def _build_uniprot_sequences():
seq_file = resource_manager.get_create_resource_file('swissprot',
cached=True)
iso_file = resource_manager.get_create_resource_file('isoforms',
cached=True)
logger.info("Loading Swissprot sequences...")
sp_seq = load_fasta_sequences(seq_file)
logger.info("Loading Uniprot isoform sequences...")
iso_seq = load_fasta_sequences(iso_file)
sp_seq.update(iso_seq)
return sp_seq
def _build_uniprot_entries():
up_entries_file = resource_manager.get_create_resource_file('up')
sc_entries_file = resource_manager.get_create_resource_file('up_sars_cov2')
uniprot_gene_name = {}
uniprot_mnemonic = {}
uniprot_mnemonic_reverse = {}
uniprot_mgi = {}
uniprot_rgd = {}
uniprot_mgi_reverse = {}
uniprot_rgd_reverse = {}
uniprot_length = {}
uniprot_features = {}
uniprot_reviewed = set()
files = [up_entries_file, sc_entries_file]
for file in files:
with open(file, 'r') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
for row in csv_rows:
up_id, gene_name, up_mnemonic, rgd, mgi, length, reviewed, \
features_json = row
# Store the entry in the reviewed set
if reviewed == 'reviewed':
uniprot_reviewed.add(up_id)
uniprot_gene_name[up_id] = gene_name
uniprot_mnemonic[up_id] = up_mnemonic
uniprot_mnemonic_reverse[up_mnemonic] = up_id
uniprot_length[up_id] = int(length)
if mgi:
mgi_ids = mgi.split(';')
if mgi_ids:
uniprot_mgi[up_id] = mgi_ids[0]
uniprot_mgi_reverse[mgi_ids[0]] = up_id
if rgd:
rgd_ids = rgd.split(';')
if rgd_ids:
uniprot_rgd[up_id] = rgd_ids[0]
uniprot_rgd_reverse[rgd_ids[0]] = up_id
uniprot_features[up_id] = [
feature_from_json(feat)
for feat in json.loads(features_json)
]
# Build a dict of features by feature ID
features_by_id = {}
for up_id, feats in uniprot_features.items():
for feat in feats:
features_by_id[feat.id] = feat
return (uniprot_gene_name, uniprot_mnemonic, uniprot_mnemonic_reverse,
uniprot_mgi, uniprot_rgd, uniprot_mgi_reverse, uniprot_rgd_reverse,
uniprot_length, uniprot_reviewed, uniprot_features, features_by_id)
def _build_hgnc_mappings():
hgnc_file = resource_manager.get_create_resource_file('hgnc')
with gzip.open(hgnc_file, 'rt', encoding='utf-8') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
hgnc_name_to_id = {}
hgnc_id_to_up = {}
up_to_hgnc_id = {}
entrez_to_up = {}
up_to_entrez = {}
for row in csv_rows:
hgnc_id = row[0][5:]
hgnc_status = row[3]
if hgnc_status == 'Approved':
hgnc_name = row[1]
hgnc_name_to_id[hgnc_name] = hgnc_id
# Uniprot
uniprot_id = row[6]
if uniprot_id:
hgnc_id_to_up[hgnc_id] = uniprot_id
uniprot_ids = uniprot_id.split(', ')
for upid in uniprot_ids:
up_to_hgnc_id[upid] = hgnc_id
# Entrez
entrez_id = row[5]
if entrez_id:
for upid in uniprot_ids:
up_to_entrez[upid] = entrez_id
entrez_to_up[entrez_id] = uniprot_id
return hgnc_name_to_id, hgnc_id_to_up, up_to_hgnc_id, \
entrez_to_up, up_to_entrez
def _build_human_mouse_rat():
hgnc_file = resource_manager.get_create_resource_file('hgnc')
with gzip.open(hgnc_file, 'rt', encoding='utf-8') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
uniprot_mouse = {}
uniprot_rat = {}
for row in csv_rows:
human_id, mgi_id, rgd_id = row[6:9]
