:
t = md.load_pdb('http://www.rcsb.org/pdb/files/%s.pdb' % pdbid)
t = t.atom_slice(t.top.select_atom_indices('minimal'))
assert_(call_dssp(tmpdir, t), md.compute_dssp(t, simplified=False)[0])
def _execute(self, directory, available_resources):
import mdtraj
from paprika.evaluator import Setup
from simtk.openmm import app
mdtraj_trajectory = mdtraj.load_pdb(self.complex_file_path)
atom_indices_by_role = _atom_indices_by_role(
self.substance, self.complex_file_path
)
host_atom_indices = atom_indices_by_role[Component.Role.Receptor]
host_trajectory = mdtraj_trajectory.atom_slice(host_atom_indices)
host_file_path = os.path.join(directory, "host_input.pdb")
host_trajectory.save(host_file_path)
host_structure = Setup.prepare_host_structure(host_file_path)
self.output_coordinate_path = os.path.join(directory, "output.pdb")
with open(self.output_coordinate_path, "w") as file:
app.PDBFile.writeFile(
host_structure.topology, host_structure.positions, file, True
)
def test_1():
for fn in ['1bpi.pdb', '1vii.pdb', '4K6Q.pdb', '1am7_protein.pdb']:
t = md.load_pdb(get_fn(fn))
t = t.atom_slice(t.top.select_atom_indices('minimal'))
f = lambda : assert_(call_dssp(t), md.compute_dssp(t, simplified=False)[0])
f.description = 'test_1: %s' % fn
yield f
def res_seq(pdb):
p = mt.load_pdb(pdb)
seq = list(p.topology.residues)
seq = [str(x)[:3] for x in seq if len(str(x)) >= 3]
return list(map(long2short, seq))
def test_4():
t = md.load_pdb(get_fn('1am7_protein.pdb'))
a = md.compute_dssp(t, simplified=True)
b = md.compute_dssp(t, simplified=False)
assert len(a) == len(b)
assert len(a[0]) == len(b[0])
assert list(np.unique(a[0])) == ['C', 'E', 'H']
def main():
inp, topf, sel, outn, dt = parse_args()
if os.path.isdir(inp):
xtcfs = sorted([os.path.join(inp, _) for _ in os.listdir(inp)])
if topf is None:
print("Eorror: top file is required!")
exit(0)
else:
meta = pd.read_pickle(inp)
topf = meta["top_fn"].values[0]
xtcfs = meta["traj_fn"].values
top = md.load_pdb(topf)
ndx = top.top.select(sel)
top = top.atom_slice(ndx)
xtcall = None
for xtcf in xtcfs:
xtc = md.load_xtc(xtcf, top=topf, atom_indices=ndx, stride=dt)
if xtcall is None:
xtcall = xtc
else:
xtc.topology = xtcall.topology
xtcall = xtcall.join(xtc)
outtop = "%s.pdb" % outn
outdcd = "%s.dcd" % outn
top.save_pdb(outtop)
xtcall.superpose(top)
xtcall.save_dcd(outdcd)
def make_perturbed_traj(traj_path, save_path, f_perturb): traj = mdtraj.load_pdb(traj_path) mean_dist = np.mean( np.sum( (traj.xyz[0] - traj.xyz[0].mean(0))**2, axis = 1), axis = 0) x = np.random.rand(traj.xyz.shape[1], traj.xyz.shape[2])*2 - 1 shifts = x * f_perturb * mean_dist traj.xyz += shifts traj.save(save_path)
def test_refine_explicit_md_short():
with integrationtest_context(set_up_project_stage='solvated'):
targetid = 'EGFR_HUMAN_D0'
templateid = 'KC1D_HUMAN_D0_4KB8_D'
refine_explicit_md(process_only_these_targets=[targetid],
process_only_these_templates=[templateid],
sim_length=2.0 * unit.femtosecond,
nsteps_per_iteration=1,
verbose=True)
explicit_metadata_filepath = os.path.join(
default_project_dirnames.models, targetid,
'refine_explicit_md-meta0.yaml')
explicit_model_filepath = os.path.join(default_project_dirnames.models,
targetid, templateid,
