def __init__(self, consensus_string: str, min_match_length: int, alignment: MSA):
self._match_intervals: Intervals = list()
self._non_match_intervals: Intervals = list()
self.mml = min_match_length
if len(consensus_string) < self.mml:
# In this case, a match of less than the min_match_length gets counted
# as a match (usually, it counts as a non_match)
it_type = IntervalType.Match
if any(map(is_non_match, consensus_string)):
it_type = IntervalType.NonMatch
self._append(Interval(it_type, 0, len(consensus_string) - 1))
else:
cur_interval = self._new_interval(consensus_string[0], 0)
for i, letter in enumerate(consensus_string[1:], start=1):
if is_type(letter, cur_interval.type):
cur_interval.modify_by(0, 1) # simple interval extension
else:
new_interval = self._add_interval(cur_interval, alignment)
if new_interval is None:
cur_interval = self._new_interval(letter, i)
else:
cur_interval = new_interval
self._add_interval(cur_interval, alignment, end=True)
self.enforce_multisequence_nonmatch_intervals(
self._match_intervals, self._non_match_intervals, alignment
)
self.enforce_alignment_interval_bijection(
self._match_intervals,
self._non_match_intervals,
alignment.get_alignment_length(),
)
def get_sub_alignment_by_list_id(
self, id_list: List[str], alignment: MSA, interval=None
):
list_records = [record for record in alignment if record.id in id_list]
sub_alignment = MSA(list_records)
if interval is not None:
sub_alignment = sub_alignment[:, interval[0] : interval[1] + 1]
return sub_alignment
def make_alignment(seqs: List[str], ids: List[str] = None) -> MSA:
seq_lengths = set(map(len, seqs))
assert (
len(seq_lengths) == 1
), "Sequences are not the same length, does not represent an alignment"
if ids is None:
seqrecords = [
SeqRecord(Seq(seq), id=f"s{i}") for i, seq in enumerate(seqs)
]
else:
seqrecords = [SeqRecord(Seq(seq), id=ID) for seq, ID in zip(seqs, ids)]
return MSA(seqrecords)
def get_consensus(cls, alignment: MSA):
"""Produces a 'consensus string' from an MSA: at each position of the
MSA, the string has a base if all aligned sequences agree, and a "*" if not.
IUPAC ambiguous bases result in non-consensus and are later expanded in the prg.
N results in consensus at that position unless they are all N."""
consensus_string = ""
for i in range(alignment.get_alignment_length()):
column = set([record.seq[i] for record in alignment])
column = column.difference({"N"})
if (len(ambiguous_bases.intersection(column)) > 0
or len(column) != 1 or column == {"-"}):
consensus_string += NONMATCH
else:
consensus_string += column.pop()
return consensus_string