def get_signatures(pdbfile, amino, property_list, nbrs):
assert os.path.isfile(pdbfile)
assert isinstance(property_list, list)
assert nbrs > 2
assert all([
p in cg.CoarseGrainProperty.supported_properties()
for p in property_list
])
pdblist = sd.read_pdb(pdbfile)
assert len(pdblist) > 0
chain = list(pdblist[0].keys())[0]
pdb = pdblist[0][chain]
try:
ca_trace = sd.pdb_to_catrace(pdb)
except:
return np.array([])
assert isinstance(pdb, sd.PDBStructure)
residue_ids = [r for r in pdb.residue_ids if r > 0][1:-1]
residue_ids = [
r for r in residue_ids if sd.get_amino(pdb.residue_name(r)) == amino
]
residue_ids = cg.filter_complete_residues(pdb, residue_ids)
vol = cg.Volume3D(size=cg.VolumeSignatureGrid.size(),
dim=cg.VolumeSignatureGrid.dim())
if len(residue_ids) > 0:
vol_signature = []
for r in residue_ids:
vol_signature.append(vol.get_signature(pdb, r, reset=True))
nbr_signature = cg.cg_neighbor_signature(ca_trace,
residues=residue_ids,
properties=property_list,
topn=nbrs)
vol_signature, nbr_signature = np.array(vol_signature), np.array(
nbr_signature)
if vol_signature.shape[0] == nbr_signature.shape[0]:
return np.concatenate((nbr_signature, vol_signature), axis=1)
return np.array([])
type=str,
required=True)
args = parser.parse_args()
if not os.path.isdir(args.pdb_dir):
raise Exception("Invalid pdb directory location [%s]" % args.pdb_dir)
if not os.path.isdir(os.path.dirname(args.output)):
raise Exception("Output directory does not exists [%s]" %
os.path.dirname(args.output))
if os.path.isfile(args.output):
warnings.warn("output file will be overwritten [%s]" % args.output)
amino = sd.get_amino(args.amino.upper())
properties = args.properties if len(
args.properties) > 0 else cg.CoarseGrainProperty.default_properties()
logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger('prepare_volume_feature_vectors')
pdb_files = [
os.path.join(args.pdb_dir, f) for f in os.listdir(args.pdb_dir)
if f.endswith('.pdb')
]
all_features = None
for i, filename in enumerate(pdb_files):
sig = get_signatures(filename,
amino,
property_list=properties,
import os
import regressor as reg
import coarse_graining as cg
import structural_dynamics as sd
if __name__ == "__main__":
restype, chain_length = "ARG", 10
model_file = os.path.join(os.path.dirname(__file__), 'data', 'arg_chain')
pdb_file = os.path.join(os.path.dirname(__file__), 'data', 'structure.pdb')
assert os.path.isdir(model_file) and os.path.isfile(pdb_file)
ca_trace = sd.read_trajectory_catrace(pdb_file)[0]['B']
model = reg.ChainedMLP(chain_length=chain_length, chained_incr=3)
model.load(model_file)
results = cg.place_atoms(ca_trace, model, sd.get_amino(restype))
resids = sorted([int(k) for k in results.keys()])
ca_coords = [ca_trace.xyz(r) for r in resids]
entity, atom_id = [], 1
for i, r in enumerate(resids):
for atom, coords in results[r].items():
entity.append({
'resname': restype,
'resid': r,
'atomname': atom,
'atomid': atom_id,
'x': coords['x'],
'y': coords['y'],
'z': coords['z']
})
atom_id += 1
entity.append({
'resname': restype,
def placement_vector(self, aa_type):
aa = get_amino(aa_type).name(one_letter_code=False)
if aa in self.__model_track:
res_ids = sorted(self.__model_track[aa])
return self.__residue_lookup.placement_vector(res_ids)
return None
def aligning_matrix(self, aa_type):
aa = get_amino(aa_type).name(one_letter_code=False)
if aa in self.__model_track:
res_ids = sorted(self.__model_track[aa])
return self.__residue_lookup.aligning_matrix(res_ids)
return None
def residue_ids(self, aa_type):
aa = get_amino(aa_type).name(one_letter_code=False)
if aa in self.__model_track:
return sorted(self.__model_track[aa])
return []