'explicit-refined.pdb.gz')
explicit_energies_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid,
'explicit-energies.txt')
explicit_log_filepath = os.path.join(default_project_dirnames.models,
targetid, templateid,
'explicit-log.yaml')
assert all(
map(os.path.exists, [
explicit_model_filepath, explicit_energies_filepath,
explicit_log_filepath
]))
with open(explicit_log_filepath) as explicit_log_file:
explicit_log = yaml.load(explicit_log_file)
assert explicit_log.get('finished') is True
assert explicit_log.get('successful') is True
explicit_model_traj = mdtraj.load_pdb(explicit_model_filepath)
def metal_searcher(file):
metal_searcher_results = {}
metal_searcher_results['one_atom_residue_names'] = []
metal_searcher_results['ok_file_count'] = 0
metal_searcher_results['error_file_count'] = 0
print(file)
try:
traj = md.load_pdb(file)
metal_searcher_results['ok_file_count'] = 1
except:
metal_searcher_results['error_file_count'] = 1
return metal_searcher_results
topo = traj.topology
one_atom_residue_atoms = [atom for atom in topo.atoms if atom.name == atom.residue.name]
if one_atom_residue_atoms:
for atom in one_atom_residue_atoms:
metal_searcher_results['one_atom_residue_names'].append(str(atom.name))
print(metal_searcher_results)
return metal_searcher_results
def fasta_dict(fasta_dir, prot_ids):
fasta_seq_dict = {}
for p in prot_ids:
if os.path.exists(os.path.join(os.path.join(fasta_dir, p+"_protein.pdb"))) and\
not os.path.exists(os.path.join(os.path.join(fasta_dir, p+"_protein.pdb"))):
try:
traj = mt.load_pdb(os.path.join(os.path.join(fasta_dir, p+"_protein.pdb")))
fasta = "".join(traj.topology.to_fasta())
print("Converting PDB to fasta")
with open(os.path.join(os.path.join(fasta_dir, p+"_protein.pdb")), 'w') as tof:
tof.write(">%s\n" % p)
tof.write("%s\n" % fasta)
tof.close()
except:
print("Doesn't find pdb or fasta file......")
fn = os.path.join(os.path.join(fasta_dir, p+".fasta"))
if os.path.exists(fn):
fasta_seq_dict[p] = get_fasta_seq(fn)
else:
print("fasta file not exists: ", fn)
fasta_seq_dict[p] = "X"
return fasta_seq_dict
def loader(file):
print(file)
try:
traj = md.load_pdb(file)
except:
return file
def test_1():
for fn in ['1bpi.pdb', '1vii.pdb', '4K6Q.pdb', '1am7_protein.pdb']:
t = md.load_pdb(get_fn(fn))
t = t.atom_slice(t.top.select_atom_indices('minimal'))
f = lambda : assert_(call_dssp(t), md.compute_dssp(t, simplified=False)[0])
f.description = 'test_1: %s' % fn
yield f
def metal_searcher(file):
metal_searcher_results = {}
metal_searcher_results['one_atom_residue_names'] = []
metal_searcher_results['ok_file_count'] = 0
metal_searcher_results['error_file_count'] = 0
print(file)
try:
traj = md.load_pdb(file)
metal_searcher_results['ok_file_count'] = 1
except:
metal_searcher_results['error_file_count'] = 1
return metal_searcher_results
topo = traj.topology
one_atom_residue_atoms = [
atom for atom in topo.atoms if atom.name == atom.residue.name
]
if one_atom_residue_atoms:
for atom in one_atom_residue_atoms:
metal_searcher_results['one_atom_residue_names'].append(
str(atom.name))
print(metal_searcher_results)
return metal_searcher_results
def test_solvate_existing_structure_protocol():
"""Tests solvating a single methanol molecule in water."""
