Merge pull request #48 from danielparton/master

1.0.5 release

docs/cli_docs.rst

@@ -22,7 +22,7 @@ The ``ensembler`` tool is operated via a number of subcommands, which should be
   ensembler cluster
   ensembler refine_implicit
   ensembler solvate
-  ensembler refine_implicit
+  ensembler refine_explicit
   ensembler package_models
 
 Furthermore, the ``ensembler quickmodel`` subcommand allows the entire modeling

docs/examples.rst

@@ -19,7 +19,7 @@ Example using the quickmodel function
 
 ::
 
-  $ ensembler quickmodel --target_uniprot_entry_name EGFR_HUMAN --uniprot_domain_regex '^Protein kinase' --template_pdbids 4KB8,4AF3 --no-loopmodel
+  $ ensembler quickmodel --target_uniprot_entry_name EGFR_HUMAN --uniprot_domain_regex '^Protein kinase' --template_pdbids 1M14,4AF3 --no-loopmodel
 
 Models human EGFR onto two templates selected via PDB IDs. The ``quickmodel`` function executes the entire modeling pipeline in one go, and is designed to work with only a few targets and templates. For generating larger numbers of models (such as entire protein families), the main pipeline functions should be used.
 

ensembler/core.py

@@ -254,6 +254,9 @@ def __init__(self, manual_overrides_filepath=None):
         else:
             manual_overrides_yaml = {}
 
+        if type(manual_overrides_yaml) is not dict:
+            manual_overrides_yaml = {}
+
         self.target = TargetManualOverrides(manual_overrides_yaml)
         self.template = TemplateManualOverrides(manual_overrides_yaml)
         self.refinement = RefinementManualOverrides(manual_overrides_yaml)

ensembler/refinement.py

@@ -62,6 +62,8 @@ def refine_implicit_md(
     if (sim_length / timestep) < nsteps_per_iteration:
         nsteps_per_iteration = int(sim_length / timestep)
 
+    niterations = int((sim_length / timestep) / nsteps_per_iteration)
+
     models_dir = os.path.abspath(ensembler.core.default_project_dirnames.models)
 
     targets, templates_resolved_seq = ensembler.core.get_targets_and_templates()
@@ -85,8 +87,6 @@ def refine_implicit_md(
     kB = unit.MOLAR_GAS_CONSTANT_R
     kT = kB * temperature
 
-    niterations = int((sim_length / timestep) / nsteps_per_iteration)
-
     def simulate_implicit_md():
 
         logger.debug("Reading model...")
@@ -719,10 +719,10 @@ def refine_explicit_md(
         rigidWater = True, # rigid water
         removeCMMotion = False, # remove center-of-mass motion
         sim_length=100.0 * unit.picoseconds,
-        timestep=2.0 * unit.femtoseconds, # timestep
-        temperature=300.0 * unit.kelvin, # simulation temperature
-        pressure=1.0 * unit.atmospheres, # simulation pressure
-        collision_rate=20.0 / unit.picoseconds, # Langevin collision rate
+        timestep=2.0 * unit.femtoseconds,   # timestep
+        temperature=300.0 * unit.kelvin,   # simulation temperature
+        pressure=1.0 * unit.atmospheres,   # simulation pressure
+        collision_rate=20.0 / unit.picoseconds,   # Langevin collision rate
         barostat_period=50,
         minimization_tolerance=10.0 * unit.kilojoules_per_mole / unit.nanometer,
         minimization_steps=20,
@@ -744,6 +744,7 @@ def refine_explicit_md(
     if (sim_length / timestep) < nsteps_per_iteration:
         nsteps_per_iteration = int(sim_length / timestep)
 
+    niterations = int((sim_length / timestep) / nsteps_per_iteration)
 
     if process_only_these_templates:
         selected_template_indices = [i for i, seq in enumerate(templates_resolved_seq) if seq.id in process_only_these_templates]
@@ -764,8 +765,6 @@ def refine_explicit_md(
     kB = unit.MOLAR_GAS_CONSTANT_R
     kT = kB * temperature
 
-    niterations = int((sim_length / timestep) / nsteps_per_iteration)
-
     def solvate_pdb(pdb, target_nwaters, water_model=water_model):
         """
         Solvate the contents of a PDB file, ensuring it has exactly 'target_nwaters' waters.

ensembler/tests/test_tools.py

@@ -1,14 +1,46 @@
+import os
 from ensembler.tools.inspect import LoopmodelLogs
 from ensembler.tools.mktraj import MkTrajImplicitStart
+from ensembler.tools.quick_model import QuickModel
+from ensembler.core import default_project_dirnames
+from simtk import unit
+from ensembler.utils import enter_temp_dir
 from ensembler.tests.integrationtest_utils import integrationtest_context
+from nose.plugins.attrib import attr
 
