def validate_residues(self):
from mmtbx.conformation_dependent_library import generate_protein_threes
from mmtbx.rotamer import ramachandran_eval, rotamer_eval
# this is so we generate rama_eval only once
rama_eval = ramachandran_eval.RamachandranEval()
rota_eval = rotamer_eval.RotamerEval()
rotamer_id = rotamer_eval.RotamerID() # loads in the rotamer names
threes = generate_protein_threes(hierarchy=self.pdb_hierarchy,
include_non_linked=True,
backbone_only=False,
geometry=None)
for i, three in enumerate(threes):
if i == 0:
self.residues.append(
ValidationResidue(three,
rama_eval,
rota_eval,
rotamer_id,
index=0))
self.residues.append(
ValidationResidue(three, rama_eval, rota_eval, rotamer_id))
if three.end:
self.residues.append(
ValidationResidue(three,
rama_eval,
rota_eval,
rotamer_id,
index=2))
def exercise_phi_psi_extraction():
for n_prox, raw_records in [
([0, 0], rec_1_residue),
([0, 0], rec_2_residues),
([4, 2], rec_3_residues),
([6, 4], rec_4_residues),
([0, 0], rec_2_chains),
([0, 0], rec_2_segids),
([8, 4], rec_2_acs_edge),
([8, 4], rec_2_acs_middle),
([6, 4], rec_4_residues_isertions),
([12, 10], pdb_1yjp),
([8, 4], pdb_1yjp_minus_4),
([4, 2], rec_3_res_ac_h),
([8, 4], rec_2_acs_middle_one_atom_1),
([8, 4], rec_2_acs_middle_one_atom_2),
([8, 4], rec_2_acs_middle_one_atom_3),
]:
tmp_hierarchy = iotbx.pdb.input(
source_info=None,
lines=flex.split_lines(raw_records)).construct_hierarchy()
for opp in range(2):
proxies = []
for three in generate_protein_threes(
hierarchy=tmp_hierarchy,
geometry=None):
ppp = three.get_dummy_dihedral_proxies(only_psi_phi_pairs=opp)
print three,'ppp',len(ppp)
proxies.extend(ppp)
print len(proxies), n_prox
assert len(proxies) == n_prox[opp], \
"Expected %d, got %d" % (
n_prox[opp],
len(proxies),
)
def update_restraints(
hierarchy,
geometry, # restraints_manager,
current_geometry=None, # xray_structure!!
sites_cart=None,
cdl_proxies=None,
ideal=True,
esd=True,
esd_factor=1.0,
log=None,
verbose=False,
):
global registry
registry = RestraintsRegistry()
if current_geometry:
assert not sites_cart
sites_cart = current_geometry.sites_cart()
if sites_cart:
pdb_atoms = hierarchy.atoms()
# XXX PDB_TRANSITION VERY SLOW
for j_seq, atom in enumerate(pdb_atoms):
atom.xyz = sites_cart[j_seq]
threes = None
average_updates = 0
total_updates = 0
for threes in generate_protein_threes(
hierarchy,
geometry,
omega_cdl=True,
# verbose=verbose,
):
threes.apply_updates = apply_updates
if threes.cis_group():
if verbose and 0:
print "cis " * 20
print threes
continue
restraint_values = get_restraint_values(threes)
if restraint_values is None:
continue
if restraint_values[0] == "I":
average_updates += 1
else:
total_updates += 1
threes.apply_updates(threes, restraint_values, cdl_proxies, ideal=ideal, esd=esd, esd_factor=esd_factor)
if registry.n:
threes.apply_average_updates(registry)
assert 0
geometry.reset_internals()
if verbose and threes and threes.errors:
if log:
log.write(" Residues not completely updated with CDL restraints\n\n")
for line in threes.errors:
if log:
log.write("%s\n" % line)
else:
print line
return geometry
def exercise_phi_psi_extraction():
for n_prox, raw_records in [
([0, 0], rec_1_residue),
([0, 0], rec_2_residues),
([4, 2], rec_3_residues),
([6, 4], rec_4_residues),
([0, 0], rec_2_chains),
([0, 0], rec_2_segids),
([8, 4], rec_2_acs_edge),
([8, 4], rec_2_acs_middle),
([6, 4], rec_4_residues_isertions),
([12, 10], pdb_1yjp),
([8, 4], pdb_1yjp_minus_4),
([4, 2], rec_3_res_ac_h),
([8, 4], rec_2_acs_middle_one_atom_1),
([8, 4], rec_2_acs_middle_one_atom_2),
([8, 4], rec_2_acs_middle_one_atom_3),
]:
tmp_hierarchy = iotbx.pdb.input(
source_info=None,
lines=flex.split_lines(raw_records)).construct_hierarchy()
for opp in range(2):
proxies = []
for three in generate_protein_threes(hierarchy=tmp_hierarchy,
geometry=None):
ppp = three.get_dummy_dihedral_proxies(only_psi_phi_pairs=opp)
print(three, 'ppp', len(ppp))
proxies.extend(ppp)
print(len(proxies), n_prox)
assert len(proxies) == n_prox[opp], \
"Expected %d, got %d" % (
n_prox[opp],
len(proxies),
)
def extract_proxies(self):
self.proxies = ext.shared_phi_psi_proxy()
from mmtbx.conformation_dependent_library import generate_protein_threes
selected_h = self.pdb_hierarchy.select(self.bool_atom_selection)
n_seq = flex.max(selected_h.atoms().extract_i_seq())
for three in generate_protein_threes(
hierarchy=selected_h,
geometry=None):
rc = three.get_phi_psi_atoms()
if rc is None: continue
phi_atoms, psi_atoms = rc
rama_key = three.get_ramalyze_key()
i_seqs = [atom.i_seq for atom in phi_atoms] + [psi_atoms[-1].i_seq]
resnames = three.get_resnames()
r_name = resnames[1]
assert rama_key in range(6)
text_rama_key = ramalyze.res_types[rama_key]
assert text_rama_key in ["general", "glycine", "cis-proline", "trans-proline",
"pre-proline", "isoleucine or valine"]
proxy = ext.phi_psi_proxy(
residue_name=r_name,
residue_type=text_rama_key,
i_seqs=i_seqs)
if not is_proxy_present(self.proxies, n_seq, proxy):
self.proxies.append(proxy)
print >> self.log, ""
print >> self.log, " %d Ramachandran restraints generated." % (
self.get_n_proxies())
def extract_proxies(self, log):
self.proxies = ext.shared_phi_psi_proxy()
from mmtbx.conformation_dependent_library import generate_protein_threes
selected_h = self.pdb_hierarchy.select(self.bool_atom_selection)
n_seq = flex.max(selected_h.atoms().extract_i_seq())
for three in generate_protein_threes(hierarchy=selected_h,
geometry=None):
rc = three.get_phi_psi_atoms()
if rc is None: continue
phi_atoms, psi_atoms = rc
rama_key = three.get_ramalyze_key()
i_seqs = [atom.i_seq for atom in phi_atoms] + [psi_atoms[-1].i_seq]
resnames = three.get_resnames()
r_name = resnames[1]
assert rama_key in range(6)
text_rama_key = ramalyze.res_types[rama_key]
assert text_rama_key in [
"general", "glycine", "cis-proline", "trans-proline",
"pre-proline", "isoleucine or valine"
]
proxy = ext.phi_psi_proxy(residue_name=r_name,
residue_type=text_rama_key,
i_seqs=i_seqs)
if not is_proxy_present(self.proxies, n_seq, proxy):
self.proxies.append(proxy)
