def process_insertion_in_exon(human_seq, spec_seq, ordinal, frame):
'''
Annotate insertion. Insertion can be a 'real' insertion, or
marked as a frameshift (meaning that it is 1, 2, 4 or 5
base pairs long)
@return: AlignmentExonPiece
'''
if len(spec_seq) in (1,2,4,5):
exon_type = "frameshift"
else:
exon_type = "insertion"
al_exon = AlignmentExonPiece(exon_type, ordinal, human_seq, spec_seq)
al_exon.set_frame(frame)
return al_exon
def translate_ensembl_exons (ensembl_exons):
cdna = ""
#determine the frame
first_exon = ensembl_exons[0]
if first_exon.relative_start == 0:
new_frame = first_exon.frame
else:
frame_complement = (first_exon.relative_start - first_exon.frame) % 3
if frame_complement == 1:
new_frame = 2
elif frame_complement == 2:
new_frame = 1
else:
new_frame = 0
#create the cdna
for ee in ensembl_exons:
new_sequence = ee.sequence[ee.relative_start:ee.relative_stop]
cdna += new_sequence
al_exon = AlignmentExonPiece("ensembl", -1, "", cdna)
al_exon.set_frame(new_frame)
return cdna[new_frame:].translate()
def process_insertion_free_region(human_seq, spec_seq, frame, ordinal, alignment_start):
'''
Process a region of an alignment without any insertions in the
reference species exon. There can exist deletions however.
If the deletions are 1 or 2 bases long, they are replaced
by N (regarded as an assembly error)
'''
al_exons = []
spec_seq = re.sub("([ATGCN])(-{1,2})([ATGCN])", replace_slash, spec_seq)
# determine the number of coding sequences interspersed with gaps
number_of_ungapped_sequences = len(re.split("-+", str(spec_seq)))
if number_of_ungapped_sequences == 1:
al_piece = AlignmentExonPiece("coding", ordinal, human_seq, spec_seq)
al_piece.set_frame(frame)
al_piece.set_alignment_locations(alignment_start, alignment_start + len(spec_seq))
return [al_piece]
# create a pattern to load the sequences
pattern_string = "([ATGCN]+)"
for i in range (0, number_of_ungapped_sequences-1):
pattern_string += "(-+)([ATGCN]+)"
pattern = re.compile (pattern_string)
sequences = re.match (pattern, str(spec_seq))
# auxiliary variables
in_coding = True
start, stop = 0,0
coding_len = 0
alignment_stop = alignment_start
for seq in sequences.groups():
stop += len(seq)
alignment_stop += len(seq)
species = seq
human = human_seq[start:start + len(seq)]
# if we're not in the coding region, there is nothing to process
# non-coding here means deletion in the referent exon
if in_coding:
if coding_len == 0:
new_frame = frame
else:
frame_status = abs(coding_len - frame) % 3
if frame_status == 1:
new_frame = 2
elif frame_status == 2:
new_frame = 1
else:
new_frame = 0
exon = AlignmentExonPiece("coding", ordinal, human, species)
exon.set_frame(new_frame)
exon.set_alignment_locations(alignment_start, alignment_stop)
al_exons.append(exon)
# there is a deletion. Remember this also - if the deletion is
# one of two bases long, we pad it with Ns and translate regularly
else:
exon = AlignmentExonPiece("deletion", ordinal, human, species)
al_exons.append(exon)
in_coding = not in_coding
start = stop
alignment_start = alignment_stop
coding_len += len(seq)
ordinal += 1
return al_exons