 
+@attr('unit')
 def test_loopmodel_logs():
     with integrationtest_context(set_up_project_stage='templates_modeled_loops'):
         loopmodel_logs = LoopmodelLogs()
         loopmodel_logs.add_missing_resis_data()
 
 
+@attr('unit')
 def test_mktraj_implicit_start():
     with integrationtest_context(set_up_project_stage='refined_explicit'):
         MkTrajImplicitStart(targetid='EGFR_HUMAN_D0', loglevel='debug')
+
+
+@attr('network')
+@attr('slow')
+def test_quick_model():
+    uniprot_entry_name = 'EGFR_HUMAN'
+    uniprot_domain_regex = '^Protein kinase'
+    pdbids = ['4AF3']
+    chainids = {
+        '4AF3': ['A']
+    }
+    with enter_temp_dir():
+        QuickModel(
+            target_uniprot_entry_name=uniprot_entry_name,
+            uniprot_domain_regex=uniprot_domain_regex,
+            pdbids=pdbids,
+            chainids=chainids,
+            loopmodel=False,
+            sim_length=2.0*unit.femtoseconds,
+        )
+        assert os.path.exists(os.path.join(
+            default_project_dirnames.models,
+            'EGFR_HUMAN_D0', 'AURKB_HUMAN_D0_4AF3_A', 'explicit-refined.pdb.gz'
+        ))

ensembler/tools/quick_model.py

@@ -6,14 +6,16 @@
 import ensembler.modeling
 import ensembler.refinement
 import ensembler.packaging
+from ensembler.core import get_templates_resolved_seq
 from simtk import unit
 
 
 class QuickModel(object):
     def __init__(self, targetid=None, templateids=None, target_uniprot_entry_name=None,
                  uniprot_domain_regex=None, pdbids=None, chainids=None, template_uniprot_query=None,
                  template_seqid_cutoff=None, loopmodel=True, package_for_fah=False, nfahclones=None,
-                 structure_dirs=None, sim_length=100*unit.picoseconds):
+                 structure_dirs=None, sim_length=100*unit.picoseconds
+                 ):
         """
         Run this after having set up targets and templates with the appropriate ensembler commands.
 
@@ -53,14 +55,19 @@ def __init__(self, targetid=None, templateids=None, target_uniprot_entry_name=No
 
         if not self.targetid:
             if not self.target_uniprot_entry_name or not self.uniprot_domain_regex:
-                raise Exception('If no targetid is passed, must specify target_uniprot_entry_name and uniprot_domain_regex.')
+                raise Exception(
+                    'If no targetid is passed, must specify target_uniprot_entry_name and '
+                    'uniprot_domain_regex.'
+                )
             uniprot_query_string = 'mnemonic:%s' % self.target_uniprot_entry_name
-            gather_targets_obj = ensembler.initproject.GatherTargetsFromUniProt(uniprot_query_string, uniprot_domain_regex=self.uniprot_domain_regex)
+            gather_targets_obj = ensembler.initproject.GatherTargetsFromUniProt(
+                uniprot_query_string, uniprot_domain_regex=self.uniprot_domain_regex
+            )
             self.targetid = gather_targets_obj.targets[0].id
 
         existing_templates = False
         if os.path.exists(os.path.join(ensembler.core.default_project_dirnames.templates, 'templates-resolved-seq.fa')):
-            all_templates_resolved_seq, all_templates_full_seq = ensembler.core.get_templates()
+            all_templates_resolved_seq = get_templates_resolved_seq()
             if len(all_templates_resolved_seq) > 0:
                 existing_templates = True
 
@@ -69,7 +76,7 @@ def __init__(self, targetid=None, templateids=None, target_uniprot_entry_name=No
                 raise Exception('No existing templates found.')
         elif self.pdbids:
             ensembler.initproject.gather_templates_from_pdb(self.pdbids, self.uniprot_domain_regex, chainids=self.chainids, structure_dirs=self.structure_dirs)
-            templates_resolved_seq, templates_full_seq = ensembler.core.get_templates()
+            templates_resolved_seq = get_templates_resolved_seq()
             self.templateids = [t.id for t in templates_resolved_seq]
         elif self.template_uniprot_query:
             ensembler.initproject.gather_templates_from_uniprot(self.template_uniprot_query, self.uniprot_domain_regex, structure_dirs=self.structure_dirs)

setup.py

@@ -3,8 +3,8 @@
 from setuptools import setup
 
 ##########################
-VERSION = "1.0.4"
-ISRELEASED = False
+VERSION = "1.0.5"
+ISRELEASED = True
 __version__ = VERSION
 ##########################