def annotate_fusion(ref_f, input_f, output_f):
"""
Align fusion juncrions to gene annotations
"""
print('Start to annotate fusion junctions...')
genes, gene_info = parse_ref1(ref_f) # gene annotations
fusions, fusion_index = parse_bed(input_f) # fusion junctions
total = set()
with open(output_f, 'w') as outf:
for chrom in genes:
# overlap gene annotations with fusion juncrions
result = Interval.overlapwith(genes[chrom].interval,
fusions[chrom])
for itl in result:
# extract gene annotations
iso = list(filter(lambda x: x.startswith('iso'), itl[2:]))
# for each overlapped fusion junction
for fus in itl[(2 + len(iso)):]:
reads = fus.split()[1]
fus_start, fus_end = fusion_index[fus]
edge_annotations = [] # first or last exon flag
for iso_id in iso:
g, i, c, s = iso_id.split()[1:]
start = gene_info[iso_id][0][0]
end = gene_info[iso_id][-1][-1]
# fusion junction excesses boundaries of gene
# annotation
if fus_start < start - 10 or fus_end > end + 10:
continue
(fusion_info,
index,
edge) = map_fusion_to_iso(fus_start,
fus_end, s,
gene_info[iso_id])
if fusion_info:
fus_start_str = str(fus_start)
fus_end_str = str(fus_end)
bed_info = '\t'.join([chrom, fus_start_str,
fus_end_str,
'FUSIONJUNC/%s' % reads,
'0', s, fus_start_str,
fus_start_str, '0,0,0'])
bed = '\t'.join([bed_info, fusion_info, g, i,
index])
if not edge: # not first or last exon
outf.write(bed + '\n')
total.add(fus)
else: # first or last exon
edge_annotations.append(bed)
if edge_annotations: # first or last exon
for bed in edge_annotations:
outf.write(bed + '\n')
total.add(fus)
print('Annotated %d fusion junctions!' % len(total))
def testOverlapwith(self):
r = Interval.overlapwith(self.c, self.d)
self.assertListEqual(r, [[3, 7, 'I', 'a', 'b', 'e', 'd', 'f'], [10, 12, 'II', 'd', 'x'], [16, 20, 'III', 'x', 'h', 'i'],
[23, 25, 'IV', 'x']], 'Failed in Overlapwith')