def __init__(self,
pdb_hierarchy,
nontrans_only=False,
out=sys.stdout,
quiet=True):
validation.__init__(self)
self.residue_count = [0, 0]
#[OMEGA_GENERAL, OMEGA_PRO]
self.omega_count = [[0,0,0], [0,0,0]]
#[OMEGA_GENERAL, OMEGA_PRO], then
#[OMEGALYZE_TRANS, OMEGALYZE_CIS, OMEGALYZE_TWISTED]
from mmtbx.validation import utils
from scitbx.array_family import flex
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)
prev_rezes, next_rezes = None, None
prev_resid = None
cur_resseq = None
next_resseq = None
for model in pdb_hierarchy.models():
for chain in model.chains():
prev_rezes, next_rezes = None, None
prev_resid = None
cur_resseq = None
next_resseq = None
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
residues = list(chain.residue_groups())
for i, residue_group in enumerate(residues):
# The reason I pass lists of atom_groups to get_phi and get_psi is to
# deal with the particular issue where some residues have an A alt
# conf that needs some atoms from a "" alt conf to get calculated
# correctly. See 1jxt.pdb for examples. This way I can search both
# the alt conf atoms and the "" atoms if necessary.
prev_atom_list, next_atom_list, atom_list = None, None, None
if cur_resseq is not None:
prev_rezes = rezes
prev_resseq = cur_resseq
rezes = construct_residues(residues[i])
cur_resseq = residue_group.resseq_as_int()
cur_icode = residue_group.icode.strip()
if (i > 0):
#check for insertion codes
if (cur_resseq == residues[i-1].resseq_as_int()) :
if (cur_icode == '') and (residues[i-1].icode.strip() == '') :
continue
elif (cur_resseq != (residues[i-1].resseq_as_int())+1):
continue
for atom_group in residue_group.atom_groups():
alt_conf = atom_group.altloc
if rezes is not None:
atom_list = rezes.get(alt_conf)
if prev_rezes is not None:
prev_atom_list = prev_rezes.get(alt_conf)
if (prev_atom_list is None):
prev_keys = sorted(prev_rezes.keys())
prev_atom_list = prev_rezes.get(prev_keys[0])
omega=get_omega(prev_atom_list, atom_list)
highest_mc_b = get_highest_mc_b(prev_atom_list, atom_list)
if omega is not None:
resname = atom_group.resname[0:3]
coords = get_center(atom_group)
if resname == "PRO":
res_type = OMEGA_PRO
else:
res_type = OMEGA_GENERAL
self.residue_count[res_type] += 1
omega_type = find_omega_type(omega)
is_nontrans = False
if omega_type == OMEGALYZE_CIS or omega_type == OMEGALYZE_TWISTED:
self.n_outliers += 1
is_nontrans = True
self.omega_count[res_type][omega_type] += 1
markup_atoms = [None, None, None, None] #for kinemage markup
if is_nontrans:
for a in prev_atom_list:
if a is None: continue
a_ = atom(pdb_atom=a)
if a.name.strip() == "CA":
markup_atoms[0] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
elif a.name.strip() == "C":
markup_atoms[1] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
for a in atom_list:
if a is None: continue
a_ = atom(pdb_atom=a)
if a.name.strip() == "N":
markup_atoms[2] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
elif a.name.strip() == "CA":
markup_atoms[3] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
#------------
#prevres=residues[i-1]
#find prev res identities for printing
prev_alts = []
prev_resnames = {}
for ag in residues[i-1].atom_groups():
prev_alts.append(ag.altloc)
prev_resnames[ag.altloc] = ag.resname
if alt_conf in prev_alts:
prev_altloc = alt_conf
else:
if len(prev_alts) > 1:
prev_altloc = prev_alts[1]
else:
prev_altloc = prev_alts[0]
prev_resname = prev_resnames[prev_altloc]
#done finding prev res identities
result = omega_result(
chain_id=chain_id,
resseq=residue_group.resseq,
icode=residue_group.icode,
resname=atom_group.resname,
altloc=atom_group.altloc,
prev_resseq=residues[i-1].resseq,
prev_icode=residues[i-1].icode,
prev_resname=prev_resname,
prev_altloc=prev_altloc,
segid=None,
omega=omega,
omega_type=omega_type,
res_type=res_type,
is_nontrans=is_nontrans,
outlier=is_nontrans,
highest_mc_b=highest_mc_b,
xyz=coords,