import mdtraj
methanol_component = Component("CO")
methanol_substance = Substance()
methanol_substance.add_component(methanol_component, ExactAmount(1))
water_substance = Substance()
water_substance.add_component(Component("O"), MoleFraction(1.0))
with tempfile.TemporaryDirectory() as temporary_directory:
build_methanol_coordinates = BuildCoordinatesPackmol("build_methanol")
build_methanol_coordinates.max_molecules = 1
build_methanol_coordinates.substance = methanol_substance
build_methanol_coordinates.execute(temporary_directory,
ComputeResources())
methanol_residue_name = build_methanol_coordinates.assigned_residue_names[
methanol_component.identifier]
solvate_coordinates = SolvateExistingStructure("solvate_methanol")
solvate_coordinates.max_molecules = 9
solvate_coordinates.substance = water_substance
solvate_coordinates.solute_coordinate_file = (
build_methanol_coordinates.coordinate_file_path)
solvate_coordinates.execute(temporary_directory, ComputeResources())
solvated_system = mdtraj.load_pdb(
solvate_coordinates.coordinate_file_path)
assert solvated_system.n_residues == 10
assert solvated_system.top.residue(0).name == methanol_residue_name
def test_refine_explicit_md_short():
with integrationtest_context(set_up_project_stage='solvated'):
targetid = 'EGFR_HUMAN_D0'
templateid = 'KC1D_HUMAN_D0_4KB8_D'
refine_explicit_md(
process_only_these_targets=[targetid],
process_only_these_templates=[templateid],
sim_length=2.0*unit.femtosecond,
nsteps_per_iteration=1,
verbose=True
)
explicit_metadata_filepath = os.path.join(
default_project_dirnames.models, targetid, 'refine_explicit_md-meta0.yaml'
)
explicit_model_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'explicit-refined.pdb.gz'
)
explicit_energies_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'explicit-energies.txt'
)
explicit_log_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'explicit-log.yaml'
)
assert all(map(
os.path.exists,
[explicit_model_filepath, explicit_energies_filepath, explicit_log_filepath]
))
with open(explicit_log_filepath) as explicit_log_file:
explicit_log = yaml.load(explicit_log_file)
assert explicit_log.get('finished') is True
assert explicit_log.get('successful') is True
explicit_model_traj = mdtraj.load_pdb(explicit_model_filepath)
def __init__(self, PDB_filename, free_energy, expdata_filename=None, use_log_normal_distances=False,
dloggamma=np.log(1.01), gamma_min=0.2, gamma_max=10.0):
"""Initialize the class.
INPUTS
conf A molecular structure as an msmbuilder Conformation() object.
NOTE: For cases where the structure is an ensemble (say, from clustering)
and the modeled NOE distances and coupling constants are averaged,
the structure itself can just be a placeholder with the right atom names
and numbering
free_energy The (reduced) free energy f = beta*F of this conformation
"""
self.PDB_filename = PDB_filename
self.expdata_filename = expdata_filename
self.conf = mdtraj.load_pdb(PDB_filename)
# Convert the coordinates from nm to Angstrom units
self.conf.xyz = self.conf.xyz*10.0
# The (reduced) free energy f = beta*F of this structure, as predicted by modeling
self.free_energy = free_energy
# Flag to use log-normal distance errors log(d/d0)
self.use_log_normal_distances = use_log_normal_distances
# Store info about gamma^(-1/6) scaling parameter array
self.dloggamma = dloggamma
self.gamma_min = gamma_min
self.gamma_max = gamma_max
self.allowed_gamma = np.exp(np.arange(np.log(self.gamma_min), np.log(self.gamma_max), self.dloggamma))
# Store distance restraint info
self.distance_restraints = []
self.distance_equivalency_groups = {}
self.ambiguous_groups = [] # list of pairs of group indices, e.g.: [ [[1,2,3],[4,5,6]], [[7],[8]], ...]
