def exercise () :
from mmtbx.building.alternate_conformations import conformer_generation
from mmtbx.monomer_library import server
import iotbx.pdb.hierarchy
generate_inputs()
params = master_phil().extract()
mon_lib_srv = server.server()
pdb_in = iotbx.pdb.hierarchy.input(file_name="shear_frag_single.pdb")
hierarchy = pdb_in.hierarchy
pdb_atoms = hierarchy.atoms()
pdb_atoms.reset_i_seq()
sites_cart = pdb_atoms.extract_xyz()
xrs = pdb_in.input.xray_structure_simple()
models = []
prev_res = next_res = next_next_res = None
for chain in hierarchy.only_model().chains() :
residue_groups = chain.residue_groups()
n_rg = len(residue_groups) # should be 4
for i_res, residue_group in enumerate(residue_groups) :
sites_orig = sites_cart.deep_copy()
next_res = next_next_res = None
if (i_res < (n_rg - 1)) :
next_res = residue_groups[i_res+1].atom_groups()[0]
if (i_res < (n_rg - 2)) :
next_next_res = residue_groups[i_res+2].atom_groups()[0]
atom_groups = residue_group.atom_groups()
primary_conf = atom_groups[0]
out = StringIO()
confs = []
for conf in conformer_generation.generate_single_residue_confs(
atom_group=primary_conf,
sites_cart=sites_cart.deep_copy(),
mon_lib_srv=mon_lib_srv,
params=params.torsion_search,
prev_residue=prev_res,
next_residue=next_res,
next_next_residue=next_next_res,
backrub=False,
shear=True) :
conf.show_summary(out=out)
confs.append(conf)
prev_res = primary_conf
if (confs is None) :
continue
if (i_res == 1) :
assert (""" A ILE 7 None 4.0 mt""")
for conf in confs :
sites_new = sites_cart.set_selected(conf.sites_selection,
conf.sites_selected())
pdb_atoms.set_xyz(sites_new)
models.append(hierarchy.only_model().detached_copy())
new_hierarchy = iotbx.pdb.hierarchy.root()
for i_model, conf in enumerate(models) :
conf.id = str(i_model + 1)
new_hierarchy.append_model(conf)
open("shear_frag_naive_ensemble.pdb", "w").write(
new_hierarchy.as_pdb_string())
def exercise () :
for module in ["reduce", "probe", "phenix_regression"] :
if (not libtbx.env.has_module(module)) :
print "%s not available, skipping" % module
return
from mmtbx.command_line import validation_summary
from iotbx import file_reader
import iotbx.pdb.hierarchy
regression_pdb = libtbx.env.find_in_repositories(
relative_path="phenix_regression/pdb/pdb1jxt.ent",
test=os.path.isfile)
out = StringIO()
summary = validation_summary.run(args=[regression_pdb], out=out)
assert approx_equal(summary.clashscore, 13.597, eps=0.001), "clashscore %s is not 13.597(0.0001)" % summary.clashscore
ss = easy_pickle.dumps(summary)
sss = easy_pickle.loads(ss)
out_1 = StringIO()
out_2 = StringIO()
summary.show(out=out_1)
sss.show(out=out_2)
assert out_1.getvalue() == out_2.getvalue()
pdb_in = file_reader.any_file(regression_pdb)
hierarchy = pdb_in.file_object.hierarchy
new_hierarchy = iotbx.pdb.hierarchy.root()
for i in range(5) :
model = hierarchy.only_model().detached_copy()
model.id = str(i+1)
new_hierarchy.append_model(model)
open("tst_validation_summary.pdb", "w").write(new_hierarchy.as_pdb_string())
out2 = StringIO()
summary = validation_summary.run(args=["tst_validation_summary.pdb"],
out=out2)
assert (type(summary).__name__ == 'ensemble')
print "OK"
def as_pdb_ensemble(self, log=None):
"""
Create a multi-MselfDEL PDB hierarchy with the trial results after filtering
by density and rotamer score.