markup_atoms=markup_atoms)
if is_nontrans or not nontrans_only: #(not nontrans_only or is_nontrans)
self.results.append(result)
if is_nontrans:
i_seqs = atom_group.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0)) #This assert copied from ramalyze
self._outlier_i_seqs.extend(i_seqs)
def __init__(self, pdb_hierarchy,
data_version="8000",
outliers_only=False,
show_errors=False,
out=sys.stdout,
quiet=False):
validation.__init__(self)
self.n_allowed = 0
self.n_favored = 0
from mmtbx.rotamer.sidechain_angles import SidechainAngles
from mmtbx.rotamer import rotamer_eval
from mmtbx.rotamer.rotamer_eval import RotamerID
from mmtbx.validation import utils
self.data_version = data_version
# if self.data_version == "500": self.outlier_threshold = 0.01
if self.data_version == "8000": self.outlier_threshold = OUTLIER_THRESHOLD
else: raise ValueError(
"data_version given to RotamerEval not recognized (%s)." % data_version)
sidechain_angles = SidechainAngles(show_errors)
rotamer_evaluator = rotamer_eval.RotamerEval(
data_version=data_version)
rotamer_id = rotamer_eval.RotamerID() # loads in the rotamer names
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
current_rotamers = {}
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
for rg in chain.residue_groups():
all_dict = construct_complete_sidechain(rg)
for atom_group in rg.atom_groups():
coords = get_center(atom_group)
resname = atom_group.resname
occupancy = get_occupancy(atom_group)
kwargs = {
"chain_id" : chain_id,
"resseq" : rg.resseq,
"icode" : rg.icode,
"altloc" : atom_group.altloc,
"resname" : resname,
"xyz" : coords,
"occupancy" : occupancy,
}
atom_dict = all_dict.get(atom_group.altloc)
res_key = get_residue_key(atom_group=atom_group)
try:
chis = sidechain_angles.measureChiAngles(
atom_group,
atom_dict)#.get(conformer.altloc))
except AttributeError:
if show_errors:
kwargs['incomplete'] = True
result = rotamer(**kwargs)
print >> out, '%s is missing some sidechain atoms' % \
result.id_str()
self.results.append(result)
continue
if (chis is not None):
if None in chis:
continue
cur_res = resname.lower().strip()
if cur_res == 'mse':
cur_res = 'met'
value = rotamer_evaluator.evaluate(cur_res, chis)
if value is not None:
self.n_total += 1
kwargs['score'] = value * 100
wrap_chis = rotamer_id.wrap_chis(resname.strip(), chis,
symmetry=False)
sym_chis = wrap_chis[:]
sym_chis = rotamer_id.wrap_sym(resname.strip(), sym_chis)
evaluation = self.evaluateScore(value)
kwargs['evaluation'] = evaluation
if evaluation == "OUTLIER":
kwargs['outlier'] = True
kwargs['rotamer_name'] = evaluation
else:
kwargs['outlier'] = False
kwargs['rotamer_name'] = rotamer_id.identify(resname,
wrap_chis)
#deal with unclassified rotamers
if kwargs['rotamer_name'] == '':
kwargs['rotamer_name'] = "UNCLASSIFIED"
while (len(wrap_chis) < 4):
wrap_chis.append(None)
kwargs['chi_angles'] = wrap_chis
result = rotamer(**kwargs)
if (result.is_outlier()) or (not outliers_only):
self.results.append(result)
out_count, out_percent = self.get_outliers_count_and_fraction()
self.out_percent = out_percent * 100.0
def __init__(self,
pdb_hierarchy,
nontrans_only=False,
out=sys.stdout,
quiet=True):
validation.__init__(self)
self.residue_count = [0, 0]
#[OMEGA_GENERAL, OMEGA_PRO]
self.omega_count = [[0,0,0], [0,0,0]]
#[OMEGA_GENERAL, OMEGA_PRO], then
#[OMEGALYZE_TRANS, OMEGALYZE_CIS, OMEGALYZE_TWISTED]
from mmtbx.validation import utils
from scitbx.array_family import flex
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)
prev_rezes, next_rezes = None, None
prev_resid = None
cur_resseq = None
next_resseq = None
for model in pdb_hierarchy.models():
for chain in model.chains():
prev_rezes, next_rezes = None, None
prev_resid = None
cur_resseq = None
next_resseq = None
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
residues = list(chain.residue_groups())
for i, residue_group in enumerate(residues):
