def UIDs2JSON(base_uri, uids, fout):
""" Uses uniprot library from Bosco Ho (https://github.com/boscoh/uniprot)."""
import uniprot ## Bosco Ho (https://github.com/boscoh/uniprot)
uniprot_data=uniprot.batch_uniprot_metadata(uids, None)
for uid in uniprot_data.keys():
for key in uniprot_data[uid].keys():
if key in ('accs', 'sequence', 'go', 'description'): #keep simple
del uniprot_data[uid][key]
json_txt=json.dumps(uniprot_data, sort_keys=True, indent=2)
fout.write(json_txt+'\n')
return
def get_uniprot_loc(seq_id):
# input should be a string in any case
if not isinstance(seq_id, str):
return None
# if there are less than 3 spaces
elif ' ' not in seq_id:
# get all protein info
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
prot_content = uniprot_data.get(seq_id)
prot_content = str(prot_content.get('comment'))
else:
prot_content = seq_id
if prot_content == None:
return None
# looking for location
loc_header = 'SUBCELLULAR LOCATION: '
loc_idx = prot_content.find(loc_header)
# if the location is present
if loc_idx > -1:
# isolating location information (not including header)
# now left side begins with the location
prot_content = prot_content[loc_idx + len(loc_header):]
# making string a bit shorter by going up to next exclamation mark
next_exclam = prot_content.find('!')
prot_content = prot_content[:next_exclam]
# finding where all punctuation marks occur in what's remaining
punct_df = pd.DataFrame(list(string.punctuation), columns=['mark'])
punct_df['pos'] = punct_df.mark.apply(lambda x: prot_content.find(x))
# ignoring punctuations that do not occur
punct_df = punct_df[punct_df.pos > -1]
# finding the mark that occurs first
punct_df = punct_df.sort('pos').reset_index(drop=True)
first_punct = punct_df.pos[punct_df.index == 0].values[0]
# getting location; removing leading and trailing spaces
location = prot_content[:first_punct].strip()
return location
else:
return None
def insert_best_seqid_column(params):
csv = params['csv']
uniprot_ids_header = params['uniprot_ids_header']
delimiter = params['delimiter']
out_csv = params['output_csv']
uniprot_cache = params['cache_dir']
if not csv:
raise IOError('No file selected')
if not os.path.isfile(csv):
raise IOError(csv + ' not found')
headers = get_headers(csv)
if uniprot_ids_header not in headers:
s = "Column header '%s' not found, available headers:\n" % uniprot_ids_header
for header in headers:
s += ' ' + header + '\n'
raise IOError(s)
logging('Reading %s\n' % csv)
entries = read_csv(csv)
all_seqids = []
for entry in entries:
tokens = entry[uniprot_ids_header].split(delimiter)
entry['seqids'] = [s.strip() for s in tokens]
all_seqids.extend(entry['seqids'])
logging('Found %d potential Uniprot IDs\n' % len(all_seqids))
uniprot_data = uniprot.batch_uniprot_metadata(
all_seqids, uniprot_cache)
for entry in entries:
best_seqid = uniprot.sort_seqids_by_uniprot(
entry['seqids'], uniprot_data)[0]
entry['best_seqid'] = best_seqid
entry['is_reviewed'] = False
if best_seqid in uniprot_data:
entry['is_reviewed'] = \
uniprot_data[best_seqid]['is_reviewed']
logging('Writing ')
logging('%s\n' % os.path.abspath(out_csv), lambda: open_file(out_csv))
headers = ['best_seqid', 'is_reviewed'] + get_headers(csv)
rows = [headers]
for entry in entries:
rows.append([entry[h] for h in headers])
write_csv(out_csv, rows)
def insert_best_seqid_column(params):
csv = params['csv']
uniprot_ids_header = params['uniprot_ids_header']
delimiter = params['delimiter']
out_csv = params['output_csv']
uniprot_cache = params['cache_dir']
if not csv:
raise IOError('No file selected')
if not os.path.isfile(csv):
raise IOError(csv + ' not found')
headers = get_headers(csv)
if uniprot_ids_header not in headers:
s = "Column header '%s' not found, available headers:\n" % uniprot_ids_header
for header in headers:
s += ' ' + header + '\n'
raise IOError(s)
logging('Reading %s\n' % csv)
entries = read_csv(csv)
all_seqids = []
for entry in entries:
tokens = entry[uniprot_ids_header].split(delimiter)
entry['seqids'] = [s.strip() for s in tokens]
all_seqids.extend(entry['seqids'])
logging('Found %d potential Uniprot IDs\n' % len(all_seqids))
uniprot_data = uniprot.batch_uniprot_metadata(all_seqids, uniprot_cache)
