GenoTEX / download /get_gene_synonyms.py
Liu-Hy's picture
Add files using upload-large-folder tool
32677ff verified
raw
history blame
2.57 kB
import gzip
import json
def build_gene_info_dict(gene_info_path, tax_id='9606'):
"""
Return a dict of:
gene_id -> {
"official_symbol": str,
"synonyms": set of str
}
"""
gene_dict = {}
with gzip.open(gene_info_path, 'rt') as f:
for line in f:
if line.startswith("#"):
continue
cols = line.strip().split('\t')
# typical columns: tax_id, GeneID, Symbol, LocusTag, Synonyms, ...
if len(cols) < 6:
continue
tid, gene_id, official_symbol, locus_tag, synonyms_str = cols[:5]
if tid != tax_id:
continue
synonyms = set(synonyms_str.split('|')) if synonyms_str else set()
gene_dict[gene_id] = {
"official_symbol": official_symbol,
"synonyms": synonyms
}
return gene_dict
def build_synonym_dict(gene_info_path):
"""
From gene_id -> {offical_symbol, synonyms},
build a dictionary from uppercase alias -> official symbol,
avoiding collisions that overwrite official symbols of different genes.
"""
alias2official = {}
gene_dict = build_gene_info_dict(gene_info_path)
# Sort keys so we always process in a consistent order (optional).
# But for big data, might skip sorting to save time.
for gid, info in gene_dict.items():
official_symbol = info["official_symbol"]
synonyms = info["synonyms"]
# Combine synonyms and the official symbol itself
# so that official symbols also map to themselves
all_aliases = synonyms.union({official_symbol})
for alias in all_aliases:
alias_upper = alias.upper()
# If the alias is already taken as an official symbol for a different gene,
# check if there's a conflict:
if alias_upper in alias2official and alias2official[alias_upper] != official_symbol\
and alias2official[alias_upper] == alias_upper:
# We have discovered the same alias points to two distinct official symbols.
# If the symbol was previously identified as an official symbol, skip overwriting.
continue
else:
alias2official[alias_upper] = official_symbol
return alias2official
if __name__ == "__main__":
synonym_dict = build_synonym_dict("/home/techt/Downloads/gene_info.gz")
with open("../metadata/gene_synonym.json", "w") as f:
json.dump(synonym_dict, f)