def fetch_ligand(self, ligand_id, smiles):
"""
fetch ligands with Ligand model
requires: ligand id, ligand id type, ligand name
requires: source_file name
"""
l = None
try:
if ligand_id in self.ligand_cache:
l = self.ligand_cache[ligand_id]
else:
l = Ligand.objects.filter(properities__web_links__index=ligand_id).first()
if l:
cid = l.properities.web_links.filter(web_resource__slug = 'pubchem').first()
if cid:
cid = cid.index
else:
l = None
else:
l = get_or_make_ligand(smiles, 'SMILES', ligand_id, )
except Exception as msg:
l = None
# print('ligand_id---',l,'\n end')
return l
def fetch_ligand(self, ligand_id, ligand_type, ligand_name, source_file):
"""
fetch ligands with Ligand model
requires: ligand id, ligand id type, ligand name
requires: source_file name
"""
l = None
try:
if ligand_id in self.ligand_cache:
l = self.ligand_cache[ligand_id]
else:
l = get_or_make_ligand(ligand_id, ligand_type, ligand_name)
self.ligand_cache[ligand_id] = l
except Exception as msg:
l = None
# print('ligand_id---',l,'\n end')
return l
def fetch_ligand(self, ligand_id, ligand_type, ligand_name, source_file):
"""
fetch ligands with Ligand model
requires: ligand id, ligand id type, ligand name
requires: source_file name
"""
l = None
if str(ligand_id) in self.ligand_cache:
if ligand_id in self.ligand_cache[str(ligand_id)]:
l = self.ligand_cache[str(ligand_id)][ligand_id]
else:
self.ligand_cache[str(ligand_id)] = {}
if not l:
try:
l = get_or_make_ligand(
ligand_id, ligand_type, str(ligand_name))
except Exception as msg:
l = None
self.ligand_cache[str(ligand_name), ligand_id] = l
self.mylog.exception("Protein fetching error | module: fetch_ligand. Row # is : ",
ligand_name, ligand_type, ligand_id, source_file)
return l
def fetch_ligand(self, ligand_id, ligand_type, ligand_name, source_file):
"""
fetch ligands with Ligand model
requires: ligand id, ligand id type, ligand name
requires: source_file name
"""
l = None
try:
if ligand_id in self.ligand_cache:
l = self.ligand_cache[ligand_id]
else:
l = get_or_make_ligand(ligand_id, ligand_type, ligand_name)
self.ligand_cache[ligand_id] = l
if l == None:
l = self.create_empty_ligand(ligand_name)
except:
web_resource = WebResource.objects.get(slug='pubchem')
try:
l = Ligand.objects.get(properities__web_links__web_resource=web_resource,
properities__web_links__index=ligand_id)
except:
l = self.create_empty_ligand(ligand_name)
# print('null ligand', l)
return l
def fetch_ligand(self, ligand_id, smiles):
"""
fetch ligands with Ligand model
requires: ligand id, ligand id type, ligand name
requires: source_file name
"""
l = None
try:
if ligand_id in self.ligand_cache:
l = self.ligand_cache[ligand_id]
else:
l = Ligand.objects.get(name=ligand_id)
if l:
return l
else:
l = get_or_make_ligand(
smiles,
'SMILES',
ligand_id,
)
except Exception as msg:
l = None
# print('ligand_id---',l,'\n end')
return l
def main_func(self, positions, iteration,count,lock):
#####Create chembl compound link and connect it to the corresponding ligand/cid#####
if iteration==0:
# First load makes sure ligands are there
list_of_chembl_ids = self.chembl_mol_ids
while count.value<len(list_of_chembl_ids):
with lock:
chembl_ligand = list_of_chembl_ids[count.value]
count.value +=1
if count.value % 1000 == 0:
print('{} Status {} out of {}'.format(
datetime.datetime.strftime(datetime.datetime.now(), '%Y-%m-%d %H:%M:%S'), count.value, len(list_of_chembl_ids)))
l = Ligand.objects.filter(properities__web_links__web_resource__slug = 'chembl_ligand', properities__web_links__index=chembl_ligand).first()
if l:
cid = l.properities.web_links.filter(web_resource__slug = 'pubchem').first()
if cid:
cid = cid.index
else:
l = None
# make sure code blow is run
if not l:
# if l already has chembl link, assume all is good.
if chembl_ligand not in self.chembl_cid_dict.keys():
cids, not_found = self.find_cid_for_chembl(chembl_ligand)
if not_found:
print('SKIPPED: Could not determine CID',chembl_ligand,cids)
continue
else:
cids = self.chembl_cid_dict[chembl_ligand]
temp = str(cids).split(';') #perhaps we should load all of the CIDs
cid = str(temp[0])
l = get_or_make_ligand(cid,'PubChem CID') #call the first cid if there are more than one
if not l:
print('SKIPPED: Ligand not found in PubChem', cid)
continue
if not l.properities.web_links.filter(web_resource__slug = 'pubchem',index = cid).exists():
# NO CID FOR LIGAND! Rare cases where SMILES was used for initial look up
wl, created = WebLink.objects.get_or_create(index=cid, web_resource=self.wr_pubchem)
l.properities.web_links.add(wl)
if not l.properities.web_links.filter(web_resource__slug = 'chembl_ligand',index = chembl_ligand).exists():
wl, created = WebLink.objects.get_or_create(index=chembl_ligand, web_resource=self.wr)
l.properities.web_links.add(wl)
###### Vendor stuff ######
if not len(l.properities.vendors.all()):
# If it has some, assume they are all loaded
cache_dir = ['pubchem', 'cid', 'vendors']
url = 'https://pubchem.ncbi.nlm.nih.gov/rest/pug_view/categories/compound/$index/JSON/'
vendors = fetch_from_web_api(url, cid, cache_dir)
if vendors:
for vendor_data in vendors['SourceCategories']['Categories'][0]['Sources'] :
lv, created = LigandVendors.objects.get_or_create(slug = slugify(vendor_data['SourceName']))
lv.name = vendor_data['SourceName']
if 'SourceURL' in vendor_data:
lv.url = vendor_data['SourceURL']
lv.save()
if 'SID' in vendor_data:
#print (vendor_data['SID'])
lvls = LigandVendorLink.objects.filter(sid = vendor_data['SID'] )
if not lvls.exists():
lvl = LigandVendorLink()
lvl.vendor = lv
lvl.lp = l.properities
lvl.sid = vendor_data['SID']
if 'RegistryID' in vendor_data:
lvl.vendor_external_id = vendor_data['RegistryID']
if 'SourceRecordURL' in vendor_data:
lvl.url = vendor_data['SourceRecordURL']
else:
continue
lvl.save()
elif iteration==1:
# Third load loads the exp (based on ligand/assay)
header = self.header_dict
skipped = 0
non_p = []
wr_chembl_assays = WebResource.objects.get(slug='chembl_assays')
while count.value<len(self.data):
with lock:
record = self.data[count.value]
count.value +=1
if count.value % 10000 == 0:
print('{} Status {} out of {}'.format(
datetime.datetime.strftime(datetime.datetime.now(), '%Y-%m-%d %H:%M:%S'), count.value, len(self.data)))
target = record[header['target_chembl_id']]
assay_id = record[header['assay_chembl_id']]
assay, created = ChemblAssay.objects.get_or_create(assay_id=assay_id)
if created:
wl, created = WebLink.objects.get_or_create(index=assay_id, web_resource=wr_chembl_assays)
assay.web_links.add(wl)
ligand =record[header['molecule_chembl_id']]
p = Protein.objects.filter(web_links__index = target, web_links__web_resource__slug = 'chembl').first()
if not p:
if not target in non_p:
non_p.append(target)
print('Not found protein!',target)
continue
ls = Ligand.objects.filter(properities__web_links__index=ligand, properities__web_links__web_resource__slug = 'chembl_ligand', canonical=True)
if not ls.exists():