for entry in entries:
best_seqid = uniprot.sort_seqids_by_uniprot(entry['seqids'],
uniprot_data)[0]
entry['best_seqid'] = best_seqid
entry['is_reviewed'] = False
if best_seqid in uniprot_data:
entry['is_reviewed'] = \
uniprot_data[best_seqid]['is_reviewed']
logging('Writing ')
logging('%s\n' % os.path.abspath(out_csv), lambda: open_file(out_csv))
headers = ['best_seqid', 'is_reviewed'] + get_headers(csv)
rows = [headers]
for entry in entries:
rows.append([entry[h] for h in headers])
write_csv(out_csv, rows)
def get_prot_seq(seq_id):
print seq_id
# just in case uniprot tries to sever the connection
try:
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
except:
try:
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
except:
try:
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
except:
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
prot_content = uniprot_data.get(seq_id)
if prot_content == None:
return None
sequence = str(prot_content.get('sequence'))
return sequence
def main():
file_rawdata = "data/MS_ISG15_Raw.tsv"
MS_dict = read_MS_rawdata(file_rawdata)
all_uniprot_ids = sorted(MS_dict.keys())
processed_ids = [
f.rsplit('_')[1].rsplit('.')[0] for f in glob.glob('Fasta/*.fa')
]
uniprot_ids = sorted(
list(set(all_uniprot_ids).difference(set(processed_ids))))
print "All IDs: %i" % len(all_uniprot_ids)
print "Processed IDs: %i" % len(processed_ids)
print "Remaining IDs: %i" % len(uniprot_ids)
uniprot_data = uniprot.batch_uniprot_metadata(uniprot_ids[0:150], 'cache')
parse_structure(uniprot_data)
parse_ortholog(uniprot_ids)
def get_uniprot_str(seq_id, just_gene=False):
# get all protein info
uniprot_data = uniprot.batch_uniprot_metadata([seq_id])
prot_content = uniprot_data.get(seq_id)
if prot_content == None:
return None
if not just_gene:
prot_content = json.dumps(prot_content.get('comment'))
else:
prot_content = json.dumps(prot_content.get('gene'))[1:-1]
# stripping out all '\n' carefully
prot_content = strip_slantn(prot_content)
return prot_content
def parse_uniprot_PDB(filename, MS_dict, dict_asa, outfile):
uniprot_id = filename.rsplit('_')[1].rsplit('.')[0]
uniprot_data = uniprot.batch_uniprot_metadata([uniprot_id], 'cache')
uniprot_seq = uniprot_data[uniprot_id]['sequence']
positions = list(set(MS_dict[uniprot_id]['Position']))
PDBs = [line.rstrip() for line in open(filename, 'r').readlines()]
for PDB in PDBs:
file_PDB_zip = 'PDB/' + PDB + '.pdb.gz'
file_PDB = 'PDB/' + PDB + '.pdb'
file_dssp = 'dssp/' + PDB + '.dssp'
fetchPDB(PDB)
if len(glob.glob(PDB + '.pdb.gz')) == 0: continue
os.system('mv ' + PDB + '.pdb.gz PDB/')
os.system('gunzip -f ' + file_PDB_zip)
os.system('/usr/local/Cellar/dssp/2.1.0/bin/mkdssp ' + file_PDB +
' > ' + file_dssp)
dict_dssp = reading_dssp(file_dssp, dict_asa)
atoms = parsePDB(file_PDB)
get_RSA_all_lys(atoms, dict_dssp, uniprot_id, PDB, outfile)
chain_ids = sorted(list(set(atoms.getChids())))
for chain_id in chain_ids:
chain_atoms = atoms.select('chain ' + chain_id)
top_aln = align_uniprot_PDB(uniprot_seq, atoms, chain_id)
if top_aln == 'NA':
print "No alignment"
continue
uniprot_seq_aln, pdb_seq_aln, score, begin, end = top_aln
for position in positions:
if position < int(begin): continue
if position > int(end): continue
pdb_resi, pdb_aa = identify_position(uniprot_seq_aln,
pdb_seq_aln, position,
uniprot_id)
if pdb_resi == 'NA': continue
pdb_resi = sorted(list(set(
chain_atoms.getResnums())))[pdb_resi - 1]
residueID = chain_id + '-' + str(pdb_resi)
if residueID not in dict_dssp.keys(): continue
RSA = dict_dssp[residueID]['RSA']
outfile.write("\t".join(
map(str, [
'ISG15', uniprot_id, position, PDB, chain_id, pdb_resi,
RSA
])) + "\n")
def protein_seq_update_celery_nofunction( full_batch = False ):
proteins = None # Protein.objects.extra(where=["CHAR_LENGTH(sequence) = 0"])
if full_batch:
proteins = Protein.objects.all()
else:
proteins = Protein.objects.extra(where=["CHAR_LENGTH(sequence) = 0"])
uniprot_data = uniprot.batch_uniprot_metadata( [ b.prot_id for b in proteins ] )
for key in uniprot_data.keys():
defaults = {}
try:
defaults['sequence'] = uniprot_data[key]['sequence']
except:
pass
try:
defaults['description'] = uniprot_data[key]['description']
