def _test_profile_split(self, sequences, length):
counts = utils.counts(sequences, length)
profile = klib.Profile(utils.as_array(counts, length))
left, right = profile.split()
assert len(left) == len(right)
assert sum(left) + sum(right) == sum(counts.values()) * 2
indices_left = {}
indices_right = {}
indices_palindrome = {}
for s, c in counts.items():
r = utils.reverse_complement(s)
if s < r:
indices_left[utils.count_index(s)] = c * 2
elif s > r:
indices_right[utils.count_index(r)] = counts[s] * 2
else:
indices_palindrome[utils.count_index(s)] = c
assert ([c for c in left if c > 0] == [
c for i, c in sorted(
list(indices_left.items()) + list(indices_palindrome.items()))
])
assert ([c for c in right if c > 0] == [
c for i, c in sorted(
list(indices_right.items()) + list(indices_palindrome.items()))
])
def _test_profile_split(self, sequences, length):
counts = utils.counts(sequences, length)
profile = klib.Profile(utils.as_array(counts, length))
left, right = profile.split()
assert len(left) == len(right)
assert sum(left) + sum(right) == sum(counts.values()) * 2
indices_left = {}
indices_right = {}
indices_palindrome = {}
for s, c in counts.items():
r = utils.reverse_complement(s)
if s < r:
indices_left[utils.count_index(s)] = c * 2
elif s > r:
indices_right[utils.count_index(r)] = counts[s] * 2
else:
indices_palindrome[utils.count_index(s)] = c
assert ([c for c in left if c > 0] ==
[c for i, c in sorted(list(indices_left.items()) +
list(indices_palindrome.items()))])
assert ([c for c in right if c > 0] ==
[c for i, c in sorted(list(indices_right.items()) +
list(indices_palindrome.items()))])
def test_profile_freq_diff_matrix(self):
counts = utils.counts(utils.SEQUENCES, 4)
profile = klib.Profile(utils.as_array(counts, 4))
freq_diffs = profile._freq_diff_matrix()
total = sum(counts.values())
for left in itertools.product('ACGT', repeat=4):
for right in itertools.product('ACGT', repeat=4):
left = ''.join(left)
right = ''.join(right)
freq_diff = freq_diffs[utils.count_index(left)][utils.count_index(right)]
if counts[right] > 0:
assert freq_diff == abs(counts[left] - counts[right]) / total
else:
assert freq_diff == 0
def test_profile_ratios_matrix(self):
counts = utils.counts(utils.SEQUENCES, 4)
profile = klib.Profile(utils.as_array(counts, 4))
ratios = profile._ratios_matrix()
total = sum(counts.values())
for left in itertools.product('ACGT', repeat=4):
for right in itertools.product('ACGT', repeat=4):
left = ''.join(left)
right = ''.join(right)
ratio = ratios[utils.count_index(left)][utils.count_index(right)]
try:
assert ratio == counts[left] / counts[right] / total
except ZeroDivisionError:
assert ratio == -1.0
def test_profile_ratios_matrix(self):
counts = utils.counts(utils.SEQUENCES, 4)
profile = klib.Profile(utils.as_array(counts, 4))
ratios = profile._ratios_matrix()
total = sum(counts.values())
for left in itertools.product('ACGT', repeat=4):
for right in itertools.product('ACGT', repeat=4):
left = ''.join(left)
right = ''.join(right)
ratio = ratios[utils.count_index(left)][utils.count_index(
right)]
try:
assert ratio == counts[left] / counts[right] / total
except ZeroDivisionError:
assert ratio == -1.0
def test_profile_freq_diff_matrix(self):
counts = utils.counts(utils.SEQUENCES, 4)
profile = klib.Profile(utils.as_array(counts, 4))
freq_diffs = profile._freq_diff_matrix()
total = sum(counts.values())
for left in itertools.product('ACGT', repeat=4):
for right in itertools.product('ACGT', repeat=4):
left = ''.join(left)
right = ''.join(right)
freq_diff = freq_diffs[utils.count_index(left)][
utils.count_index(right)]
if counts[right] > 0:
assert freq_diff == abs(counts[left] -
counts[right]) / total
else:
assert freq_diff == 0