def test_monoexonic(self):
exon = self.gff_lines[1]
transcript_line = self.gff_lines[0]
transcript_line.end = exon.end
model = Transcript(transcript_line)
model.add_exon(exon)
model.finalize()
fasta = self.fasta[model.chrom][model.start - 1:model.end]
tcheck = TranscriptChecker(model.copy(), fasta, strand_specific=False)
tcheck.check_strand()
self.assertIsNone(tcheck.strand)
tcheck = TranscriptChecker(model.copy(), fasta, strand_specific=True)
tcheck.check_strand()
self.assertEqual(tcheck.strand, "+")
neg = model.copy()
neg.strand = "-"
tcheck = TranscriptChecker(neg.copy(), fasta, strand_specific=False)
tcheck.check_strand()
self.assertIsNone(tcheck.strand)
tcheck = TranscriptChecker(neg.copy(), fasta, strand_specific=True)
tcheck.check_strand()
self.assertEqual(tcheck.strand, "-")
def test_monoexonic(self):
exon = self.gff_lines[1]
transcript_line = self.gff_lines[0]
transcript_line.end = exon.end
model = Transcript(transcript_line)
model.add_exon(exon)
model.finalize()
fasta = self.fasta[model.chrom][model.start - 1: model.end]
tcheck = TranscriptChecker(model.copy(), fasta, strand_specific=False)
tcheck.check_strand()
self.assertIsNone(tcheck.strand)
tcheck = TranscriptChecker(model.copy(), fasta, strand_specific=True)
tcheck.check_strand()
self.assertEqual(tcheck.strand, "+")
neg = model.copy()
neg.strand = "-"
tcheck = TranscriptChecker(neg.copy(), fasta, strand_specific=False)
tcheck.check_strand()
self.assertIsNone(tcheck.strand)
tcheck = TranscriptChecker(neg.copy(), fasta, strand_specific=True)
tcheck.check_strand()
self.assertEqual(tcheck.strand, "-")
def test_monoexonic_suspicious(self):
"""A monoexonic transcript should never appear as a suspicious transcript, in terms of splicing."""
exon = self.gff_lines[1]
transcript_line = self.gff_lines[0]
transcript_line.end = exon.end
model = Transcript(transcript_line)
model.add_exon(exon)
model.finalize()
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
del model.attributes["mixed_splices"]
self.assertFalse(model.suspicious_splicing)
model.attributes["canonical_number"] = 0
self.assertFalse(model.suspicious_splicing)
del model.attributes["canonical_number"]
model.attributes["canonical_on_reverse_strand"] = True
self.assertFalse(model.suspicious_splicing)
model.attributes["canonical_on_reverse_strand"] = False
self.assertFalse(model.suspicious_splicing)
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
del model.attributes["mixed_splices"]
del model.attributes["canonical_on_reverse_strand"]
model.attributes["canonical_number"] = 0
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
model.attributes["canonical_on_reverse_strand"] = True
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
del model.attributes["canonical_on_reverse_strand"]
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
def test_monoexonic_suspicious(self):
"""A monoexonic transcript should never appear as a suspicious transcript, in terms of splicing."""
exon = self.gff_lines[1]
transcript_line = self.gff_lines[0]
transcript_line.end = exon.end
model = Transcript(transcript_line)
model.add_exon(exon)
model.finalize()
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
del model.attributes["mixed_splices"]
self.assertFalse(model.suspicious_splicing)
model.attributes["canonical_number"] = 0
self.assertFalse(model.suspicious_splicing)
del model.attributes["canonical_number"]
model.attributes["canonical_on_reverse_strand"] = True
self.assertFalse(model.suspicious_splicing)
model.attributes["canonical_on_reverse_strand"] = False
self.assertFalse(model.suspicious_splicing)
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
del model.attributes["mixed_splices"]
del model.attributes["canonical_on_reverse_strand"]
model.attributes["canonical_number"] = 0
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
model.attributes["canonical_on_reverse_strand"] = True
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
del model.attributes["canonical_on_reverse_strand"]
model.attributes["mixed_splices"] = "6positive,1negative"
self.assertFalse(model.suspicious_splicing)
self.assertFalse(model.only_non_canonical_splicing)
def launch(args):
"""
Simple launcher script.
:param args: the argparse Namespace
"""
if hasattr(args, "region") and args.region is not None:
try:
args.chrom, args.start, args.end = args.region
except ValueError as exc:
raise ValueError("{0} {1}".format(exc, args.region))
if args.start >= args.end:
raise ValueError("Start greater than end: {0}\t{1}".format(
args.start, args.end))
transcript = None
with GTF(args.gtf) as gtf:
for row in gtf:
if row.chrom != args.chrom:
continue
else:
if row.is_transcript is True:
if transcript is not None and \
transcript.start >= args.start and transcript.end <= args.end:
print(transcript.format("gtf"), file=args.out)
transcript = None
if args.assume_sorted is True and row.start > args.end:
break
transcript = Transcript(row)
else:
transcript.add_exon(row)
if transcript is not None and transcript.start >= args.start and transcript.end <= args.end:
print(transcript.format("gtf"), file=args.out)
def parse_prediction(args, genes, positions, queue_logger):
"""
This function performs the real comparison between the reference and the prediction.
