def add_swissmodel(taxid=9606):
def fix(p):
global_settings.verbose = True
p.parse_all(mode='serial')
assert len(p.sequence) > 0, 'Darn. Sequence is zero AA long'
p.dump()
global_settings.verbose = False
print(
f'************************ {taxid} *************************************'
)
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
if os.path.splitext(pf)[1] != '.p':
continue
try:
p = ProteinGatherer().load(file=os.path.join(path, pf))
except:
p = ProteinGatherer(uniprot=pf.replace('.p', ''))
fix(p)
if len(p.sequence) == 0:
try:
fix(p)
except Exception:
traceback.print_exc(file=sys.stdout)
else:
p.parse_swissmodel()
p.dump()
def hotfix_swiss(taxid=9606):
# The database stores the data differently to what the data in the indices say!
# https://swissmodel.expasy.org/repository/uniprot/P31946.pdb?from=1&to=232&template=2bq0&provider=pdb
# https://swissmodel.expasy.org/repository/uniprot/P31946.pdb?sort=seqsim&provider=pdb&template=2bq0&range=1-232
def fix(p):
global_settings.verbose = True
p.parse_all(mode='serial')
assert len(p.sequence) > 0, 'Darn. Sequence is zero AA long'
p.dump()
global_settings.verbose = False
print(
f'************************ {taxid} *************************************'
)
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
if os.path.splitext(pf)[1] != '.p':
continue
try:
p = ProteinGatherer().load(file=os.path.join(path, pf))
except:
p = ProteinGatherer(uniprot=pf.replace('.p', ''))
fix(p)
if len(p.sequence) == 0:
try:
fix(p)
except Exception:
traceback.print_exc(file=sys.stdout)
else:
for model in p.swissmodel:
model.url = re.sub(r'from\=(\d+)&to\=(\d+)',
r'sort=seqsim&range=\1-\2', model.url)
p.dump()
def mini_gene_data():
genes = '''DOCK180
DOCK2
DOCK3
DOCK4
DOCK5
DOCK6
DOCK7
DOCK8
DOCK9
DOCK10
DOCK11
'''.split()
data = {}
from pprint import PrettyPrinter
pprint = PrettyPrinter().pprint
namedex = json.load(open('data/human_prot_namedex.json'))
for uni in set(namedex.values()):
g = ProteinGatherer(uniprot=uni).parse_uniprot()
data[g.gene_name] = {
'name': g.gene_name,
'uniprot': g.uniprot,
'len': len(g),
'domains': {
k: g.features[k]
for k in ('active site', 'modified residue',
'topological domain', 'domain', 'region of interest',
'transmembrane region') if k in g.features
},
'disease': g.diseases
}
#print(g.gene_name,g.uniprot,len(g))
json.dump(data, open('map.json', 'w'))
def fix_empty(taxid=9606):
from collections import Counter
global_settings.verbose = False
glitchy = 0
fine = 0
fixed = 0
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
p = ProteinGatherer().load(file=os.path.join(path, pf))
if len(p.sequence) == 0:
print('****************************************')
print(f'Attempting to fix {p.gene_name}')
try:
global_settings.verbose = True
p.parse_uniprot()
p.parse_swissmodel()
p.compute_params()
p.parse_gnomAD()
p.get_PTM()
assert len(p.sequence) > 0, 'Darn. Sequence is zero AA long'
p.dump()
fixed += 1
global_settings.verbose = False
except Exception:
traceback.print_exc(file=sys.stdout)
glitchy += 1
else:
fine += 1
print('****************************************')
print(f'Fine: {fine:,}, Fixed {fixed:,}, Glitchy: {glitchy:,}')
def reparse_gene(name):
human = json.load(
open(
os.path.join(global_settings.dictionary_folder,
'taxid9606-names2uniprot.json')))
target = human[name]
p = ProteinGatherer(uniprot=target)
p.parse_uniprot()
print(p.sequence)
def how_many_empty(taxid=9606):
from collections import Counter
global_settings.verbose = False
empty = 0
full = 0
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
p = ProteinGatherer().load(file=os.path.join(path, pf))
if len(p.sequence) == 0:
print(p)
empty += 1
else:
full += 1
print(full, empty)
def iterate_taxon(taxid=9606):
"""
This is an ad hoc fix to fix humans or similar. For full deployment use ProteomeParser.
:param taxid:
:return:
"""
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
try:
protein = ProteinGatherer().load(file=os.path.join(path, pf))
protein.gnomAD = []
protein.parse_gnomAD()
protein.get_PTM()
protein.compute_params()
protein.dump()
#michelanglo_protein.get_offsets().parse_gnomAD().compute_params()
#michelanglo_protein.dump()
except:
pass
#### workspace!
if 1 == 1:
#os.mkdir(os.path.join(ProteinCore.settings.temp_folder, 'PDB'))
#describe('P01112')
#analyse('P62873')
#how_many_empty()
#fix_empty()
#compress()
#parse_uniprot(
#inspect_offsets('P01133')
#touch_offsets()
#fix_all_offsets()
#all_swiss()
fix_all_offsets()
elif 1 == 9:
p = ProteinGatherer(taxid='9606', uniprot='P62873').load()
print(p.gnomAD)
print(p.parse_gnomAD())
print(p.gnomAD)
print(p.features['PSP_modified_residues'])
from michelanglo_protein.generate.split_phosphosite import Phoshosite
#ph = Phoshosite().split().write('phosphosite')
p = ProteinGatherer(taxid='9606', uniprot='P62879').load()
print(':B and (' +
' or '.join([str(m.x) for m in p.gnomAD if m.homozygous]) + ')')
print([m for m in p.gnomAD if m.homozygous])
print(' '.join([str(m.description.split()[0]) for m in p.gnomAD]) + ')')
print([m for m in p.gnomAD if m.homozygous])
elif 1 == 0:
iterate_taxon('9606')
namedexfile=os.path.join(global_settings.dictionary_folder,
'taxid9606-names2uniprot.json'),
folder=os.path.join(global_settings.temp_folder, 'gnomAD')).split()
if __name__ == '__main__':
global_settings.verbose = True #False
global_settings.error_tolerant = True
global_settings.startup(data_folder='../protein-data')
global_settings.retrieve_references(ask=False, refresh=False)
## Phosphosite
#Phosphosite().split().write()
message('Phosphosite split')
## Uniprot
UniprotMasterReader(first_n_protein=0)
message('Uniprot split')
# gnomAD data needs to be split up after that the dictionaries are made.
# _gnomad()
message('gnomAD split')
taxid = 9606 # that's humans
path = os.path.join(global_settings.pickle_folder, f'taxid{taxid}')
for pf in os.listdir(path):
try:
protein = ProteinGatherer().load(file=os.path.join(path, pf))
protein.gnomAD = []
protein.parse_gnomAD()
protein.get_PTM()
protein.dump()
except:
pass
message('Done.')