except:
pass
prot, _ = Protein.objects.update_or_create( prot_id = key, defaults = defaults )
return 'Protein sequences updated'
def retrieve_uniprot_meta(uniprotIDs_file):
with open(uniprotIDs_file, 'r') as f:
id_list = [uni_id.strip() for uni_id in f]
meta_dict = uniprot.batch_uniprot_metadata(id_list)
return meta_dict
# Clean up caches
os.system('rm cache*')
# Example 1 - reading a fasta file
seqids, fastas = uniprot.read_fasta('example.fasta')
pprint.pprint(seqids, indent=2)
# Example 2 - map identifiers for RefSeq to Uniprot
seqids = "NP_000508.1 NP_001018081.3".split()
pairs = uniprot.batch_uniprot_id_mapping_pairs('P_REFSEQ_AC', 'ACC', seqids)
pprint.pprint(pairs, indent=2)
# Example 2 - get UniProt metadata
uniprot_seqids = [j for i, j in pairs]
uniprot_data = uniprot.batch_uniprot_metadata(uniprot_seqids, 'cache')
pprint.pprint(uniprot_data, indent=2)
# Example 3 - parse for isoforms in metadata
text = open('cache/metadata.0.txt').read()
uniprot_data = uniprot.parse_isoforms(text)
pprint.pprint(uniprot_data)
# Example 4 - chaining commands to map seqids
seqids = "EFG_MYCA1 YP_885981.1 ENSG00000196176 Q91ZU6-8".split()
uniprot_data = uniprot.get_metadata_with_some_seqid_conversions(
seqids, 'cache2')
pprint.pprint(uniprot_data, indent=2)
# Example 4 - chaining commands to map seqids
seqids = "EFG_MYCA1 YP_885981.1 ENSG00000196176 Q91ZU6-8".split()
# Example 3 - sequential identifier mapping to UniProt
# identifiers using robust but slow method
seqids = """
EFG_MYCA1 YP_885981.1 CpC231_1796
""".split()
mapping = uniprot.sequentially_convert_to_uniprot_id(seqids, "cache.json")
uniprot_seqids = mapping.values()
# Example 4 - get UniProt metadata
uniprot_data = uniprot.batch_uniprot_metadata(uniprot_seqids, "cache2.txt")
pprint.pprint(uniprot_data, indent=2)
for l in open("cache2.txt"):
print l.strip()
uniprot.write_fasta("example.output.fasta", uniprot_data, uniprot_seqids)
# Example 5 - chaining commands to make your own
# special mapper
def map_to_refseq(seqids):
uniprot_mapping = uniprot.sequentially_convert_to_uniprot_id(seqids, "func.cache.json")
# Load hand-curated CycD putative target list
filename = '/data/cycd_targets/cycd_target_uniprot.txt'
targetIDs = pd.read_csv(filename)
already_seen = pd.concat((targetIDs['Entry'], entries))
# Load hit list from PSSM
filename = '/data/cycd_targets/hsap_proteome/hsap_hits>20.csv'
targetIDs = pd.read_csv(filename, sep='\t')
entries = targetIDs['Entry']
# Do a merge to see what's not already seen in the hand-curated list
merged = set(entries) - set(already_seen)
# Fetch and write as FASTA
out_name = '/data/cycd_targets/hsap_hits>20.fasta'
upData = uniprot.batch_uniprot_metadata(merged, 'cache')
uniprot.write_fasta(out_name, upData, merged)
split_fastas(out_name)
PSIPRED_DIR = '/data/cycd_targets/cycd_target_uniprot_individuals'
seqs = []
for filename in os.listdir(PSIPRED_DIR):
if filename.endswith('.ss2'):
print 'Working on ', filename
#Load PSIPRED VFORMAT in a sane way to extract only relevant info
df = pd.read_csv(os.path.join(PSIPRED_DIR, filename),
header=0,
delim_whitespace=True,
# Clean up caches
os.system('rm cache*')
# Example 1 - reading a fasta file
seqids, fastas = uniprot.read_fasta('example.fasta')
pprint.pprint(seqids, indent=2)
# Example 2 - map identifiers for RefSeq to Uniprot
seqids = "NP_000508.1 NP_001018081.3".split()
pairs = uniprot.batch_uniprot_id_mapping_pairs(
'P_REFSEQ_AC', 'ACC', seqids)
pprint.pprint(pairs, indent=2)
# Example 2 - get UniProt metadata
uniprot_seqids = [j for i,j in pairs]
uniprot_data = uniprot.batch_uniprot_metadata(
uniprot_seqids, 'cache')
pprint.pprint(uniprot_data, indent=2)
# Example 3 - parse for isoforms in metadata
text = open('cache/metadata.0.txt').read()
uniprot_data = uniprot.parse_isoforms(text)
pprint.pprint(uniprot_data)
# Example 4 - chaining commands to map seqids
seqids = "EFG_MYCA1 YP_885981.1 ENSG00000196176 Q91ZU6-8".split()
uniprot_data = uniprot.get_metadata_with_some_seqid_conversions(
seqids, 'cache2')
pprint.pprint(uniprot_data, indent=2)
# Example 4 - chaining commands to map seqids
seqids = "EFG_MYCA1 YP_885981.1 ENSG00000196176 Q91ZU6-8".split()
