def tst_02():
# Test for remove_clashes
import mmtbx.model
from mmtbx.clashes import remove_clashes
import iotbx.pdb
import sys
pdb_inp = iotbx.pdb.input(lines=pdb_poor.splitlines(), source_info='None')
model = mmtbx.model.manager(model_input=pdb_inp, )
model.set_log(log=null_out())
print "\n", "-" * 79
print " Summary of input model statistics "
print "-" * 79
model.get_restraints_manager()
geometry = model.geometry_statistics()
geometry.show(log=sys.stdout)
rc = remove_clashes(model=model)
print "\n", "-" * 79
print "Starting residues: %d " % (
rc.model.get_hierarchy().overall_counts().n_residues)
print "Side-chains removed: %d Residues removed: %d" % (
rc.side_chains_removed, rc.residues_removed)
print "Final residues: %d " % (
rc.new_model.get_hierarchy().overall_counts().n_residues)
rc.new_model.set_log(log=null_out())
rc.new_model.get_restraints_manager()
new_geometry = rc.new_model.geometry_statistics()
new_geometry.show(log=sys.stdout)
assert rc.side_chains_removed == 1
assert rc.residues_removed == 0
def show_geo(fn, prefix):
pdb_inp = iotbx.pdb.input(file_name=fn)
model = mmtbx.model.manager(model_input=pdb_inp,
build_grm=True,
log=null_out())
mes = "%s %s" % (prefix, model.geometry_statistics(
use_hydrogens=False).show_short())
return model, mes
def run(args):
assert len(args) == 1
# Read file into pdb_input class
inp = iotbx.pdb.input(file_name=args[0])
# create a model manager
# Catch Sorry about MTRIX here.
model = mmtbx.model.manager(
model_input=inp,
restraint_objects=
None, # these are ligands if any [('fname', cif_object), ()]
log=null_out(),
)
print("=" * 80)
print("number of atoms with MTRIX multiplication:",
model.get_number_of_atoms())
show_ss_counts(model)
# Expand with BIOMT if needed. MTRIX are already expanded by default
# Catch case when both MTRIX and BIOMT present, or other Sorry raised by
# BIOMT handling.
# LIMITATION: this should be done before any selections made on model.manager
double_counter = 0
try:
model.expand_with_BIOMT_records()
except Sorry as e:
if str(e).startswith("Model has been already expanded"):
double_counter += 1
print("=" * 80)
print("number of atoms with BIOMT multiplication:",
model.get_number_of_atoms())
show_ss_counts(model)
# Get default params
pdb_int_params = mmtbx.model.manager.get_default_pdb_interpretation_params(
)
# Set whatever you want
pdb_int_params.pdb_interpretation.secondary_structure.protein.enabled = True
pdb_int_params.pdb_interpretation.ncs_search.enabled = True
pdb_int_params.pdb_interpretation.ncs_search.residue_match_radius = 999
pdb_int_params.pdb_interpretation.clash_guard.nonbonded_distance_threshold = None