print >> log, ""
print >> log, " %d Ramachandran restraints generated." % (
self.get_n_proxies())
def run():
filename = 'tst_multi.pdb'
f = open(filename, 'w')
f.write(pdb_lines)
f.close()
pdb_inp = pdb.input(filename)
pdb_hierarchy = pdb_inp.construct_hierarchy()
from mmtbx.conformation_dependent_library.tst_rdl import \
get_geometry_restraints_manager
geometry_restraints_manager = get_geometry_restraints_manager(filename)
pdb_hierarchy.reset_i_seq_if_necessary()
refine = [
False, # -179
True, # -44
False, # 86
True, # -22
False, # -179
]
refine += [True] * 5
refine += [False] * 14
omegalyze = [
False,
False,
False,
True,
False,
]
omegalyze += [True] * 3
omegalyze += [False] * 16
from mmtbx.conformation_dependent_library import generate_protein_threes
for i, threes in enumerate(
generate_protein_threes(
pdb_hierarchy,
geometry_restraints_manager,
cdl_class=True,
#verbose=verbose,
)):
print(i, threes)
print(' omega %5.1f' % threes.get_omega_value())
print(" cis? %-5s %s" %
(threes.cis_group(), threes.cis_group(limit=30)))
print(" trans? %-5s %s" %
(threes.trans_group(), threes.trans_group(limit=30)))
print(" rama %s" % threes.get_ramalyze_key())
print(' conf %s' % threes.is_pure_main_conf())
assert threes.cis_group() == refine[i], '%s!=%s' % (threes.cis_group(),
refine[i])
assert threes.cis_group(limit=30) == omegalyze[i]
for j in range(0, 181, 10):
i += 1
print(" %3d %-5s %-8s %-5s" % (
j,
threes._define_omega_a_la_duke_using_limit(j) == 'cis',
threes._define_omega_a_la_duke_using_limit(j, limit=30),
refine[i],
))
assert (threes._define_omega_a_la_duke_using_limit(j) == 'cis'
) == refine[i]
def extract_proxies(self, hierarchy):
self.hierarchy = hierarchy
selected_h = hierarchy.select(self.bool_atom_selection)
n_seq = flex.max(selected_h.atoms().extract_i_seq())
# Drop all previous proxies
self._oldfield_proxies = ext.shared_phi_psi_proxy()
self._emsley_proxies = ext.shared_phi_psi_proxy()
# it would be great to save rama_eval, but the fact that this is called in
# pdb_interpretation, not in mmtbx.model makes it impossible
if self.need_filtering:
self.rama_eval = rama_eval()
for three in generate_protein_threes(hierarchy=selected_h,
geometry=None):
rc = three.get_phi_psi_atoms()
if rc is None: continue
rama_key = three.get_ramalyze_key()
if self.need_filtering:
angles = three.get_phi_psi_angles()
rama_score = self.rama_eval.get_score(rama_key, angles[0],
angles[1])
r_evaluation = self.rama_eval.evaluate_score(
rama_key, rama_score)
phi_atoms, psi_atoms = rc
i_seqs = [atom.i_seq for atom in phi_atoms] + [psi_atoms[-1].i_seq]
resnames = three.get_resnames()
r_name = resnames[1]
assert rama_key in range(6)
text_rama_key = ramalyze.res_types[rama_key]
assert text_rama_key in [
"general", "glycine", "cis-proline", "trans-proline",
"pre-proline", "isoleucine or valine"
]
proxy = ext.phi_psi_proxy(residue_name=r_name,
residue_type=text_rama_key,
i_seqs=i_seqs)
# pick where to put...
if self.params.rama_potential == "oldfield":
if self.need_filtering:
if r_evaluation == ramalyze.RAMALYZE_FAVORED:
self.append_oldfield_proxies(proxy, n_seq)
elif r_evaluation == ramalyze.RAMALYZE_ALLOWED and self.params.restrain_rama_allowed:
self.append_oldfield_proxies(proxy, n_seq)
elif r_evaluation == ramalyze.RAMALYZE_OUTLIER and self.params.restrain_rama_outliers:
self.append_oldfield_proxies(proxy, n_seq)
elif self.params.restrain_allowed_outliers_with_emsley:
self.append_emsley_proxies(proxy, n_seq)
else:
self.append_oldfield_proxies(proxy, n_seq)
else: # self.params.rama_potential == "emsley":
self.append_emsley_proxies(proxy, n_seq)
print("", file=self.log)
print(" %d Ramachandran restraints generated." %
(self.get_n_proxies()),
file=self.log)
print(" %d Oldfield and %d Emsley." %
(self.get_n_oldfield_proxies(), self.get_n_emsley_proxies()),
file=self.log)
def get_phi_psi_atoms(hierarchy):
phi_psi_atoms = []
for three in generate_protein_threes(
hierarchy=hierarchy,
geometry=None):
phi_atoms, psi_atoms = three.get_phi_psi_atoms()
rama_key = three.get_ramalyze_key()
# print "rama_key", rama_key
phi_psi_atoms.append(([phi_atoms, psi_atoms],rama_key))
return phi_psi_atoms
def get_phi_psi_atoms(hierarchy):
phi_psi_atoms = []
for three in generate_protein_threes(hierarchy=hierarchy, geometry=None):
psatoms = three.get_phi_psi_atoms()
if psatoms is not None:
phi_atoms, psi_atoms = psatoms
else:
phi_atoms, psi_atoms = None, None
rama_key = three.get_ramalyze_key()
# print "rama_key", rama_key
phi_psi_atoms.append(([phi_atoms, psi_atoms], rama_key))
return phi_psi_atoms
def test_phi_psi_key(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
#restraints_manager=restraints_manager
geometry=restraints_manager.geometry,
)
):
key = threes.get_cdl_key(force_plus_one=True)
print key, filenames[filename][1]
assert key == filenames[filename][1][i]
def test_phi_psi_key(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
#restraints_manager=restraints_manager
geometry=restraints_manager.geometry,
)
):
key = threes.get_cdl_key(force_plus_one=True)
print key, filenames[filename][1]
assert key == filenames[filename][1][i]
def add_main_chain_atoms(
hierarchy,
geometry_restraints_manager,
verbose=False,
):
from mmtbx.conformation_dependent_library import generate_protein_threes
for three in generate_protein_threes(
hierarchy,
geometry_restraints_manager,
verbose=verbose,
):
print(three)
add_main_chain_atoms_to_protein_three(three)
assert 0
def get_dihedrals_and_phi_psi(model):
from cctbx.geometry_restraints import dihedral_proxy_registry
dihedral_registry = dihedral_proxy_registry(strict_conflict_handling=True)
dihedral_registry.initialize_table()
from mmtbx.conformation_dependent_library import generate_protein_threes
grm = model._processed_pdb_file.geometry_restraints_manager()
dihedral_proxies = grm.get_dihedral_proxies().deep_copy()
for p in dihedral_proxies:
dihedral_registry.add_if_not_duplicated(p)
for three in generate_protein_threes(hierarchy=model.get_hierarchy(),
geometry=None):
proxies = three.get_dummy_dihedral_proxies(only_psi_phi_pairs=False)