# if no ligand matches this, then ignore -- be sure this works later.
skipped += 1
continue
for l in ls:
if len(ls)>1:
print('issue with canonical! give to munk',l,l.pk,ligand)
break
assay_experiments = AssayExperiment.objects.filter( protein=p, ligand=l, assay=assay)
if assay_experiments.exists():
assay_experiment = assay_experiments.get()
else:
assay_experiment = AssayExperiment()
assay_experiment.assay = assay
assay_experiment.ligand = l
assay_experiment.protein = p
assay_experiment.assay_type = record[header['assay_type']]
assay_experiment.pchembl_value = record[header['pchembl_value']]
assay_experiment.assay_description = record[header['assay_description']]
assay_experiment.published_value = record[header['published_value']]
assay_experiment.published_relation = record[header['published_relation']]
assay_experiment.published_type = record[header['published_type']]
assay_experiment.published_units = record[header['published_units']]
assay_experiment.standard_value = record[header['standard_value']]
assay_experiment.standard_relation = record[header['standard_relation']]
assay_experiment.standard_type = record[header['standard_type']]
assay_experiment.standard_units = record[header['standard_units']]
try:
assay_experiment.save()
except IntegrityError:
assay_experiment = AssayExperiment.objects.get( protein=p, ligand=l, assay=assay)
print('done, skipped:',skipped)
def main_func(self, positions, iteration,count,lock):
missing_proteins = {}
c = 0
skipped = 0
rows = self.data_all
while count.value<len(rows):
with lock:
r = rows[count.value]
count.value +=1
current = time.time()
c += 1
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
float(r['review'])
r['review'] = str(int(r['review']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
if r['reference'] not in self.publication_cache:
try:
wl = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
try:
wl = WebLink.objects.create(index=r['reference'],
web_resource = WebResource.objects.get(slug=pub_type))
except IntegrityError:
wl = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
try:
pub = Publication.objects.get(web_link=wl)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = wl
pub.save()
except IntegrityError:
pub = Publication.objects.get(web_link=wl)
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' + str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['reference']] = pub
else:
pub = self.publication_cache[r['reference']]
l = None
if str(r['ligand_name']) in self.ligand_cache:
if r['ligand_id'] in self.ligand_cache[str(r['ligand_name'])]:
l = self.ligand_cache[str(r['ligand_name'])][r['ligand_id']]
else:
self.ligand_cache[str(r['ligand_name'])] = {}
if not l:
try:
l = get_or_make_ligand(r['ligand_id'],r['ligand_type'],str(r['ligand_name']))
except Exception as msg:
print('Something errored with ligand, aborting entry of mutation',r['ligand_name'],r['ligand_type'],r['ligand_id'],r['source_file'])
print(msg)
traceback.print_exc()
continue
self.ligand_cache[str(r['ligand_name'])][r['ligand_id']] = l
protein=Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein=protein.get()
elif r['protein'] not in missing_proteins:
# Can contain code to try to figure out what protein it is.
pass
else:
missing_proteins[r['protein']] += 1
continue
res=Residue.objects.filter(protein_conformation__protein=protein,amino_acid=r['mutation_from'],sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res=res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'])
# print('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'],r['source_file'])
skipped += 1
continue
self.logger.info('Parsed '+str(c)+' bias data entries. Skipped '+str(skipped))
sorted_missing_proteins = sorted(missing_proteins.items(), key=operator.itemgetter(1),reverse=True)
print(missing_proteins)
def main_func(self, positions, iteration):
# filenames
if not positions[1]:
rows = self.data[positions[0]:]
else:
rows = self.data[positions[0]:positions[1]]
missing_proteins = {}
mutants_for_proteins = {}
c = 0
skipped = 0
inserted = 0
bulk_m = []
bulk_r = []
current_sheet = time.time()
for r in rows:
# print(source_file,c)
# PRINT IF ERRORS OCCUR
# self.logger.info('File '+str(r['source_file'])+' number '+str(c))
current = time.time()
c += 1
# if c%100==0:
# self.logger.info('Parsed '+str(c)+' mutant data entries')
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
float(r['review'])
r['review'] = str(int(r['review']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
if r['reference'] not in self.publication_cache:
try:
pub = Publication.objects.get(
web_link__index=r['reference'],
web_link__web_resource__slug=pub_type)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = WebLink.objects.get(
index=r['reference'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
wl = WebLink.objects.create(
index=r['reference'],
web_resource=WebResource.objects.get(
slug=pub_type))
pub.web_link = wl
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' +
str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['reference']] = pub
else:
pub = self.publication_cache[r['reference']]
# print(r['review'],r['reference'])
if r['review'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
# print(r['review'],pub_type)
if r['review']:
if r['review'] not in self.publication_cache:
try:
pub_review = Publication.objects.get(
web_link__index=r['review'],
web_link__web_resource__slug=pub_type)
except Publication.DoesNotExist:
pub_review = Publication()
try:
pub_review.web_link = WebLink.objects.get(
index=r['review'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
wl = WebLink.objects.create(
index=r['review'],
web_resource=WebResource.objects.get(
slug=pub_type))
pub_review.web_link = wl
if pub_type == 'doi':
pub_review.update_from_doi(doi=r['review'])
elif pub_type == 'pubmed':
pub_review.update_from_pubmed_data(
index=r['review'])
try:
pub_review.save()
except:
self.logger.error('error with review ' +
str(r['review']) + ' ' +
pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['review']] = pub_review
else:
pub_review = self.publication_cache[r['review']]
else:
pub_review = None
l = None
if str(r['ligand_name']) in self.ligand_cache:
if r['ligand_id'] in self.ligand_cache[str(r['ligand_name'])]:
l = self.ligand_cache[str(
r['ligand_name'])][r['ligand_id']]
else:
self.ligand_cache[str(r['ligand_name'])] = {}
if not l:
l = get_or_make_ligand(r['ligand_id'], r['ligand_type'],
str(r['ligand_name']))
self.ligand_cache[str(r['ligand_name'])][r['ligand_id']] = l
l_ref = None
if str(r['exp_mu_ligand_ref']) in self.ref_ligand_cache:
l_ref = self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])]
else:
if Ligand.objects.filter(
name=r['exp_mu_ligand_ref'], canonical=True
).exists(
): #if this name is canonical and it has a ligand record already
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=True)
elif Ligand.objects.filter(
name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=False
).exists(
): #if this matches an alias that only has "one" parent canonical name - eg distinct
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=False)
elif Ligand.objects.filter(
name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=True
).exists(
): #if this matches an alias that only has several canonical parents, must investigate, start with empty.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = False
l_ref.ambigious_alias = True
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'],
canonical=False).exists(
): #amigious_alias not specified
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False)
l_ref.ambigious_alias = False
l_ref.save()
elif r['exp_mu_ligand_ref']: #if neither a canonical or alias exists, create the records. Remember to check for canonical / alias status.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = True
l_ref.ambigious_alias = False
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
try:
l_ref.save()
except IntegrityError:
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False)
print("error failing ligand, duplicate?")
# logger.error("FAILED SAVING LIGAND, duplicate?")
else:
l_ref = None
self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])] = l_ref
protein_id = 0
residue_id = 0
protein = Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein = protein.get()
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
else:
skipped += 1
if r['protein'] in missing_proteins:
missing_proteins[r['protein']] += 1
else:
missing_proteins[r['protein']] = 1
self.logger.error('Skipped due to no protein ' +
r['protein'])
continue
res = Residue.objects.filter(
protein_conformation__protein=protein,
amino_acid=r['mutation_from'],
sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res = res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' +
r['protein'] + ' pos:' +
str(r['mutation_pos']) + ' AA:' +
r['mutation_from'])
skipped += 1
continue
if r['ligand_class']:
l_role, created = LigandRole.objects.get_or_create(
name=r['ligand_class'],
defaults={'slug': slugify(r['ligand_class'])[:50]
}) # FIXME this should not be needed
else:
l_role = None
if r['exp_type']:
exp_type_id, created = MutationExperimentalType.objects.get_or_create(
type=r['exp_type'])
else:
exp_type_id = None
if r['exp_func']:
exp_func_id, created = MutationFunc.objects.get_or_create(
func=r['exp_func'])
else:
exp_func_id = None
if r['exp_mu_effect_ligand_prop'] or r['exp_mu_effect_qual']:
exp_qual_id, created = MutationQual.objects.get_or_create(
qual=r['exp_mu_effect_qual'],
prop=r['exp_mu_effect_ligand_prop'])
else:
exp_qual_id = None
if r['opt_type'] or r['opt_wt'] or r['opt_mu'] or r[
'opt_sign'] or r['opt_percentage'] or r['opt_qual'] or r[
'opt_agonist']:
exp_opt_id, created = MutationOptional.objects.get_or_create(
type=r['opt_type'],
wt=r['opt_wt'],
mu=r['opt_mu'],
sign=r['opt_sign'],
percentage=r['opt_percentage'],
qual=r['opt_qual'],
agonist=r['opt_agonist'])
else:
exp_opt_id = None
try:
mutation, created = Mutation.objects.get_or_create(
amino_acid=r['mutation_to'], protein=protein, residue=res)
except IntegrityError:
mutation = Mutation.objects.get(amino_acid=r['mutation_to'],
protein=protein,
residue=res)
logtypes = ['pEC50', 'pIC50', 'pK']
foldchange = 0
typefold = ''
if r['exp_wt_value'] != 0 and r[
'exp_mu_value_raw'] != 0: #fix for new format
if re.match("(" + ")|(".join(logtypes) + ")",
r['exp_type']): #-log values!