self.ndistances = 0
# Store dihedral restraint info
self.dihedral_restraints = []
self.dihedral_equivalency_groups = {}
self.dihedral_ambiguity_groups = {}
self.ndihedrals = 0
# Create a KarplusRelation object
self.karplus = KarplusRelation()
# variables to store pre-computed SSE and effective degrees of freedom (d.o.f.)
self.sse_distances = np.array([0.0 for gamma in self.allowed_gamma])
self.Ndof_distances = None
self.sse_dihedrals = None
self.Ndof_dihedrals = None
self.betas = None # if reference is used, an array of N_j betas for each distance
self.neglog_reference_priors = None
self.sum_neglog_reference_priors = 0.0
# If an experimental data file is given, load in the information
self.expdata_filename = expdata_filename
if expdata_filename != None:
self.load_expdata(expdata_filename)
def test_topology(get_fn):
top = md.load_pdb(get_fn('native.pdb')).topology
with HDF5TrajectoryFile(temp, 'w') as f:
f.topology = top
with HDF5TrajectoryFile(temp) as f:
assert f.topology == top
def test_load_trajectory(get_fn):
# Compare a TNG file to the PDB file it was created from.
pdbtraj = md.load_pdb(get_fn('frame0.pdb'))
tngtraj = md.load_tng(get_fn('frame0.tng'), top=pdbtraj.topology)
eq(pdbtraj.n_frames, tngtraj.n_frames)
eq(pdbtraj.unitcell_vectors, tngtraj.unitcell_vectors)
eq(pdbtraj.xyz, tngtraj.xyz)
def test_3nch_conect():
# This has conect entries that use all available digits, good failure case.
t1 = load_pdb(get_fn('3nch.pdb.gz'))
top, bonds = t1.top.to_dataframe()
bonds = dict(((a, b), 1) for (a, b) in bonds)
eq(bonds[19782, 19783], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19784], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19785], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19786], 1) # Check that last SO4 molecule has right bonds
def _store_highest_seqid_model(self):
models_sorted = self.df.sort('seqid').templateid
for modelid in models_sorted:
model_filepath = os.path.join(self.models_target_dir, modelid, self.model_filename)
if os.path.exists(model_filepath):
self.ref_modelid = modelid
self.ref_model_filepath = model_filepath
break
self.ref_model_traj = mdtraj.load_pdb(self.ref_model_filepath)
def _align_structures(self):
for irow in range(self.nrows):
row_first_col = irow * self.ncols
self._ref_traj = mdtraj.load_pdb(self.structure_filepaths[row_first_col])
self._heavy_atoms = self._ref_traj.topology.select('not name H')
for icol in range(self.ncols):
alignment_index = row_first_col + icol
structure_filepath = self.structure_filepaths[alignment_index]
self._align_structure(structure_filepath, alignment_index)
def __init__(self, PDB_filename, lam, free_energy, data=None,
use_log_normal_noe=False, dloggamma=np.log(1.01), gamma_min=0.2,
gamma_max=10.0):
"""Initialize the Restraint class.
INPUTS
PDB_filename A topology file (*.pdb)
lam lambda value (between 0 and 1)
free_energy The (reduced) free energy f = beta*F of this conformation
data input data for BICePs (both model and exp)
use_log_normal_distances Not sure what's this...
