def _make_read_unmapped(rec: AlignedSegment) -> None:
"""Removes mapping information from a read."""
if rec.is_reverse:
quals = rec.query_qualities
quals.reverse()
rec.query_sequence = dnautils.reverse_complement(rec.query_sequence)
rec.query_qualities = quals
rec.is_reverse = False
rec.reference_id = sam.NO_REF_INDEX
rec.reference_start = sam.NO_REF_POS
rec.cigar = None
rec.mapping_quality = 0
rec.template_length = 0
rec.is_duplicate = False
rec.is_secondary = False
rec.is_supplementary = False
rec.is_proper_pair = False
rec.is_unmapped = True
def convert_events_to_softclipping(
read: pysam.AlignedSegment,
orientation: str,
max_event_size: int,
min_anchor_size: Optional[int] = None,
) -> pysam.AlignedSegment:
"""
given an alignment, simplifies the alignment by grouping everything past the first anchor and including the
first event considered too large and unaligning them turning them into softclipping
"""
if min_anchor_size is None:
min_anchor_size = max_event_size
if orientation == ORIENT.LEFT:
event_size = 0
adjusted_cigar = []
anchor = 0
for state, count in read.cigar:
if state == CIGAR.M:
raise NotImplementedError('match v mismatch must be specified')
elif anchor < min_anchor_size:
if state == CIGAR.EQ:
anchor += count
elif state in EVENT_STATES:
event_size += count
if event_size > max_event_size:
break
else:
event_size = 0
adjusted_cigar.append((state, count))
if event_size > max_event_size:
while adjusted_cigar[-1][0] in EVENT_STATES:
del adjusted_cigar[-1]
aligned = sum([y for x, y in adjusted_cigar if x in QUERY_ALIGNED_STATES] + [0])
sc = len(read.query_sequence) - aligned
adjusted_cigar.append((CIGAR.S, sc))
read = copy(read)
read.cigar = adjusted_cigar
elif orientation == ORIENT.RIGHT:
# more complicated than left b/c need to also adjust the start position
event_size = 0
anchor = 0
adjusted_cigar = []
for state, count in read.cigar[::-1]: # first event from the right
if state == CIGAR.M:
raise NotImplementedError('match v mismatch must be specified')
elif anchor < min_anchor_size:
if state == CIGAR.EQ:
anchor += count
elif state in EVENT_STATES:
event_size += count
if event_size > max_event_size:
break
else:
event_size = 0
adjusted_cigar.append((state, count))
if event_size > max_event_size:
while adjusted_cigar[-1][0] in EVENT_STATES:
del adjusted_cigar[-1]
originally_refaligned = sum(
[y for x, y in read.cigar if x in REFERENCE_ALIGNED_STATES] + [0]
)
refaligned = sum([y for x, y in adjusted_cigar if x in REFERENCE_ALIGNED_STATES] + [0])
aligned = sum([y for x, y in adjusted_cigar if x in QUERY_ALIGNED_STATES] + [0])
sc = len(read.query_sequence) - aligned
adjusted_cigar = [(CIGAR.S, sc)] + adjusted_cigar[::-1]
read = copy(read)
read.cigar = adjusted_cigar
read.reference_start += originally_refaligned - refaligned
else:
raise ValueError('orientation must be specified', orientation)
return read