foldchange = round(
math.pow(10, -r['exp_mu_value_raw']) /
pow(10, -r['exp_wt_value']), 3)
typefold = r['exp_type'] + "_log"
else:
foldchange = round(
r['exp_mu_value_raw'] / r['exp_wt_value'], 3)
typefold = r['exp_type'] + "_not_log"
if foldchange < 1 and foldchange != 0:
foldchange = -round((1 / foldchange), 3)
elif r['fold_effect'] != 0:
foldchange = round(r['fold_effect'], 3)
if foldchange < 1: foldchange = -round((1 / foldchange), 3)
raw_experiment = self.insert_raw(r)
bulk = MutationExperiment(
refs=pub,
review=pub_review,
protein=protein,
residue=res,
ligand=l,
ligand_role=l_role,
ligand_ref=l_ref,
#raw = raw_experiment, #raw_experiment, OR None
optional=exp_opt_id,
exp_type=exp_type_id,
exp_func=exp_func_id,
exp_qual=exp_qual_id,
mutation=mutation,
wt_value=r['exp_wt_value'], #
wt_unit=r['exp_wt_unit'],
mu_value=r['exp_mu_value_raw'],
mu_sign=r['exp_mu_effect_sign'],
foldchange=foldchange)
# mut_id = obj.id
bulk_r.append(raw_experiment)
bulk_m.append(bulk)
inserted += 1
end = time.time()
diff = round(end - current, 2)
#print(diff)
self.logger.info('Parsed ' + str(c) +
' mutant data entries. Skipped ' + str(skipped))
current = time.time()
raws = MutationRaw.objects.bulk_create(bulk_r)
for i, me in enumerate(bulk_m):
me.raw = raws[i]
MutationExperiment.objects.bulk_create(bulk_m)
end = time.time()
diff = round(end - current, 2)
current_sheet
diff_2 = round(end - current_sheet, 2)
print("overall", diff_2, "bulk", diff, len(bulk_m), "skipped",
str(skipped))
def main_func(self, positions, iteration, count, lock):
# filenames
# if not positions[1]:
# rows = self.data[positions[0]:]
# else:
# rows = self.data[positions[0]:positions[1]]
missing_proteins = {}
mutants_for_proteins = {}
wrong_uniport_ids = {}
c = 0
skipped = 0
inserted = 0
bulk_m = []
bulk_r = []
current_sheet = time.time()
rows = self.data_all
while count.value < len(rows):
with lock:
r = rows[count.value]
count.value += 1
# for r in rows:
# print(r['source_file'],c)
# PRINT IF ERRORS OCCUR
#self.logger.info('File '+str(r['source_file'])+' number '+str(c))
current = time.time()
c += 1
# if c%100==0:
# self.logger.info('Parsed '+str(c)+' mutant data entries')
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
float(r['review'])
r['review'] = str(int(r['review']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
if r['reference'] not in self.publication_cache:
try:
wl = WebLink.objects.get(index=r['reference'],
web_resource__slug=pub_type)
except WebLink.DoesNotExist:
try:
wl = WebLink.objects.create(
index=r['reference'],
web_resource=WebResource.objects.get(
slug=pub_type))
except IntegrityError:
wl = WebLink.objects.get(index=r['reference'],
web_resource__slug=pub_type)
try:
pub = Publication.objects.get(web_link=wl)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = wl
pub.save()
except IntegrityError:
pub = Publication.objects.get(web_link=wl)
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' +
str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['reference']] = pub
else:
pub = self.publication_cache[r['reference']]
# print(r['review'],r['reference'])
if r['review'].isdigit(): #assume pubmed
pub_type = 'pubmed'
elif r['review'].startswith('http'):
pub_type = 'raw_link'
else: #assume doi
pub_type = 'doi'
# print(r['review'],pub_type)
if r['review']:
if r['review'] not in self.publication_cache:
try:
wl = WebLink.objects.get(index=r['review'],
web_resource__slug=pub_type)
except WebLink.DoesNotExist:
try:
wl = WebLink.objects.create(
index=r['review'],
web_resource=WebResource.objects.get(
slug=pub_type))
except IntegrityError:
wl = WebLink.objects.get(
index=r['review'], web_resource__slug=pub_type)
try:
pub_review = Publication.objects.get(web_link=wl)
except Publication.DoesNotExist:
pub_review = Publication()
try:
pub_review.web_link = wl
pub_review.save()
except IntegrityError:
pub_review = Publication.objects.get(web_link=wl)
if pub_type == 'doi':
pub_review.update_from_doi(doi=r['review'])
elif pub_type == 'pubmed':
pub_review.update_from_pubmed_data(
index=r['review'])
try:
pub_review.save()
except:
self.logger.error('error with review ' +
str(r['review']) + ' ' +
pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['review']] = pub_review
else:
pub_review = self.publication_cache[r['review']]
else:
pub_review = None
l = None
if str(r['ligand_name']) in self.ligand_cache:
if r['ligand_id'] in self.ligand_cache[str(r['ligand_name'])]:
l = self.ligand_cache[str(
r['ligand_name'])][r['ligand_id']]
else:
self.ligand_cache[str(r['ligand_name'])] = {}
if not l:
try:
l = get_or_make_ligand(r['ligand_id'], r['ligand_type'],
str(r['ligand_name']))
except Exception as msg:
print(
'Something errored with ligand, aborting entry of mutation',
r['ligand_name'], r['ligand_type'], r['ligand_id'],
r['source_file'])
print(msg)
traceback.print_exc()
continue
self.ligand_cache[str(r['ligand_name'])][r['ligand_id']] = l
l_ref = None
if str(r['exp_mu_ligand_ref']) in self.ref_ligand_cache:
l_ref = self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])]
else:
if Ligand.objects.filter(
name=r['exp_mu_ligand_ref'], canonical=True
).exists(
): #if this name is canonical and it has a ligand record already
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=True)
elif Ligand.objects.filter(
name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=False
).exists(
): #if this matches an alias that only has "one" parent canonical name - eg distinct
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=False)
elif Ligand.objects.filter(
name=r['exp_mu_ligand_ref'],
canonical=False,
ambigious_alias=True
).exists(
): #if this matches an alias that only has several canonical parents, must investigate, start with empty.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = False
l_ref.ambigious_alias = True
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'],
canonical=False).exists(
): #amigious_alias not specified
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False)
l_ref.ambigious_alias = False
l_ref.save()
elif r['exp_mu_ligand_ref']: #if neither a canonical or alias exists, create the records. Remember to check for canonical / alias status.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = True
l_ref.ambigious_alias = False
try:
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
except IntegrityError:
if Ligand.objects.filter(name=r['exp_mu_ligand_ref'],
canonical=True).exists():
l_ref = Ligand.objects.get(
name=r['exp_mu_ligand_ref'], canonical=True)
else:
l_ref = Ligand.objects.get(
name=r['exp_mu_ligand_ref'], canonical=False)
# print("error failing ligand, duplicate?")
try:
l_ref.save()
except IntegrityError:
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'],
canonical=False)
# print("error failing ligand, duplicate?")
# logger.error("FAILED SAVING LIGAND, duplicate?")
else:
l_ref = None
self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])] = l_ref
protein_id = 0
residue_id = 0
protein = Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein = protein.get()
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
elif r['protein'] not in missing_proteins:
try:
r['protein'] = wrong_uniport_ids[r['protein']]
real_uniprot = wrong_uniport_ids[r['protein']]
protein = Protein.objects.get(entry_name=r['protein'])
# print('fetched with lookup table',r['protein'])
except:
# look for it as uniprot
protein = Protein.objects.filter(
web_links__web_resource__slug='uniprot',
web_links__index=r['protein'].upper())
if protein.exists():
protein = protein.get()
real_uniprot = protein.entry_name
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
else:
# Try to lookup in uniprot to catch typing errors / variants in entry_name
url = 'http://www.uniprot.org/uniprot/$index.xml'
cache_dir = ['uniprot', 'id']
uniprot_protein = fetch_from_web_api(url,
r['protein'],
cache_dir,
xml=True)
try:
real_uniprot = uniprot_protein.find(
'.//{http://uniprot.org/uniprot}name'
).text.lower()
protein = Protein.objects.get(
entry_name=real_uniprot)
except:
skipped += 1
if r['protein'] in missing_proteins:
missing_proteins[r['protein']] += 1
else:
missing_proteins[r['protein']] = 1
# print('Skipped due to no protein '+ r['protein'])
self.logger.error(
'Skipped due to no protein ' +
r['protein'])
continue
wrong_uniport_ids[r['protein']] = protein.entry_name
r['protein'] = real_uniprot
else:
missing_proteins[r['protein']] += 1
continue
res = Residue.objects.filter(
protein_conformation__protein=protein,
amino_acid=r['mutation_from'],
sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res = res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' +
r['protein'] + ' pos:' +
str(r['mutation_pos']) + ' AA:' +
r['mutation_from'])
# print('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'],r['source_file'])
skipped += 1
continue
if r['ligand_class']:
try:
l_role, created = LigandRole.objects.get_or_create(
name=r['ligand_class'],
defaults={'slug': slugify(r['ligand_class'])[:50]
}) # FIXME this should not be needed
except Exception as e:
if LigandRole.objects.filter(
slug=slugify(r['ligand_class'])[:50]).exists():
l_role = LigandRole.objects.get(
slug=slugify(r['ligand_class'])[:50])
if l_role.name == slugify(r['ligand_class'])[:50]:
#if name of role is same as slug, then it was created by constructs script, replace it
l_role.name = r['ligand_class']
l_role.save()
else:
print(e)
print("Error with", r['ligand_class'],
slugify(r['ligand_class'])[:50])
l_role, created = LigandRole.objects.get_or_create(
slug=slugify(r['ligand_class'])
[:50]) # FIXME this should not be needed
else:
l_role = None
if r['exp_type']:
exp_type_id, created = MutationExperimentalType.objects.get_or_create(
type=r['exp_type'])
else:
exp_type_id = None
if r['exp_func']:
exp_func_id, created = MutationFunc.objects.get_or_create(
func=r['exp_func'])
else:
exp_func_id = None
if r['exp_mu_effect_ligand_prop'] or r['exp_mu_effect_qual']:
exp_qual_id, created = MutationQual.objects.get_or_create(
qual=r['exp_mu_effect_qual'],
prop=r['exp_mu_effect_ligand_prop'])
else:
exp_qual_id = None
# if r['opt_type'] or r['opt_wt'] or r['opt_mu'] or r['opt_sign'] or r['opt_percentage'] or r['opt_qual'] or r['opt_agonist']:
# exp_opt_id, created = MutationOptional.objects.get_or_create(type=r['opt_type'], wt=r['opt_wt'], mu=r['opt_mu'], sign=r['opt_sign'], percentage=r['opt_percentage'], qual=r['opt_qual'], agonist=r['opt_agonist'])
# else:
# exp_opt_id = None
try:
mutation, created = Mutation.objects.get_or_create(
amino_acid=r['mutation_to'], protein=protein, residue=res)
except IntegrityError:
mutation = Mutation.objects.get(amino_acid=r['mutation_to'],
protein=protein,
residue=res)
logtypes = ['pEC50', 'pIC50', 'pK']
foldchange = 0
typefold = ''
if r['exp_wt_value'] != 0 and r[
'exp_mu_value_raw'] != 0: #fix for new format
if re.match("(" + ")|(".join(logtypes) + ")",
r['exp_type']): #-log values!
