def __set_fasta_seq(self, fasta_seq):
if isinstance(fasta_seq, Seq.Seq):
self.__fasta_seq = pyfaidx.Sequence(name=self.id, seq=str(fasta_seq))
elif isinstance(fasta_seq, str):
self.__fasta_seq = pyfaidx.Sequence(name=self.id, seq=str(fasta_seq))
assert len(self.__fasta_seq) == len(str(fasta_seq))
elif isinstance(fasta_seq, pyfaidx.Sequence):
self.__fasta_seq = fasta_seq
else:
raise ValueError("Unkown type: {}".format(type(fasta_seq)))
def tests():
seq_a = pyfaidx.Sequence('contig1', 'ATGCGT', start=1, end=6)
print(seq_a, seq_a.start, seq_a.end)
seq_b = pyfaidx.Sequence('contig1', 'GCGAGT', start=13, end=18)
seq_a = connectFragment(seq_a, seq_b)
# Connect fragment after original
print(seq_a, seq_a.start, seq_a.end)
# Wrong contig
seq_c = pyfaidx.Sequence('different_contig1', 'CCCCCC', start=7, end=12)
seq_a = connectFragment(seq_a, seq_c)
seq_c = pyfaidx.Sequence('contig1', 'CCCCCC', start=7, end=12)
seq_a = connectFragment(seq_a, seq_c)
# Substitute fragment in centre
print(seq_a, seq_a.start, seq_a.end)
seq_a = pyfaidx.Sequence('contig1', 'ATGCGT', start=1, end=6)
seq_b = pyfaidx.Sequence('contig1', 'GCGAGT', start=13, end=18)
seq_b.percentIdentity = 97.1
seq_b = connectFragment(seq_b, seq_a)
# Connect fragment before original
print(seq_b, seq_b.start, seq_b.end, seq_b.percentIdentity)
seq_c = pyfaidx.Sequence('contig1', 'CCC', start=5, end=7)
seq_c.percentIdentity = 97.2
seq_b = connectFragment(seq_b, seq_c)
print(seq_b, seq_b.start, seq_b.end, seq_b.percentIdentity)
seq_c = pyfaidx.Sequence('contig1', 'AATTT', start=8, end=12)
seq_b = connectFragment(seq_b, seq_c)
print(seq_b, seq_b.start, seq_b.end, len(seq_b))
def test_init(self):
with self.assertRaises(ValueError):
tcheck = TranscriptChecker(self.model, None)
for wrong_splices in ["AGGT", None, 100]:
with self.assertRaises(ValueError):
tcheck = TranscriptChecker(self.model, self.model_fasta, canonical_splices=wrong_splices)
tcheck = TranscriptChecker(self.model, self.model_fasta)
self.assertEqual(tcheck.cdna_length, 1718)
self.assertEqual(sorted(tcheck.exons), sorted([(exon.start, exon.end) for exon in self.exons]))
self.assertEqual(tcheck.fasta_seq, self.model_fasta)
with self.subTest(initializer=Bio.Seq.Seq):
_ = TranscriptChecker(self.model, Bio.Seq.Seq(str(self.model_fasta)))
with self.subTest(initializer=str):
_ = TranscriptChecker(self.model, str(self.model_fasta))
with self.subTest(initializer=pyfaidx.Sequence):
_ = TranscriptChecker(self.model, pyfaidx.Sequence(seq=str(self.model_fasta), name=tcheck.id))
# Now check initializing with a GFF/GTF line
for out_format in ["gtf", "gff3"]:
with self.subTest(out_format=out_format):
line = self.model.format(out_format).split("\n")[0]
try:
tcheck = TranscriptChecker(line, self.model_fasta)
except ValueError as exc:
raise ValueError(line)
def match_seq(rec: pd.Series, sequences: pyfaidx.Fasta) -> pyfaidx.Sequence:
"""Given a feature in a GTF/GFF read in by gtfparse, match_seq() will extract the corresponding
DNA sequence and create a new pyfaidx.Sequence object
Parameters
----------
rec : :class:`~pandas.Series`
Information for a feature (i.e. gene, exon, etc...). Requires the following indices: strand,
gene_name, feature, strand, start, end, seq_hash
sequences : :class:`~pyfaidx.Sequence`
Object containing sequences to match against the positions in the index.
Returns
-------
:class:`~pyfaidx.Sequence object` with annotation from `rec` and sequence information from
`sequences`.
