def transcripts_from_gtf_lines(lines, attr_defs=None):
transcripts = collections.OrderedDict()
for feature in GTFFeature.parse(lines, attr_defs):
# feature = GTFFeature.parse(line, attr_defs)
# skip gene annotation in gtf files
if "transcript_id" not in feature.attrs:
continue
t_id = feature.attrs["transcript_id"]
if t_id not in transcripts:
if feature.feature_type != "transcript":
raise GTFError(
"Feature type '%s' found before 'transcript' record: %s" %
(feature.feature_type, str(feature)))
t = Transcript()
t.chrom = feature.seqid
t.start = feature.start
t.end = feature.end
# convert from string strand notation ("+", "-", ".")
# to integer (0, 1)
t.strand = strand_str_to_int(feature.strand)
t.exons = []
t.attrs = feature.attrs
transcripts[t_id] = t
else:
t = transcripts[t_id]
if feature.feature_type == "exon":
t.exons.append(Exon(feature.start, feature.end))
# sort transcript exons by genomic position
for t in transcripts.itervalues():
t.exons.sort()
return transcripts.values()
def to_formatted_gtf(lines, gtf_file, attr_defs=None):
transcripts = collections.OrderedDict()
for line in lines:
feature = GTFFeature.from_string(line, attr_defs)
# skip gene annotation in gtf files
if "transcript_id" not in feature.attrs:
continue
t_id = feature.attrs["transcript_id"]
# extract exon information
if t_id not in transcripts:
if feature.feature_type == "exon":
t = Transcript()
t.chrom = feature.seqid
t.start = feature.start
t.end = feature.end
t.strand = strand_str_to_int(feature.strand)
t.exons = [Exon(feature.start, feature.end)]
t.attrs = dict()
for each_attr in feature.attrs:
if 'exon' not in each_attr.lower():
t.attrs.update({each_attr: feature.attrs[each_attr]})
transcripts[t_id] = t
else:
t.start = t.start if t.start <= feature.start else feature.start
t.end = t.end if t.end >= feature.end else feature.end
t.exons.append(Exon(feature.start, feature.end))
with open(gtf_file, 'w') as gtf_output:
for each_tr in transcripts:
each_tr_obj = transcripts[each_tr]
for each_feature in each_tr_obj.to_gtf_features():
gtf_output.write(
'{gtf_line}\n'.format(gtf_line=str(each_feature)))
def transcripts_from_gtf_lines(lines, attr_defs=None):
transcripts = collections.OrderedDict()
for line in lines:
feature = GTFFeature.from_string(line, attr_defs)
t_id = feature.attrs["transcript_id"]
if t_id not in transcripts:
if feature.feature_type != "transcript":
raise GTFError("Feature type '%s' found before 'transcript' record: %s" %
(feature.feature_type, str(feature)))
t = Transcript()
t.chrom = feature.seqid
t.start = feature.start
t.end = feature.end
# convert from string strand notation ("+", "-", ".")
# to integer (0, 1)
t.strand = strand_str_to_int(feature.strand)
t.exons = []
t.attrs = feature.attrs
transcripts[t_id] = t
else:
t = transcripts[t_id]
if feature.feature_type == "exon":
t.exons.append(Exon(feature.start, feature.end))
# sort transcript exons by genomic position
for t in transcripts.itervalues():
t.exons.sort()
return transcripts.values()
def to_gtf_features(self, source=None, score=1000):
if source is None:
source = 'assemblyline'
# transcript feature
f = GTFFeature()
f.seqid = self.chrom
f.source = source
f.feature_type = 'transcript'
f.start = self.start
f.end = self.end
f.score = score
f.strand = strand_int_to_str(self.strand)
f.phase = '.'
f.attrs = self.attrs
features = [f]
# exon features
for i, e in enumerate(self.exons):
f = GTFFeature()
f.seqid = self.chrom
f.source = source
f.feature_type = 'exon'
f.start = e.start
f.end = e.end
f.score = score
f.strand = strand_int_to_str(self.strand)
f.phase = '.'
f.attrs = self.attrs.copy()
f.attrs["exon_number"] = i
features.append(f)
return features
def to_gtf_features(self, source=None, score=1000):
if source is None:
source = 'assemblyline'
# transcript feature
f = GTFFeature()
f.seqid = self.chrom
f.source = source
f.feature_type = 'transcript'
f.start = self.start
f.end = self.end
f.score = score
f.strand = strand_int_to_str(self.strand)
f.phase = '.'
f.attrs = self.attrs
features = [f]
# exon features
for i,e in enumerate(self.exons):
f = GTFFeature()
f.seqid = self.chrom
f.source = source
f.feature_type = 'exon'
f.start = e.start
f.end = e.end
f.score = score
f.strand = strand_int_to_str(self.strand)
f.phase = '.'
f.attrs = self.attrs.copy()
f.attrs["exon_number"] = i
features.append(f)
return features