#pdb_int_params.pdb_interpretation.nonbonded_weight = None
# set the params. Note, that GRM would be dropped, even if it was already
# constructed. In this example it is not yet constructed.
model.set_pdb_interpretation_params(params=pdb_int_params)
grm = model.get_restraints_manager()
# Not clear which one should be used at the moment
gs = model.geometry_statistics()
gs.show()
# The second way
msi = model.get_model_statistics_info()
msi.show_remark_3()
def get_statistics(self, model):
# should we shift here? No
# should we multiply NCS here? No
geometry = model.geometry_statistics().result()
motifs = phi_psi_2.get_results(model=model, log=null_out())['counts']
mcounts = motifs.get_counts()
res = {}
for key, value in mcounts.iteritems():
res[key] = value.percent
geometry.merge(group_args(**res))
return geometry
def get_statistics(self, model):
# should we shift here? No
# should we multiply NCS here? No
geometry = model.geometry_statistics().result()
motifs = phi_psi_2.results_manager(model=model, log=null_out()).get_motif_count()
mcounts = motifs.get_counts()
res = {}
# TODO is mcounts a dict ? If not consider changing back
for key, value in six.iteritems(mcounts):
res[key] = value.percent
geometry.merge(group_args(**res))
return geometry
def run():
rmsd_dirs = ["0.3/", "0.6/", "0.9/", "1.2/", "1.5/"]
base_names = ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"]
ref = get_reference()
tmp_cntr1 = 0
print "Hbond analysis"
print "model min max mean recovered bond_rmsd rama_favored clashscore"
for sub_root in [
"./perturbed/", "./cctbx_opt/", "./xtb_opt_stpmax_0.3/",
"./xtb_opt_stpmax_0.4/", "./xtb_opt_stpmax_0.5/",
"./terachem_opt_final/"
]:
print sub_root
for rmsd_dir in rmsd_dirs:
h_bonds = flex.double()
rmsd_bonds = flex.double()
rama_fav = flex.double()
clashscore = flex.double()
cntr = 0
for fn in base_names:
file_name = sub_root + rmsd_dir + fn + ".pdb"
if (not os.path.exists(file_name)): assert 0, file_name
model = get_model(file_name)
assert ref.h.is_similar_hierarchy(model.get_hierarchy())
g = model.geometry_statistics(use_hydrogens=False).result()
rmsd_bonds.append(g.bond.mean)
rama_fav.append(g.ramachandran.favored)
clashscore.append(g.clash.score)
#print g.bond.mean, g.clash.score, g.rotamer.outliers, g.c_beta.outliers, \
# g.ramachandran.outliers, g.ramachandran.allowed, g.ramachandran.favored
sites_cart = model.get_sites_cart()
if (sub_root.count("perturbed") > 0):
tmp_cntr1 += 1
assert approx_equal(
float(rmsd_dir.replace("/", "")),
flex.mean(
flex.sqrt((ref.sites_cart - sites_cart).dot())),
0.01)
for pair in ref.h_bonds_i_seqs:
d = dist(sites_cart[pair[0]], sites_cart[pair[1]])
h_bonds.append(d)
cntr += 1
assert cntr == 10, cntr
sel = h_bonds < 2.3
sel &= h_bonds > 1.6
if (h_bonds.size() > 0):
print rmsd_dir, "%8.3f %8.3f %8.3f"%h_bonds.min_max_mean().as_tuple(), \
" %7.2f"%(sel.count(True)*100./(len(ref.h_bonds_i_seqs)*cntr)),\
" %6.4f %7.2f %4.2f"%(flex.mean(rmsd_bonds), flex.mean(rama_fav), flex.mean(clashscore))
else:
print rmsd_dir, "N/A"
#
assert tmp_cntr1 == 50, tmp_cntr1
def exercise_2():
inp = iotbx.pdb.input(lines=pdb_str_1, source_info=None)
model = mmtbx.model.manager(
model_input = inp)
model.setup_scattering_dictionaries(scattering_table="wk1995")
g_stat = model.geometry_statistics()
def run(
pdb_filename,
model_completion=True,
keep_alt_loc=False,
skip_validation=False,
calculate_charge=False,
append_to_end_of_model=False,
neutron_option=None,
hydrogen_atom_occupancies=0.,
):
#
# do all *.pdb in a directory
#
if os.path.isdir(pdb_filename):
loop_over_dir(pdb_filename)
return
#
# fetch a PDB code
#
if pdb_filename.find(".pdb") == -1:
if not os.path.exists('%s.pdb' % pdb_filename):
run_fetch_pdb(pdb_filename)
pdb_filename = '%s.pdb' % pdb_filename
if 1: # read file option
ppf = hierarchy_utils.get_processed_pdb(pdb_filename=pdb_filename)
elif 0: # raw records example
pdb_inp = iotbx.pdb.input(pdb_filename)
hierarchy = pdb_inp.construct_hierarchy()
raw_records = []
