def annotate(df, cols):
""" get longest transcript for gene column or regions """
def get_regions(row, cols):
prefix = '' if row[cols[0]].startswith('chr') else 'chr'
return "{}{}:{}-{}".format(prefix, row[cols[0]], row[cols[1]],
row[cols[2]])
llama = LlamaEnsembl()
ucsc = UCSCapi()
dd = {'query': [], 'transcripts': []}
if len(cols) == 1:
genes = df[cols[0]].values
for gene in genes:
chrom, start, end = llama.get_gene_pos(gene)
res = ucsc.query("chr{}:{}-{}".format(chrom, start, end))
dd['query'].append(gene)
dd['transcripts'].append(res.longest(gene)['transcript'])
mergecol = cols[0]
elif len(cols) == 3:
dd['genes'] = []
df['region'] = df.apply(get_regions, axis=1, args=[cols])
regions = df['region'].values
for region in regions:
res = ucsc.query(region)
for gene in res.genes():
dd['genes'].append(gene)
dd['query'].append(region)
dd['transcripts'].append(res.longest(gene)['transcript'])
mergecol = 'region'
ndf = df.merge(pd.DataFrame(dd),
left_on=mergecol,
right_on='query',
how='outer')
return ndf[~ndf.duplicated()]
def test_gene_transcripts():
llama = LlamaEnsembl()
ucsc = UCSCapi()
gene = 'BRCA1'
chrom, start, end = llama.get_gene_pos(gene)
res = ucsc.query("chr{}:{}-{}".format(chrom, start, end))
transcript = res.longest(gene)
print(transcript['transcript'])