try:
foldchange = round(
math.pow(10, -r['exp_mu_value_raw']) /
pow(10, -r['exp_wt_value']), 3)
except:
print(r)
typefold = r['exp_type'] + "_log"
elif "%" == r['exp_wt_unit']:
# if % then it's a difference case, then lower value is bad. Otherwise it's conc and lower is better
foldchange = round(
r['exp_wt_value'] / r['exp_mu_value_raw'], 3)
else:
foldchange = round(
r['exp_mu_value_raw'] / r['exp_wt_value'], 3)
typefold = r['exp_type'] + "_not_log"
if foldchange > 0 and foldchange < 1 and foldchange != 0:
foldchange = -round((1 / foldchange), 3)
elif r['fold_effect'] != 0:
foldchange = round(r['fold_effect'], 3)
if foldchange < 1: foldchange = -round((1 / foldchange), 3)
r['fold_effect'] = foldchange
raw_experiment = self.insert_raw(r)
# raw_experiment.save()
bulk = MutationExperiment(
refs=pub,
review=pub_review,
submitting_group=r['submitting_group'],
data_container=r['data_container'],
data_container_number=r['data_container_number'],
protein=protein,
residue=res,
ligand=l,
ligand_role=l_role,
ligand_ref=l_ref,
# raw = raw_experiment, #raw_experiment, OR None
# optional = exp_opt_id,
exp_type=exp_type_id,
exp_func=exp_func_id,
exp_qual=exp_qual_id,
mutation=mutation,
wt_value=r['exp_wt_value'], #
wt_unit=r['exp_wt_unit'],
mu_value=r['exp_mu_value_raw'],
mu_sign=r['exp_mu_effect_sign'],
foldchange=foldchange,
opt_receptor_expression=r['opt_receptor_expression'],
opt_basal_activity=r['opt_basal_activity'],
opt_gain_of_activity=r['opt_gain_of_activity'],
opt_ligand_emax=r['opt_ligand_emax'],
opt_agonist=r['opt_agonist'],
)
# for line,val in r.items():
# val = str(val)
# if len(val)>100:
# print(line,"too long",val)
# mut_id = obj.id
bulk_r.append(raw_experiment)
bulk_m.append(bulk)
# try:
# bulk.save()
# except Exception as e:
# print(e)
# print(r)
# break
#print('saved ',r['source_file'])
inserted += 1
end = time.time()
diff = round(end - current, 2)
#print(diff)
self.logger.info('Parsed ' + str(c) +
' mutant data entries. Skipped ' + str(skipped))
current = time.time()
raws = MutationRaw.objects.bulk_create(bulk_r)
for i, me in enumerate(bulk_m):
me.raw = raws[i]
MutationExperiment.objects.bulk_create(bulk_m)
end = time.time()
diff = round(end - current, 2)
# current_sheet
diff_2 = round(end - current_sheet, 2)
print("overall", diff_2, "bulk", diff, len(bulk_m), "skipped",
str(skipped))
sorted_missing_proteins = sorted(missing_proteins.items(),
key=operator.itemgetter(1),
reverse=True)
def add_construct(d):
#delete if already name there
Construct.objects.filter(name = d['construct_crystal']['pdb_name']).delete()
protein = Protein.objects.filter(entry_name=d['construct_crystal']['uniprot']).get()
structure = Structure.objects.filter(pdb_code__index=d['construct_crystal']['pdb'].upper()).get()
protein_conformation = structure.protein_conformation
construct = Construct()
construct.protein = protein
construct.name = d['construct_crystal']['pdb_name']
construct.json = json.dumps(d, indent=4, separators=(',', ': '))
construct.structure = structure
#CrystalInfo
crystal = CrystalInfo()
crystal.resolution = structure.resolution
crystal.pdb_data = structure.pdb_data
crystal.pdb_code = structure.pdb_code.index
crystal.save()
construct.crystal = crystal
#Contact INFO
if 'contact_info' in d:
construct.contributor, created = ContributorInfo.objects.get_or_create(name = d['contact_info']['name_cont'],
pi_email = d['contact_info']['pi_email'],
pi_name = d['contact_info']['pi_name'],
urls = d['contact_info']['url'],
date = datetime.datetime.strptime(d['contact_info']['date'], '%m/%d/%Y').strftime('%Y-%m-%d'),
address = d['contact_info']['address'])
construct.save()
#MUTATIONS
for mutation in d['mutations']:
if 'type' not in mutation:
mutation['type'] = ''
if 'remark' not in mutation:
mutation['remark'] = ''
mut = ConstructMutation.objects.create(sequence_number=mutation['pos'],wild_type_amino_acid=mutation['wt'],mutated_amino_acid=mutation['mut'],mutation_type=mutation['type'],remark=mutation['remark'])
construct.mutations.add(mut)
#DELETIONS
insert_deletions = {}
for deletion in d['deletions']:
if 'start' in deletion:
dele = ConstructDeletion.objects.create(start=deletion['start'],end=deletion['end'])
else:
dele = ConstructDeletion.objects.create(start=deletion['pos'],end=deletion['pos'])
construct.deletions.add(dele)
if deletion['origin']!='user':
id = deletion['origin'].split('_')[1]
insert_deletions[id] = deletion
#INSERTIONS (AUX)
for name,aux in d['auxiliary'].items():
id = name.replace('aux','')
aux_type,created = ConstructInsertionType.objects.get_or_create(name=aux['type'],subtype=aux['subtype'])
insert = ConstructInsertion.objects.create(insert_type=aux_type,presence=aux['presence'],position=aux['position']+"_"+id)
if insert.presence == 'YES' and insert.position.startswith('Within Receptor'):
#need to fetch range
if 'start' in aux:
insert.start = aux['start']
insert.end = aux['start']
else:
insert.start = insert_deletions[id]['start']
insert.end = insert_deletions[id]['end']
insert.save()
construct.insertions.add(insert)
#MODIFICATIONS
for modification in d['modifications']:
mod = ConstructModification.objects.create(modification=modification['type'],position_type=modification['position'][0],
pos_start=modification['position'][1][0],
pos_end=modification['position'][1][1],remark=modification['remark'] )
construct.modifications.add(mod)
#EXPRESSION
if 'expression' in d:
if 'expr_method' in d['expression']:
if 'expr_remark' not in d['expression']:
d['expression']['expr_remark'] = ''
construct.expression,created = ExpressionSystem.objects.get_or_create(expression_method=d['expression']['expr_method'],
host_cell_type=d['expression']['host_cell_type'],
host_cell=d['expression']['host_cell'],
remarks=d['expression']['expr_remark'])
#solubilization
if 'solubilization' in d:
if 'deterg_type' in d['solubilization']:
c_list = ChemicalList()
list_name,created = ChemicalListName.objects.get_or_create(name='Solubilization')
c_list.name = list_name
c_list.save()
ct, created = ChemicalType.objects.get_or_create(name='detergent')
chem, created = Chemical.objects.get_or_create(name=d['solubilization']['deterg_type'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=d['solubilization']['deterg_concentr'], concentration_unit=d['solubilization']['deterg_concentr_unit'], chemical=chem)
c_list.chemicals.add(cc)
ct, created = ChemicalType.objects.get_or_create(name='additive')
chem, created = Chemical.objects.get_or_create(name=d['solubilization']['solub_additive'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=d['solubilization']['additive_concentr'], concentration_unit=d['solubilization']['addit_concentr_unit'], chemical=chem)
c_list.chemicals.add(cc)
solubilization = Solubilization.objects.create(chemical_list = c_list)
construct.solubilization = solubilization
construct.save()
#Purification
purification = Purification.objects.create()
for puri,step in d['solubilization'].items():
if not puri.startswith(('chem_enz_treatment','sol_remark')):
continue
else:
s,created = PurificationStep.objects.get_or_create(name=step)
purification.steps.add(s)
construct.purification = purification
construct.save()
#CRYSTALLIZATION
if 'crystallization' in d:
if 'crystal_type' in d['crystallization']:
c = Crystallization()
sub_name = "" if 'lcp_lipid' not in d['crystallization'] else d['crystallization']['lcp_lipid']
c_type, created = CrystallizationTypes.objects.get_or_create(name=d['crystallization']['crystal_type'], sub_name=sub_name)
c_method, created = CrystallizationMethods.objects.get_or_create(name=d['crystallization']['crystal_method'])
c.crystal_type = c_type
c.crystal_method = c_method
if 'crystal_remark' in d['crystallization']:
c.remarks = d['crystallization']['crystal_remark']
c.temp = d['crystallization']['temperature']
if d['crystallization']['ph']=='single_ph':
c.ph_start = d['crystallization']['ph_single']
c.ph_end = d['crystallization']['ph_single']
else:
c.ph_start = d['crystallization']['ph_range_one']
c.ph_end = d['crystallization']['ph_range_two']
c.protein_conc = d['crystallization']['protein_concentr']
c.protein_conc_unit = d['crystallization']['protein_conc_unit']
c.save()
#MAKE LISTS
c_list = ChemicalList()
list_name,created = ChemicalListName.objects.get_or_create(name='crystallization_chemical_components')
c_list.name = list_name
c_list.save()
for chemical in d['crystallization']['chemical_components']:
ct, created = ChemicalType.objects.get_or_create(name='crystallization_chemical_components')
chem, created = Chemical.objects.get_or_create(name=chemical['component'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=chemical['value'], concentration_unit=chemical['unit'], chemical=chem)
c_list.chemicals.add(cc)
c.chemical_lists.add(c_list)
if d['crystallization']['crystal_type']=='lipidic cubic phase': #make list of LCP stuff
c_list = ChemicalList()
# c_list.name = d['crystallization']['lcp_lipid']
list_name,created = ChemicalListName.objects.get_or_create(name='LCP')
c_list.name = list_name
c_list.save()
ct, created = ChemicalType.objects.get_or_create(name='LCP Lipid additive')
chem, created = Chemical.objects.get_or_create(name=d['crystallization']['lcp_add'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=d['crystallization']['lcp_conc'], concentration_unit=d['crystallization']['lcp_conc_unit'], chemical=chem)
c_list.chemicals.add(cc)
c.chemical_lists.add(c_list)
#DETERGENT
if 'detergent' in d['crystallization']:
c_list = ChemicalList()
list_name,created = ChemicalListName.objects.get_or_create(name='Detergent')
c_list.name = list_name
c_list.save()
ct, created = ChemicalType.objects.get_or_create(name='detergent')
chem, created = Chemical.objects.get_or_create(name=d['crystallization']['detergent'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=d['crystallization']['deterg_conc'], concentration_unit=d['crystallization']['deterg_conc_unit'], chemical=chem)
c_list.chemicals.add(cc)
c.chemical_lists.add(c_list)
#LIPID
if 'lipid' in d['crystallization']:
c_list = ChemicalList()
list_name,created = ChemicalListName.objects.get_or_create(name='Lipid')
c_list.name = list_name
c_list.save()
ct, created = ChemicalType.objects.get_or_create(name='lipid')
chem, created = Chemical.objects.get_or_create(name=d['crystallization']['lipid'], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=d['crystallization']['lipid_concentr'], concentration_unit=d['crystallization']['lipid_concentr_unit'], chemical=chem)
c_list.chemicals.add(cc)
c.chemical_lists.add(c_list)
#Use ligand function to get ligand if it exists or otherwise create. Lots of checks for inchi/smiles/name
ligand = get_or_make_ligand(d['construct_crystal']['ligand_id'],d['construct_crystal']['ligand_id_type'],d['construct_crystal']['ligand_name'])
if 'ligand_activity' not in d['construct_crystal']:
d['construct_crystal']['ligand_activity'] = 'unknown'
if ligand and 'ligand_activity' in d['construct_crystal']:
role_slug = slugify(d['construct_crystal']['ligand_activity'])
try:
lr, created = LigandRole.objects.get_or_create(slug=role_slug,
defaults={'name': d['construct_crystal']['ligand_activity']})
except IntegrityError:
lr = LigandRole.objects.get(slug=role_slug)
if ligand:
ligand_c = CrystallizationLigandConc()
ligand_c.construct_crystallization = c
ligand_c.ligand = ligand
if lr:
ligand_c.ligand_role = lr
if 'ligand_conc' in d['construct_crystal']:
ligand_c.ligand_conc = d['construct_crystal']['ligand_conc']
if 'ligand_conc_unit' in d['construct_crystal']:
ligand_c.ligand_conc_unit = d['construct_crystal']['ligand_conc_unit']
ligand_c.save()
c.ligands.add(ligand_c)
construct.crystallization = c
construct.save()
def create_mutant_data(self, filenames):
self.logger.info('CREATING MUTANT DATA')
# what files should be parsed?
if not filenames:
filenames = os.listdir(self.structure_data_dir)
missing_proteins = {}
mutants_for_proteins = {}
for source_file in filenames:
source_file_path = os.sep.join([self.structure_data_dir, source_file])
if os.path.isfile(source_file_path) and source_file[0] != '.':
self.logger.info('Reading file {}'.format(source_file_path))
# read the yaml file
if source_file[-4:]=='xlsx' or source_file[-3:]=='xls':
rows = self.loaddatafromexcel(source_file_path)
rows = self.analyse_rows(rows)
elif source_file[-4:]=='yaml':
rows = yaml.load(open(source_file_path, 'r'))
temp = []
for r in rows:
d = {}
d['reference'] = r['pubmed']
d['protein'] = r['entry_name'].replace("__","_").lower()
d['mutation_pos'] = r['seq']
d['mutation_from'] = r['from_res']
d['mutation_to'] = r['to_res']
d['ligand_name'] = ''
d['ligand_type'] = ''
d['ligand_id'] = ''
d['ligand_class'] = ''
d['exp_type'] = ''
d['exp_func'] = ''
d['exp_wt_value'] = 0
d['exp_wt_unit'] = ''
d['exp_mu_effect_sign'] = ''
d['exp_mu_value_raw'] = 0
d['fold_effect'] = 0
d['exp_mu_effect_qual'] = ''
d['exp_mu_effect_ligand_prop'] = ''
d['exp_mu_ligand_ref'] = ''
d['opt_type'] = ''
d['opt_wt'] = 0
d['opt_mu'] = 0
d['opt_sign'] = ''
d['opt_percentage'] = 0
d['opt_qual'] = ''
d['opt_agonist'] = ''
if len(d['mutation_to'])>1 or len(d['mutation_from'])>1: #if something is off with amino acid
continue
temp.append(d)
rows = temp
else:
self.logger.info('unknown format'.source_file)
continue
c = 0
skipped = 0
inserted = 0
for r in rows:
c += 1
if c%1000==0:
self.logger.info('Parsed '+str(c)+' mutant data entries')
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
try:
pub = Publication.objects.get(web_link__index=r['reference'], web_link__web_resource__slug=pub_type)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
wl = WebLink.objects.create(index=r['reference'],
web_resource = WebResource.objects.get(slug=pub_type))
pub.web_link = wl
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' + str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
l = get_or_make_ligand(r['ligand_id'],r['ligand_type'],str(r['ligand_name']))
if Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=True).exists(): #if this name is canonical and it has a ligand record already
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=True)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False).exists(): #if this matches an alias that only has "one" parent canonical name - eg distinct
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=True).exists(): #if this matches an alias that only has several canonical parents, must investigate, start with empty.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = False
l_ref.ambigious_alias = True
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
elif r['exp_mu_ligand_ref']: #if neither a canonical or alias exists, create the records. Remember to check for canonical / alias status.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = True
l_ref.ambigious_alias = False
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
else:
l_ref = None
protein_id = 0
residue_id = 0
protein=Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein=protein.get()
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
else:
skipped += 1
if r['protein'] in missing_proteins:
missing_proteins[r['protein']] += 1
else:
missing_proteins[r['protein']] = 1
self.logger.error('Skipped due to no protein '+ r['protein'])
continue
res=Residue.objects.filter(protein_conformation__protein=protein,amino_acid=r['mutation_from'],sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res=res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'])
skipped += 1
continue
if r['ligand_class']:
l_role, created = LigandRole.objects.get_or_create(name=r['ligand_class'],
defaults={'slug': slugify(r['ligand_class'])[:50]}) # FIXME this should not be needed
else:
l_role = None
if r['exp_type']:
exp_type_id, created = MutationExperimentalType.objects.get_or_create(type=r['exp_type'])
else:
exp_type_id = None
if r['exp_func']:
exp_func_id, created = MutationFunc.objects.get_or_create(func=r['exp_func'])
else:
exp_func_id = None
if r['exp_mu_effect_ligand_prop'] or r['exp_mu_effect_qual']:
exp_qual_id, created = MutationQual.objects.get_or_create(qual=r['exp_mu_effect_qual'], prop=r['exp_mu_effect_ligand_prop'])
else:
exp_qual_id = None
if r['opt_type'] or r['opt_wt'] or r['opt_mu'] or r['opt_sign'] or r['opt_percentage'] or r['opt_qual'] or r['opt_agonist']:
exp_opt_id, created = MutationOptional.objects.get_or_create(type=r['opt_type'], wt=r['opt_wt'], mu=r['opt_mu'], sign=r['opt_sign'], percentage=r['opt_percentage'], qual=r['opt_qual'], agonist=r['opt_agonist'])
else:
exp_opt_id = None
mutation, created = Mutation.objects.get_or_create(amino_acid=r['mutation_to'],protein=protein, residue=res)
logtypes = ['pEC50','pIC50','pK']
foldchange = 0
typefold = ''
if r['exp_wt_value']!=0 and r['exp_mu_value_raw']!=0: #fix for new format
if re.match("(" + ")|(".join(logtypes) + ")", r['exp_type']): #-log values!