# The reason I pass lists of atom_groups to get_phi and get_psi is to
# deal with the particular issue where some residues have an A alt
# conf that needs some atoms from a "" alt conf to get calculated
# correctly. See 1jxt.pdb for examples. This way I can search both
# the alt conf atoms and the "" atoms if necessary.
prev_atom_list, next_atom_list, atom_list = None, None, None
if cur_resseq is not None:
prev_rezes = rezes
prev_resseq = cur_resseq
rezes = construct_residues(residues[i])
cur_resseq = residue_group.resseq_as_int()
cur_icode = residue_group.icode.strip()
if (i > 0):
#check for insertion codes
if (cur_resseq == residues[i-1].resseq_as_int()) :
if (cur_icode == '') and (residues[i-1].icode.strip() == '') :
continue
elif (cur_resseq != (residues[i-1].resseq_as_int())+1):
continue
for atom_group in residue_group.atom_groups():
alt_conf = atom_group.altloc
if rezes is not None:
atom_list = rezes.get(alt_conf)
if prev_rezes is not None:
prev_atom_list = prev_rezes.get(alt_conf)
if (prev_atom_list is None):
prev_keys = sorted(prev_rezes.keys())
prev_atom_list = prev_rezes.get(prev_keys[0])
omega=get_omega(prev_atom_list, atom_list)
highest_mc_b = get_highest_mc_b(prev_atom_list, atom_list)
if omega is not None:
resname = atom_group.resname[0:3]
coords = get_center(atom_group)
if resname == "PRO":
res_type = OMEGA_PRO
else:
res_type = OMEGA_GENERAL
self.residue_count[res_type] += 1
omega_type = find_omega_type(omega)
is_nontrans = False
if omega_type == OMEGALYZE_CIS or omega_type == OMEGALYZE_TWISTED:
self.n_outliers += 1
is_nontrans = True
self.omega_count[res_type][omega_type] += 1
markup_atoms = [None, None, None, None] #for kinemage markup
if is_nontrans:
for a in prev_atom_list:
if a is None: continue
a_ = atom(pdb_atom=a)
if a.name.strip() == "CA":
markup_atoms[0] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
elif a.name.strip() == "C":
markup_atoms[1] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
for a in atom_list:
if a is None: continue
a_ = atom(pdb_atom=a)
if a.name.strip() == "N":
markup_atoms[2] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
elif a.name.strip() == "CA":
markup_atoms[3] = kin_atom(
id_str=a_.atom_group_id_str(),xyz=a_.xyz)
#------------
#prevres=residues[i-1]
#find prev res identities for printing
prev_alts = []
prev_resnames = {}
for ag in residues[i-1].atom_groups():
prev_alts.append(ag.altloc)
prev_resnames[ag.altloc] = ag.resname
if alt_conf in prev_alts:
prev_altloc = alt_conf
else:
if len(prev_alts) > 1:
prev_altloc = prev_alts[1]
else:
prev_altloc = prev_alts[0]
prev_resname = prev_resnames[prev_altloc]
#done finding prev res identities
result = omega_result(
chain_id=chain_id,
resseq=residue_group.resseq,
icode=residue_group.icode,
resname=atom_group.resname,
altloc=atom_group.altloc,
prev_resseq=residues[i-1].resseq,
prev_icode=residues[i-1].icode,
prev_resname=prev_resname,
prev_altloc=prev_altloc,
segid=None,
omega=omega,
omega_type=omega_type,
res_type=res_type,
is_nontrans=is_nontrans,
outlier=is_nontrans,
highest_mc_b=highest_mc_b,
xyz=coords,
markup_atoms=markup_atoms)
if is_nontrans or not nontrans_only: #(not nontrans_only or is_nontrans)
self.results.append(result)
if is_nontrans:
i_seqs = atom_group.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0)) #This assert copied from ramalyze
self._outlier_i_seqs.extend(i_seqs)
def __init__(self,
pdb_hierarchy,
nontrans_only=False,
out=sys.stdout,
quiet=True):
validation.__init__(self)
self.residue_count = [0, 0]
#[OMEGA_GENERAL, OMEGA_PRO]
self.omega_count = [[0, 0, 0], [0, 0, 0]]
#[OMEGA_GENERAL, OMEGA_PRO], then
#[OMEGALYZE_TRANS, OMEGALYZE_CIS, OMEGALYZE_TWISTED]
from mmtbx.validation import utils
from scitbx.array_family import flex
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)
first_conf_altloc = None
prev_chain_id = None
for twores in generate_protein_fragments(
pdb_hierarchy,
length=2,
geometry=None,
include_non_standard_peptides=True):
main_residue = twores[
1] #this is the relevant residue for id-ing cis-Pro
conf_altloc = get_conformer_altloc(twores)
prevres_altloc, mainres_altloc = get_local_omega_altlocs(twores)
twores_altloc = prevres_altloc or mainres_altloc #default '' evals False
chain = main_residue.parent().parent()
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
if chain_id != prev_chain_id: #if we've moved to a new chain...
