def test_should_return_if_msa_is_valid(self):
# setup
problem = MSA(score_list=[])
problem.identifiers = ['seq1', 'seq2', 'seq3']
problem.number_of_variables = 3
msa_valid = MSASolution(problem,
msa=[('seq1', 'AC---TGAC'),
('seq2', 'AT--CT--C'),
('seq3', 'AAC---TGC')])
msa_not_valid = MSASolution(problem,
msa=[('seq1', 'A'), ('seq2', 'A'),
('seq3', 'AA')])
# check
self.assertTrue(msa_valid.is_valid_msa())
self.assertFalse(msa_not_valid.is_valid_msa())
def do_mutation(self, solution: MSASolution) -> MSASolution:
if random.random() <= self.probability:
# Select one random sequence from all
for seq in range(solution.number_of_variables):
gaps_group = solution.gaps_groups[seq]
if len(gaps_group) >= 4:
random_gaps_group = random.randrange(
0,
len(gaps_group) - 2, 2)
shift_to = -1 if random.randint(0, 1) == 0 else 1
gaps_group[random_gaps_group] += shift_to
gaps_group[random_gaps_group + 1] += shift_to
solution.merge_gaps_groups()
if self.remove_full_of_gap_columns:
solution.remove_full_of_gaps_columns()
# Sanity check: alignment is valid (same length for all sequences)
if not solution.is_valid_msa():
raise Exception("Mutated solution is not valid! {0}".format(
solution.decode_alignment_as_list_of_pairs()))
return solution
def do_mutation(self, solution: MSASolution) -> MSASolution:
if random.random() <= self.probability:
if solution.number_of_variables >= 1:
seq = random.randint(0, solution.number_of_variables - 1)
else:
seq = 0
gaps_group = solution.gaps_groups[seq]
if len(gaps_group) >= 4:
random_gaps_group = random.randrange(0, len(gaps_group) - 2, 2)
right_is_closest = False
if not right_is_closest:
to_add = gaps_group[random_gaps_group +
3] - gaps_group[random_gaps_group +
2] + 1
gaps_group[random_gaps_group + 1] += to_add
del gaps_group[random_gaps_group + 3]
del gaps_group[random_gaps_group + 2]
solution.merge_gaps_groups()
if self.remove_full_of_gap_columns:
solution.remove_full_of_gaps_columns()
# Sanity check: alignment is valid (same length for all sequences)
if not solution.is_valid_msa():
raise Exception("Mutated solution is not valid! {0}".format(
solution.decode_alignment_as_list_of_pairs()))
return solution
def do_mutation(self, solution: MSASolution) -> MSASolution:
if random.random() <= self.probability:
for i in range(solution.number_of_variables):
gaps_group = solution.gaps_groups[i]
if len(gaps_group) >= 4:
random_gaps_group = random.randrange(
0,
len(gaps_group) - 2, 2)
right_is_closest = False
if not right_is_closest:
diff = (gaps_group[random_gaps_group + 3] - gaps_group[random_gaps_group + 2]) - \
(gaps_group[random_gaps_group + 1] - gaps_group[random_gaps_group])
if diff < 0:
# diff < 0 means that gaps group 2 is shorter than gaps group 1, thus we need to decrease
# the length of the gaps group 1
diff = -1 * diff
gaps_group[random_gaps_group + 1] -= diff
gaps_group[random_gaps_group + 3] += diff
# displace gaps group 2 one position to the left
gaps_group[random_gaps_group + 2] -= diff
gaps_group[random_gaps_group + 3] -= diff
elif diff > 0:
# diff > 0 means that gaps group 2 is larger than gaps group 1, thus we need to increase
# the length of the gaps group 1
gaps_group[random_gaps_group + 1] += diff
gaps_group[random_gaps_group + 3] -= diff
# displace gaps group 2 one position to the right
gaps_group[random_gaps_group + 2] += diff
gaps_group[random_gaps_group + 3] += diff
if self.remove_full_of_gap_columns:
solution.remove_full_of_gaps_columns()
# Sanity check: alignment is valid (same length for all sequences)
if not solution.is_valid_msa():
raise Exception("Mutated solution is not valid! {0}".format(
solution.decode_alignment_as_list_of_pairs()))
return solution
def do_mutation(self, solution: MSASolution) -> MSASolution:
if random.random() <= self.probability:
length_of_alignment = solution.get_length_of_alignment()
for seq_index in range(solution.number_of_variables):
point = random.randint(0, length_of_alignment - 1)
solution.add_gap_to_sequence_at_index(seq_index, point)
if self.remove_full_of_gap_columns:
solution.remove_full_of_gaps_columns()
# Sanity check: alignment is valid (same length for all sequences)
if not solution.is_valid_msa():
raise Exception("Mutated solution is not valid! {0}".format(
solution.decode_alignment_as_list_of_pairs()))
return solution