:returns: the new multi-model PDB hierarchy, or if no models passed the
filtering stage, None
"""
if (log is None) : log = null_out()
import iotbx.pdb.hierarchy
root = iotbx.pdb.hierarchy.root()
trials_and_models = self.get_filtered_trials(include_pdb_hierarchies=True,
log=log)
n_kept = 0
for k, (trial, hierarchy) in enumerate(trials_and_models):
n_kept += 1
new_model = hierarchy.only_model().detached_copy()
new_model.id = str(n_kept)
print >> log, "MODEL %d:" % (k+1)
trial.show_summary(prefix=" ", out=log)
root.append_model(new_model)
if (n_kept == 0):
return None
else :
return root
def exercise():
for module in ["reduce", "probe", "phenix_regression"]:
if not libtbx.env.has_module(module):
print "%s not available, skipping" % module
return
from mmtbx.command_line import validation_summary
from iotbx import file_reader
import iotbx.pdb.hierarchy
regression_pdb = libtbx.env.find_in_repositories(
relative_path="phenix_regression/pdb/pdb1jxt.ent", test=os.path.isfile
)
out = StringIO()
summary = validation_summary.run(args=[regression_pdb], out=out)
assert approx_equal(summary.clashscore, 13.597, eps=0.0001)
ss = easy_pickle.dumps(summary)
sss = easy_pickle.loads(ss)
out_1 = StringIO()
out_2 = StringIO()
summary.show(out=out_1)
sss.show(out=out_2)
assert out_1.getvalue() == out_2.getvalue()
pdb_in = file_reader.any_file(regression_pdb)
hierarchy = pdb_in.file_object.hierarchy
new_hierarchy = iotbx.pdb.hierarchy.root()
for i in range(5):
model = hierarchy.only_model().detached_copy()
model.id = str(i + 1)
new_hierarchy.append_model(model)
open("tst_validation_summary.pdb", "w").write(new_hierarchy.as_pdb_string())
out2 = StringIO()
summary = validation_summary.run(args=["tst_validation_summary.pdb"], out=out2)
assert type(summary).__name__ == "ensemble"
print "OK"
def get_chain_and_ranges(hierarchy):
"""
Helper function to get info from hierarchy to create spec object.
Used in get_ncs_info_as_spec
hierarchy is already selected
"""
c_ids = [c.id for c in hierarchy.only_model().chains()]
assert len(set(c_ids)) == 1
c_id = c_ids[0]
rgs_all = []
for chain in hierarchy.chains():
rgs_all += list(chain.residue_groups())
ranges = []
in_range = False
first_id = None
last_id = None
for rg in rgs_all:
resseq = int(rg.resseq)
if in_range:
if resseq - last_id > 1:
# terminate range, start new one
ranges.append([first_id, last_id])
first_id = resseq
last_id = resseq
else:
# continue range
last_id = resseq
else:
# start new range
first_id = resseq
last_id = resseq
in_range = True
# dumping rest:
ranges.append([first_id, last_id])
return c_id, ranges, len(c.residue_groups())
def exercise_rejoin():
from mmtbx.building.alternate_conformations.tst_build_simple import pdb_raw
pdb_in = iotbx.pdb.hierarchy.input(pdb_string=pdb_raw)
hierarchy = pdb_in.hierarchy
params = alternate_conformations.rejoin_phil.extract()
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy,
params=params,
model_error_ml=0.5,
log=null_out())
assert (n_modified == 3), n_modified # Gln5
# split residue 6 without changing coordinates, set occupancy very low
chain = hierarchy.only_model().chains()[0]
rg6 = chain.residue_groups()[5]
ag = rg6.only_atom_group().detached_copy()
for atom in ag.atoms():
atom.occ = 0.05
for atom in rg6.atoms():
atom.occ = 0.95
rg6.only_atom_group().altloc = 'A'
ag.altloc = 'B'
rg6.append_atom_group(ag)
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy.deep_copy(),
params=params,
model_error_ml=0.5,
log=null_out())
assert (n_modified == 1), n_modified
# now with higher B-factors for all atoms
for atom in hierarchy.atoms():
atom.b = atom.b * 10
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy, params=params, log=null_out())
assert (n_modified == 1), n_modified
def as_pdb_ensemble (self, log=None) :
"""
Create a multi-MselfDEL PDB hierarchy with the trial results after filtering
by density and rotamer score.
