def main():
sampletable = (download_geo_metadata().pipe(update_fbgns).pipe(
update_gene_symbols).pipe(reorder_columns))
out_file = PROJECT_DIR / "timecourse-aln-wf/config/sampletable.tsv"
logger.info(f"Writing out sample table: {out_file}")
sampletable.to_csv(out_file, sep="\t", index=False)
def download_geo_metadata() -> pd.DataFrame:
logger.info("Querying GEO for {}".format(GSE))
tmpDir = TemporaryDirectory()
gse = GEOparse.get_GEO(GSE, destdir=tmpDir.name, silent=True)
## Pull out sample attributes and build data frame
attributes = []
for gsm, dat in gse.gsms.items():
try:
# Parse sample title
attrs = re.match(
r"^DRSC_(?P<plate_id>plate\d)_(?P<time_point>day\d)_(?P<plate_row>[A-H])(?P<plate_column>\d+)_(?P<drsc>DRSC(\d+|NA))_(?P<fbgn>(FBgn\d+|NA))_(?P<symbol>.*?)$",
dat.metadata["title"][0],
).groupdict()
attrs["GEO"] = gsm
# Get SRX accession from linkout
for x in dat.metadata["relation"]:
match = re.match(r"(\w+):.*[\/=](\w+\d+)$", x)
if match:
k, v = match.groups()
attrs[k] = v
# Expand out SRR accessions
for srr in get_srrs(attrs["SRA"]):
attrs["samplename"] = srr
attrs["Run"] = srr
attributes.append(attrs)
except AttributeError:
print(gsm, dat.metadata["title"])
df = pd.DataFrame(attributes).rename(columns={
"SRA": "SRX"
}).set_index(["samplename", "SRX"])
return df
def get_current_flybase_annotations():
# Import config set up references
logger.info("Loading config: {}".format(CONFIG))
with open(CONFIG) as fh:
config = yaml.load(fh)
assembly = config["assembly"]
tag = config["aligner"]["tag"]
REF = os.path.join(os.environ["REFERENCES_DIR"], assembly, tag)
# load flybase annotations
FB_ANNO = os.path.join(REF,
"fb_annotation/dmel_{}.fb_annotation".format(tag))
logger.info("Loading FlyBase annotation file: {}".format(FB_ANNO))
fb = pd.read_table(FB_ANNO)[[
"primary_FBgn", "gene_symbol", "secondary_FBgn"
]]
fb.rename(columns={
"primary_FBgn": "FBgn",
"gene_symbol": "symbol"
},
inplace=True)
## Make map of old fbgn to current fbgn and current fbgn to current symbol
fbgns = {}
genes = {}
for i, record in fb.iterrows():
fbgn = record.FBgn
symbol = record.symbol
fbgn2 = record.secondary_FBgn
fbgns[fbgn] = fbgn
genes[fbgn] = symbol
if isinstance(fbgn2, str):
for f2 in fbgn2.strip().split(","):
fbgns[f2] = fbgn
from s2rnai.logger import logger
def srr_iter(srx):
"""Create generator to return SRR given an SRX."""
res = Entrez.efetch(db='sra', id=srx)
xml = res.read()
root = ElementTree.fromstring(xml)
for run in root.iter('RUN'):
yield run.get('accession')
res.close()
# Import config set up references
logger.info('Loading config: {}'.format(CONFIG))
with open(CONFIG) as fh:
config = yaml.load(fh)
assembly = config['assembly']
tag = config['aligner']['tag']
REF = os.path.join(os.environ['REFERENCES_DIR'], assembly, tag)
# load flybase annotations
FB_ANNO = os.path.join(REF, 'fb_annotation/dmel_{}.fb_annotation'.format(tag))
logger.info('Loading FlyBase annotation file: {}'.format(FB_ANNO))
fb = pd.read_table(FB_ANNO)[['primary_FBgn', 'gene_symbol', 'secondary_FBgn']]
fb.rename(
columns={
'primary_FBgn': 'FBgn',
'gene_symbol': 'symbol'
if __name__ == '__main__':
# Import config and set up references
with open('../config/config.yml') as fh:
config = yaml.load(fh)
assembly = config['assembly']
tag = config['aligner']['tag']
REF = os.path.join(os.environ['REFERENCES_DIR'], assembly, tag)
# Download gff file from FlyBase
url = config['references'][config['assembly']][
config['gtf']['tag']]['gtf']['url'].replace('gtf', 'gff')
fname = os.path.join('../data/external/FlyBase', os.path.basename(url))
if not os.path.exists(fname):
logger.info('Createing FlyBase directory.')
os.makedirs(os.path.dirname(fname), exist_ok=True)
with open(fname, 'wb') as fh:
logger.info('Downloading GFF')
response = urllib.request.urlretrieve(url, fname)
# Output filename to store drsc gff info
drsc_fname = fname.replace('.gff.gz', '.drsc.gff')
# use grep to filter our DRSC, this is the fastest way
if not os.path.exists(drsc_fname):
logger.info('Filtering GFF')
cmd = 'gunzip -c {fname} | grep "DRSC_dsRNA" | grep "RNAi_reagent" > {drsc_fname}'.format(
fname=fname, drsc_fname=drsc_fname)
subprocess.run(cmd, shell=True)