def run(self):
project_strucids = paf.get_strucids_from_project(self.protein)
# get crystal names for protein according to db
conn, c = dbf.connectDB()
c.execute('select crystal_name from proasis_hits where protein=%s', (self.protein,))
rows = c.fetchall()
crystal_list = []
for row in rows:
crystal_list.append(str(row[0]))
crystal_list = list(set(crystal_list))
db_strucids = []
# status_list = []
# good_list = []
file_checks = {'crystal': [], 'bound_state': [], 'mod_date': [], 'pdb': [], 'mtz': [], '2fofc': [], 'fofc': [],
'ligs': []}
# get info for crystals identified
for crystal in list(set(crystal_list)):
c.execute(
"select strucid, bound_conf, modification_date from proasis_hits "
"where crystal_name like %s and strucid NOT LIKE ''",
(crystal,))
bound_list = []
strucid_list = []
mod_date_list = []
rows = c.fetchall()
for row in rows:
strucid = str(row[0])
strucid_list.append(strucid)
bound_conf = str(row[1])
bound_list.append(bound_conf)
modification_date = str(row[2])
mod_date_list.append(modification_date)
unique_bound = list(set(bound_list))
unique_modification_date = list(set(mod_date_list))
unique_strucids = list(set(strucid_list))
for ids in unique_strucids:
db_strucids.append(ids)
c.execute("select strucid from proasis_leads where protein=%s and strucid!=''", (self.protein,))
rows = c.fetchall()
for row in rows:
db_strucids.append(str(row[0]))
# # get info about crystals in proasis_hits (ones identified as in refinement) that haven't made it to
# # proasis
#
# if sum([len(unique_modification_date), len(unique_bound), len(unique_strucids)]) == 0:
#
# c.execute(
# "select bound_conf, modification_date, exists_pdb, exists_mtz, exists_2fofc, exists_fofc, "
# "ligand_list from proasis_hits where crystal_name like %s",
# (crystal,))
# rows = c.fetchall()
#
# for row in rows:
# file_checks['crystal'].append(crystal)
# file_checks['bound_state'].append(str(row[0]))
# file_checks['mod_date'].append(str(row[1]))
# file_checks['pdb'].append(str(row[2]))
# file_checks['mtz'].append(str(row[3]))
# file_checks['2fofc'].append(str(row[4]))
# file_checks['fofc'].append(str(row[5]))
# file_checks['ligs'].append(str(row[6]))
# clear up mismatching entries
in_common = list(set(db_strucids) & set(project_strucids))
for strucid in db_strucids:
if strucid not in in_common:
print(self.protein + ': ' + strucid + ' found in database but not in proasis')
print('removing entry from db...')
c.execute("UPDATE proasis_hits set strucid='' where strucid=%s", (strucid,))
conn.commit()
c.execute("UPDATE proasis_leads set strucid='' where strucid=%s", (strucid,))
conn.commit()
print('\n')
for strucid in project_strucids:
if strucid not in in_common:
print(self.protein + ': ' + strucid + ' found in proasis but not in db')
print('removing entry from proasis...')
paf.delete_structure(strucid)
print('\n')