:returns: the new multi-model PDB hierarchy, or if no models passed the
filtering stage, None
"""
if (log is None) : log = null_out()
import iotbx.pdb.hierarchy
root = iotbx.pdb.hierarchy.root()
trials_and_models = self.get_filtered_trials(include_pdb_hierarchies=True,
log=log)
n_kept = 0
for k, (trial, hierarchy) in enumerate(trials_and_models) :
n_kept += 1
new_model = hierarchy.only_model().detached_copy()
new_model.id = str(n_kept)
print >> log, "MODEL %d:" % (k+1)
trial.show_summary(prefix=" ", out=log)
root.append_model(new_model)
if (n_kept == 0) :
return None
else :
return root
def join_hierarchies (hierarchies, model_ids=None) :
import iotbx.pdb.hierarchy
root = iotbx.pdb.hierarchy.root()
for k, hierarchy in enumerate(hierarchies) :
model_id = k + 1
if (model_ids is not None) :
model_id = model_ids[k]
model = iotbx.pdb.hierarchy.model(id=str(model_id))
root.append_model(model)
chain = hierarchy.only_model().only_chain().detached_copy()
model.append_chain(chain)
return root
def join_hierarchies(hierarchies, model_ids=None):
import iotbx.pdb.hierarchy
root = iotbx.pdb.hierarchy.root()
for k, hierarchy in enumerate(hierarchies):
model_id = k + 1
if (model_ids is not None):
model_id = model_ids[k]
model = iotbx.pdb.hierarchy.model(id=str(model_id))
root.append_model(model)
chain = hierarchy.only_model().only_chain().detached_copy()
model.append_chain(chain)
return root
def exercise_rejoin () :
from mmtbx.building.alternate_conformations.tst_build_simple import pdb_raw
pdb_in = iotbx.pdb.hierarchy.input(pdb_string=pdb_raw)
hierarchy = pdb_in.hierarchy
params = alternate_conformations.rejoin_phil.extract()
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy,
params=params,
model_error_ml=0.5,
log=null_out())
assert (n_modified == 3), n_modified # Gln5
# split residue 6 without changing coordinates, set occupancy very low
chain = hierarchy.only_model().chains()[0]
rg6 = chain.residue_groups()[5]
ag = rg6.only_atom_group().detached_copy()
for atom in ag.atoms() :
atom.occ = 0.05
for atom in rg6.atoms() :
atom.occ = 0.95
rg6.only_atom_group().altloc = 'A'
ag.altloc = 'B'
rg6.append_atom_group(ag)
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy.deep_copy(),
params=params,
model_error_ml=0.5,
log=null_out())
assert (n_modified == 1), n_modified
# now with higher B-factors for all atoms
for atom in hierarchy.atoms() :
atom.b = atom.b * 10
n_modified = alternate_conformations.rejoin_split_single_conformers(
pdb_hierarchy=hierarchy,
params=params,
log=null_out())
assert (n_modified == 1), n_modified
def exercise_synthetic () :
from mmtbx.regression import tst_build_alt_confs
pdb_in = iotbx.pdb.hierarchy.input(pdb_string=tst_build_alt_confs.pdb_raw)
xrs = pdb_in.input.xray_structure_simple()
fc = abs(xrs.structure_factors(d_min=1.5).f_calc())
flags = fc.resolution_filter(d_min=1.6).generate_r_free_flags()
ls = fc.lone_set(other=flags)
# case 1: no work set in high-res shell
flags2 = ls.array(data=flex.bool(ls.size(), True))
flags_all = flags.concatenate(other=flags2)
mtz_out = fc.as_mtz_dataset(column_root_label="F")
mtz_out.add_miller_array(flags_all, column_root_label="FreeR_flag")
mtz_out.mtz_object().write("tst_molprobity_1.mtz")
open("tst_molprobity_1.pdb", "w").write(tst_build_alt_confs.pdb_raw)
args = [
"tst_molprobity_1.pdb",
"tst_molprobity_1.mtz",
"--kinemage",
"--maps",
"flags.clashscore=False",
"flags.xtriage=True",
]
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# case 2: no test set in high-res shell
flags2 = ls.array(data=flex.bool(ls.size(), False))
flags_all = flags.concatenate(other=flags2)
mtz_out = fc.as_mtz_dataset(column_root_label="F")
mtz_out.add_miller_array(flags_all, column_root_label="FreeR_flag")
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# case 3: multi-MODEL structure
# XXX This is not a very sophisticated test - it only ensures that the
# program does not crash. We need a test for expected output...
hierarchy = pdb_in.hierarchy
model2 = hierarchy.only_model().detached_copy()
hierarchy.append_model(model2)
hierarchy.models()[0].id = "1"
hierarchy.models()[1].id = "2"
open("tst_molprobity_multi_model.pdb", "w").write(hierarchy.as_pdb_string())
args = [
"tst_molprobity_multi_model.pdb",
"tst_molprobity_1.mtz",
"--kinemage",
"--maps",
"flags.clashscore=False",
]
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# test rotamer distributions
open("tst_molprobity_misc1.pdb", "w").write(tst_build_alt_confs.pdb_raw)
args = [
"tst_molprobity_1.pdb",
"rotamer_library=8000",
"outliers_only=False",
"flags.clashscore=False",
]
out = StringIO()
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
result.show(outliers_only=False, out=out)
assert (""" A 7 TYR m-80 93.10 299.6,92.0""" in
out.getvalue()), out.getvalue()
def refit_residues(pdb_hierarchy,
cif_objects,
fmodel,
use_rotamers=True,
anneal=False,
verbose=True,
allow_modified_residues=False,
out=sys.stdout):
"""
Use real-space refinement tools to fit newly extended residues.