first_conf_altloc = conf_altloc #...reset reference altloc
prev_chain_id = chain_id
if (conf_altloc != first_conf_altloc) and twores_altloc == '':
#skip non-alternate residues unless this is the first time thru a chain
continue
omega_atoms = get_omega_atoms(twores)
#omega_atoms is the list [CA1 C1 N2 CA2], with None for missing atoms
if None in omega_atoms:
continue
omega = get_omega(omega_atoms)
if omega is None: continue
omega_type = find_omega_type(omega)
if omega_type == OMEGALYZE_TRANS:
is_nontrans = False
else:
is_nontrans = True
self.n_outliers += 1
if main_residue.resname == "PRO": res_type = OMEGA_PRO
else: res_type = OMEGA_GENERAL
self.residue_count[res_type] += 1
self.omega_count[res_type][omega_type] += 1
highest_mc_b = get_highest_mc_b(twores[0].atoms(),
twores[1].atoms())
coords = get_center(main_residue)
markup_atoms = []
for omega_atom in omega_atoms:
markup_atoms.append(
kin_atom(omega_atom.parent().id_str(), omega_atom.xyz))
result = omega_result(
model_id=twores[0].parent().parent().parent().id,
chain_id=chain_id,
resseq=main_residue.resseq,
icode=main_residue.icode,
resname=main_residue.resname,
altloc=mainres_altloc,
prev_resseq=twores[0].resseq,
prev_icode=twores[0].icode,
prev_resname=twores[0].resname,
prev_altloc=prevres_altloc,
segid=None,
omega=omega,
omega_type=omega_type,
res_type=res_type,
is_nontrans=is_nontrans,
outlier=is_nontrans,
highest_mc_b=highest_mc_b,
xyz=coords,
markup_atoms=markup_atoms)
if is_nontrans or not nontrans_only: #(not nontrans_only or is_nontrans)
self.results.append(result)
if is_nontrans:
i_seqs = main_residue.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0)
) #This assert copied from ramalyze
self._outlier_i_seqs.extend(i_seqs)
self.results.sort(key=lambda x: x.model_id + ':' + x.id_str())
def __init__(self,
pdb_hierarchy,
outliers_only=False,
out=sys.stdout,
collect_ideal=False,
quiet=False):
validation.__init__(self)
self._outlier_i_seqs = flex.size_t()
self.beta_ideal = {}
relevant_atom_names = {
" CA ": None,
" N ": None,
" C ": None,
" CB ": None
} # FUTURE: set
output_list = []
self.stats = group_args(n_results=0,
n_weighted_results=0,
n_weighted_outliers=0)
from mmtbx.validation import utils
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
for rg in chain.residue_groups():
for i_cf, cf in enumerate(rg.conformers()):
for i_residue, residue in enumerate(cf.residues()):
if (residue.resname == "GLY"):
continue
is_first = (i_cf == 0)
is_alt_conf = False
relevant_atoms = {}
for atom in residue.atoms():
if (atom.name in relevant_atom_names):
relevant_atoms[atom.name] = atom
if (len(atom.parent().altloc) != 0):
is_alt_conf = True
if ((is_first or is_alt_conf)
and len(relevant_atoms) == 4):
result = calculate_ideal_and_deviation(
relevant_atoms=relevant_atoms,
resname=residue.resname)
dev = result.deviation
dihedralNABB = result.dihedral
betaxyz = result.ideal
if (dev is None): continue
resCB = relevant_atoms[" CB "]
self.stats.n_results += 1
self.stats.n_weighted_results += resCB.occ
if (dev >= 0.25 or outliers_only == False):
if (dev >= 0.25):
self.n_outliers += 1
self.stats.n_weighted_outliers += resCB.occ
self._outlier_i_seqs.append(atom.i_seq)
if (is_alt_conf):
altchar = cf.altloc
else:
altchar = " "
res = residue.resname.lower()
sub = chain.id
if (len(sub) == 1):
sub = " " + sub
result = cbeta(chain_id=chain_id,
resname=residue.resname,
resseq=residue.resseq,
icode=residue.icode,
altloc=altchar,
xyz=resCB.xyz,
occupancy=resCB.occ,
deviation=dev,
dihedral_NABB=dihedralNABB,
ideal_xyz=betaxyz,
outlier=(dev >= 0.25))
self.results.append(result)
key = result.id_str()
if (collect_ideal):
self.beta_ideal[key] = betaxyz
def run(args):
"""
I suggest adding here:
cctbx_project/mmtbx/validation/regression/tst_mp_geo.py
test cases with just .pdb, without arguments, etc.