# good_structures = {'crystal': [], 'bound_state': [], 'mod_date': [], 'strucid': []}
# if len(set([len(unique_modification_date), len(unique_bound), len(unique_strucids)])) == 1:
# status_list.append(0)
# good_list.append(crystal)
# for i in range(0, len(unique_bound)):
# good_structures['crystal'].append(crystal)
# good_structures['bound_state'].append(unique_bound[i])
# good_structures['mod_date'].append(unique_modification_date[i])
# good_structures['strucid'].append(unique_strucids[i])
# for i in range(0, len(good_structures['strucid'])):
# if good_structures['strucid'][i] not in project_strucids:
# print('missing or incorrect strucid in db for '
# + str(good_structures['crystal'][i] + ' (' + str(good_structures['strucid'][i]) + ')'))
# for key in file_checks.keys():
# if '0' in file_checks[key]:
# error_list.append(str('missing ' + str(key) + ' file!'))
#
# if 'None' in file_checks[key]:
# error_list.append(str('None value found for ' + str(key)))
#
# elif len(set([len(unique_modification_date), len(unique_bound), len(unique_strucids)])) > 1:
# status_list.append(1)
#
# error_frame = pd.DataFrame.from_dict(file_checks)
# cols = ['crystal', 'bound_state', 'mod_date', 'ligs', 'mtz', 'pdb', '2fofc', 'fofc']
# error_frame = error_frame[cols]
# error_frame.sort_values(by=['crystal'], inplace=True)
#
# good_frame = pd.DataFrame.from_dict(good_structures)
# cols = ['crystal', 'bound_state', 'mod_date', 'strucid']
# good_frame = good_frame[cols]
# good_frame.sort_values(by=['crystal'], inplace=True)
#
#
# clean up repeats
repeats = find_proasis_repeats(self.protein)
for i, x in enumerate(repeats['crystal']):
bound_list = repeats['bound_confs'][i]
strucids = repeats['strucids'][i]
if len(bound_list)==len(strucids):
if len(list(set(bound_list)))==1:
print(str('identical uploaded structures: ' + str(strucids)) + ' (' + x + ')')
print('removing repeat structures from proasis, and updating database...')
to_delete_strucs=strucids[1:]
to_delete_confs=bound_list[1:]
for j in range(0, len(to_delete_strucs)):
c.execute('DELETE FROM proasis_hits WHERE strucid=%s and bound_conf=%s', (to_delete_strucs[j],
to_delete_confs[j]))
conn.commit()
paf.delete_structure(to_delete_strucs[j])
pd.DataFrame.from_dict(repeats).to_csv('test.csv')
def run(self):
fail_count = 0
# select anything 'in refinement' (3) or above
refinement = Refinement.objects.filter(outcome__gte=3)
# set up info for each entry that matches the filter
for obj in refinement:
# set up blank fields for entries in proasis hits table
bound_conf = ''
files = []
mtz = ''
two_fofc = ''
fofc = ''
mod_date = ''
proasis_hit_entry = ''
entry = ''
confs = []
ligand_list = []
# if there is a pdb file named in the bound_conf field, use it as the upload structure for proasis
if obj.bound_conf:
if os.path.isfile(obj.bound_conf):
bound_conf = obj.bound_conf
# otherwise, use the most recent pdb file (according to soakdb)
elif obj.pdb_latest:
if os.path.isfile(obj.pdb_latest):
# if this is from a refinement folder, find the bound-state pdb file, rather than the ensemble
if 'Refine' in obj.pdb_latest:
search_path = '/'.join(obj.pdb_latest.split('/')[:-1])
files = glob.glob(
str(search_path + '/refine*bound*.pdb'))
if len(files) == 1:
bound_conf = files[0]
else:
# if can't find bound state, just use the latest pdb file
bound_conf = obj.pdb_latest
else:
# no pdb = no proasis upload (same for mtz, two_fofc and fofc)
# TODO: Turn this into a function instead of repeating file check
fail_count += 1
continue
mtz = db_functions.check_file_status('refine.mtz', bound_conf)
two_fofc = db_functions.check_file_status('2fofc.map', bound_conf)
fofc = db_functions.check_file_status('fofc.map', bound_conf)
if not mtz[0] or not two_fofc[0] or not fofc[0]:
fail_count += 1
continue
# if a suitable pdb file is found, then search for ligands
if bound_conf:
try:
pdb_file = open(bound_conf, 'r')
ligand_list = []
for line in pdb_file:
if "LIG" in line:
try:
# ligands identified by 'LIG', with preceeding '.' for alt conf letter
lig_string = re.search(r".LIG.......",
line).group()
# just use lig string instead of separating into list items (to handle altconfs)
# TODO: This has changed from a list of ['LIG','RES','ID'] to string. Check usage
ligand_list.append(lig_string)