"""
from mmtbx.refinement.real_space import fit_residue
from mmtbx.rotamer import rotamer_eval
import mmtbx.monomer_library.server
from mmtbx import building
from scitbx.array_family import flex
mon_lib_srv = mmtbx.monomer_library.server.server()
rotamer_manager = rotamer_eval.RotamerEval()
ppdb_out = box_out = out
if (not verbose):
ppdb_out = null_out()
box_out = null_out()
make_sub_header("Processing new model", out=ppdb_out)
processed_pdb = building.reprocess_pdb(
pdb_hierarchy=pdb_hierarchy,
crystal_symmetry=fmodel.f_obs().crystal_symmetry(),
cif_objects=cif_objects,
out=ppdb_out)
print("", file=ppdb_out)
hierarchy = processed_pdb.all_chain_proxies.pdb_hierarchy
xrs = processed_pdb.xray_structure()
grm_geometry = processed_pdb.geometry_restraints_manager()
make_sub_header("Fitting residues", out=out)
target_map = fmodel.map_coefficients(
map_type="2mFo-DFc", exclude_free_r_reflections=True).fft_map(
resolution_factor=0.25).apply_sigma_scaling().real_map_unpadded()
unit_cell = xrs.unit_cell()
for chain in hierarchy.only_model().chains():
if (not chain.is_protein()) and (not allow_modified_residues): continue
residue_groups = chain.residue_groups()
for i_rg, residue_group in enumerate(residue_groups):
prev_res = next_res = None
atom_groups = residue_group.atom_groups()
if (len(atom_groups) > 1): continue
residue = atom_groups[0]
atoms = residue.atoms()
atoms.reset_tmp()
segids = atoms.extract_segid()
if (segids.all_eq("XXXX")):
sites_start = atoms.extract_xyz()
def get_two_fofc_mean(residue):
sum = 0
n_atoms = 0
for atom in residue.atoms():
if (not atom.element.strip() in ["H", "D"]):
site_frac = unit_cell.fractionalize(
site_cart=atom.xyz)
sum += target_map.eight_point_interpolation(
site_frac)
n_atoms += 1
assert (n_atoms > 0)
return sum / n_atoms
sites_start = atoms.extract_xyz().deep_copy()
two_fofc_mean_start = get_two_fofc_mean(residue)
refit = fit_residue.run_with_minimization(
target_map=target_map,
residue=residue,
xray_structure=xrs,
mon_lib_srv=mon_lib_srv,
rotamer_manager=rotamer_manager,
geometry_restraints_manager=grm_geometry,
real_space_gradients_delta=fmodel.f_obs().d_min() * 0.25,
rms_bonds_limit=0.01,
rms_angles_limit=1.0,
backbone_sample_angle=20,
allow_modified_residues=allow_modified_residues)
two_fofc_mean_end = get_two_fofc_mean(residue)
sites_end = atoms.extract_xyz()
flag = ""
if (two_fofc_mean_end > two_fofc_mean_start):
flag = " <-- keep"
xrs = refit.xray_structure
else:
atoms.set_xyz(sites_start)
for atom in atoms:
atom.tmp = 1
print(" residue '%s' : rmsd=%5.3f 2fofc_start=%5.3f 2fofc_end=%5.3f%s" \
% (residue.id_str(), sites_end.rms_difference(sites_start),
two_fofc_mean_start, two_fofc_mean_end, flag), file=out)
return hierarchy, xrs
def exercise():
from mmtbx.building.alternate_conformations import density_sampling
from mmtbx.utils import fmodel_simple
from mmtbx.monomer_library import server
from iotbx import file_reader
import iotbx.pdb.hierarchy
generate_inputs()
fmodel_params = """
high_resolution = 1.2
r_free_flags_fraction = 0.1
add_sigmas = True
pdb_file = ser_frag.pdb
output {
label = F
type = *real complex
file_name = ser_frag.mtz
}
fmodel.k_sol = 0.3
fmodel.b_sol = 20
"""
open("ser_frag_fmodel.eff", "w").write(fmodel_params)
assert (easy_run.fully_buffered("phenix.fmodel ser_frag_fmodel.eff").