for p in proxies:
dihedral_registry.add_if_not_duplicated(p)
return dihedral_registry.proxies
def get_phi_psi_dict(pdb_hierarchy):
rc = {}
for i, three in enumerate(
generate_protein_threes(hierarchy=pdb_hierarchy, geometry=None)):
phi_psi_angles = three.get_phi_psi_angles()
is_alt_conf = ' '
relevant_atoms = {}
for atom in three[1].atoms():
if (atom.name in relevant_atom_names):
if (len(atom.parent().altloc) != 0):
is_alt_conf = atom.parent().altloc
break
id_str = '|%s:%s|' % (three[1].id_str(), is_alt_conf)
rc[id_str] = phi_psi_angles
return rc
def test_average(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
geometry=restraints_manager.geometry,
)
):
if threes.registry.n:
atoms = hierarchy.atoms()
for key in threes.registry.n:
print key
for atom in key:
print atoms[atom].quote()
#assert threes.registry.n.keys() == filenames[filename][3]
assert filenames[filename][3][0] in threes.registry.n
def test_average(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
geometry=restraints_manager.geometry,
)
):
if threes.registry.n:
atoms = hierarchy.atoms()
for key in threes.registry.n:
print key
for atom in key:
print atoms[atom].quote()
#assert threes.registry.n.keys() == filenames[filename][3]
assert filenames[filename][3][0] in threes.registry.n
def get_complete_dihedral_proxies_2(model, log=None):
from six.moves import cStringIO as StringIO
from cctbx.geometry_restraints import dihedral_proxy_registry
from mmtbx.conformation_dependent_library import generate_protein_threes
if log is None:
log = StringIO
dihedral_registry = dihedral_proxy_registry(strict_conflict_handling=True)
dihedral_registry.initialize_table()
grm = model.get_restraints_manager().geometry
dihedral_proxies = grm.get_dihedral_proxies().deep_copy()
for p in dihedral_proxies:
dihedral_registry.add_if_not_duplicated(p)
for three in generate_protein_threes(hierarchy=model.get_hierarchy(),
geometry=None):
proxies = three.get_dummy_dihedral_proxies(only_psi_phi_pairs=False)
for p in proxies:
dihedral_registry.add_if_not_duplicated(p)
return dihedral_registry.proxies
def get_phi_psi_atoms(hierarchy, omega=False):
phi_psi_atoms = []
for three in generate_protein_threes(hierarchy=hierarchy,
geometry=None,
cdl_class=True):
psatoms = three.get_phi_psi_atoms()
if psatoms is not None:
phi_atoms, psi_atoms = psatoms
else:
phi_atoms, psi_atoms = None, None
rama_key = three.get_ramalyze_key()
# print "rama_key", rama_key
if omega:
phi_psi_atoms.append(
([phi_atoms, psi_atoms], rama_key, three.get_omega_value()))
else:
phi_psi_atoms.append(([phi_atoms, psi_atoms], rama_key))
return phi_psi_atoms
def get_dihedrals_and_phi_psi(processed_pdb_file):
from cctbx.geometry_restraints import dihedral_proxy_registry
dihedral_registry = dihedral_proxy_registry(
strict_conflict_handling=True)
dihedral_registry.initialize_table()
from mmtbx.conformation_dependent_library import generate_protein_threes
grm = processed_pdb_file.geometry_restraints_manager()
dihedral_proxies = grm.get_dihedral_proxies().deep_copy()
for p in dihedral_proxies:
dihedral_registry.add_if_not_duplicated(p)
for three in generate_protein_threes(
hierarchy=processed_pdb_file.all_chain_proxies.pdb_hierarchy,
geometry=None):
proxies = three.get_dummy_dihedral_proxies(
only_psi_phi_pairs=False)
for p in proxies:
dihedral_registry.add_if_not_duplicated(p)
return dihedral_registry.proxies
def validate_residues(self) :
from mmtbx.conformation_dependent_library import generate_protein_threes
from mmtbx.rotamer import ramachandran_eval,rotamer_eval
# this is so we generate rama_eval only once
rama_eval = ramachandran_eval.RamachandranEval()
rota_eval = rotamer_eval.RotamerEval()
rotamer_id = rotamer_eval.RotamerID() # loads in the rotamer names
threes = generate_protein_threes(
hierarchy = self.pdb_hierarchy,
include_non_linked=True,
backbone_only=False,
geometry=None)
for i,three in enumerate(threes) :
if i == 0 :
self.residues.append(ValidationResidue(three,rama_eval,
rota_eval,rotamer_id,index=0))
self.residues.append(ValidationResidue(three,rama_eval,
rota_eval,rotamer_id))
if three.end :
self.residues.append(ValidationResidue(three,rama_eval,
rota_eval,rotamer_id,index=2))
def test_cdl_lookup(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
#restraints_manager=restraints_manager
geometry=restraints_manager.geometry,
)
):
res_type_group = cdl_utils.get_res_type_group(
threes[1].resname,
threes[2].resname,
)
key = threes.get_cdl_key(force_plus_one=True)
key = key[-2:]
restraint_values = omega_database[res_type_group][key]
print i, key, restraint_values[:4], filenames[filename][2]
del threes.registry.n
threes.registry.n = {}
assert restraint_values[:4] == filenames[filename][2][i]
def test_cdl_lookup(hierarchy,
filename,
restraints_manager,
):
for i, threes in enumerate(cdl.generate_protein_threes(
hierarchy,
#restraints_manager=restraints_manager
geometry=restraints_manager.geometry,
)
):
res_type_group = cdl_utils.get_res_type_group(
threes[1].resname,
threes[2].resname,
)
key = threes.get_cdl_key(force_plus_one=True)
key = key[-2:]
print 'res_type_group',res_type_group,key
restraint_values = omega_database[res_type_group][key]
print i, key, restraint_values[:4], filenames[filename][2]
del threes.registry.n
threes.registry.n = {}
assert restraint_values[:4] == filenames[filename][2][i]
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)
# =========================================================================
# change keys in pickle to Python 3 string
# very temporary fix until pickle is updated
if sys.version_info.major == 3:
from libtbx.utils import to_str
for key in list(self.db.keys()):
self.db[to_str(key)] = self.db[key]
for subkey in list(self.db[key].keys()):
self.db[to_str(key)][to_str(subkey)] = self.db[key][subkey]
# =========================================================================
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"]
def __init__ (self,
pdb_hierarchy,
outliers_only=False,
show_errors=False,
out=sys.stdout,
quiet=False) :