foldchange = round(math.pow(10,-r['exp_mu_value_raw'])/pow(10,-r['exp_wt_value']),3);
typefold = r['exp_type']+"_log"
else:
foldchange = round(r['exp_mu_value_raw']/r['exp_wt_value'],3);
typefold = r['exp_type']+"_not_log"
if foldchange<1 and foldchange!=0:
foldchange = -round((1/foldchange),3)
elif r['fold_effect']!=0:
foldchange = round(r['fold_effect'],3);
if foldchange<1: foldchange = -round((1/foldchange),3);
raw_experiment = self.insert_raw(r)
obj, created = MutationExperiment.objects.get_or_create(
refs=pub,
protein=protein,
residue=res,
ligand=l,
ligand_role=l_role,
ligand_ref = l_ref,
raw = raw_experiment,
optional = exp_opt_id,
exp_type=exp_type_id,
exp_func=exp_func_id,
exp_qual = exp_qual_id,
mutation=mutation,
wt_value=r['exp_wt_value'], #
wt_unit=r['exp_wt_unit'],
mu_value = r['exp_mu_value_raw'],
mu_sign = r['exp_mu_effect_sign'],
foldchange = foldchange
)
mut_id = obj.id
inserted += 1
self.logger.info('Parsed '+str(c)+' mutant data entries. Skipped '+str(skipped))
sorted_missing_proteins = sorted(missing_proteins.items(), key=operator.itemgetter(1),reverse=True)
sorted_mutants_for_proteins = sorted(mutants_for_proteins.items(), key=operator.itemgetter(1),reverse=True)
self.logger.info('COMPLETED CREATING MUTANTS')
def main_func(self, positions, iteration):
# filenames
if not positions[1]:
rows = self.data[positions[0]:]
else:
rows = self.data[positions[0]:positions[1]]
missing_proteins = {}
mutants_for_proteins = {}
c = 0
skipped = 0
inserted = 0
bulk_m = []
bulk_r = []
current_sheet = time.time()
for r in rows:
# print(source_file,c)
# PRINT IF ERRORS OCCUR
# self.logger.info('File '+str(r['source_file'])+' number '+str(c))
current = time.time()
c += 1
# if c%100==0:
# self.logger.info('Parsed '+str(c)+' mutant data entries')
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
float(r['review'])
r['review'] = str(int(r['review']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
if r['reference'] not in self.publication_cache:
try:
pub = Publication.objects.get(web_link__index=r['reference'], web_link__web_resource__slug=pub_type)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
wl = WebLink.objects.create(index=r['reference'],
web_resource = WebResource.objects.get(slug=pub_type))
pub.web_link = wl
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' + str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['reference']] = pub
else:
pub = self.publication_cache[r['reference']]
# print(r['review'],r['reference'])
if r['review'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
# print(r['review'],pub_type)
if r['review']:
if r['review'] not in self.publication_cache:
try:
pub_review = Publication.objects.get(web_link__index=r['review'], web_link__web_resource__slug=pub_type)
except Publication.DoesNotExist:
pub_review = Publication()
try:
pub_review.web_link = WebLink.objects.get(index=r['review'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
wl = WebLink.objects.create(index=r['review'],
web_resource = WebResource.objects.get(slug=pub_type))
pub_review.web_link = wl
if pub_type == 'doi':
pub_review.update_from_doi(doi=r['review'])
elif pub_type == 'pubmed':
pub_review.update_from_pubmed_data(index=r['review'])
try:
pub_review.save()
except:
self.logger.error('error with review ' + str(r['review']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['review']] = pub_review
else:
pub_review = self.publication_cache[r['review']]
else:
pub_review = None
l = None
if str(r['ligand_name']) in self.ligand_cache:
if r['ligand_id'] in self.ligand_cache[str(r['ligand_name'])]:
l = self.ligand_cache[str(r['ligand_name'])][r['ligand_id']]
else:
self.ligand_cache[str(r['ligand_name'])] = {}
if not l:
l = get_or_make_ligand(r['ligand_id'],r['ligand_type'],str(r['ligand_name']))
self.ligand_cache[str(r['ligand_name'])][r['ligand_id']] = l
l_ref = None
if str(r['exp_mu_ligand_ref']) in self.ref_ligand_cache:
l_ref = self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])]
else:
if Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=True).exists(): #if this name is canonical and it has a ligand record already
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=True)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False).exists(): #if this matches an alias that only has "one" parent canonical name - eg distinct
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=True).exists(): #if this matches an alias that only has several canonical parents, must investigate, start with empty.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = False
l_ref.ambigious_alias = True
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False).exists(): #amigious_alias not specified
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False)
l_ref.ambigious_alias = False
l_ref.save()
elif r['exp_mu_ligand_ref']: #if neither a canonical or alias exists, create the records. Remember to check for canonical / alias status.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = True
l_ref.ambigious_alias = False
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
try:
l_ref.save()
except IntegrityError:
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False)
print("error failing ligand, duplicate?")
# logger.error("FAILED SAVING LIGAND, duplicate?")
else:
l_ref = None
self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])] = l_ref
protein_id = 0
residue_id = 0
protein=Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein=protein.get()
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
else:
skipped += 1
if r['protein'] in missing_proteins:
missing_proteins[r['protein']] += 1
else:
missing_proteins[r['protein']] = 1
self.logger.error('Skipped due to no protein '+ r['protein'])
continue
res=Residue.objects.filter(protein_conformation__protein=protein,amino_acid=r['mutation_from'],sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res=res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'])
skipped += 1
continue
if r['ligand_class']:
l_role, created = LigandRole.objects.get_or_create(name=r['ligand_class'],
defaults={'slug': slugify(r['ligand_class'])[:50]}) # FIXME this should not be needed
else:
l_role = None
if r['exp_type']:
exp_type_id, created = MutationExperimentalType.objects.get_or_create(type=r['exp_type'])
else:
exp_type_id = None
if r['exp_func']:
exp_func_id, created = MutationFunc.objects.get_or_create(func=r['exp_func'])
else:
exp_func_id = None
if r['exp_mu_effect_ligand_prop'] or r['exp_mu_effect_qual']:
exp_qual_id, created = MutationQual.objects.get_or_create(qual=r['exp_mu_effect_qual'], prop=r['exp_mu_effect_ligand_prop'])
else:
exp_qual_id = None
if r['opt_type'] or r['opt_wt'] or r['opt_mu'] or r['opt_sign'] or r['opt_percentage'] or r['opt_qual'] or r['opt_agonist']:
exp_opt_id, created = MutationOptional.objects.get_or_create(type=r['opt_type'], wt=r['opt_wt'], mu=r['opt_mu'], sign=r['opt_sign'], percentage=r['opt_percentage'], qual=r['opt_qual'], agonist=r['opt_agonist'])
else:
exp_opt_id = None
try:
mutation, created = Mutation.objects.get_or_create(amino_acid=r['mutation_to'],protein=protein, residue=res)
except IntegrityError:
mutation = Mutation.objects.get(amino_acid=r['mutation_to'],protein=protein, residue=res)
logtypes = ['pEC50','pIC50','pK']
foldchange = 0
typefold = ''
if r['exp_wt_value']!=0 and r['exp_mu_value_raw']!=0: #fix for new format
if re.match("(" + ")|(".join(logtypes) + ")", r['exp_type']): #-log values!