"""
master_phil = get_master_phil()
import iotbx.phil
input_objects = iotbx.phil.process_command_line_with_files(
args=args, master_phil=master_phil, pdb_file_def="mp_geo.pdb")
work_params = input_objects.work.extract()
assert len(work_params.mp_geo.pdb) == 1, "Need a model file to run"
file_name = work_params.mp_geo.pdb[0]
out_file = None
if work_params.mp_geo.out_file != None:
out_file = work_params.mp_geo.out_file
do_bonds_and_angles = work_params.mp_geo.bonds_and_angles
do_kinemage = work_params.mp_geo.kinemage
do_rna_backbone = work_params.mp_geo.rna_backbone
outliers_only = work_params.mp_geo.outliers_only
use_cdl = work_params.mp_geo.cdl
log = StringIO()
basename = os.path.basename(file_name)
if out_file == None:
import sys
out = sys.stdout
else:
if do_bonds_and_angles:
out = file(out_file, 'w')
elif do_kinemage:
out = file(out_file, 'a')
elif do_rna_backbone:
out = file(out_file, 'w')
restraints_loading_flags = {}
restraints_loading_flags["use_neutron_distances"] = False
from mmtbx.validation import utils
params = pdb_interpretation.master_params.extract()
params.restraints_library.cdl = use_cdl
params.clash_guard.nonbonded_distance_threshold = None
processed_pdb_file = pdb_interpretation.process(
params=params,
mon_lib_srv=server.server(),
ener_lib=server.ener_lib(),
file_name=file_name,
strict_conflict_handling=True,
restraints_loading_flags=restraints_loading_flags,
force_symmetry=True,
substitute_non_crystallographic_unit_cell_if_necessary=True,
log=log)
grm = processed_pdb_file.geometry_restraints_manager()
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=processed_pdb_file.all_chain_proxies.pdb_hierarchy)
if do_bonds_and_angles or do_kinemage:
rc = get_bond_and_angle_outliers(
pdb_hierarchy=processed_pdb_file.all_chain_proxies.pdb_hierarchy,
xray_structure=processed_pdb_file.xray_structure(),
geometry_restraints_manager=grm,
use_segids=use_segids,
outliers_only=outliers_only)
#get chain types
chain_types = {}
for chain in processed_pdb_file.all_chain_proxies.\
pdb_hierarchy.models()[0].chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
main_conf = chain.conformers()[0]
if chain_types.get(chain_id) not in ["NA", "PROTEIN"]:
if (main_conf.is_na()):
chain_types[chain_id] = "NA"
elif (main_conf.is_protein()):
chain_types[chain_id] = "PROTEIN"
else:
chain_types[chain_id] = "UNK"
outliers = []
#bonds
#for result in rc.bonds.results:
for result in sorted(
rc.bonds.results,
key=lambda x:
(x.atoms_info[0].resseq, get_altloc(atoms_info=x.atoms_info),
get_atoms_str(atoms_info=x.atoms_info))):
atom_info = result.atoms_info[0]
# label:chain:number:ins:alt:type:measure:value:sigmas:class
atoms_str = get_atoms_str(atoms_info=result.atoms_info)
altloc = get_altloc(atoms_info=result.atoms_info)
chain_id = atom_info.chain_id
outliers.append([
chain_id, atom_info.resseq, atom_info.icode, altloc,
atom_info.resname, atoms_str, result.model, result.score,
chain_types[atom_info.chain_id]
])
#angles
#for result in rc.angles.results:
for result in sorted(
rc.angles.results,
key=lambda x:
(x.atoms_info[0].resseq, get_altloc(atoms_info=x.atoms_info),
get_atoms_str(atoms_info=x.atoms_info))):
atom_info = result.atoms_info[0]
# label:chain:number:ins:alt:type:measure:value:sigmas:class
atoms_str = get_atoms_str(atoms_info=result.atoms_info)
altloc = get_altloc(atoms_info=result.atoms_info)
chain_id = atom_info.chain_id
outliers.append([
chain_id, atom_info.resseq, atom_info.icode, altloc,
atom_info.resname, atoms_str, result.model, result.score,
chain_types[atom_info.chain_id]
])
if do_bonds_and_angles:
for outlier in outliers:
print >> out, "%s:%2s:%s:%s:%s:%s:%s:%.3f:%.3f:%s" % (
basename, outlier[0], outlier[1], outlier[2], outlier[3],
outlier[4], outlier[5], outlier[6], outlier[7], outlier[8])
elif do_kinemage:
print >> out, rc.bonds.kinemage_header
for result in rc.bonds.results:
print >> out, result.as_kinemage()
print >> out, rc.angles.kinemage_header
for result in rc.angles.results:
