def unpickle_miller_arrays(file_name):
result = easy_pickle.load(file_name)
# Python 3 pickle fix
# =========================================================================
if sys.version_info.major == 3:
result = easy_pickle.fix_py2_pickle(result)
# =========================================================================
if (isinstance(result, miller.array)):
return [result]
result = list(result)
for miller_array in result:
if (not isinstance(miller_array, miller.array)):
return None
return result
def load_db(file_name=None):
if (file_name is None):
file_name = libtbx.env.find_in_repositories(
relative_path="chem_data/polygon_data/all_mvd.pickle",
test=os.path.isfile)
assert os.path.isfile(file_name)
database_dict = easy_pickle.load(file_name)
# Python 3 pickle fix
# =========================================================================
if sys.version_info.major == 3:
database_dict = easy_pickle.fix_py2_pickle(database_dict)
# =========================================================================
return database_dict
def __init__(self, models, log):
db_path = libtbx.env.find_in_repositories(
relative_path="chem_data/rama_z/top8000_rama_z_dict.pkl",
test=os.path.isfile)
self.log = log
# this takes ~0.15 seconds, so I don't see a need to cache it somehow.
self.db = easy_pickle.load(db_path)
# Python 3 pickle fix
# =========================================================================
if sys.version_info.major == 3:
self.db = easy_pickle.fix_py2_pickle(self.db)
# =========================================================================
self.calibration_values = {
'H': (-0.045355950779513175, 0.1951165524439217),
'S': (-0.0425581278436754, 0.20068584887814633),
'L': (-0.018457764754231075, 0.15788374669456848),
'W': (-0.016806654295023003, 0.12044960331869274)
}
self.residue_counts = {"H": 0, "S": 0, "L": 0}
self.z_score = {"H": None, "S": None, "L": None, 'W': None}
self.means = {"H": {}, "S": {}, "L": {}}
self.stds = {"H": {}, "S": {}, "L": {}}
self.phi_step = 4
self.psi_step = 4
self.n_phi_half = 45
self.n_psi_half = 45
# this is needed to disable e.g. selection functionality when
# multiple models are present
self.n_models = len(models)
self.res_info = []
for model in models:
if model.get_hierarchy().models_size() > 1:
hierarchy = iotbx.pdb.hierarchy.root()
m = model.get_hierarchy().models()[0].detached_copy()
hierarchy.append_model(m)
asc = hierarchy.atom_selection_cache()
else:
hierarchy = model.get_hierarchy()
asc = model.get_atom_selection_cache()
sec_str_master_phil = iotbx.phil.parse(sec_str_master_phil_str)
ss_params = sec_str_master_phil.fetch().extract()
ss_params.secondary_structure.protein.search_method = "from_ca"
ss_params.secondary_structure.from_ca_conservative = True
ssm = ss_manager(
hierarchy,
atom_selection_cache=asc,
geometry_restraints_manager=None,
sec_str_from_pdb_file=None,
# params=None,
params=ss_params.secondary_structure,
was_initialized=False,
mon_lib_srv=None,
verbose=-1,
log=null_out(),
# log=sys.stdout,
)
filtered_ann = ssm.actual_sec_str.deep_copy()
filtered_ann.remove_short_annotations(
helix_min_len=4,
sheet_min_len=4,
keep_one_stranded_sheets=True)
self.helix_sel = asc.selection(
filtered_ann.overall_helices_selection())
self.sheet_sel = asc.selection(
filtered_ann.overall_sheets_selection())
used_atoms = set()
for three in generate_protein_threes(hierarchy=hierarchy,
geometry=None):
main_residue = three[1]
phi_psi_atoms = three.get_phi_psi_atoms()
if phi_psi_atoms is None:
continue
phi_atoms, psi_atoms = phi_psi_atoms
key = [x.i_seq for x in phi_atoms] + [psi_atoms[-1].i_seq]
key = "%s" % key
if key not in used_atoms:
phi, psi = three.get_phi_psi_angles()
rkey = three.get_ramalyze_key()
resname = main_residue.resname
ss_type = self._figure_out_ss(three)
self.res_info.append(
["", rkey, resname, ss_type, phi, psi])
self.residue_counts[ss_type] += 1
used_atoms.add(key)
self.residue_counts["W"] = self.residue_counts[
"H"] + self.residue_counts["S"] + self.residue_counts["L"]