raise_if_errors().return_code == 0)
assert os.path.isfile("ser_frag.mtz")
mtz_in = file_reader.any_file("ser_frag.mtz")
f_obs = mtz_in.file_server.miller_arrays[0]
flags = mtz_in.file_server.miller_arrays[1]
flags = flags.customized_copy(data=(flags.data() == 1))
mon_lib_srv = server.server()
pdb_in = iotbx.pdb.hierarchy.input(file_name="ser_frag_single.pdb")
hierarchy = pdb_in.hierarchy
pdb_atoms = hierarchy.atoms()
pdb_atoms.reset_i_seq()
sites_cart = pdb_atoms.extract_xyz()
xrs = pdb_in.input.xray_structure_simple()
fmodel = fmodel_simple(f_obs=f_obs,
xray_structures=[xrs],
scattering_table="n_gaussian",
r_free_flags=flags,
skip_twin_detection=True)
models = []
prev_res = next_res = next_next_res = None
for chain in hierarchy.only_model().chains():
residue_groups = chain.residue_groups()
n_rg = len(residue_groups)
for i_res, residue_group in enumerate(residue_groups):
sites_orig = sites_cart.deep_copy()
next_res = next_next_res = None
if (i_res < (n_rg - 1)):
next_res = residue_groups[i_res + 1].atom_groups()[0]
if (i_res < (n_rg - 2)):
next_next_res = residue_groups[i_res + 2].atom_groups()[0]
atom_groups = residue_group.atom_groups()
primary_conf = atom_groups[0]
out = StringIO()
confs = density_sampling.screen_residue(
residue=primary_conf,
prev_residue=prev_res,
next_residue=next_res,
next_next_residue=next_next_res,
sites_cart=sites_cart,
fmodel=fmodel,
mon_lib_srv=mon_lib_srv,
params=None,
backrub=True,
shear=False,
verbose=True,
out=out)
prev_res = primary_conf
if (confs is None):
continue
# TODO tweak density sampling to allow a backrubbed conformer with a
# chi1 t rotamer for Ser 99
if (i_res == 1):
assert (""" A SER 99 20.0 None t"""
in out.getvalue())
for conf in confs:
sites_new = sites_cart.set_selected(conf.sites_selection,
conf.sites_selected())
pdb_atoms.set_xyz(sites_new)
models.append(hierarchy.only_model().detached_copy())
confs = density_sampling.screen_residue(
residue=primary_conf,
prev_residue=prev_res,
next_residue=next_res,
next_next_residue=next_next_res,
sites_cart=sites_cart,
fmodel=fmodel,
mon_lib_srv=mon_lib_srv,
params=None,
backrub=False,
out=out)
if (i_res == 1):
print len(confs)
new_hierarchy = iotbx.pdb.hierarchy.root()
for i_model, conf in enumerate(models):
conf.id = str(i_model + 1)
new_hierarchy.append_model(conf)
open("ser_frag_guided_ensemble.pdb",
"w").write(new_hierarchy.as_pdb_string())
def refit_residues (
pdb_hierarchy,
cif_objects,
fmodel,
use_rotamers=True,
anneal=False,
verbose=True,
allow_modified_residues=False,
out=sys.stdout) :
"""
Use real-space refinement tools to fit newly extended residues.
"""
from mmtbx.refinement.real_space import fit_residue
from mmtbx.rotamer import rotamer_eval
import mmtbx.monomer_library.server
from mmtbx import building
from scitbx.array_family import flex
mon_lib_srv = mmtbx.monomer_library.server.server()
rotamer_manager = rotamer_eval.RotamerEval()
ppdb_out = box_out = out
if (not verbose) :
ppdb_out = null_out()
box_out = null_out()
make_sub_header("Processing new model", out=ppdb_out)
processed_pdb = building.reprocess_pdb(
pdb_hierarchy=pdb_hierarchy,
crystal_symmetry=fmodel.f_obs().crystal_symmetry(),
cif_objects=cif_objects,
out=ppdb_out)
print >> ppdb_out, ""
hierarchy = processed_pdb.all_chain_proxies.pdb_hierarchy
xrs = processed_pdb.xray_structure()
grm_geometry = processed_pdb.geometry_restraints_manager()
make_sub_header("Fitting residues", out=out)
target_map = fmodel.map_coefficients(
map_type="2mFo-DFc",
exclude_free_r_reflections=True).fft_map(
resolution_factor=0.25).apply_sigma_scaling().real_map_unpadded()
unit_cell = xrs.unit_cell()
for chain in hierarchy.only_model().chains() :
if (not chain.is_protein()) and (not allow_modified_residues) : continue
residue_groups = chain.residue_groups()
for i_rg, residue_group in enumerate(residue_groups) :
prev_res = next_res = None