print >> out, result.as_kinemage()
out.close()
elif do_rna_backbone:
from mmtbx.validation import utils
rna_bb = utils.get_rna_backbone_dihedrals(processed_pdb_file)
print >> out, rna_bb
if out_file is not None:
out.close()
def __init__ (self, pdb_hierarchy,
outliers_only=False,
out=sys.stdout,
collect_ideal=False,
quiet=False) :
validation.__init__(self)
self._outlier_i_seqs = flex.size_t()
self.beta_ideal = {}
relevant_atom_names = {
" CA ": None, " N ": None, " C ": None, " CB ": None} # FUTURE: set
output_list = []
from mmtbx.validation import utils
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
for rg in chain.residue_groups():
for i_cf,cf in enumerate(rg.conformers()):
for i_residue,residue in enumerate(cf.residues()):
if (residue.resname == "GLY") :
continue
is_first = (i_cf == 0)
is_alt_conf = False
relevant_atoms = {}
for atom in residue.atoms():
if (atom.name in relevant_atom_names):
relevant_atoms[atom.name] = atom
if (len(atom.parent().altloc) != 0):
is_alt_conf = True
if ((is_first or is_alt_conf) and len(relevant_atoms) == 4):
result = calculate_ideal_and_deviation(
relevant_atoms=relevant_atoms,
resname=residue.resname)
dev = result.deviation
dihedralNABB = result.dihedral
betaxyz = result.ideal
if (dev is None) : continue
if(dev >=0.25 or outliers_only==False):
if(dev >=0.25):
self.n_outliers+=1
self._outlier_i_seqs.append(atom.i_seq)
if (is_alt_conf):
altchar = cf.altloc
else:
altchar = " "
res=residue.resname.lower()
sub=chain.id
if(len(sub)==1):
sub=" "+sub
resCB = relevant_atoms[" CB "]
result = cbeta(
chain_id=chain_id,
resname=residue.resname,
resseq=residue.resseq,
icode=residue.icode,
altloc=altchar,
xyz=resCB.xyz,
occupancy=resCB.occ,
deviation=dev,
dihedral_NABB=dihedralNABB,
ideal_xyz=betaxyz,
outlier=(dev >= 0.25))
self.results.append(result)
key = result.id_str()
if (collect_ideal) :
self.beta_ideal[key] = betaxyz
def __init__(self,
pdb_hierarchy,
outliers_only=False,
out=sys.stdout,
collect_ideal=False,
apply_phi_psi_correction=False,
display_phi_psi_correction=False,
quiet=False):
validation.__init__(self)
self._outlier_i_seqs = flex.size_t()
self.beta_ideal = {}
output_list = []
self.stats = group_args(n_results=0,
n_weighted_results=0,
n_weighted_outliers=0)
if apply_phi_psi_correction:
phi_psi_angles = get_phi_psi_dict(pdb_hierarchy)
new_outliers = 0
outliers_removed = 0
total_residues = 0
from mmtbx.validation import utils
use_segids = utils.use_segids_in_place_of_chainids(
hierarchy=pdb_hierarchy)
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
for rg in chain.residue_groups():
for i_cf, cf in enumerate(rg.conformers()):
for i_residue, residue in enumerate(cf.residues()):
if (residue.resname == "GLY"):
continue
is_first = (i_cf == 0)
is_alt_conf = False
relevant_atoms = {}
for atom in residue.atoms():
if (atom.name in relevant_atom_names):
relevant_atoms[atom.name] = atom
if (len(atom.parent().altloc) != 0):
is_alt_conf = True
if ((is_first or is_alt_conf)
and len(relevant_atoms) == 4):
result = calculate_ideal_and_deviation(
relevant_atoms=relevant_atoms,
resname=residue.resname)
dev = result.deviation
dihedralNABB = result.dihedral
betaxyz = result.ideal
if (dev is None): continue
resCB = relevant_atoms[" CB "]
self.stats.n_results += 1
self.stats.n_weighted_results += resCB.occ
if (is_alt_conf):
altchar = cf.altloc
else:
altchar = " "
if apply_phi_psi_correction:
total_residues += 1
id_str = '|%s:%s|' % (residue.id_str(),
altchar)
phi_psi = phi_psi_angles.get(id_str, None)
if phi_psi:
rc = cbd_utils.get_phi_psi_correction(
result,
residue,
phi_psi,
display_phi_psi_correction=
display_phi_psi_correction,
)
if rc:
dev, dihedralNABB, start, finish = rc
if start and not finish:
outliers_removed += 1
elif not start and finish:
new_outliers += 1
if (dev >= 0.25 or outliers_only == False):
if (dev >= 0.25):
self.n_outliers += 1
self.stats.n_weighted_outliers += resCB.occ
self._outlier_i_seqs.append(atom.i_seq)
res = residue.resname.lower()
sub = chain.id
if (len(sub) == 1):
