def test_phred_conversions():
assert util.phred_to_prob(40) == 0.9999
assert util.phred_to_prob(30) == 0.999
assert util.phred_to_prob(20) == 0.99
assert util.phred_to_prob(10) == 0.9
assert util.prob_to_phred(0.9999) == 40
assert util.prob_to_phred(0.999) == 30
assert util.prob_to_phred(0.99) == 20
assert util.prob_to_phred(0.9) == 10
def mut_sequence(self, record, orientation):
"""Introduce substitution errors to a sequence
If a random probability is higher than the probability of the basecall
being correct, introduce a substitution error
Args:
record (SeqRecord): a read record with error scores
orientation (string): orientation of the read. Can be 'forward' or
'reverse'
Returns:
Seq: a sequence
"""
# get the right subst_matrix
if orientation == 'forward':
nucl_choices = self.subst_choices_for
elif orientation == 'reverse':
nucl_choices = self.subst_choices_rev
mutable_seq = record.seq.tomutable()
quality_list = record.letter_annotations["phred_quality"]
position = 0
for nucl, qual in zip(mutable_seq, quality_list):
if random.random() > util.phred_to_prob(qual) \
and nucl.upper() not in 'RYWSMKHBVDN':
mutable_seq[position] = str(
np.random.choice(
nucl_choices[position][nucl.upper()][0],
p=nucl_choices[position][nucl.upper()][1]))
position += 1
return mutable_seq.toseq()
def gen_phred_scores(self, mean_quality, orientation):
"""Generate a normal distribution, transform to phred scores
Generate a list of phred score according to a normal distribution
centered around the ErrorModel quality
Args:
mean_quality (int): mean phred score
Returns:
list: list of phred scores following a normal distribution
"""
norm = [min(q, 0.9999) for q in np.random.normal(
util.phred_to_prob(mean_quality), 0.01, self.read_length)]
phred = [util.prob_to_phred(p) for p in norm]
return phred