atom_groups = residue_group.atom_groups()
if (len(atom_groups) > 1) : continue
residue = atom_groups[0]
atoms = residue.atoms()
atoms.reset_tmp()
segids = atoms.extract_segid()
if (segids.all_eq("XXXX")) :
sites_start = atoms.extract_xyz()
def get_two_fofc_mean (residue) :
sum = 0
n_atoms = 0
for atom in residue.atoms() :
if (not atom.element.strip() in ["H","D"]) :
site_frac = unit_cell.fractionalize(site_cart=atom.xyz)
sum += target_map.eight_point_interpolation(site_frac)
n_atoms += 1
assert (n_atoms > 0)
return sum / n_atoms
sites_start = atoms.extract_xyz().deep_copy()
two_fofc_mean_start = get_two_fofc_mean(residue)
refit = fit_residue.run_with_minimization(
target_map=target_map,
residue=residue,
xray_structure=xrs,
mon_lib_srv=mon_lib_srv,
rotamer_manager=rotamer_manager,
geometry_restraints_manager=grm_geometry,
real_space_gradients_delta=fmodel.f_obs().d_min()*0.25,
rms_bonds_limit=0.01,
rms_angles_limit=1.0,
backbone_sample_angle=20,
allow_modified_residues=allow_modified_residues)
two_fofc_mean_end = get_two_fofc_mean(residue)
sites_end = atoms.extract_xyz()
flag = ""
if (two_fofc_mean_end > two_fofc_mean_start) :
flag = " <-- keep"
xrs = refit.xray_structure
else :
atoms.set_xyz(sites_start)
for atom in atoms :
atom.tmp = 1
print >> out, \
" residue '%s' : rmsd=%5.3f 2fofc_start=%5.3f 2fofc_end=%5.3f%s" \
% (residue.id_str(), sites_end.rms_difference(sites_start),
two_fofc_mean_start, two_fofc_mean_end, flag)
return hierarchy, xrs
scattering_table="n_gaussian")
params = cmdline.params
working_phil = master_phil.format(python_object=params)
master_phil.fetch_diff(source=working_phil).show(out=out)
fmodel = cmdline.fmodel
hierarchy = cmdline.pdb_hierarchy
sele_cache = hierarchy.atom_selection_cache()
assert (params.selection is not None)
selection = sele_cache.selection(params.selection)
assert (selection.count(True) > 0)
have_results = False
pdb_atoms = hierarchy.atoms()
sites_cart = pdb_atoms.extract_xyz()
ensembles = []
t1 = time.time()
for chain in hierarchy.only_model().chains() :
prev_residue = next_residue = None
residue_groups = chain.residue_groups()
n_groups = len(residue_groups)
for i_res, residue_group in enumerate(residue_groups) :
if (i_res < n_groups - 1) :
next_residue = residue_groups[i_res+1].atom_groups()[0]
i_seqs = residue_group.atoms().extract_i_seq()
if (selection.select(i_seqs).all_eq(True)) :
atom_group = residue_group.only_atom_group()
if (get_class(atom_group.resname) != "common_amino_acid") :
continue
confs = density_sampling.screen_residue(
residue=atom_group,
prev_residue=prev_residue,
next_residue=next_residue,
def exercise_synthetic () :
from mmtbx.regression import tst_build_alt_confs
pdb_in = iotbx.pdb.hierarchy.input(pdb_string=tst_build_alt_confs.pdb_raw)
xrs = pdb_in.input.xray_structure_simple()
fc = abs(xrs.structure_factors(d_min=1.5).f_calc())
flags = fc.resolution_filter(d_min=1.6).generate_r_free_flags()
ls = fc.lone_set(other=flags)
# case 1: no work set in high-res shell
flags2 = ls.array(data=flex.bool(ls.size(), True))
flags_all = flags.concatenate(other=flags2)
mtz_out = fc.as_mtz_dataset(column_root_label="F")
mtz_out.add_miller_array(flags_all, column_root_label="FreeR_flag")
mtz_out.mtz_object().write("tst_molprobity_1.mtz")
open("tst_molprobity_1.pdb", "w").write(tst_build_alt_confs.pdb_raw)
args = [
"tst_molprobity_1.pdb",
"tst_molprobity_1.mtz",
"--kinemage",
"--maps",
"flags.clashscore=False",
"flags.xtriage=True",
]
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# case 2: no test set in high-res shell
flags2 = ls.array(data=flex.bool(ls.size(), False))
flags_all = flags.concatenate(other=flags2)
mtz_out = fc.as_mtz_dataset(column_root_label="F")
mtz_out.add_miller_array(flags_all, column_root_label="FreeR_flag")
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# case 3: multi-MODEL structure