sub = " " + sub
result = cbeta(chain_id=chain_id,
resname=residue.resname,
resseq=residue.resseq,
icode=residue.icode,
altloc=altchar,
xyz=resCB.xyz,
occupancy=resCB.occ,
deviation=dev,
dihedral_NABB=dihedralNABB,
ideal_xyz=betaxyz,
outlier=(dev >= 0.25))
self.results.append(result)
key = result.id_str()
if (collect_ideal):
self.beta_ideal[key] = betaxyz
if apply_phi_psi_correction:
print('''
Outliers removed : %5d
New outliers : %5d
Num. of outliers : %5d
Num. of residues : %5d
''' % (
outliers_removed,
new_outliers,
self.n_outliers,
total_residues,
))
def __init__(
self, pdb_hierarchy, data_version="8000", outliers_only=False, show_errors=False, out=sys.stdout, quiet=False
):
validation.__init__(self)
self.n_allowed = 0
self.n_favored = 0
from mmtbx.rotamer.sidechain_angles import SidechainAngles
from mmtbx.rotamer import rotamer_eval
from mmtbx.rotamer.rotamer_eval import RotamerID
from mmtbx.validation import utils
self.data_version = data_version
# if self.data_version == "500": self.outlier_threshold = 0.01
if self.data_version == "8000":
self.outlier_threshold = 0.003
else:
raise ValueError("data_version given to RotamerEval not recognized (%s)." % data_version)
sidechain_angles = SidechainAngles(show_errors)
rotamer_evaluator = rotamer_eval.RotamerEval(data_version=data_version)
rotamer_id = rotamer_eval.RotamerID() # loads in the rotamer names
use_segids = utils.use_segids_in_place_of_chainids(hierarchy=pdb_hierarchy)
current_rotamers = {}
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
for rg in chain.residue_groups():
all_dict = construct_complete_sidechain(rg)
for atom_group in rg.atom_groups():
coords = get_center(atom_group)
resname = atom_group.resname
occupancy = get_occupancy(atom_group)
kwargs = {
"chain_id": chain_id,
"resseq": rg.resseq,
"icode": rg.icode,
"altloc": atom_group.altloc,
"resname": resname,
"xyz": coords,
"occupancy": occupancy,
}
atom_dict = all_dict.get(atom_group.altloc)
res_key = get_residue_key(atom_group=atom_group)
try:
chis = sidechain_angles.measureChiAngles(atom_group, atom_dict) # .get(conformer.altloc))
except AttributeError:
if show_errors:
kwargs["incomplete"] = True
result = rotamer(**kwargs)
print >> out, "%s is missing some sidechain atoms" % result.id_str()
self.results.append(result)
continue
if chis is not None:
if None in chis:
continue
cur_res = resname.lower().strip()
if cur_res == "mse":
cur_res = "met"
value = rotamer_evaluator.evaluate(cur_res, chis)
if value is not None:
self.n_total += 1
kwargs["score"] = value * 100
wrap_chis = rotamer_id.wrap_chis(resname.strip(), chis, symmetry=False)
sym_chis = wrap_chis[:]
sym_chis = rotamer_id.wrap_sym(resname.strip(), sym_chis)
evaluation = self.evaluateScore(value)
kwargs["evaluation"] = evaluation
if evaluation == "OUTLIER":
kwargs["outlier"] = True
kwargs["rotamer_name"] = evaluation
else:
kwargs["outlier"] = False
kwargs["rotamer_name"] = rotamer_id.identify(resname, wrap_chis)
# deal with unclassified rotamers
if kwargs["rotamer_name"] == "":
kwargs["rotamer_name"] = "UNCLASSIFIED"
while len(wrap_chis) < 4:
wrap_chis.append(None)
kwargs["chi_angles"] = wrap_chis
result = rotamer(**kwargs)
if (result.is_outlier()) or (not outliers_only):
self.results.append(result)
out_count, out_percent = self.get_outliers_count_and_fraction()
self.out_percent = out_percent * 100.0
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
from mmtbx.validation import utils
import mmtbx.rotamer
from mmtbx.rotamer import ramachandran_eval
from scitbx.array_family import flex
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 = []
r = ramachandran_eval.RamachandranEval()
prev_rezes, next_rezes = None, None
prev_resid = None
cur_resseq = None
next_resseq = None
for model in pdb_hierarchy.models():
for chain in model.chains():
if use_segids:
chain_id = utils.get_segid_as_chainid(chain=chain)
else:
chain_id = chain.id
residues = list(chain.residue_groups())
for i, residue_group in enumerate(residues):
