def molprobity_validation(targetid, ensembler_stage=None, loglevel=None):
set_loglevel(loglevel)
valid_model_ids = []
if mpistate.rank == 0:
if ensembler_stage is None:
ensembler_stage = get_most_advanced_ensembler_modeling_stage(
targetid)
valid_model_ids = get_valid_model_ids(ensembler_stage, targetid)
if ensembler_stage is None:
ensembler_stage = mpistate.comm.bcast(ensembler_stage, root=0)
valid_model_ids = mpistate.comm.bcast(valid_model_ids, root=0)
nvalid_model_ids = len(valid_model_ids)
model_structure_filename = model_filenames_by_ensembler_stage[
ensembler_stage]
models_target_dir = os.path.join(default_project_dirnames.models, targetid)
molprobity_results_filepath = os.path.join(
models_target_dir,
'validation_scores_sorted-molprobity-{}'.format(ensembler_stage))
molprobity_scores_sublist = []
for model_index in range(mpistate.rank, nvalid_model_ids, mpistate.size):
model_id = valid_model_ids[model_index]
logger.debug('MPI process {} working on model {}'.format(
mpistate.rank, model_id))
molprobity_score = run_molprobity_oneline_analysis_and_write_results(
targetid,
model_id,
ensembler_stage,
model_structure_filename=model_structure_filename,
models_target_dir=models_target_dir,
)
molprobity_scores_sublist.append((model_id, molprobity_score))
molprobity_scores_gathered_list = mpistate.comm.gather(
molprobity_scores_sublist, root=0)
if mpistate.rank == 0:
molprobity_scores_list_of_tuples = [
item for sublist in molprobity_scores_gathered_list
for item in sublist
]
molprobity_scores_sorted = sorted(molprobity_scores_list_of_tuples,
key=lambda x: x[1])
write_molprobity_scores_list(molprobity_scores_sorted,
molprobity_results_filepath)
def molprobity_validation(targetid, ensembler_stage=None, loglevel=None):
set_loglevel(loglevel)
valid_model_ids = []
if mpistate.rank == 0:
if ensembler_stage is None:
ensembler_stage = get_most_advanced_ensembler_modeling_stage(targetid)
valid_model_ids = get_valid_model_ids(ensembler_stage, targetid)
if ensembler_stage is None:
ensembler_stage = mpistate.comm.bcast(ensembler_stage, root=0)
valid_model_ids = mpistate.comm.bcast(valid_model_ids, root=0)
nvalid_model_ids = len(valid_model_ids)
model_structure_filename = model_filenames_by_ensembler_stage[ensembler_stage]
models_target_dir = os.path.join(default_project_dirnames.models, targetid)
molprobity_results_filepath = os.path.join(
models_target_dir, "validation_scores_sorted-molprobity-{}".format(ensembler_stage)
)
molprobity_scores_sublist = []
for model_index in range(mpistate.rank, nvalid_model_ids, mpistate.size):
model_id = valid_model_ids[model_index]
logger.debug("MPI process {} working on model {}".format(mpistate.rank, model_id))
molprobity_score = run_molprobity_oneline_analysis_and_write_results(
targetid,
model_id,
ensembler_stage,
model_structure_filename=model_structure_filename,
models_target_dir=models_target_dir,
)
molprobity_scores_sublist.append((model_id, molprobity_score))
molprobity_scores_gathered_list = mpistate.comm.gather(molprobity_scores_sublist, root=0)
if mpistate.rank == 0:
molprobity_scores_list_of_tuples = [item for sublist in molprobity_scores_gathered_list for item in sublist]
molprobity_scores_sorted = sorted(molprobity_scores_list_of_tuples, key=lambda x: x[1])
write_molprobity_scores_list(molprobity_scores_sorted, molprobity_results_filepath)
def __init__(self,
targetid,
ensembler_stage=None,
traj_filepath=None,
topol_filepath=None,
models_data_filepath=None,
process_only_these_templates=None,
loglevel=None,
run_main=True):
"""Makes a trajectory for a given target, using mdtraj. The trajectory can be used with other
software, e.g. for visualization with PyMOL or VMD.