# Optimization hint: make it possible to pass
# ramachandran_eval.RamachandranEval() from outside.
# Better - convert this to using mmtbx.model.manager where
# RamachandranEval is already available.
validation.__init__(self)
self.n_allowed = 0
self.n_favored = 0
self.n_type = [ 0 ] * 6
self._outlier_i_seqs = flex.size_t()
pdb_atoms = pdb_hierarchy.atoms()
all_i_seqs = pdb_atoms.extract_i_seq()
if (all_i_seqs.all_eq(0)) :
pdb_atoms.reset_i_seq()
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
analysis = ""
output_list = []
count_keys = []
uniqueness_keys = []
r = ramachandran_eval.RamachandranEval()
##if use_segids:
## chain_id = utils.get_segid_as_chainid(chain=chain)
## else:
## chain_id = chain.id
for three in generate_protein_threes(hierarchy=pdb_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
phi = get_dihedral(phi_atoms)
psi = get_dihedral(psi_atoms)
coords = get_center(main_residue) #should find the CA of the center residue
if (phi is not None and psi is not None):
res_type = RAMA_GENERAL
#self.n_total += 1
if (main_residue.resname[0:3] == "GLY"):
res_type = RAMA_GLYCINE
elif (main_residue.resname[0:3] == "PRO"):
is_cis = is_cis_peptide(three)
if is_cis:
res_type = RAMA_CISPRO
else:
res_type = RAMA_TRANSPRO
elif (three[2].resname == "PRO"):
res_type = RAMA_PREPRO
elif (main_residue.resname[0:3] == "ILE" or \
main_residue.resname[0:3] == "VAL"):
res_type = RAMA_ILE_VAL
#self.n_type[res_type] += 1
value = r.evaluate(res_types[res_type], [phi, psi])
ramaType = self.evaluateScore(res_type, value)
is_outlier = ramaType == RAMALYZE_OUTLIER
c_alphas = None
# XXX only save kinemage data for outliers
if is_outlier :
c_alphas = get_cas_from_three(three)
assert (len(c_alphas) == 3)
markup = self.as_markup_for_kinemage(c_alphas)
else:
markup = None
result = ramachandran(
model_id=main_residue.parent().parent().parent().id,
chain_id=main_residue.parent().parent().id,
resseq=main_residue.resseq,
icode=main_residue.icode,
resname=main_residue.resname,
#altloc=main_residue.parent().altloc,
altloc=get_altloc_from_three(three),
segid=None, # XXX ???
phi=phi,
psi=psi,
rama_type=ramaType,
res_type=res_type,
score=value*100,
outlier=is_outlier,
xyz=coords,
markup=markup)
#if result.chain_id+result.resseq+result.icode not in count_keys:
result_key = result.model_id+result.chain_id+result.resseq+result.icode
if result.altloc in ['','A'] and result_key not in count_keys:
self.n_total += 1
self.n_type[res_type] += 1
self.add_to_validation_counts(ramaType)
count_keys.append(result_key)
if (not outliers_only or is_outlier) :
if (result.altloc != '' or
result_key not in uniqueness_keys):
#the threes/conformers method results in some redundant result
# calculations in structures with alternates. Using the
# uniqueness_keys list prevents redundant results being added to
# the final list
self.results.append(result)
uniqueness_keys.append(result_key)
if is_outlier :
i_seqs = main_residue.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0))
self._outlier_i_seqs.extend(i_seqs)
self.results.sort(key=lambda r: r.model_id+r.id_str())
out_count, out_percent = self.get_outliers_count_and_fraction()
fav_count, fav_percent = self.get_favored_count_and_fraction()
self.out_percent = out_percent * 100.0
self.fav_percent = fav_percent * 100.0
def extract_proxies(self, hierarchy):
def _get_motifs():
from phenix.programs.phi_psi_2 import results_manager as pp2
pp2_manager = pp2(model=None, log=self.log)
phi_psi_2_motifs = pp2_manager.get_overall_motif_count_and_output(
None,
self.hierarchy,
return_rama_restraints=True,
)
return phi_psi_2_motifs
phi_psi_2_motifs = None
favored = ramalyze.RAMALYZE_FAVORED
allowed = ramalyze.RAMALYZE_ALLOWED
outlier = ramalyze.RAMALYZE_OUTLIER
self.hierarchy = hierarchy
bool_atom_selection = self._determine_bool_atom_selection(hierarchy)
selected_h = hierarchy.select(bool_atom_selection)
n_seq = flex.max(selected_h.atoms().extract_i_seq())
# Drop all previous proxies
self._oldfield_proxies = ext.shared_phi_psi_proxy()
self._emsley_proxies = ext.shared_phi_psi_proxy()
self._emsley8k_proxies = ext.shared_phi_psi_proxy()
self._phi_psi_2_proxies = ext.shared_phi_psi_proxy()
# it would be great to save rama_eval, but the fact that this is called in
# pdb_interpretation, not in mmtbx.model makes it impossible
self.rama_eval = rama_eval()
outl = []
for three in generate_protein_threes(hierarchy=selected_h,
geometry=None):
rc = three.get_phi_psi_atoms()
if rc is None: continue
rama_key = three.get_ramalyze_key()
angles = three.get_phi_psi_angles()
rama_score = self.rama_eval.get_score(rama_key, angles[0],
angles[1])
r_eval = self.rama_eval.evaluate_score(rama_key, rama_score)
phi_atoms, psi_atoms = rc
i_seqs = [atom.i_seq for atom in phi_atoms] + [psi_atoms[-1].i_seq]
resnames = three.get_resnames()
r_name = resnames[1]
assert rama_key in range(6)
text_rama_key = ramalyze.res_types[rama_key]
assert text_rama_key in [
"general", "glycine", "cis-proline", "trans-proline",
"pre-proline", "isoleucine or valine"
]
# pick where to put...
ev_match_dict = {
favored: self.params.favored,
allowed: self.params.allowed,
outlier: self.params.outlier
}
r_type = ev_match_dict[r_eval]
if r_type == 'oldfield':
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=1) # XXX Not used in oldfield
self.append_oldfield_proxies(proxy, n_seq)
### THIS IS CRUEL. REMOVE ONCE favored/allowed/outlier are made multiple!
if (self.params.inject_emsley8k_into_oldfield_favored):
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=5)
self.append_emsley8k_proxies(proxy, n_seq)
###
elif r_type == 'emsley':
weight = self.params.emsley.weight
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=weight)
self.append_emsley_proxies(proxy, n_seq)
elif r_type == 'emsley8k':
if (r_eval is favored):
weight = self.params.emsley8k.weight_favored
elif (r_eval is allowed):
weight = self.params.emsley8k.weight_allowed
elif (r_eval is outlier):
weight = self.params.emsley8k.weight_outlier
else:
raise RuntimeError("Rama eveluation failed.")
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=weight)
self.append_emsley8k_proxies(proxy, n_seq)
elif r_type == 'phi_psi_2':
from phenix.pdb_tools.phi_psi_2_data import get_phi_psi_key_for_rama_proxy
if phi_psi_2_motifs is None: phi_psi_2_motifs = _get_motifs()
if (r_eval is favored):
strategy = self.params.phi_psi_2.favored_strategy
elif (r_eval is allowed):
strategy = self.params.phi_psi_2.allowed_strategy
elif (r_eval is outlier):
strategy = self.params.phi_psi_2.outlier_strategy
else:
raise RuntimeError("Rama eveluation failed.")
