def compress(taxid=9606, target='../gpickle'):
if not os.path.exists(target):
os.mkdir(target)
source = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(source):
p = ProteinCore().load(file=os.path.join(source, pf))
p.gdump(file=os.path.join(target, os.path.splitext(pf)[0] + '.gzp'))
def compress(
taxid=9606,
target='/home/matteo/Coding/Michelanglo/MichelaNGLo-human-protein-data/gpickle'
):
if not os.path.exists(target):
os.mkdir(target)
source = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(source):
p = ProteinCore().load(file=os.path.join(source, pf))
p.gdump(file=os.path.join(target, os.path.splitext(pf)[0] + '.gzp'))
def inspect_offsets(uniprot):
print('***************** inspect_offsets *******************************')
p = ProteinCore(taxid='9606', uniprot=uniprot).load()
for s in p.pdbs:
print(s.code)
print(s.chain_definitions)
print(s._get_sifts())
def random_view(request):
while True:
name = random.choice(list(human.keys()))
uniprot = human[name]
protein = ProteinCore(taxid=9606, uniprot=uniprot).load()
if protein.pdbs:
pdb = random.choice(protein.pdbs)
elif protein.swissmodel:
pdb = random.choice(protein.swissmodel)
else:
continue
i = random.randint(pdb.x, pdb.y)
try:
return {
'name':
name,
'uniprot':
uniprot,
'taxid':
'9606',
'species':
'human',
'mutation':
f'p.{protein.sequence[i-1]}{i}{random.choice(Mutation.aa_list)}'
}
except IndexError:
log.error(
f'Impossible... pdb.x out of bounds in unicode for gene {uniprot}'
)
continue
def random_view(self):
while True:
name = random.choice(list(human.keys()))
uniprot = human[name]
protein = ProteinCore(taxid=9606, uniprot=uniprot).load()
if protein.pdbs:
pdb = random.choice(protein.pdbs)
elif protein.swissmodel:
pdb = random.choice(protein.swissmodel)
else:
continue
try:
i = random.randint(pdb.x, pdb.y)
# the to_resn cannot be the same as original or *
to_resn = random.choice(
list(
set(Mutation.aa_list) -
{'*', protein.sequence[i - 1]}))
return {
'name': name,
'uniprot': uniprot,
'taxid': '9606',
'species': 'human',
'mutation': f'p.{protein.sequence[i - 1]}{i}{to_resn}'
}
except IndexError:
log.error(
f'Impossible... pdb.x out of bounds in unicode for gene {uniprot}'
)
continue
def get_transcript(request):
enst = request.params['enst']
mutation = request.params['mutation']
mapper = ENSTMapper(enst)
if mapper.is_full_match():
return {'uniprot': mapper.uniprot, 'mutation': mutation}
else:
p = ProteinCore(uniprot=mapper.uniprot, taxid=9606).load()
return {
'uniprot': mapper.uniprot,
'mutation': mapper.convert(p.sequence, mutation)
}
def touch_offsets(taxid=9606):
overview = []
global_settings.verbose = False
source = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(source):
p = ProteinCore().load(file=os.path.join(source, pf))
for s in p.pdbs:
if s.type != 'rcsb':
continue
details = s._get_sifts()
v = []
for detail in details:
## clean rows
for k in ('PDB_BEG', 'PDB_END', 'RES_END', 'RES_BEG', 'SP_BEG',
'SP_END'):
if k == 'None' or k is None:
detail[k] = None
elif isinstance(detail[k], int):
pass # this means so test is being done.
else:
r = re.search(
'(-?\d+)', detail[k]
) # str().isdigit() does not like negatives.
if r is None:
detail[k] = None
else:
detail[k] = int(
r.group(1)) # yes. py int is signed
## get offset
if detail['PDB_BEG'] is not None: ##nice.