"""
try:
rev: bool = bool(rec["strand"] == "-")
seq = pyfaidx.Sequence(
name=f"{rec['gene_name']}_"
f"{rec['feature']}_"
f"{rec['strand']}_"
f"{rec['start']}_"
f"{rec['end']}_"
f"{rec['seq_hash']}",
seq=sequences.get_seq(name=rec["seqname"],
start=rec["start"],
end=rec["end"],
rc=rev).seq,
)
return seq
except ValueError:
print(f"problem with {rec['gene_name']} {rec['start']} "
f"{rec['end']} {rec['seqname']} {rec['strand']}")
def pad_transcripts(self):
"""
"""
try:
self.fai = pyfaidx.Fasta(self.json_conf["reference"]["genome"])
except KeyError:
raise KeyError(self.json_conf.keys())
five_graph = self.define_graph(self.transcripts, self.__share_extreme, three_prime=False)
three_graph = self.define_graph(self.transcripts, self.__share_extreme, three_prime=True)
five_comm = deque(sorted(self.find_communities(five_graph),
key=lambda clique: min(self[_].start for _ in clique)))
three_comm = deque(sorted(self.find_cliques(three_graph),
key=lambda clique: max(self[_].end for _ in clique),
reverse=True))
five_found = set()
# First do the 5' end
__to_modify = dict()
while len(five_comm) > 0:
comm = five_comm.popleft()
comm = deque(sorted(list(set.difference(set(comm), five_found)),
key=lambda tid: self[tid].start))
if len(comm) == 1:
continue
first = comm.popleft()
five_found.add(first)
comm_start = self[first].start
# self[first].strip_cds()
for tid in comm:
if ((self[tid].start - comm_start + 1) <
self.json_conf["pick"]["alternative_splicing"]["ts_distance"] and
len([_ for _ in self.splices if comm_start <= _ <= self[tid].start]) <
self.json_conf["pick"]["alternative_splicing"]["ts_max_splices"] and
self[tid].start > comm_start):
__to_modify[tid] = [comm_start, False]
five_found.add(tid)
else:
continue
comm = deque([_ for _ in comm if _ not in five_found])
if comm:
five_comm.appendleft(comm)
# Then do the 3' end
three_found = set()
while len(three_comm) > 0:
comm = three_comm.popleft()
comm = deque(sorted(list(set.difference(set(comm), three_found)),
key=lambda tid: self[tid].end, reverse=True))
if len(comm) == 1:
continue
first = comm.popleft()
three_found.add(first)
comm_end = self[first].end
for tid in comm:
if ((self[tid].end - comm_end + 1) <
self.json_conf["pick"]["alternative_splicing"]["ts_distance"] and
len([_ for _ in self.splices if self[tid].end <= _ <= comm_end]) <
self.json_conf["pick"]["alternative_splicing"]["ts_max_splices"] and
self[tid].end < comm_end):
if tid in __to_modify:
__to_modify[tid][1] = comm_end
else:
__to_modify[tid] = [False, comm_end]
three_found.add(tid)
else:
continue
comm = deque([_ for _ in comm if _ not in three_found ])
if comm:
three_comm.appendleft(comm)
# Now we can do the proper modification
for tid in __to_modify:
new_transcript = self[tid].copy()
old_length = new_transcript.cdna_length
# First get the ORFs
if new_transcript.combined_cds_length > 0:
internal_orfs = list(new_transcript.get_internal_orf_beds())
else:
internal_orfs = []
# Remove the CDS and unfinalize
new_transcript.strip_cds()
new_transcript.unfinalize()
upstream = 0
downstream = 0
if __to_modify[tid][0]:
__new_exon = (__to_modify[tid][0], new_transcript.exons[0][1])
upstream = new_transcript.start - __to_modify[tid][0]
new_transcript.start = __to_modify[tid][0]
new_transcript.remove_exon(new_transcript.exons[0])
new_transcript.add_exon(__new_exon)
new_transcript.exons = sorted(new_transcript.exons)
if __to_modify[tid][1]:
__new_exon = (new_transcript.exons[-1][0], __to_modify[tid][1])
downstream = __to_modify[tid][1] - new_transcript.end
new_transcript.end = __to_modify[tid][1]
new_transcript.remove_exon(new_transcript.exons[-1])
new_transcript.add_exon(__new_exon)
new_transcript.exons = sorted(new_transcript.exons)
# Now for the difficult part
if internal_orfs and (__to_modify[tid][1] or __to_modify[tid][0]):
self.logger.warning("Enlarging the ORFs for TID %s (%s)",
tid, __to_modify[tid])
new_orfs = []
seq = ''
for exon in new_transcript.exons:
seq += self.fai[self.chrom][exon[0] - 1:exon[1]].seq
seq = pyfaidx.Sequence(tid, seq)
self.logger.warning("For TID %s we have new length %d, old length %d, exons:\n%s",
tid, len(seq), old_length, new_transcript.exons)
if self.strand == "-":
seq = seq.reverse.complement
upstream, downstream = downstream, upstream
for orf in internal_orfs:
self.logger.warning("Old ORF: %s", str(orf))
orf.expand(seq, upstream, downstream)
self.logger.warning("New ORF: %s", str(orf))
new_orfs.append(orf)
from ..utilities.log_utils import create_default_logger
new_transcript.logger = create_default_logger("TEMP")
new_transcript.logger.setLevel("DEBUG")
new_transcript.load_orfs(new_orfs)
new_transcript.logger.setLevel("WARNING")
# Now finalize again
new_transcript.finalize()
self.transcripts[tid] = new_transcript