# The reason I pass lists of atom_groups to get_phi and get_psi is to
# deal with the particular issue where some residues have an A alt
# conf that needs some atoms from a "" alt conf to get calculated
# correctly. See 1jxt.pdb for examples. This way I can search both
# the alt conf atoms and the "" atoms if necessary.
prev_atom_list, next_atom_list, atom_list = None, None, None
if cur_resseq is not None:
prev_rezes = rezes
prev_resseq = cur_resseq
rezes = construct_complete_residues(residues[i])
cur_resseq = residue_group.resseq_as_int()
cur_icode = residue_group.icode.strip()
if (i > 0):
#check for insertion codes
if (cur_resseq == residues[i-1].resseq_as_int()) :
if (cur_icode == '') and (residues[i-1].icode.strip() == '') :
continue
elif (cur_resseq != (residues[i-1].resseq_as_int())+1):
continue
if (i < len(residues)-1):
#find next residue
if residue_group.resseq_as_int() == \
residues[i+1].resseq_as_int():
if (cur_icode == '') and (residues[i+1].icode.strip() == '') :
continue
elif residue_group.resseq_as_int() != \
(residues[i+1].resseq_as_int())-1:
continue
next_rezes = construct_complete_residues(residues[i+1])
next_resid = residues[i+1].resseq_as_int()
else:
next_rezes = None
next_resid = None
for atom_group in residue_group.atom_groups():
alt_conf = atom_group.altloc
if rezes is not None:
atom_list = rezes.get(alt_conf)
if prev_rezes is not None:
prev_atom_list = prev_rezes.get(alt_conf)
if (prev_atom_list is None):
prev_keys = sorted(prev_rezes.keys())
prev_atom_list = prev_rezes.get(prev_keys[0])
if next_rezes is not None:
next_atom_list = next_rezes.get(alt_conf)
if (next_atom_list is None):
next_keys = sorted(next_rezes.keys())
next_atom_list = next_rezes.get(next_keys[0])
phi = get_phi(prev_atom_list, atom_list)
psi = get_psi(atom_list, next_atom_list)
coords = get_center(atom_group)
if (phi is not None and psi is not None):
res_type = RAMA_GENERAL
self.n_total += 1
if (atom_group.resname[0:3] == "GLY"):
res_type = RAMA_GLYCINE
elif (atom_group.resname[0:3] == "PRO"):
is_cis = is_cis_peptide(prev_atom_list, atom_list)
if is_cis:
res_type = RAMA_CISPRO
else:
res_type = RAMA_TRANSPRO
elif (isPrePro(residues, i)):
res_type = RAMA_PREPRO
elif (atom_group.resname[0:3] == "ILE" or \
atom_group.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 = []
for atoms in [prev_atom_list, atom_list, next_atom_list] :
for a in atoms :
if (a.name.strip() == "CA") :
a_ = atom(pdb_atom=a)
c_alphas.append(c_alpha(
id_str=a_.atom_group_id_str(),
xyz=a_.xyz))
assert (len(c_alphas) == 3)
result = ramachandran(
chain_id=chain_id,
resseq=residue_group.resseq,
icode=residue_group.icode,
resname=atom_group.resname,
altloc=atom_group.altloc,
segid=None, # XXX ???
phi=phi,
psi=psi,
rama_type=ramaType,
res_type=res_type,
score=value*100,
outlier=is_outlier,
xyz=coords,
c_alphas=c_alphas)
if (not outliers_only or is_outlier) :
self.results.append(result)
if is_outlier :
i_seqs = atom_group.atoms().extract_i_seq()
assert (not i_seqs.all_eq(0))
self._outlier_i_seqs.extend(i_seqs)
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