dloggamma gamma is in log space
gamma_min min value of gamma
gamma_max max value of gamma
"""
# Store restraint info
self.restraints = [] # a list of data container objects for each restraint (e.g. NMR_Chemicalshift_Ca())
self.n = 0
self.sse = 0
self.Ndof = None
# used for exponential reference potential
self.betas = None
self.neglog_exp_ref = None
self.sum_neglog_exp_ref = 0.0
# used for Gaussian reference potential
self.ref_sigma = None
self.ref_mean = None
self.neglog_gau_ref = None
self.sum_neglog_gau_ref = 0.0
#NOTE: This is where we appended the code from previous
self.PDB_filename = PDB_filename
self.data = data
self.conf = mdtraj.load_pdb(PDB_filename)
# Convert the coordinates from nm to Angstrom units
self.conf.xyz = self.conf.xyz*10.0
# The (reduced) free energy f = beta*F of this structure, as predicted by modeling
self.free_energy = lam*free_energy
# Store info about gamma^(-1/6) scaling parameter array
self.dloggamma = dloggamma
self.gamma_min = gamma_min
self.gamma_max = gamma_max
self.allowed_gamma = np.exp(np.arange(np.log(self.gamma_min), np.log(self.gamma_max), self.dloggamma))
# Flag to use log-normal distance errors log(d/d0)
self.use_log_normal_noe = use_log_normal_noe
# Create a KarplusRelation object
self.karplus = KarplusRelation()
def get_renumbered_topol_resnums(target):
models_target_dir = os.path.join(default_project_dirnames.models, target.id)
renumbered_resnums = {}
for topol_type in ['implicit', 'explicit']:
topol_path = os.path.join(models_target_dir, 'topol-renumbered-{}.pdb'.format(topol_type))
if not os.path.exists(topol_path):
continue
traj = mdtraj.load_pdb(topol_path)
res_numbers = [resi.resSeq for resi in traj.top.residues]
renumbered_resnums[topol_type] = res_numbers
logger.info('Will use renumbered residues from {} for target {}'.format(topol_path, target.id))
return renumbered_resnums
def add_bonds_to_pdb(infile, topfile):
t = md.load_pdb(infile)
if t.n_residues == 1:
return
first = t.top.residue(0)
t.restrict_atoms(range(first.n_atoms))
gtop = app.GromacsTopFile(topfile)._currentMoleculeType
top, bonds = t.top.to_dataframe()
bonds = np.array([(row[0],row[1]) for row in gtop.bonds],'int')
bonds = bonds -1
t.top = md.Topology.from_dataframe(top, bonds)
t.save(infile)
def haszn(i):
try:
traj = md.load_pdb(i)
except:
return None
topo = traj.topology
atoms = [str(x) for x in topo.atoms]
atomszn = [x for x in atoms if "ZN" in x]
if atomszn:
znpdbs.append(i)
def metal_scanner(file):
contains_metal = False
try:
traj = md.load_pdb(file)
except:
return None
if [atom for atom in traj.top.atoms if atom.name == 'ZN']:
contains_metal = True
print(contains_metal)
return contains_metal
def metal_scanner(file):
contains_metal = [False, file]
try:
traj = md.load_pdb(file)
except:
return [None, file]
if [atom for atom in traj.top.atoms if atom.name == 'ZN' and atom.residue.name == 'ZN']:
contains_metal[0] = True
print(contains_metal)
return contains_metal
def file_reader(file):
global ok_file_count
global error_file_count
try:
traj = md.load_pdb(file)
with lock:
ok_file_count.value += 1
except:
with lock:
error_file_count.value += 1
print(file)
print(ok_file_count.value)
def __init__(self, PDB_filename, ref, dlogsigma=np.log(1.02),
sigma_min=0.05, sigma_max=20.0, use_global_ref_sigma=True):
"""Initialize the Restraint class.
:param str PDB_filename: A topology file (*.pdb)
:param str ref: Reference potential.
:param float default=np.log(1.02) dlogsigma:
:param float sigma_min: default = 0.05
:param float sigma_max: default = 20.0
:param bool default=True use_global_ref_sigma: """
# Store restraint info
self.restraints = [] # a list of data container objects for each restraint (e.g. NMR_Chemicalshift_Ca())
# Conformational Information
self.PDB_filename = PDB_filename
self.conf = mdtraj.load_pdb(PDB_filename)
# Convert the coordinates from nm to Angstrom units
self.conf.xyz = self.conf.xyz*10.0
# used for exponential reference potential
self.betas = None
self.neglog_exp_ref = None
self.sum_neglog_exp_ref = 0.0
# used for Gaussian reference potential
self.ref_sigma = None
self.ref_mean = None
self.neglog_gaussian_ref = None
self.sum_neglog_gaussian_ref = 0.0
self.use_global_ref_sigma = use_global_ref_sigma
self.see = None
# Storing the reference potential
self.ref = ref
# set sigma range
self.dlogsigma = dlogsigma
self.sigma_min = sigma_min
self.sigma_max = sigma_max
self.allowed_sigma = np.exp(np.arange(np.log(self.sigma_min),
np.log(self.sigma_max), self.dlogsigma))
self.sigma_index = len(self.allowed_sigma)/2
self.sigma = self.allowed_sigma[self.sigma_index]
def __init__(self, PDB_filename, ref, dlogsigma=np.log(1.02), sigma_min=0.05, sigma_max=20.0, use_global_ref_sigma=True):
"""Initialize the Restraint class.