if human_id:
if mgi_id:
mgi_id = mgi_id.split(', ')[0]
if mgi_id.startswith('MGI:'):
mgi_id = mgi_id[4:]
mouse_id = um.uniprot_mgi_reverse.get(mgi_id)
if mouse_id:
uniprot_mouse[human_id] = mouse_id
if rgd_id:
rgd_id = rgd_id.split(', ')[0]
if rgd_id.startswith('RGD:'):
rgd_id = rgd_id[4:]
rat_id = um.uniprot_rgd_reverse.get(rgd_id)
if rat_id:
uniprot_rat[human_id] = rat_id
return uniprot_mouse, uniprot_rat
def _build_refseq_uniprot():
refseq_uniprot_file = resource_manager.get_create_resource_file(
'refseq_uniprot')
refseq_up = {}
with gzip.open(refseq_uniprot_file, 'rt', encoding='utf-8') as f:
csvreader = csv.reader(f)
for refseq_id, up_id in csvreader:
if refseq_id not in refseq_up:
refseq_up[refseq_id] = []
refseq_up[refseq_id].append(up_id)
return refseq_up
def _build_uniprot_entries():
up_entries_file = resource_manager.get_create_resource_file('up')
uniprot_gene_name = {}
uniprot_mnemonic = {}
uniprot_mnemonic_reverse = {}
uniprot_mgi = {}
uniprot_rgd = {}
uniprot_mgi_reverse = {}
uniprot_rgd_reverse = {}
uniprot_length = {}
uniprot_signal_peptide = {}
uniprot_reviewed = set()
with open(up_entries_file, 'r') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
for row in csv_rows:
up_id, gene_name, up_mnemonic, rgd, mgi, length, reviewed, \
signal_peptide = row
# Store the entry in the reviewed set
if reviewed == 'reviewed':
uniprot_reviewed.add(up_id)
uniprot_gene_name[up_id] = gene_name
uniprot_mnemonic[up_id] = up_mnemonic
uniprot_mnemonic_reverse[up_mnemonic] = up_id
uniprot_length[up_id] = int(length)
if mgi:
mgi_ids = mgi.split(';')
if mgi_ids:
uniprot_mgi[up_id] = mgi_ids[0]
uniprot_mgi_reverse[mgi_ids[0]] = up_id
if rgd:
rgd_ids = rgd.split(';')
if rgd_ids:
uniprot_rgd[up_id] = rgd_ids[0]
uniprot_rgd_reverse[rgd_ids[0]] = up_id
uniprot_signal_peptide[up_id] = (None, None)
if signal_peptide:
match = re.match(r'SIGNAL (\d+) (\d+) ', signal_peptide)
if match:
beg_pos, end_pos = match.groups()
uniprot_signal_peptide[up_id] = \
(int(beg_pos), int(end_pos))
return (uniprot_gene_name, uniprot_mnemonic, uniprot_mnemonic_reverse,
uniprot_mgi, uniprot_rgd, uniprot_mgi_reverse, uniprot_rgd_reverse,
uniprot_length, uniprot_reviewed, uniprot_signal_peptide)
def _build_uniprot_sec():
# File containing secondary accession numbers mapped
# to primary accession numbers
sec_file = resource_manager.get_create_resource_file('upsec')
uniprot_sec = {}
lines = open(sec_file, 'rt').readlines()
for i, l in enumerate(lines):
if l.startswith('Secondary AC'):
entry_lines = lines[i + 2:]
for l in entry_lines:
sec_id, prim_id = l.split()
try:
uniprot_sec[sec_id].append(prim_id)
except KeyError:
uniprot_sec[sec_id] = [prim_id]
return uniprot_sec
def _get_phospho_site_dataset():
"""Read phosphosite data into dicts keyed by Uniprot ID and by site group.
Returns
-------
tuple
The first element of the tuple contains the PhosphoSite data keyed
by Uniprot ID, the second element contains data keyed by site group.
Both dicts have instances of the PhosphoSite namedtuple as values.
If the PhosphoSite data file cannot be loaded, returns (None, None).