for atom in hierarchy.atoms():
raw_records.append(atom.format_atom_record())
ppf = hierarchy_utils.get_processed_pdb(raw_records=raw_records)
else: # from hierarchy
pdb_inp = iotbx.pdb.input(pdb_filename)
hierarchy = pdb_inp.construct_hierarchy()
ppf = hierarchy_utils.get_processed_pdb(
pdb_inp=hierarchy.as_pdb_input())
if not skip_validation:
model = mmtbx.model.manager(
model_input=None,
crystal_symmetry=ppf.all_chain_proxies.pdb_inp.crystal_symmetry(),
pdb_hierarchy=ppf.all_chain_proxies.pdb_hierarchy)
initial_model_statistics = model.geometry_statistics()
# should use cctbx
if keep_alt_loc: pass
else:
hierarchy = remove_alternative_locations(
ppf.all_chain_proxies.pdb_hierarchy)
ppf.all_chain_proxies.pdb_hierarchy.shift_to_origin(
ppf.all_chain_proxies.pdb_inp.crystal_symmetry())
ppf.all_chain_proxies.pdb_hierarchy.remove_residue_groups_with_atoms_on_special_positions_selective(
ppf.all_chain_proxies.pdb_inp.crystal_symmetry())
if 0:
output = hierarchy_utils.write_hierarchy(
pdb_filename, # uses to get output filename
ppf.all_chain_proxies.pdb_inp,
ppf.all_chain_proxies.pdb_hierarchy,
"temp0")
ppf = hierarchy_utils.get_processed_pdb(pdb_filename=output)
else:
raw_records = hierarchy_utils.get_raw_records(
ppf.all_chain_proxies.pdb_inp,
ppf.all_chain_proxies.pdb_hierarchy,
)
ppf = hierarchy_utils.get_processed_pdb(raw_records=raw_records)
hetero_charges = charges.get_hetero_charges(
ppf.all_chain_proxies.pdb_inp,
ppf.all_chain_proxies.pdb_hierarchy,
)
if not hetero_charges:
# some defaults
hetero_charges = charges.default_ion_charges
inter_residue_bonds = charges.get_inter_residue_bonds(ppf)
#
# extends side chains and add hydrogens
#
if model_completion:
use_capping_hydrogens = False
fname = 'complete'
else:
use_capping_hydrogens = True
fname = 'capping'
#assert 0 # model has H
ppf = completion.complete_pdb_hierarchy(
ppf.all_chain_proxies.pdb_hierarchy,
ppf.geometry_restraints_manager(),
pdb_filename=pdb_filename,
pdb_inp=ppf.all_chain_proxies.pdb_inp,
use_capping_hydrogens=use_capping_hydrogens,
append_to_end_of_model=append_to_end_of_model,
)
if calculate_charge:
new_pdb_filename = hierarchy_utils.write_hierarchy(
pdb_filename, # uses to get output filename
ppf.all_chain_proxies.pdb_inp,
ppf.all_chain_proxies.pdb_hierarchy,
fname)
## need now inter_residue_bonds because of added hydrogens
## maybe update the bond table!!!
ppf = hierarchy_utils.get_processed_pdb(pdb_filename=new_pdb_filename)
inter_residue_bonds = charges.get_inter_residue_bonds(ppf,
verbose=True)
total_charge = charges.calculate_pdb_hierarchy_charge(
ppf.all_chain_proxies.pdb_hierarchy,
hetero_charges=hetero_charges,
inter_residue_bonds=inter_residue_bonds,
)
print("total_charge", total_charge)
# Idealize H as riding
params = mmtbx.model.manager.get_default_pdb_interpretation_params()
params.pdb_interpretation.use_neutron_distances = True
params.pdb_interpretation.restraints_library.cdl = False
h = ppf.all_chain_proxies.pdb_hierarchy
asc = h.atom_selection_cache()
sel = asc.selection("element H or element D")
model = mmtbx.model.manager(
model_input=None,
pdb_hierarchy=ppf.all_chain_proxies.pdb_hierarchy,
crystal_symmetry=ppf.all_chain_proxies.pdb_inp.crystal_symmetry(),
log=null_out())
model.process(make_restraints=True,
grm_normalization=True,
pdb_interpretation_params=params)
model.idealize_h_riding()
if neutron_option == 'all_d':
from mmtbx.ligands.ready_set_utils import perdeuterate_model_ligands
perdeuterate_model_ligands(ppf.all_chain_proxies.pdb_hierarchy)
model.set_occupancies(hydrogen_atom_occupancies, selection=sel)
## after no error getting total charge, write the completed pdb file
hierarchy_utils.write_hierarchy(
pdb_filename, # uses to get output filename
ppf.all_chain_proxies.pdb_inp,
model.get_hierarchy(),
fname)
if not skip_validation:
final_model_statistics = model.geometry_statistics()
def get_statistics(self, model):
# should we shift here? No
# should we multiply NCS here? No
return model.geometry_statistics().result()