# XXX This is not a very sophisticated test - it only ensures that the
# program does not crash. We need a test for expected output...
hierarchy = pdb_in.hierarchy
model2 = hierarchy.only_model().detached_copy()
hierarchy.append_model(model2)
hierarchy.models()[0].id = "1"
hierarchy.models()[1].id = "2"
open("tst_molprobity_multi_model.pdb", "w").write(hierarchy.as_pdb_string())
args = [
"tst_molprobity_multi_model.pdb",
"tst_molprobity_1.mtz",
"--kinemage",
"--maps",
]
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
out = StringIO()
result.show(out=out)
# test rotamer distributions
open("tst_molprobity_misc1.pdb", "w").write(tst_build_alt_confs.pdb_raw)
args = [
"tst_molprobity_1.pdb",
"rotamer_library=8000",
]
out = StringIO()
result = molprobity.run(args=args,
ignore_missing_modules=True,
out=null_out()).validation
result.show(outliers_only=False, out=out)
def exercise () :
from mmtbx.building.alternate_conformations import density_sampling
from mmtbx.utils import fmodel_simple
from mmtbx.monomer_library import server
from iotbx import file_reader
import iotbx.pdb.hierarchy
generate_inputs()
fmodel_params = """
high_resolution = 1.2
r_free_flags_fraction = 0.1
add_sigmas = True
pdb_file = shear_frag.pdb
output {
label = F
type = *real complex
file_name = shear_frag.mtz
}
fmodel.k_sol = 0.3
fmodel.b_sol = 20
"""
open("shear_frag_fmodel.eff", "w").write(fmodel_params)
assert (easy_run.fully_buffered("phenix.fmodel shear_frag_fmodel.eff"
).raise_if_errors().return_code == 0)
assert os.path.isfile("shear_frag.mtz")
mtz_in = file_reader.any_file("shear_frag.mtz")
f_obs = mtz_in.file_server.miller_arrays[0]
flags = mtz_in.file_server.miller_arrays[1]
flags = flags.customized_copy(data=(flags.data()==1))
mon_lib_srv = server.server()
pdb_in = iotbx.pdb.hierarchy.input(file_name="shear_frag_single.pdb")
hierarchy = pdb_in.hierarchy
pdb_atoms = hierarchy.atoms()
pdb_atoms.reset_i_seq()
sites_cart = pdb_atoms.extract_xyz()
xrs = pdb_in.input.xray_structure_simple()
fmodel = fmodel_simple(
f_obs=f_obs,
xray_structures=[xrs],
scattering_table="n_gaussian",
r_free_flags=flags,
skip_twin_detection=True)
models = []
prev_res = next_res = next_next_res = None
for chain in hierarchy.only_model().chains() :
residue_groups = chain.residue_groups()
n_rg = len(residue_groups) # should be 4
for i_res, residue_group in enumerate(residue_groups) :
sites_orig = sites_cart.deep_copy()
next_res = next_next_res = None
if (i_res < (n_rg - 1)) :
next_res = residue_groups[i_res+1].atom_groups()[0]
if (i_res < (n_rg - 2)) :
next_next_res = residue_groups[i_res+2].atom_groups()[0]
atom_groups = residue_group.atom_groups()
primary_conf = atom_groups[0]
out = StringIO()
confs = density_sampling.screen_residue(
residue=primary_conf,
prev_residue=prev_res,
next_residue=next_res,
next_next_residue=next_next_res,
sites_cart=sites_cart,
fmodel=fmodel,
mon_lib_srv=mon_lib_srv,
params=None,
backrub=True,
shear=True,
verbose=True,
out=out)
prev_res = primary_conf
if (confs is None) :
continue
if (i_res == 1) :
assert (""" A ILE 7 None 4.0 mt""")
for conf in confs :
sites_new = sites_cart.set_selected(conf.sites_selection,
conf.sites_selected())
pdb_atoms.set_xyz(sites_new)
models.append(hierarchy.only_model().detached_copy())
new_hierarchy = iotbx.pdb.hierarchy.root()
for i_model, conf in enumerate(models) :
conf.id = str(i_model + 1)
new_hierarchy.append_model(conf)
open("shear_frag_naive_ensemble.pdb", "w").write(
new_hierarchy.as_pdb_string())
def run(args, out=None):
if (out is None): out = sys.stdout
from mmtbx.building.alternate_conformations import density_sampling
import mmtbx.maps.utils
import mmtbx.utils
from mmtbx.monomer_library import server
import iotbx.pdb.hierarchy
get_class = iotbx.pdb.common_residue_names_get_class
mon_lib_srv = server.server()
cmdline = mmtbx.utils.cmdline_load_pdb_and_data(
args=args,
master_phil=master_phil,
process_pdb_file=False,
scattering_table="n_gaussian")
params = cmdline.params
working_phil = master_phil.format(python_object=params)
master_phil.fetch_diff(source=working_phil).show(out=out)
fmodel = cmdline.fmodel
hierarchy = cmdline.pdb_hierarchy
sele_cache = hierarchy.atom_selection_cache()
assert (params.selection is not None)
selection = sele_cache.selection(params.selection)
assert (selection.count(True) > 0)
have_results = False
pdb_atoms = hierarchy.atoms()
sites_cart = pdb_atoms.extract_xyz()
ensembles = []
t1 = time.time()
for chain in hierarchy.only_model().chains():
prev_residue = next_residue = None
residue_groups = chain.residue_groups()
n_groups = len(residue_groups)
for i_res, residue_group in enumerate(residue_groups):
if (i_res < n_groups - 1):
next_residue = residue_groups[i_res + 1].atom_groups()[0]
i_seqs = residue_group.atoms().extract_i_seq()
if (selection.select(i_seqs).all_eq(True)):
atom_group = residue_group.only_atom_group()
if (get_class(atom_group.resname) != "common_amino_acid"):
continue
confs = density_sampling.screen_residue(
residue=atom_group,
prev_residue=prev_residue,
next_residue=next_residue,
sites_cart=sites_cart,
fmodel=fmodel,
mon_lib_srv=mon_lib_srv,
params=params,
map_file_name="map_coeffs.mtz",
out=out)
ensemble = []
for conf in confs:
conf_atoms = conf.set_atom_sites(pdb_atoms)
pdb_lines = []
for atom in conf_atoms:
pdb_lines.append(atom.format_atom_record())
ensemble.append("\n".join(pdb_lines))
ensembles.append(ensemble)
prev_residue = residue_group.atom_groups()[0]
if (len(ensembles) == 0):
raise Sorry("No alternate conformations found.")