INPUTS
------
PDB_filename A topology file (*.pdb)
data input data for BICePs (both model and exp)
ref Reference potential.
"""
# Store restraint info
self.restraints = [] # a list of data container objects for each restraint (e.g. NMR_Chemicalshift_Ca())
# Conformational Information
self.PDB_filename = PDB_filename
self.conf = mdtraj.load_pdb(PDB_filename)
# Convert the coordinates from nm to Angstrom units
self.conf.xyz = self.conf.xyz*10.0
# used for exponential reference potential
self.betas = None
self.neglog_exp_ref = None
self.sum_neglog_exp_ref = 0.0
# used for Gaussian reference potential
self.ref_sigma = None
self.ref_mean = None
self.neglog_gaussian_ref = None
self.sum_neglog_gaussian_ref = 0.0
self.use_global_ref_sigma = use_global_ref_sigma
self.see = None
# Storing the reference potential
self.ref = ref
# set sigma range
self.dlogsigma = dlogsigma
self.sigma_min = sigma_min
self.sigma_max = sigma_max
# self.allowed_sigma = np.exp(np.arange(np.log(self.sigma_min),
# np.log(self.sigma_max), self.dlogsigma))
self.allowed_sigma = np.arange(self.sigma_min,self.sigma_max,self.dlogsigma)
self.sigma_index = len(self.allowed_sigma)/2
self.sigma = self.allowed_sigma[self.sigma_index]
def test_refine_implicit_md_short():
with integrationtest_context(set_up_project_stage='clustered'):
targetid = 'EGFR_HUMAN_D0'
templateid = 'KC1D_HUMAN_D0_4KB8_D'
refine_implicit_md(
process_only_these_targets=[targetid],
process_only_these_templates=[templateid],
sim_length=2.0*unit.femtosecond,
nsteps_per_iteration=1,
minimization_steps=1,
loglevel='debug'
)
implicit_metadata_filepath = os.path.join(
default_project_dirnames.models, targetid, 'refine_implicit_md-meta0.yaml'
)
implicit_model_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'implicit-refined.pdb.gz'
)
implicit_energies_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'implicit-energies.txt'
)
implicit_log_filepath = os.path.join(
default_project_dirnames.models, targetid, templateid, 'implicit-log.yaml'
)
assert all(map(
os.path.exists,
[implicit_model_filepath, implicit_energies_filepath, implicit_log_filepath]
))
with open(implicit_log_filepath) as implicit_log_file:
implicit_log = yaml.load(implicit_log_file)
assert implicit_log.get('finished') is True
assert implicit_log.get('successful') is True
assert implicit_log.get('ph') == 8.0
assert os.path.exists(implicit_metadata_filepath)
with open(implicit_metadata_filepath) as implicit_metadata_file:
implicit_metadata = yaml.load(implicit_metadata_file)
assert implicit_metadata.get('refine_implicit_md').get('custom_residue_variants') == {
'EGFR_HUMAN_D0': {49: 'ASH'}
}
implicit_model_traj = mdtraj.load_pdb(implicit_model_filepath)
resis = [resi for resi in implicit_model_traj.top.residues]
resi49 = resis[49]
resi49_atom_strings = [str(atom) for atom in resi49.atoms]
assert 'ASP50-HD2' in resi49_atom_strings