It needs the following inputs:
:param args: the Namespace with the necessary parameters
:param genes: Dictionary with the reference genes, of the form
dict[chrom][(start,end)] = [gene object]
:param positions: Dictionary with the positions of the reference genes, of the form
dict[chrom][IntervalTree]
:param queue_logger: Logger
:return:
"""
# start the class which will manage the statistics
accountant_instance = Accountant(genes, args)
assigner_instance = Assigner(genes, positions, args, accountant_instance)
transcript = None
if hasattr(args, "self") and args.self is True:
args.prediction = to_gff(args.reference.name)
ref_gff = isinstance(args.prediction, GFF3)
__found_with_orf = set()
for row in args.prediction:
if row.header is True:
continue
# queue_logger.debug("Row:\n{0:>20}".format(str(row)))
if row.is_transcript is True or row.feature == "match":
queue_logger.debug("Transcript row:\n%s", str(row))
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id):
__name = re.sub(r"\.orf[0-9]+$", "", transcript.id)
if __name not in __found_with_orf:
__found_with_orf.add(__name)
assigner_instance.get_best(transcript)
else:
pass
else:
assigner_instance.get_best(transcript)
transcript = Transcript(row, logger=queue_logger)
elif row.is_exon is True:
# Case 1: we are talking about cDNA_match and GFF
if ref_gff is True and "match" not in row.feature:
if transcript is None:
raise TypeError(
"Transcript not defined inside the GFF; line:\n{}".
format(row))
else:
queue_logger.debug("Adding exon to transcript %s: %s",
transcript.id, row)
transcript.add_exon(row)
elif ref_gff is True and "match" in row.feature:
if transcript is not None and row.id == transcript.id:
transcript.add_exon(row)
elif transcript is not None and transcript.id in row.parent:
transcript.add_exon(row)
elif transcript is None or (transcript is not None
and row.id != transcript.id):
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id) and \
(not transcript.id.endswith("orf1")):
pass
else:
assigner_instance.get_best(transcript)
queue_logger.debug("New transcript: %s", row.transcript)
transcript = Transcript(row, logger=queue_logger)
elif ref_gff is False:
if transcript is None or (transcript is not None
and transcript.id != row.transcript):
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id) and \
(not transcript.id.endswith("orf1")):
pass
else:
assigner_instance.get_best(transcript)
queue_logger.debug("New transcript: %s", row.transcript)
transcript = Transcript(row, logger=queue_logger)
transcript.add_exon(row)
else:
raise TypeError("Unmatched exon: {}".format(row))
elif row.header:
continue
else:
queue_logger.debug("Skipped row: {}".format(row))
if transcript is not None:
if re.search(r"\.orf[0-9]+$",
transcript.id) and not transcript.id.endswith("orf1"):
pass
else:
assigner_instance.get_best(transcript)
# Finish everything, including printing refmap and stats
assigner_instance.finish()
args.prediction.close()
def parse_prediction(args, genes, positions, queue_logger):
"""
This function performs the real comparison between the reference and the prediction.
It needs the following inputs:
:param args: the Namespace with the necessary parameters
:param genes: Dictionary with the reference genes, of the form
dict[chrom][(start,end)] = [gene object]
:param positions: Dictionary with the positions of the reference genes, of the form
dict[chrom][IntervalTree]
:param queue_logger: Logger
:return:
"""
# start the class which will manage the statistics
accountant_instance = Accountant(genes, args)
assigner_instance = Assigner(genes, positions, args, accountant_instance)
transcript = None
if hasattr(args, "self") and args.self is True:
args.prediction = to_gff(args.reference.name)
ref_gff = isinstance(args.prediction, GFF3)
__found_with_orf = set()
for row in args.prediction:
if row.header is True:
continue
# queue_logger.debug("Row:\n{0:>20}".format(str(row)))
if row.is_transcript is True or row.feature == "match":
queue_logger.debug("Transcript row:\n%s", str(row))
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id):
__name = re.sub(r"\.orf[0-9]+$", "", transcript.id)
if __name not in __found_with_orf:
__found_with_orf.add(__name)
assigner_instance.get_best(transcript)
else:
pass
else:
assigner_instance.get_best(transcript)
transcript = Transcript(row, logger=queue_logger)
elif row.is_exon is True:
# Case 1: we are talking about cDNA_match and GFF
if ref_gff is True and "match" not in row.feature:
if transcript is None:
raise TypeError("Transcript not defined inside the GFF; line:\n{}".format(row))
else:
queue_logger.debug("Adding exon to transcript %s: %s",
transcript.id, row)
transcript.add_exon(row)
elif ref_gff is True and "match" in row.feature:
if transcript is not None and row.id == transcript.id:
transcript.add_exon(row)
elif transcript is not None and transcript.id in row.parent:
transcript.add_exon(row)
elif transcript is None or (transcript is not None and row.id != transcript.id):
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id) and \
(not transcript.id.endswith("orf1")):
pass
else:
assigner_instance.get_best(transcript)
queue_logger.debug("New transcript: %s", row.transcript)
transcript = Transcript(row, logger=queue_logger)
elif ref_gff is False:
if transcript is None or (transcript is not None and transcript.id != row.transcript):
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id) and \
(not transcript.id.endswith("orf1")):
pass
else:
assigner_instance.get_best(transcript)
queue_logger.debug("New transcript: %s", row.transcript)
transcript = Transcript(row, logger=queue_logger)
transcript.add_exon(row)
else:
raise TypeError("Unmatched exon: {}".format(row))
elif row.header:
continue
else:
queue_logger.debug("Skipped row: {}".format(row))
if transcript is not None:
if re.search(r"\.orf[0-9]+$", transcript.id) and not transcript.id.endswith("orf1"):
pass
else:
assigner_instance.get_best(transcript)
# Finish everything, including printing refmap and stats
assigner_instance.finish()
args.prediction.close()