if strategy == 'closest':
strategy += '_%0.1f_%0.1f' % tuple(
three.get_phi_psi_angles())
pp2_key = get_phi_psi_key_for_rama_proxy(
phi_psi_2_motifs,
three,
strategy=strategy,
)
if pp2_key is None: continue
weight = 1
proxy = ext.phi_psi_proxy(residue_type=pp2_key,
i_seqs=i_seqs,
weight=weight)
outl.append([proxy.residue_type, three])
self.append_phi_psi_2_proxies(proxy, n_seq)
elif (r_type is None):
pass
else:
raise RuntimeError("Not an option: %s" % str(r_type))
print("", file=self.log)
print(" %d Ramachandran restraints generated." %
(self.get_n_proxies()),
file=self.log)
print(" %d Oldfield, %d Emsley, %d emsley8k and %d Phi/Psi/2." %
(self.get_n_oldfield_proxies(), self.get_n_emsley_proxies(),
self.get_n_emsley8k_proxies(), self.get_n_phi_psi_2_proxies()),
file=self.log)
if outl:
print(' Rama restraints by Phi/Psi/2')
for pp2, three in outl:
print(' %s : %s' % (three[1].id_str(), pp2.split('|')[0]),
file=self.log)
def fix_rama_outlier(self,
pdb_hierarchy, out_res_num_list, prefix="", minimize=True,
ss_annotation=None,
tried_rama_angles_for_chain={},
tried_final_rama_angles_for_chain={}):
def comb_pair_in_bad_pairs(comb_pair, bad_pairs):
if None in comb_pair:
return False
all_combs = [comb_pair]
all_combs.append((comb_pair[0]-20, comb_pair[1]))
all_combs.append((comb_pair[0]+20, comb_pair[1]))
all_combs.append((comb_pair[0], comb_pair[1]-20))
all_combs.append((comb_pair[0], comb_pair[1]+20))
all_c_adj = []
for p in all_combs:
new_p = p
if p[0] > 180:
new_p = (p[0]-180, p[1])
if p[0] < -180:
new_p = (p[0]+180, p[1])
if p[1] > 180:
new_p = (p[0], p[1]-180)
if p[0] < -180:
new_p = (p[0], p[1]+180)
all_c_adj.append(new_p)
for p in all_c_adj:
if p in bad_pairs:
return True
return False
original_pdb_h = pdb_hierarchy.deep_copy()
original_pdb_h.reset_atom_i_seqs()
original_pdb_h_asc = original_pdb_h.atom_selection_cache()
chain_id = original_pdb_h.only_model().only_chain().id
all_results = []
# only forward
# variants_searches = [
# #ccd_radius, change_all, change_radius, direction_forward
# ((1, False, 0, True ),1),
# # ((1, False, 0, False),1),
# ((2, False, 0, True ),1),
# # ((2, False, 0, False),1),
# ((3, False, 0, True ),2),
# # ((3, False, 0, False),2),
# ((2, True, 1, True ),1),
# # ((2, True, 1, False),1),
# ((3, True, 1, True ),2),
# # ((3, True, 1, False),2),
# ((3, True, 2, True ),3),
# # ((3, True, 2, False),3),
# ]
# only backward
# variants_searches = [
# #ccd_radius, change_all, change_radius, direction_forward
# # ((1, False, 0, True ),1),
# ((1, False, 0, False),1),
# # ((2, False, 0, True ),1),
# ((2, False, 0, False),1),
# # ((3, False, 0, True ),2),
# ((3, False, 0, False),2),
# # ((2, True, 1, True ),1),
# ((2, True, 1, False),1),
# # ((3, True, 1, True ),2),
# ((3, True, 1, False),2),
# # ((3, True, 2, True ),3),
# ((3, True, 2, False),3),
# ]
# both
variants_searches = [
#ccd_radius, change_all, change_radius, direction_forward
((1, False, 0, True ),1),
((1, False, 0, False),1),
((2, False, 0, True ),1),
((2, False, 0, False),1),
((3, False, 0, True ),2),
((3, False, 0, False),2),
((2, True, 1, True ),1),
((2, True, 1, False),1),
((3, True, 1, True ),2),
((3, True, 1, False),2),
((3, True, 2, True ),3),
((3, True, 2, False),3),
]
decided_variants = []
for variant, level in variants_searches:
if level <= self.params.variant_search_level:
decided_variants.append(variant)
for ccd_radius, change_all, change_radius, direction_forward in decided_variants:
# while ccd_radius <= 3:
fixing_omega = False
print >> self.log, " Starting optimization with radius=%d, " % ccd_radius,
print >> self.log, "change_all=%s, change_radius=%d, " % (change_all, change_radius),
print >> self.log, "direction=forward" if direction_forward else "direction=backwards"
self.log.flush()
#
(moving_h, moving_ref_atoms_iseqs, fixed_ref_atoms,
m_selection, contains_ss_element, anchor_present) = get_fixed_moving_parts(
pdb_hierarchy=pdb_hierarchy,
out_res_num_list=out_res_num_list,
# n_following=1,
# n_previous=ccd_radius+ccd_radius-1,
n_following=ccd_radius,
n_previous=ccd_radius,
ss_annotation=ss_annotation,
direction_forward=direction_forward,
log=self.log)
# print " moving_ref_atoms_iseqs", moving_ref_atoms_iseqs
print " moving_h resseqs:", [x.resseq for x in moving_h.residue_groups()]
moving_h_set = []
all_angles_combination_f = starting_conformations.get_all_starting_conformations(
moving_h,
change_radius,
n_outliers=len(out_res_num_list),
direction_forward=direction_forward,
cutoff=self.params.variant_number_cutoff,
change_all=change_all,
# log=self.log,
check_omega=self.params.make_all_trans,
)
#
# print "len(all_angles_combination_f)", len(all_angles_combination_f)
if len(all_angles_combination_f) == 0:
print "In starting conformations - outlier was fixed?"
# return result
else:
# here we should filter first ones that in
# tried_rama_angles_for_chain
filter_out = [] # [[tried values],[tried values],...]
for three in generate_protein_threes(
hierarchy=moving_h,
geometry=None):
if three[1].resseq in tried_rama_angles_for_chain.keys():
filter_out.append(tried_rama_angles_for_chain[three[1].resseq])
else:
filter_out.append((None, None))
ff_all_angles = []
print "filter_out", filter_out
for comb in all_angles_combination_f:
good = True
for comb_pair, bad_pairs in zip(comb, filter_out):
if bad_pairs == (None, None):
continue
# print "comb_pair, bad_pairs", comb_pair, bad_pairs
# if comb_pair in bad_pairs:
if comb_pair_in_bad_pairs(comb_pair, bad_pairs):
good = False
# print " Rejecting comb_pair", comb_pair
break
if good:
ff_all_angles.append(comb)
print "len(all_angles_combination_f)", len(all_angles_combination_f)
print "len(ff_all_angles)", len(ff_all_angles)
n_added = 0
n_all_combination = len(ff_all_angles)
if n_all_combination == 0:
print >> self.log, "Strange - got 0 combinations."
i_max = min(self.params.variant_number_cutoff, n_all_combination)
# assert i_max > 0
step = 0
if i_max > 1:
step = float(n_all_combination-1)/float(i_max-1)
if step < 1:
step = 1
for i in range(i_max):
comb = ff_all_angles[int(round(step*i))]
setted_h, fixed_omega = starting_conformations.set_rama_angles(
moving_h,
list(comb),
direction_forward=direction_forward,
check_omega=self.params.make_all_trans)
fixing_omega = fixing_omega or fixed_omega
moving_h_set.append(setted_h)
# print >> self.log, "Model %d, angles:" % i, comb
if self.params.make_all_trans and utils.n_bad_omegas(moving_h_set[-1]) != 0:
print "Model_%d_angles_%s.pdb" % (i, comb),
print "got ", utils.n_bad_omegas(moving_h_set[-1]), "bad omegas"
moving_h_set[-1].write_pdb_file("Model_%d_angles_%s.pdb" % (i, comb))
utils.list_omega(moving_h_set[-1], self.log)
assert 0
if len(moving_h_set) == 0:
# outlier was fixed before somehow...
# or there's a bug in get_starting_conformations
print >> self.log, "outlier was fixed before somehow"
return original_pdb_h
print "self.tried_rama_angles inside", self.tried_rama_angles
print "tried_rama_angles_for_chain", tried_rama_angles_for_chain
print "checking values", ccd_radius, change_all, change_radius, direction_forward
for i, h in enumerate(moving_h_set):
# if [x in tried_rama_angles_for_chain.keys() for x in out_res_num_list].count(True) > 0:
# print >> self.log, "Warning!!! make something here (check angles or so)"
# print >> self.log, "Skipping nonstable solution, tried previously:", (ccd_radius, change_all, change_radius, direction_forward, i)
# continue
resulting_rmsd = None
n_iter = 0
if anchor_present:
fixed_ref_atoms_coors = [x.xyz for x in fixed_ref_atoms]
# print "params to constructor", fixed_ref_atoms, h, moving_ref_atoms_iseqs
# easy_pickle.dump(file_name="crash.pkl", obj=[
# fixed_ref_atoms_coors,
# h,
# moving_ref_atoms_iseqs,
# direction_forward,
# self.params.save_states])
ccd_obj = ccd_cpp(fixed_ref_atoms_coors, h, moving_ref_atoms_iseqs)
ccd_obj.run(direction_forward=direction_forward, save_states=self.params.save_states)
resulting_rmsd = ccd_obj.resulting_rmsd
n_iter = ccd_obj.n_iter
if self.params.save_states:
states = ccd_obj.states
states.write(file_name="%s%s_%d_%s_%d_%i_states.pdb" % (chain_id, out_res_num_list[0], ccd_radius, change_all, change_radius, i))
map_target = 0
if self.reference_map is not None:
map_target = maptbx.real_space_target_simple(
unit_cell = self.xrs.crystal_symmetry().unit_cell(),
density_map = self.reference_map,
sites_cart = h.atoms().extract_xyz())
mc_rmsd = get_main_chain_rmsd_range(moving_h, h, all_atoms=True)
if self.verbose:
print >> self.log, "Resulting anchor and backbone RMSDs, mapcc, n_iter for model %d:" % i,
print >> self.log, resulting_rmsd, ",", mc_rmsd, ",", map_target, ",", n_iter
self.log.flush()
#
# setting new coordinates
#
moved_with_side_chains_h = pdb_hierarchy.deep_copy()
# setting xyz
#
for i_source, i_dest in enumerate(m_selection):
moved_with_side_chains_h.atoms()[i_dest].set_xyz(h.atoms()[i_source].xyz)
# set_xyz_smart(
# dest_h=moved_with_side_chains_h,
# source_h=h)
#
# placing side-chains
#
# moved_with_side_chains_h.write_pdb_file(
# file_name="%s_before_sc_placement_%d.pdb" % (prefix, i))
placing_range = get_res_nums_around(moved_with_side_chains_h,
center_resnum_list=out_res_num_list,
n_following=ccd_radius,
n_previous=ccd_radius,
include_intermediate=True,
avoid_ss_annot=ss_annotation)
place_side_chains(moved_with_side_chains_h, original_pdb_h, original_pdb_h_asc,
self.rotamer_manager, placing_range, self.ideal_res_dict)
# moved_with_side_chains_h.write_pdb_file(
# file_name="%s_after_sc_placement_%d.pdb" % (prefix, i))
#
# finalizing with geometry_minimization
#
# determining angles of interest
# print "Recording picked angle for outliers"
threes = generate_protein_threes(
# hierarchy=moving_h,
hierarchy=h,
geometry=None)
start_angles = []
final_angles = []
for angle_pair, three in zip(ff_all_angles[int(round(step*i))], threes):
# print "three[1].resseq in out_res_num_list, angle_pair", three[1].resseq, out_res_num_list, angle_pair
if three[1].resseq in out_res_num_list:
# if three[1].resseq not in tried_rama_angles_for_chain.keys():
# tried_rama_angles_for_chain[three[1].resseq] = []
start_angles.append((three[1].resseq, angle_pair))
ps_angles = three.get_phi_psi_angles()
final_angles.append((three[1].resseq, tuple(ps_angles)))
# tried_rama_angles_for_chain[three[1].resseq].append(angle_pair)
# print >> self.log, "Ended up with", three[1].resseq, "%.1f %.1f" % (ps_angles[0], ps_angles[1])
# print "Updated tried_rama_angles_for_chain:", tried_rama_angles_for_chain
if (not self.ccd_solution_is_duplicated(
final_angles=final_angles,
tried_final_rama_angles_for_chain=tried_final_rama_angles_for_chain)):
all_results.append((moved_with_side_chains_h.deep_copy(), mc_rmsd, resulting_rmsd, map_target, n_iter))
else:
continue
if self.ccd_solution_is_ok(
anchor_rmsd=resulting_rmsd,
mc_rmsd=mc_rmsd,
n_outliers=len(out_res_num_list),
ccd_radius=ccd_radius,
change_all_angles=change_all,
change_radius=change_radius,
contains_ss_element=contains_ss_element,
fixing_omega=fixing_omega):
print "Choosen result (mc_rmsd, anchor_rmsd, map_target, n_iter):", mc_rmsd, resulting_rmsd, map_target, n_iter
# Save to tried_ccds
for rn, angles in start_angles:
if rn not in tried_rama_angles_for_chain.keys():
tried_rama_angles_for_chain[rn] = []
tried_rama_angles_for_chain[rn].append(angles)
# Save final angles
for rn, angles in final_angles:
if rn not in tried_final_rama_angles_for_chain.keys():
tried_final_rama_angles_for_chain[rn] = []
tried_final_rama_angles_for_chain[rn].append(angles)
print >> self.log, "Ended up with", final_angles
print >> self.log, "Updated tried_rama_angles_for_chain:", tried_rama_angles_for_chain
print >> self.log, "Updated tried_final_rama_angles_for_chain:", tried_final_rama_angles_for_chain
self.log.flush()
if minimize:
print >> self.log, "minimizing..."
# moved_with_side_chains_h.write_pdb_file(
# file_name="%s_result_before_min_%d.pdb" % (prefix, i))
if self.reference_map is None:
minimize_wrapper_for_ramachandran(
hierarchy=moved_with_side_chains_h,
xrs=xrs,
original_pdb_h=original_pdb_h,
log=self.log,
grm=self.grm,
ss_annotation=self.secondary_structure_annotation)
else:
mwwm = minimize_wrapper_with_map(
pdb_h=moved_with_side_chains_h,
xrs=xrs,
target_map=self.reference_map,
grm=self.grm,
ss_annotation=self.secondary_structure_annotation,
log=self.log)
# moved_with_side_chains_h.write_pdb_file(
# file_name="%s_result_minimized_%d.pdb" % (prefix, i))
final_rmsd = get_main_chain_rmsd_range(moved_with_side_chains_h,
original_pdb_h, placing_range)
print >> self.log, "FINAL RMSD after minimization:", final_rmsd
return moved_with_side_chains_h
all_results.sort(key=lambda tup: tup[1])
if self.verbose:
print >> self.log, "ALL RESULTS:"
i = 0
for ar in all_results:
print >> self.log, ar[1:],
if ar[2] < 0.4:
# fn = "variant_%d.pdb" % i
# ar[0].write_pdb_file(file_name=fn)
# print fn
i += 1
else:
print >> self.log, " no output"
if self.params.force_rama_fixes:
# find and apply the best varian from all_results. This would be the one
# with the smallest rmsd given satisfactory closure
print >> self.log, "Applying the best found variant:",
i = 0
while i < len(all_results) and all_results[i][2] > 1.5:
i += 1
# apply
# === duplication!!!!
if i < len(all_results):
print >> self.log, all_results[i][1:]
if minimize:
print >> self.log, "minimizing..."