"""
global _data_by_up
global _data_by_site_grp
phosphosite_data_file = resource_manager.get_create_resource_file('psp')
if _data_by_up is None or _data_by_site_grp is None:
with open(phosphosite_data_file, 'r') as fh:
# Get the csv reader generator
reader = csv.reader(fh, delimiter='\t')
# Skip 4 rows
for _ in range(4):
next(reader)
# Build up a dict by protein
data_by_up = defaultdict(lambda: defaultdict(list))
data_by_site_grp = defaultdict(list)
for row in reader:
site = PhosphoSite(*row)
res_pos = site.MOD_RSD.split('-')[0]
#res_pos = res_pos[1:] # DANGEROUS: lookup based on pos alone
base_acc_id = site.ACC_ID.split('-')[0]
data_by_up[site.ACC_ID][res_pos].append(site)
# If the ID was isoform specific, add to the dict for the whole
# protein
if base_acc_id != site.ACC_ID:
data_by_up[base_acc_id][res_pos].append(site)
# Catch the handful of isoforms that have a Uniprot ID without
# the hyphen
elif site.ACC_ID in _iso_to_ref_map:
ref_id = _iso_to_ref_map[site.ACC_ID]
data_by_up[ref_id][res_pos].append(site)
# To catch additional cases, include an entry for the -1 base ID
else:
data_by_up['%s-1' % base_acc_id] = data_by_up[base_acc_id]
data_by_site_grp[site.SITE_GRP_ID].append(site)
_data_by_up = data_by_up
_data_by_site_grp = data_by_site_grp
return (_data_by_up, _data_by_site_grp)
def _build_hgnc_mappings():
hgnc_file = resource_manager.get_create_resource_file('hgnc')
with open(hgnc_file, 'r') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
hgnc_ids = {}
uniprot_ids = {}
for row in csv_rows:
hgnc_id = row[0][5:]
hgnc_status = row[3]
if hgnc_status == 'Approved':
hgnc_name = row[1]
hgnc_ids[hgnc_name] = hgnc_id
# Uniprot
uniprot_id = row[6]
uniprot_ids[hgnc_id] = uniprot_id
return hgnc_ids, uniprot_ids
def _build_uniprot_entries():
up_entries_file = resource_manager.get_create_resource_file('up')
uniprot_gene_name = {}
uniprot_mnemonic = {}
uniprot_mnemonic_reverse = {}
uniprot_mgi = {}
uniprot_rgd = {}
uniprot_mgi_reverse = {}
uniprot_rgd_reverse = {}
uniprot_length = {}
uniprot_features = {}
uniprot_reviewed = set()
organisms_by_id = {}
uniprot_entrez = {}
uniprot_entrez_reverse = {}
files = [up_entries_file]
mgi_name_to_up = {}
rgd_name_to_up = {}
for file in files:
with gzip.open(file, 'rt', encoding='utf-8') as fh:
csv_rows = csv.reader(fh, delimiter='\t')
# Skip the header row
next(csv_rows)
for row in csv_rows:
up_id, gene_name, up_mnemonic, rgd, mgi, length, reviewed, \
organism_id, entrez_id, features_json = row
# Store the entry in the reviewed set
if reviewed == 'reviewed':
uniprot_reviewed.add(up_id)
# This is to turn empty strings into explicit Nones
uniprot_gene_name[up_id] = gene_name if gene_name else None
uniprot_mnemonic[up_id] = up_mnemonic
uniprot_mnemonic_reverse[up_mnemonic] = up_id
uniprot_length[up_id] = int(length)
if mgi:
mgi_ids = mgi.split(';')
if mgi_ids:
uniprot_mgi[up_id] = mgi_ids[0]
uniprot_mgi_reverse[mgi_ids[0]] = up_id
mgi_name_to_up[gene_name] = up_id
if rgd:
rgd_ids = rgd.split(';')
if rgd_ids:
uniprot_rgd[up_id] = rgd_ids[0]
uniprot_rgd_reverse[rgd_ids[0]] = up_id
rgd_name_to_up[gene_name] = up_id
uniprot_features[up_id] = [feature_from_json(feat) for
feat in json.loads(features_json)]
organisms_by_id[up_id] = organism_id
# Entrez mappings
entrez_ids = [ei for ei in
[e.strip() for e in entrez_id.split(';')] if ei]
for eid in entrez_ids:
uniprot_entrez[up_id] = eid
uniprot_entrez_reverse[eid] = up_id
# Build a dict of features by feature ID
features_by_id = {}
for up_id, feats in uniprot_features.items():
for feat in feats:
features_by_id[feat.id] = feat
return (uniprot_gene_name, uniprot_mnemonic, uniprot_mnemonic_reverse,
uniprot_mgi, uniprot_rgd, uniprot_mgi_reverse, uniprot_rgd_reverse,
uniprot_length, uniprot_reviewed, uniprot_features, features_by_id,
organisms_by_id, uniprot_entrez, uniprot_entrez_reverse,
mgi_name_to_up, rgd_name_to_up)
def _build_refseq_sequences():
seq_file = resource_manager.get_create_resource_file('refseq_seq',
cached=True)
logger.info("Loading RefSeq protein sequences...")
seq = load_fasta_sequences(seq_file, id_delimiter=' ', id_index=0)
return seq