except:
continue
# if no ligands are found in the pdb file, no upload to proasis (checked that no strucs. had alternative
# labels in them)
except:
ligand_list = None
if not ligand_list:
continue
# get a unique list of ligands
unique_ligands = list(set(ligand_list))
# remove the first letter (alt conf) from unique ligands
lig_no_conf = [l[1:] for l in unique_ligands]
for l in lig_no_conf:
# check whether there are more than 1 entries for any of the lig strings without alt conf
if lig_no_conf.count(l) > 1:
# this is an alt conf situation - add the alt confs to the conf list
confs.extend([lig for lig in unique_ligands if l in lig])
# get the date the pdb file was modified
mod_date = misc_functions.get_mod_date(bound_conf)
if mod_date:
# if there's already an entry for that structure
if ProasisHits.objects.filter(
refinement=obj,
crystal_name=obj.crystal_name).exists():
# if there are no alternate conformations
if not confs:
# get the relevant entry
entry = ProasisHits.objects.get(
refinement=obj, crystal_name=obj.crystal_name)
# if the pdb file is older than the current one, or it has not been uploaded to proasis
if entry.modification_date < mod_date or not entry.strucid:
# delete structure and remove files to remove from proasis is strucid exists
if entry.strucid:
proasis_api_funcs.delete_structure(
entry.strucid)
entry.strucid = None
entry.save()
if self.hit_directory in entry.pdb_file:
os.remove(entry.pdb_file)
if self.hit_directory in entry.mtz:
os.remove(entry.mtz)
if self.hit_directory in entry.two_fofc:
os.remove(entry.two_fofc)
if self.hit_directory in entry.fofc:
os.remove(entry.fofc)
# otherwise, just update the relevant fields
entry.pdb_file = bound_conf
entry.modification_date = mod_date
entry.mtz = mtz[1]
entry.two_fofc = two_fofc[1]
entry.fofc = fofc[1]
entry.ligand_list = unique_ligands
entry.save()
# if there ARE alternate conformations
else:
# for each conformation
for conf in confs:
# do the same as above, but setting the altconf field too
# TODO: functionalise to add altconfs and not repeat method
entry = ProasisHits.objects.get(
refinement=obj,
crystal_name=obj.crystal_name,
altconf=conf)
if entry.modification_date < mod_date or not entry.strucid:
if entry.strucid:
proasis_api_funcs.delete_structure(
entry.strucid)
entry.strucid = None
entry.save()
if self.hit_directory in entry.pdb_file:
os.remove(entry.pdb_file)
if self.hit_directory in entry.mtz:
os.remove(entry.mtz)
if self.hit_directory in entry.two_fofc:
os.remove(entry.two_fofc)
if self.hit_directory in entry.fofc:
os.remove(entry.fofc)
entry.pdb_file = bound_conf
entry.modification_date = mod_date
entry.mtz = mtz[1]
entry.two_fofc = two_fofc[1]
entry.fofc = fofc[1]
entry.ligand_list = unique_ligands
entry.altconf = conf
entry.save()
# if there's not already an entry for that structure
else:
# if no altconfs
if not confs:
# create entry without an altconf
proasis_hit_entry = ProasisHits.objects.get_or_create(
refinement=obj,
crystal_name=obj.crystal_name,
pdb_file=bound_conf,
modification_date=mod_date,
mtz=mtz[1],
two_fofc=two_fofc[1],
fofc=fofc[1],
ligand_list=unique_ligands)
# if altconfs
if confs:
for conf in confs:
# create an entry for each altconf
# TODO: The pdb file will need to be edited later to pull out other altconfs of the same lig
proasis_hit_entry = ProasisHits.objects.get_or_create(
refinement=obj,
crystal_name=obj.crystal_name,
pdb_file=bound_conf,
modification_date=mod_date,
mtz=mtz[1],
two_fofc=two_fofc[1],
fofc=fofc[1],
ligand_list=unique_ligands,
altconf=conf)
dimple = Dimple.objects.filter(crystal_name=obj.crystal_name)
if dimple.count() == 1:
if dimple[0].reference and dimple[
0].reference.reference_pdb:
if os.path.isfile(dimple[0].reference.reference_pdb):
proasis_lead_entry = ProasisLeads.objects.get_or_create(
reference_pdb=dimple[0].reference)
else:
if ProasisLeads.objects.filter(
reference_pdb=dimple[0].reference).exists(
):
print('removing...')
proasis_lead_entry = ProasisLeads.objects.get(
reference_pdb=dimple[0].reference)
proasis_lead_entry.delete()
print(fail_count)
with self.output().open('w') as f:
f.write('')