# all_results[i][0].write_pdb_file(
# file_name="%s_result_before_min_%d.pdb" % (prefix, i))
if self.reference_map is None:
minimize_wrapper_for_ramachandran(
hierarchy=all_results[i][0],
xrs=xrs,
original_pdb_h=original_pdb_h,
log=self.log,
grm=self.grm,
ss_annotation=self.secondary_structure_annotation)
else:
mwwm = minimize_wrapper_with_map(
pdb_h=all_results[i][0],
xrs=xrs,
target_map=self.reference_map,
grm=self.grm,
ss_annotation=self.secondary_structure_annotation,
log=self.log)
# all_results[i][0].write_pdb_file(
# file_name="%s_result_minimized_%d.pdb" % (prefix, i))
final_rmsd = get_main_chain_rmsd_range(all_results[i][0],
original_pdb_h, placing_range)
print >> self.log, "FINAL RMSD after minimization:", final_rmsd
return all_results[i][0]
else:
print >> self.log, " NOT FOUND!"
for i in all_results:
print >> self.log, i[1:]
# === end of duplication!!!!
else:
print >> self.log, "Epic FAIL: failed to fix rama outlier:", out_res_num_list
print >> self.log, " Options were: (mc_rmsd, resultign_rmsd, n_iter)"
for i in all_results:
print >> self.log, i[1:]
return original_pdb_h
def extract_proxies(self, hierarchy):
favored = ramalyze.RAMALYZE_FAVORED
allowed = ramalyze.RAMALYZE_ALLOWED
outlier = ramalyze.RAMALYZE_OUTLIER
self.hierarchy = hierarchy
bool_atom_selection = self._determine_bool_atom_selection(hierarchy)
selected_h = hierarchy.select(bool_atom_selection)
n_seq = flex.max(selected_h.atoms().extract_i_seq())
# Drop all previous proxies
self._oldfield_proxies = ext.shared_phi_psi_proxy()
self._emsley_proxies = ext.shared_phi_psi_proxy()
self._emsley8k_proxies = ext.shared_phi_psi_proxy()
# it would be great to save rama_eval, but the fact that this is called in
# pdb_interpretation, not in mmtbx.model makes it impossible
for three in generate_protein_threes(hierarchy=selected_h,
geometry=None):
rc = three.get_phi_psi_atoms()
if rc is None: continue
rama_key = three.get_ramalyze_key()
angles = three.get_phi_psi_angles()
rama_score = self.rama_eval.get_score(rama_key, angles[0],
angles[1])
r_eval = self.rama_eval.evaluate_score(rama_key, rama_score)
phi_atoms, psi_atoms = rc
i_seqs = [atom.i_seq for atom in phi_atoms] + [psi_atoms[-1].i_seq]
resnames = three.get_resnames()
r_name = resnames[1]
assert rama_key in range(6)
text_rama_key = ramalyze.res_types[rama_key]
assert text_rama_key in [
"general", "glycine", "cis-proline", "trans-proline",
"pre-proline", "isoleucine or valine"
]
# pick where to put...
ev_match_dict = {
favored: self.params.favored,
allowed: self.params.allowed,
outlier: self.params.outlier
}
r_type = ev_match_dict[r_eval]
if r_type == 'oldfield':
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=1) # XXX Not used in oldfield
self.append_oldfield_proxies(proxy, n_seq)
### THIS IS CRUEL. REMOVE ONCE favored/allowed/outlier are made multiple!
if (self.params.inject_emsley8k_into_oldfield_favored):
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=5)
self.append_emsley8k_proxies(proxy, n_seq)
###
elif r_type == 'emsley':
weight = self.params.emsley.weight
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=weight)
self.append_emsley_proxies(proxy, n_seq)
elif r_type == 'emsley8k':
if (r_eval is favored):
weight = self.params.emsley8k.weight_favored
elif (r_eval is allowed):
weight = self.params.emsley8k.weight_allowed
elif (r_eval is outlier):
weight = self.params.emsley8k.weight_outlier
else:
raise RuntimeError("Rama eveluation failed.")
proxy = ext.phi_psi_proxy(residue_type=text_rama_key,
i_seqs=i_seqs,
weight=weight)
self.append_emsley8k_proxies(proxy, n_seq)
elif (r_type is None):
pass
else:
raise RuntimeError("Not an option: %s" % str(r_type))
print("", file=self.log)
print(" %d Ramachandran restraints generated." %
(self.get_n_proxies()),
file=self.log)
print(" %d Oldfield and %d Emsley and %d emsley8k." %
(self.get_n_oldfield_proxies(), self.get_n_emsley_proxies(),
self.get_n_emsley8k_proxies()),
file=self.log)
bond.distance_ideal = averages[key]/averages.n[key]
elif len(key)==3:
rkey = (key[2],key[1],key[0])
averages.n[rkey]=averages.n[key]
for angle in self.geometry.angle_proxies:
if angle.i_seqs in averages.n:
key = angle.i_seqs
if key not in averages:
assert 0
angle.angle_ideal = averages[key]/averages.n[key]
if __name__=="__main__":
import sys
from iotbx import pdb
from test_rdl import get_geometry_restraints_manager
filename=sys.argv[1]
pdb_inp = pdb.input(filename)
pdb_hierarchy = pdb_inp.construct_hierarchy()
geometry_restraints_manager = get_geometry_restraints_manager(filename)
pdb_hierarchy.reset_i_seq_if_necessary()
from mmtbx.conformation_dependent_library import generate_protein_threes
for threes in generate_protein_threes(pdb_hierarchy,
geometry_restraints_manager,
#verbose=verbose,
):
print threes
print " cis? %s" % threes.cis_group()
print " rama %s" % threes.get_ramalyze_key()
print ' conf %s' % threes.is_pure_main_conf()
print "OK"
def add_terminal_hydrogens_threes(
hierarchy,
geometry_restraints_manager,
terminate_all_N_terminals=False,
terminate_all_C_terminals=False,
use_capping_hydrogens=False,
append_to_end_of_model=False,
verbose=False,
):
from mmtbx.conformation_dependent_library import generate_protein_threes
additional_hydrogens = [] #hierarchy_utils.smart_add_atoms()
for three in generate_protein_threes(
hierarchy,
geometry_restraints_manager,
#include_non_linked=False,
backbone_only=False,
include_linked_via_restraints_manager=True,
verbose=verbose,
):
# print three
#print dir(three)
# print geometry_restraints_manager
#print dir(geometry_restraints_manager)
bond_params_table = geometry_restraints_manager.bond_params_table
# print bond_params_table
#print dir(bond_params_table)
# print 'use_capping_hydrogens',use_capping_hydrogens
def get_bonds():
bonds = {}
for i, a1 in enumerate(residue_group.atoms()):
for j, a2 in enumerate(residue_group.atoms()):
if i >= j: continue
bond = three.bond_params_table.lookup(a1.i_seq, a2.i_seq)
if bond:
bonds[(a1.i_seq, a2.i_seq)] = True
bonds[(a2.i_seq, a1.i_seq)] = True
return bonds
if use_capping_hydrogens:
for i in range(len(three)):
residue_group = three.get_residue_group_from_hierarchy(
hierarchy, i)
rc = conditional_add_cys_hg_to_atom_group(
geometry_restraints_manager, residue_group)
#assert not rc, '%s' % rc
if three.start:
residue_group = three.get_residue_group_from_hierarchy(
hierarchy, 0)
rc = add_n_terminal_hydrogens_to_residue_group(
residue_group,
bonds=get_bonds(),
use_capping_hydrogens=use_capping_hydrogens,
append_to_end_of_model=append_to_end_of_model,
)
if rc: additional_hydrogens.append(rc)
if three.end:
residue_group = three.get_residue_group_from_hierarchy(