Parameters
----------
targetid : str
e.g. 'EGFR_HUMAN_D0'
ensembler_stage : str
The Ensembler stage from which to build models, e.g. 'build_models' results in a trajectory
built from the 'model.pdb.gz' files output by the build_models command.
options: build_models|refine_implicit_md|refine_explicit_md
default: most advanced stage for which model files are available
traj_filepath : str
default: models/[targetid]/traj-[ensembler_stage].xtc
topol_filepath : str
default: models/[targetid]/traj-[ensembler_stage]-topol.pdb
models_data_filepath :
default: models/[targetid]/traj-[ensembler_stage]-data.csv
process_only_these_templates : list of str
Returns
-------
traj : mdtraj.Trajectory
df : pandas.DataFrame
models data (e.g. sequence identities):
"""
ensembler.utils.set_loglevel(loglevel)
ensembler.core.check_project_toplevel_dir()
self.models_target_dir = os.path.join(default_project_dirnames.models,
targetid)
logger.debug('Working on target %s' % targetid)
if ensembler_stage is None:
self.ensembler_stage = get_most_advanced_ensembler_modeling_stage(
targetid)
else:
self.ensembler_stage = ensembler_stage
if traj_filepath is None:
self.traj_filepath = os.path.join(
self.models_target_dir,
'traj-{0}.xtc'.format(self.ensembler_stage))
else:
self.traj_filepath = traj_filepath
if topol_filepath is None:
self.topol_filepath = os.path.join(
self.models_target_dir,
'traj-{0}-topol.pdb'.format(self.ensembler_stage))
else:
self.topol_filepath = topol_filepath
if models_data_filepath is None:
self.models_data_filepath = os.path.join(
self.models_target_dir,
'traj-{0}-data.csv'.format(self.ensembler_stage))
else:
self.models_data_filepath = models_data_filepath
if process_only_these_templates:
self.templateids = process_only_these_templates
else:
directories = [
directory for directory in os.walk(self.models_target_dir)
]
self.templateids = directories[0][1]
if run_main:
self._gen_df()
self.df.to_csv(self.models_data_filepath,
columns=['templateid', 'seqid'])
self._construct_traj()
self._superpose()
self._write_traj()
def __init__(self, targetid, ensembler_stage=None, traj_filepath=None, topol_filepath=None,
models_data_filepath=None, process_only_these_templates=None, loglevel=None,
run_main=True):
"""Makes a trajectory for a given target, using mdtraj. The trajectory can be used with other
software, e.g. for visualization with PyMOL or VMD.
Parameters
----------
targetid : str
e.g. 'EGFR_HUMAN_D0'
ensembler_stage : str
The Ensembler stage from which to build models, e.g. 'build_models' results in a trajectory
built from the 'model.pdb.gz' files output by the build_models command.
options: build_models|refine_implicit_md|refine_explicit_md
default: most advanced stage for which model files are available
traj_filepath : str
default: models/[targetid]/traj-[ensembler_stage].xtc
topol_filepath : str
default: models/[targetid]/traj-[ensembler_stage]-topol.pdb
models_data_filepath :
default: models/[targetid]/traj-[ensembler_stage]-data.csv
process_only_these_templates : list of str
Returns
-------
traj : mdtraj.Trajectory
df : pandas.DataFrame
models data (e.g. sequence identities):
"""
ensembler.utils.set_loglevel(loglevel)
ensembler.core.check_project_toplevel_dir()
self.models_target_dir = os.path.join(default_project_dirnames.models, targetid)
logger.debug('Working on target %s' % targetid)
if ensembler_stage is None:
self.ensembler_stage = get_most_advanced_ensembler_modeling_stage(targetid)
else:
self.ensembler_stage = ensembler_stage
if traj_filepath is None:
self.traj_filepath = os.path.join(
self.models_target_dir, 'traj-{0}.xtc'.format(self.ensembler_stage)
)
else:
self.traj_filepath = traj_filepath
if topol_filepath is None:
self.topol_filepath = os.path.join(
self.models_target_dir, 'traj-{0}-topol.pdb'.format(self.ensembler_stage)
)
else:
self.topol_filepath = topol_filepath
if models_data_filepath is None:
self.models_data_filepath = os.path.join(
self.models_target_dir, 'traj-{0}-data.csv'.format(self.ensembler_stage)
)
else:
self.models_data_filepath = models_data_filepath
if process_only_these_templates:
self.templateids = process_only_these_templates
else:
directories = [ directory for directory in os.walk(self.models_target_dir) ]
self.templateids = directories[0][1]
if run_main:
self._gen_df()
self.df.to_csv(self.models_data_filepath, columns=['templateid', 'seqid'])
self._construct_traj()
self._superpose()
self._write_traj()