foldchange = round(math.pow(10,-r['exp_mu_value_raw'])/pow(10,-r['exp_wt_value']),3);
typefold = r['exp_type']+"_log"
else:
foldchange = round(r['exp_mu_value_raw']/r['exp_wt_value'],3);
typefold = r['exp_type']+"_not_log"
if foldchange<1 and foldchange!=0:
foldchange = -round((1/foldchange),3)
elif r['fold_effect']!=0:
foldchange = round(r['fold_effect'],3);
if foldchange<1: foldchange = -round((1/foldchange),3);
raw_experiment = self.insert_raw(r)
bulk = MutationExperiment(
refs=pub,
review=pub_review,
protein=protein,
residue=res,
ligand=l,
ligand_role=l_role,
ligand_ref = l_ref,
#raw = raw_experiment, #raw_experiment, OR None
optional = exp_opt_id,
exp_type=exp_type_id,
exp_func=exp_func_id,
exp_qual = exp_qual_id,
mutation=mutation,
wt_value=r['exp_wt_value'], #
wt_unit=r['exp_wt_unit'],
mu_value = r['exp_mu_value_raw'],
mu_sign = r['exp_mu_effect_sign'],
foldchange = foldchange
)
# mut_id = obj.id
bulk_r.append(raw_experiment)
bulk_m.append(bulk)
inserted += 1
end = time.time()
diff = round(end - current,2)
#print(diff)
self.logger.info('Parsed '+str(c)+' mutant data entries. Skipped '+str(skipped))
current = time.time()
raws = MutationRaw.objects.bulk_create(bulk_r)
for i,me in enumerate(bulk_m):
me.raw = raws[i]
MutationExperiment.objects.bulk_create(bulk_m)
end = time.time()
diff = round(end - current,2)
current_sheet
diff_2 = round(end - current_sheet,2)
print("overall",diff_2,"bulk",diff,len(bulk_m),"skipped",str(skipped))
def import_xtal(self):
xtals = self.parse_excel(self.annotation_file,'Xtal-methods')
xtals_list = {}
for x in xtals:
if x[1] in xtals_list:
print('pdbcode duplicate?',x[1])
xtals_list[x[1]] = x
xtal_chems = self.parse_excel(self.annotation_file,'Xtal-Chemicals')
xtals_chems_list = {}
for x in xtal_chems:
if x[1] not in xtals_chems_list:
xtals_chems_list[x[1]] = []
xtals_chems_list[x[1]].append(x)
xtal_ligands = self.parse_excel(self.annotation_file,'PDB-ligand-complex')
xtal_ligands_list = {}
for x in xtal_ligands:
if x[1] not in xtal_ligands_list:
xtal_ligands_list[x[1]] = []
xtal_ligands_list[x[1]].append(x)
missing = list(set(self.all_pdbs) - set(xtals_list.keys()))
print(sorted(missing)," do not have any Xtal-methods annotated -- add them to sheet with NONE if they have none")
missing = list(set(self.all_pdbs) - set(xtals_chems_list.keys()))
print(sorted(missing)," do not have any Xtal-Chemicals annotated -- add them to sheet with NONE if they have none")
missing = list(set(self.all_pdbs) - set(xtal_ligands_list.keys()))
print(sorted(missing)," do not have any PDB-ligand-complex annotated -- add them to sheet with NONE if they have none")
for pdb,x in xtals_list.items():
try:
construct = Construct.objects.get(structure__pdb_code__index=pdb.upper())
except:
print(pdb,'cannot find pdb construct')
continue
c = Crystallization()
c_type, created = CrystallizationTypes.objects.get_or_create(name=x[3], sub_name=x[4])
c_method, created = CrystallizationMethods.objects.get_or_create(name=x[2])
c.crystal_type = c_type
c.crystal_method = c_method
c.remarks = x[10]
c.temp = x[7]
# Some entries have it wrong here
try:
c.ph_start = float(x[8])
c.ph_end = float(x[9])
except:
c.ph_start = 0
c.ph_end = 0
c.protein_conc = x[5]
c.protein_conc_unit = x[6]
c.save()
if pdb in xtals_chems_list:
chems = xtals_chems_list[pdb]
chem_types = {}
for chem in chems:
ctype = chem[2]
if ctype not in chem_types:
chem_types[ctype] = []
chem_types[ctype].append(chem)
for ctype,chems in chem_types.items():
c_list = ChemicalList()
list_name,created = ChemicalListName.objects.get_or_create(name=ctype)
c_list.name = list_name
c_list.save()
for ch in chems:
ct, created = ChemicalType.objects.get_or_create(name=ch[4])
chem, created = Chemical.objects.get_or_create(name=ch[3], chemical_type=ct)
cc, created = ChemicalConc.objects.get_or_create(concentration=ch[5], concentration_unit=ch[6], chemical=chem)
c_list.chemicals.add(cc)
c.chemical_lists.add(c_list)
else:
print('no chems for ',pdb)
if pdb in xtal_ligands_list:
l = xtal_ligands_list[pdb][0]
ligand = get_or_make_ligand(l[7],l[6],l[2])
role_slug = slugify(l[3])
try:
lr, created = LigandRole.objects.get_or_create(slug=role_slug,
defaults={'name': l[3]})
except IntegrityError:
lr = LigandRole.objects.get(slug=role_slug)
if ligand:
ligand_c = CrystallizationLigandConc()
ligand_c.construct_crystallization = c
ligand_c.ligand = ligand
if lr:
ligand_c.ligand_role = lr
if l[4]:
ligand_c.ligand_conc = l[4]
if l[5]:
ligand_c.ligand_conc_unit = l[5]
ligand_c.save()
c.ligands.add(ligand_c)
construct.crystallization = c
construct.save()
def main_func(self, positions, iteration,count,lock):
# filenames
# if not positions[1]:
# rows = self.data[positions[0]:]
# else:
# rows = self.data[positions[0]:positions[1]]
missing_proteins = {}
mutants_for_proteins = {}
wrong_uniport_ids = {}
c = 0
skipped = 0
inserted = 0
bulk_m = []
bulk_r = []
current_sheet = time.time()
rows = self.data_all
while count.value<len(rows):
with lock:
r = rows[count.value]
count.value +=1
# for r in rows:
# print(r['source_file'],c)
# PRINT IF ERRORS OCCUR
#self.logger.info('File '+str(r['source_file'])+' number '+str(c))
current = time.time()
c += 1
# if c%100==0:
# self.logger.info('Parsed '+str(c)+' mutant data entries')
# publication
try: #fix if it thinks it's float.
float(r['reference'])
r['reference'] = str(int(r['reference']))
float(r['review'])
r['review'] = str(int(r['review']))
except ValueError:
pass
if r['reference'].isdigit(): #assume pubmed
pub_type = 'pubmed'
else: #assume doi
pub_type = 'doi'
if r['reference'] not in self.publication_cache:
try:
wl = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
try:
wl = WebLink.objects.create(index=r['reference'],
web_resource = WebResource.objects.get(slug=pub_type))
except IntegrityError:
wl = WebLink.objects.get(index=r['reference'], web_resource__slug=pub_type)
try:
pub = Publication.objects.get(web_link=wl)
except Publication.DoesNotExist:
pub = Publication()
try:
pub.web_link = wl
pub.save()
except IntegrityError:
pub = Publication.objects.get(web_link=wl)
if pub_type == 'doi':
pub.update_from_doi(doi=r['reference'])
elif pub_type == 'pubmed':
pub.update_from_pubmed_data(index=r['reference'])
try:
pub.save()
except:
self.logger.error('error with reference ' + str(r['reference']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['reference']] = pub
else:
pub = self.publication_cache[r['reference']]
# print(r['review'],r['reference'])
if r['review'].isdigit(): #assume pubmed
pub_type = 'pubmed'
elif r['review'].startswith('http'):
pub_type = 'raw_link'
else: #assume doi
pub_type = 'doi'
# print(r['review'],pub_type)
if r['review']:
if r['review'] not in self.publication_cache:
try:
wl = WebLink.objects.get(index=r['review'], web_resource__slug=pub_type)
except WebLink.DoesNotExist:
try:
wl = WebLink.objects.create(index=r['review'],
web_resource = WebResource.objects.get(slug=pub_type))
except IntegrityError:
wl = WebLink.objects.get(index=r['review'], web_resource__slug=pub_type)
try:
pub_review = Publication.objects.get(web_link=wl)
except Publication.DoesNotExist:
pub_review = Publication()
try:
pub_review.web_link = wl
pub_review.save()
except IntegrityError:
pub_review = Publication.objects.get(web_link=wl)
if pub_type == 'doi':
pub_review.update_from_doi(doi=r['review'])
elif pub_type == 'pubmed':
pub_review.update_from_pubmed_data(index=r['review'])
try:
pub_review.save()
except:
self.logger.error('error with review ' + str(r['review']) + ' ' + pub_type)
continue #if something off with publication, skip.
self.publication_cache[r['review']] = pub_review
else:
pub_review = self.publication_cache[r['review']]
else:
pub_review = None
l = None
if str(r['ligand_name']) in self.ligand_cache:
if r['ligand_id'] in self.ligand_cache[str(r['ligand_name'])]:
l = self.ligand_cache[str(r['ligand_name'])][r['ligand_id']]
else:
self.ligand_cache[str(r['ligand_name'])] = {}
if not l:
try:
l = get_or_make_ligand(r['ligand_id'],r['ligand_type'],str(r['ligand_name']))
except Exception as msg:
print('Something errored with ligand, aborting entry of mutation',r['ligand_name'],r['ligand_type'],r['ligand_id'],r['source_file'])
print(msg)
traceback.print_exc()
continue
self.ligand_cache[str(r['ligand_name'])][r['ligand_id']] = l
l_ref = None
if str(r['exp_mu_ligand_ref']) in self.ref_ligand_cache:
l_ref = self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])]
else:
if Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=True).exists(): #if this name is canonical and it has a ligand record already
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=True)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False).exists(): #if this matches an alias that only has "one" parent canonical name - eg distinct
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=False)
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False, ambigious_alias=True).exists(): #if this matches an alias that only has several canonical parents, must investigate, start with empty.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = False
l_ref.ambigious_alias = True
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
l_ref.save()
elif Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=False).exists(): #amigious_alias not specified
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False)
l_ref.ambigious_alias = False
l_ref.save()
elif r['exp_mu_ligand_ref']: #if neither a canonical or alias exists, create the records. Remember to check for canonical / alias status.