offset = detail['SP_BEG'] - detail['PDB_BEG']
if offset and detail['PDB_BEG'] == detail['RES_BEG']:
v.append('off-start')
elif offset:
v.append('off-unstart')
elif detail['SP_BEG'] != 1:
v.append('no-off-unstart')
else:
v.append('no-off-start')
elif detail['PDB_END'] is not None:
offset = detail['SP_BEG'] - (
detail['PDB_END'] -
(detail['SP_END'] - detail['SP_BEG']))
if offset and detail['PDB_END'] == detail['RES_END']:
v.append('off-start')
elif offset:
v.append('off-unstart')
elif detail['SP_BEG'] != 1:
v.append('no-off-unstart')
else:
v.append('no-off-start')
elif detail['SP_BEG'] == 1:
offset = 0
v.append('no-off-start')
elif detail['RES_BEG'] == 1:
# This is problematic. This means that there are unresolved residues at the N & C termini.
# This can go either way.
v.append('RES1')
offset = 0
else:
v.append('RESn')
offset = 0
if 'RESn' in v or 'RES1' in v:
offset = 0
c = Counter(v).most_common()
overview.append('+'.join(sorted(set(v))))
print(Counter(overview).most_common())
def fix_offsets(file):
"""
This method fixes the offsets of a file. and saves.
:param file: fullpath.
:return:
"""
p = ProteinCore().load(file=file)
lines = []
for s in p.pdbs:
if s.type != 'rcsb':
continue
details = s._get_sifts()
s.chain_definitions = []
s.offsets = {}
for detail in details:
## clean rows
for k in ('PDB_BEG', 'PDB_END', 'RES_END', 'RES_BEG', 'SP_BEG',
'SP_END'):
if k == 'None' or k is None:
detail[k] = None
elif isinstance(detail[k], int):
pass # this means so test is being done.
else:
r = re.search(
'(-?\d+)',
detail[k]) # str().isdigit() does not like negatives.
if r is None:
detail[k] = None
else:
detail[k] = int(r.group(1)) # yes. py int is signed
## get offset
if detail['PDB_BEG'] is not None: ##nice.
offset = detail['SP_BEG'] - detail['PDB_BEG']
elif detail['PDB_END'] is not None:
offset = detail['SP_BEG'] - (
detail['PDB_END'] - (detail['SP_END'] - detail['SP_BEG']))
elif detail['SP_BEG']:
try:
if detail['SP_PRIMARY'] == p.uniprot:
offset = s.get_offset_from_PDB(detail, p.sequence)
else:
seq = ProteinCore(
uniprot=detail['SP_PRIMARY']).load().sequence
offset = s.get_offset_from_PDB(detail, seq)
except: ## Pymol subclasses BaseException.
offset = 0
else:
offset = 0
detail['offset'] = offset
lines.append(
f"{s.code}\t{detail['CHAIN']}\t{detail['SP_PRIMARY']}\t{offset}"
)
s.chain_definitions.append({
'chain': detail['CHAIN'],
'uniprot': detail['SP_PRIMARY'],
'x': detail["SP_BEG"],
'y': detail["SP_END"],
'offset': offset,
'range': f'{detail["SP_BEG"]}-{detail["SP_END"]}',
'name': None,
'description': None
})
s.offsets[detail['CHAIN']] = offset
try:
if s.chain != '*':
detail = next(
filter(lambda x: s.chain == x['chain'],
s.chain_definitions))
s.offset = detail['offset']
except:
pass
if p.pdbs:
p.dump()
return '\n'.join(lines)
def describe(uniprot):
print('***************** DESCRIPTION *******************************')
p = ProteinCore(taxid='9606',
uniprot=uniprot).load() # gnb1 P62873 gnb2 P62879
pprint(p.asdict())
def get_data(uniprot: str) -> ProteinCore:
return ProteinCore(taxid=9606, uniprot=uniprot).load()
# p = ProteinCore(taxid='9606', uniprot='P01112').load()
# p = ProteinCore(taxid='3562', uniprot='Q8GT36').load()
# print(sum(p.properties['kd'])/len(p))
# print(sum(p.properties['Flex']) / len(p))
# from create import message
# download_swissmodel()
# all_swiss(fx=hotfix_swiss)
# message('Reswissed!')
# fix_offsets('../protein-data/pickle/taxid9606/Q13586.p')
# print('**************************************')
taxid = 9606
gene = 'P04637' # TP53
# describe(gene)
p = ProteinCore(taxid=taxid,
uniprot=gene).load() # gnb1 P62873 gnb2 P62879
print(p.features.keys())
print(p.features['PSP_modified_residues'])
exit()
# #pprint(p.__dict__)
# t = [s for s in p.pdbs if s.code.lower() == '1a1u'][0]
# print(str(t))
# print(t.offset)
# print(t.chain_definitions)
# sifts = t._get_sifts()
# print(sifts)
# print(t.get_offset_from_PDB(sifts[0], p.sequence))
# p.parse_uniprot()
# p.parse_swissmodel()