hierarchy, 2)
rc = add_c_terminal_oxygens_to_residue_group(
residue_group,
bonds=get_bonds(),
use_capping_hydrogens=use_capping_hydrogens,
append_to_end_of_model=append_to_end_of_model,
)
if rc: additional_hydrogens.append(rc)
return additional_hydrogens
def __init__ (self,
pdb_hierarchy,
outliers_only=False,
show_errors=False,
out=sys.stdout,
quiet=False) :
validation.__init__(self)
self.n_allowed = 0
self.n_favored = 0
self.n_type = [ 0 ] * 6
self._outlier_i_seqs = flex.size_t()
pdb_atoms = pdb_hierarchy.atoms()
all_i_seqs = pdb_atoms.extract_i_seq()
if (all_i_seqs.all_eq(0)) :
pdb_atoms.reset_i_seq()
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
analysis = ""
output_list = []
count_keys = []
uniqueness_keys = []
r = ramachandran_eval.RamachandranEval()
##if use_segids:
## chain_id = utils.get_segid_as_chainid(chain=chain)
## else:
## chain_id = chain.id
for three in generate_protein_threes(hierarchy=pdb_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
phi = get_dihedral(phi_atoms)
psi = get_dihedral(psi_atoms)
coords = get_center(main_residue) #should find the CA of the center residue
if (phi is not None and psi is not None):
res_type = RAMA_GENERAL
#self.n_total += 1
if (main_residue.resname[0:3] == "GLY"):
res_type = RAMA_GLYCINE
elif (main_residue.resname[0:3] == "PRO"):
is_cis = is_cis_peptide(three)
if is_cis:
res_type = RAMA_CISPRO
else:
res_type = RAMA_TRANSPRO
elif (three[2].resname == "PRO"):
res_type = RAMA_PREPRO
elif (main_residue.resname[0:3] == "ILE" or \
main_residue.resname[0:3] == "VAL"):
res_type = RAMA_ILE_VAL
#self.n_type[res_type] += 1
value = r.evaluate(res_types[res_type], [phi, psi])
ramaType = self.evaluateScore(res_type, value)
is_outlier = ramaType == RAMALYZE_OUTLIER
c_alphas = None
# XXX only save kinemage data for outliers
if is_outlier :
c_alphas = get_cas_from_three(three)
assert (len(c_alphas) == 3)
markup = self.as_markup_for_kinemage(c_alphas)
else:
markup = None
result = ramachandran(
chain_id=main_residue.parent().parent().id,
resseq=main_residue.resseq,
icode=main_residue.icode,
resname=main_residue.resname,
#altloc=main_residue.parent().altloc,
altloc=get_altloc_from_three(three),
segid=None, # XXX ???
phi=phi,
psi=psi,
rama_type=ramaType,
res_type=res_type,
score=value*100,
outlier=is_outlier,
xyz=coords,
markup=markup)
#if result.chain_id+result.resseq+result.icode not in count_keys:
if result.altloc in ['','A'] and result.chain_id+result.resseq+result.icode not in count_keys:
self.n_total += 1
self.n_type[res_type] += 1
self.add_to_validation_counts(ramaType)
count_keys.append(result.chain_id+result.resseq+result.icode)
if (not outliers_only or is_outlier) :
if (result.altloc != '' or
result.chain_id+result.resseq+result.icode not in uniqueness_keys):
#the threes/conformers method results in some redundant result
# calculations in structures with alternates. Using the
# uniqueness_keys list prevents redundant results being added to
# the final list
self.results.append(result)
uniqueness_keys.append(result.chain_id+result.resseq+result.icode)
if is_outlier :
i_seqs = main_residue.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0))
self._outlier_i_seqs.extend(i_seqs)
self.results.sort(key=lambda r: r.id_str())
out_count, out_percent = self.get_outliers_count_and_fraction()
fav_count, fav_percent = self.get_favored_count_and_fraction()
self.out_percent = out_percent * 100.0
self.fav_percent = fav_percent * 100.0
def update_restraints(
hierarchy,
geometry, # restraints_manager,
current_geometry=None, # xray_structure!!
sites_cart=None,
cdl_proxies=None,
ideal=True,
esd=True,
esd_factor=1.,
log=None,
verbose=False,
):
global registry
registry = RestraintsRegistry()
if current_geometry:
assert not sites_cart
sites_cart = current_geometry.sites_cart()
if sites_cart:
pdb_atoms = hierarchy.atoms()
# XXX PDB_TRANSITION VERY SLOW
for j_seq, atom in enumerate(pdb_atoms):
atom.xyz = sites_cart[j_seq]
threes = None
average_updates = 0
total_updates = 0
for threes in generate_protein_threes(
hierarchy,
geometry,
cdl_class=True,
omega_cdl=True,
#verbose=verbose,
):
threes.apply_updates = apply_updates
if threes.cis_group(omega_cdl=True):
if verbose and 0:
print 'cis ' * 20
print threes
continue
restraint_values = get_restraint_values(threes)
if restraint_values is None: continue
if restraint_values[0] == "I":
average_updates += 1
else:
total_updates += 1
threes.apply_updates(
threes,
restraint_values,
cdl_proxies,
ideal=ideal,
esd=esd,
esd_factor=esd_factor,
)
if registry.n:
threes.apply_average_updates(registry)
assert 0
geometry.reset_internals()
if verbose and threes and threes.errors:
if log:
log.write(
" Residues not completely updated with CDL restraints\n\n")
for line in threes.errors:
if log:
log.write("%s\n" % line)
else:
print line
return geometry
bond.distance_ideal = averages[key]/averages.n[key]
elif len(key)==3:
rkey = (key[2],key[1],key[0])
averages.n[rkey]=averages.n[key]
for angle in self.geometry.angle_proxies:
if angle.i_seqs in averages.n:
key = angle.i_seqs
if key not in averages:
assert 0
angle.angle_ideal = averages[key]/averages.n[key]
if __name__=="__main__":
import sys
from iotbx import pdb
from test_rdl import get_geometry_restraints_manager
filename=sys.argv[1]
pdb_inp = pdb.input(filename)
pdb_hierarchy = pdb_inp.construct_hierarchy()
geometry_restraints_manager = get_geometry_restraints_manager(filename)
pdb_hierarchy.reset_i_seq_if_necessary()
from mmtbx.conformation_dependent_library import generate_protein_threes
for threes in generate_protein_threes(pdb_hierarchy,
geometry_restraints_manager,
#verbose=verbose,
):
print threes
print " cis? %s" % threes.cis_group()
print " rama %s" % threes.get_ramalyze_key()
print ' conf %s' % threes.is_pure_main_conf()
print "OK"
def __init__(self, model, log):
db_path = libtbx.env.find_in_repositories(
relative_path="chem_data/rama_z/top8000_rama_z_dict.pkl",
test=os.path.isfile)
rmsd_path = libtbx.env.find_in_repositories(
relative_path="chem_data/rama_z/rmsd.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)
self.rmsd_estimator = easy_pickle.load(rmsd_path)
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.interpolation_fs = {"H": {}, "S": {}, "L": {}}
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
self.res_info = []
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
self.ssm = ss_manager(model.get_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 = self.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=model.get_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"]
for i in self.res_info:
print(i, file=self.log)