lp = LigandProperities()
lp.save()
l_ref = Ligand()
l_ref.properities = lp
l_ref.name = r['exp_mu_ligand_ref']
l_ref.canonical = True
l_ref.ambigious_alias = False
try:
l_ref.save()
l_ref.load_by_name(r['exp_mu_ligand_ref'])
except IntegrityError:
if Ligand.objects.filter(name=r['exp_mu_ligand_ref'], canonical=True).exists():
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=True)
else:
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False)
# print("error failing ligand, duplicate?")
try:
l_ref.save()
except IntegrityError:
l_ref = Ligand.objects.get(name=r['exp_mu_ligand_ref'], canonical=False)
# print("error failing ligand, duplicate?")
# logger.error("FAILED SAVING LIGAND, duplicate?")
else:
l_ref = None
self.ref_ligand_cache[str(r['exp_mu_ligand_ref'])] = l_ref
protein_id = 0
residue_id = 0
protein=Protein.objects.filter(entry_name=r['protein'])
if protein.exists():
protein=protein.get()
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
elif r['protein'] not in missing_proteins:
try:
r['protein'] = wrong_uniport_ids[r['protein']]
real_uniprot = wrong_uniport_ids[r['protein']]
protein=Protein.objects.get(entry_name=r['protein'])
# print('fetched with lookup table',r['protein'])
except:
# look for it as uniprot
protein=Protein.objects.filter(web_links__web_resource__slug='uniprot', web_links__index=r['protein'].upper())
if protein.exists():
protein=protein.get()
real_uniprot = protein.entry_name
if r['protein'] in mutants_for_proteins:
mutants_for_proteins[r['protein']] += 1
else:
mutants_for_proteins[r['protein']] = 1
else:
# Try to lookup in uniprot to catch typing errors / variants in entry_name
url = 'http://www.uniprot.org/uniprot/$index.xml'
cache_dir = ['uniprot', 'id']
uniprot_protein = fetch_from_web_api(url, r['protein'], cache_dir, xml = True)
try:
real_uniprot = uniprot_protein.find('.//{http://uniprot.org/uniprot}name').text.lower()
protein=Protein.objects.get(entry_name=real_uniprot)
except:
skipped += 1
if r['protein'] in missing_proteins:
missing_proteins[r['protein']] += 1
else:
missing_proteins[r['protein']] = 1
# print('Skipped due to no protein '+ r['protein'])
self.logger.error('Skipped due to no protein '+ r['protein'])
continue
wrong_uniport_ids[r['protein']] = protein.entry_name
r['protein'] = real_uniprot
else:
missing_proteins[r['protein']] += 1
continue
res=Residue.objects.filter(protein_conformation__protein=protein,amino_acid=r['mutation_from'],sequence_number=r['mutation_pos']) #FIXME MAKE AA CHECK
if res.exists():
res=res.get()
else:
self.logger.error('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'])
# print('Skipped due to no residue or mismatch AA ' + r['protein'] + ' pos:'+str(r['mutation_pos']) + ' AA:'+r['mutation_from'],r['source_file'])
skipped += 1
continue
if r['ligand_class']:
try:
l_role, created = LigandRole.objects.get_or_create(name=r['ligand_class'],
defaults={'slug': slugify(r['ligand_class'])[:50]}) # FIXME this should not be needed
except Exception as e:
if LigandRole.objects.filter(slug=slugify(r['ligand_class'])[:50]).exists():
l_role = LigandRole.objects.get(slug=slugify(r['ligand_class'])[:50])
if l_role.name == slugify(r['ligand_class'])[:50]:
#if name of role is same as slug, then it was created by constructs script, replace it
l_role.name = r['ligand_class']
l_role.save()
else:
print(e)
print("Error with",r['ligand_class'],slugify(r['ligand_class'])[:50] )
l_role, created = LigandRole.objects.get_or_create(slug=slugify(r['ligand_class'])[:50]) # FIXME this should not be needed
else:
l_role = None
if r['exp_type']:
exp_type_id, created = MutationExperimentalType.objects.get_or_create(type=r['exp_type'])
else:
exp_type_id = None
if r['exp_func']:
exp_func_id, created = MutationFunc.objects.get_or_create(func=r['exp_func'])
else:
exp_func_id = None
if r['exp_mu_effect_ligand_prop'] or r['exp_mu_effect_qual']:
exp_qual_id, created = MutationQual.objects.get_or_create(qual=r['exp_mu_effect_qual'], prop=r['exp_mu_effect_ligand_prop'])
else:
exp_qual_id = None
# if r['opt_type'] or r['opt_wt'] or r['opt_mu'] or r['opt_sign'] or r['opt_percentage'] or r['opt_qual'] or r['opt_agonist']:
# exp_opt_id, created = MutationOptional.objects.get_or_create(type=r['opt_type'], wt=r['opt_wt'], mu=r['opt_mu'], sign=r['opt_sign'], percentage=r['opt_percentage'], qual=r['opt_qual'], agonist=r['opt_agonist'])
# else:
# exp_opt_id = None
try:
mutation, created = Mutation.objects.get_or_create(amino_acid=r['mutation_to'],protein=protein, residue=res)
except IntegrityError:
mutation = Mutation.objects.get(amino_acid=r['mutation_to'],protein=protein, residue=res)
logtypes = ['pEC50','pIC50','pK']
foldchange = 0
typefold = ''
if r['exp_wt_value']!=0 and r['exp_mu_value_raw']!=0: #fix for new format
if re.match("(" + ")|(".join(logtypes) + ")", r['exp_type']): #-log values!
try:
foldchange = round(math.pow(10,-r['exp_mu_value_raw'])/pow(10,-r['exp_wt_value']),3);
except:
print(r)
typefold = r['exp_type']+"_log"
elif "%"==r['exp_wt_unit']:
# if % then it's a difference case, then lower value is bad. Otherwise it's conc and lower is better
foldchange = round(r['exp_wt_value']/r['exp_mu_value_raw'],3);
else:
foldchange = round(r['exp_mu_value_raw']/r['exp_wt_value'],3);
typefold = r['exp_type']+"_not_log"
if foldchange>0 and foldchange<1 and foldchange!=0:
foldchange = -round((1/foldchange),3)
elif r['fold_effect']!=0:
foldchange = round(r['fold_effect'],3);
if foldchange<1: foldchange = -round((1/foldchange),3);
r['fold_effect'] = foldchange
raw_experiment = self.insert_raw(r)
# raw_experiment.save()
bulk = MutationExperiment(
refs=pub,
review=pub_review,
submitting_group = r['submitting_group'],
data_container = r['data_container'],
data_container_number = r['data_container_number'],
protein=protein,
residue=res,
ligand=l,
ligand_role=l_role,
ligand_ref = l_ref,
# raw = raw_experiment, #raw_experiment, OR None
# optional = exp_opt_id,
exp_type=exp_type_id,
exp_func=exp_func_id,
exp_qual = exp_qual_id,
mutation=mutation,
wt_value=r['exp_wt_value'], #
wt_unit=r['exp_wt_unit'],
mu_value = r['exp_mu_value_raw'],
mu_sign = r['exp_mu_effect_sign'],
foldchange = foldchange,
opt_receptor_expression = r['opt_receptor_expression'],
opt_basal_activity = r['opt_basal_activity'],
opt_gain_of_activity = r['opt_gain_of_activity'],
opt_ligand_emax = r['opt_ligand_emax'],
opt_agonist = r['opt_agonist'],
)
# for line,val in r.items():
# val = str(val)
# if len(val)>100:
# print(line,"too long",val)
# mut_id = obj.id
bulk_r.append(raw_experiment)
bulk_m.append(bulk)
# try:
# bulk.save()
# except Exception as e:
# print(e)
# print(r)
# break
#print('saved ',r['source_file'])
inserted += 1
end = time.time()
diff = round(end - current,2)
#print(diff)
self.logger.info('Parsed '+str(c)+' mutant data entries. Skipped '+str(skipped))
current = time.time()
raws = MutationRaw.objects.bulk_create(bulk_r)
for i,me in enumerate(bulk_m):
me.raw = raws[i]
MutationExperiment.objects.bulk_create(bulk_m)
end = time.time()
diff = round(end - current,2)
current_sheet
diff_2 = round(end - current_sheet,2)
print("overall",diff_2,"bulk",diff,len(bulk_m),"skipped",str(skipped))
sorted_missing_proteins = sorted(missing_proteins.items(), key=operator.itemgetter(1),reverse=True)