t2 = time.time()
print("search time: %.1fs" % (t2 - t1), file=out)
for i_ens, ensemble in enumerate(ensembles):
ensemble_hierarchy = iotbx.pdb.hierarchy.root()
for k, model_str in enumerate(ensemble):
input = iotbx.pdb.hierarchy.input(pdb_string=model_str)
model = input.hierarchy.only_model().detached_copy()
model.id = str(k + 1)
ensemble_hierarchy.append_model(model)
f = open("ensemble_%d.pdb" % (i_ens + 1), "w")
f.write(ensemble_hierarchy.as_pdb_string())
f.close()
print("wrote ensemble_%d.pdb" % (i_ens + 1))
if (params.coot):
easy_run.call(
"coot --pdb %s --auto map_coeffs.mtz --pdb ensemble.pdb" %
params.input.pdb.file_name[0])
def run (args, out=None ):
if (out is None) : out = sys.stdout
from mmtbx.building.alternate_conformations import density_sampling
import mmtbx.maps.utils
import mmtbx.utils
from mmtbx.monomer_library import server
import iotbx.pdb.hierarchy
get_class = iotbx.pdb.common_residue_names_get_class
mon_lib_srv = server.server()
cmdline = mmtbx.utils.cmdline_load_pdb_and_data(
args=args,
master_phil=master_phil,
process_pdb_file=False,
scattering_table="n_gaussian")
params = cmdline.params
working_phil = master_phil.format(python_object=params)
master_phil.fetch_diff(source=working_phil).show(out=out)
fmodel = cmdline.fmodel
hierarchy = cmdline.pdb_hierarchy
sele_cache = hierarchy.atom_selection_cache()
assert (params.selection is not None)
selection = sele_cache.selection(params.selection)
assert (selection.count(True) > 0)
have_results = False
pdb_atoms = hierarchy.atoms()
sites_cart = pdb_atoms.extract_xyz()
ensembles = []
t1 = time.time()
for chain in hierarchy.only_model().chains() :
prev_residue = next_residue = None
residue_groups = chain.residue_groups()
n_groups = len(residue_groups)
for i_res, residue_group in enumerate(residue_groups) :
if (i_res < n_groups - 1) :
next_residue = residue_groups[i_res+1].atom_groups()[0]
i_seqs = residue_group.atoms().extract_i_seq()
if (selection.select(i_seqs).all_eq(True)) :
atom_group = residue_group.only_atom_group()
if (get_class(atom_group.resname) != "common_amino_acid") :
continue
confs = density_sampling.screen_residue(
residue=atom_group,
prev_residue=prev_residue,
next_residue=next_residue,
sites_cart=sites_cart,
fmodel=fmodel,
mon_lib_srv=mon_lib_srv,
params=params,
map_file_name="map_coeffs.mtz",
out=out)
ensemble = []
for conf in confs :
conf_atoms = conf.set_atom_sites(pdb_atoms)
pdb_lines = []
for atom in conf_atoms :
pdb_lines.append(atom.format_atom_record())
ensemble.append("\n".join(pdb_lines))
ensembles.append(ensemble)
prev_residue = residue_group.atom_groups()[0]
if (len(ensembles) == 0) :
raise Sorry("No alternate conformations found.")
t2 = time.time()
print >> out, "search time: %.1fs" % (t2-t1)
for i_ens, ensemble in enumerate(ensembles) :
ensemble_hierarchy = iotbx.pdb.hierarchy.root()
for k, model_str in enumerate(ensemble) :
input = iotbx.pdb.hierarchy.input(pdb_string=model_str)
model = input.hierarchy.only_model().detached_copy()
model.id = str(k+1)
ensemble_hierarchy.append_model(model)
f = open("ensemble_%d.pdb" % (i_ens+1), "w")
f.write(ensemble_hierarchy.as_pdb_string())
f.close()
print "wrote ensemble_%d.pdb" % (i_ens+1)
if (params.coot) :
easy_run.call("coot --pdb %s --auto map_coeffs.mtz --pdb ensemble.pdb" %
params.input.pdb.file_name[0])
