def test_transform_1(self):
"""simple het & hom indels"""
refseq = self.fa0
mutseq = DipSeq(refseq.seqid + '.mut',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
mutseq.transform(refseq, self.vcf1)
insertion_1 = range(4, 7)
self.assertEqual(
''.join(base(mutseq.seqA[POS - 1]) for POS in insertion_1), "AGG")
self.assertEqual(list(mutseq.relA[POS - 1] for POS in insertion_1),
[4, 4, 4])
insertion_2 = range(5, 9)
self.assertEqual(
''.join(base(mutseq.seqB[POS - 1]) for POS in insertion_2), "CTTT")
self.assertEqual(list(mutseq.relB[POS - 1] for POS in insertion_2),
[5, 5, 5, 5])
deletion_1A = range(15, 17)
deletion_1B = range(16, 18)
self.assertEqual(
''.join(base(mutseq.seqA[POS - 1]) for POS in deletion_1A), "CC")
self.assertEqual(list(mutseq.relA[POS - 1] for POS in deletion_1A),
[13, 16])
self.assertEqual(
''.join(base(mutseq.seqB[POS - 1]) for POS in deletion_1B), "CC")
self.assertEqual(list(mutseq.relB[POS - 1] for POS in deletion_1B),
[13, 16])
def test_mutagen(self):
"""tests both mutagen and transform methods"""
# Another case where we take a shortcut by simply asserting that the
# data in the test resource files mut_vcf and mut_seq are valid "by
# inspection" and require that the test run matches them every time.
# In this case cross-referencing the randomly generated mutations in
# the vcf with the transformed sequence file is what's required. You
# should eyeball it yourself — it's interesting!
refseq = self.fa1
vcf = VCF("sample1")
mut_rate = 0.01
homo_frac = 0.333333
indel_frac = 0.15
indel_extend = 0.3
max_insertion = 1000
qasim.reseed(12345678) # deterministic iff we set the seed
DipSeq.mutagen(refseq, vcf, mut_rate, homo_frac, indel_frac,
indel_extend, max_insertion)
out = StringIO()
vcf.write(out)
with open(self.mut_vcf) as fh:
self.assertEqual(out.getvalue(), ''.join(fh.readlines()))
mutseq = DipSeq(refseq.seqid + '.mut',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
mutseq.transform(refseq, vcf)
out = StringIO()
mutseq.write(out)
with open(self.mut_seq) as fh:
self.assertEqual(out.getvalue(), ''.join(fh.readlines()))
def test_transform_4(self):
"""disallow mutations overlapping deletions in same vcf"""
refseq = self.fa0
mutseq = DipSeq(refseq.seqid + '.mut',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
expected_msg = EXCEPT_MUT % {'POS': 7, 'OLDPOS': 5}
with self.assertRaisesRegex(Exception, expected_msg):
mutseq.transform(refseq, self.vcf4)
def test_transform_2(self):
"""complex overlapping mutations"""
refseq = self.fa0
mutseq = DipSeq(refseq.seqid + '.mut',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
mutseq.transform(refseq, self.vcf2)
out = StringIO()
mutseq.write(out)
# we take a bit of a shorcut and rather than testing all the
# explicit logic of the transformation we just test equality to
# this output that we assert is valid "by inspection".
self.assertEqual(out.getvalue(),
(">TEST.mut.0 small fasta for testing\n"
"AAAAGGCCGAAACCCC\n"
"1234445690123456\n"
">TEST.mut.1 small fasta for testing\n"
"AAAAGGCTTTCAAAACCCC\n"
"1234445555690123456\n"))
def test_transform_3(self):
"""overlapping mutations in somatic mode"""
refseq = self.fa0
grmseq = DipSeq(refseq.seqid + '.grm',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
grmseq.transform(refseq, self.vcfgrm)
out = StringIO()
grmseq.write(out)
self.assertEqual(out.getvalue(),
(">TEST.grm.0 small fasta for testing\n"
"AAAAGGCCCCAAAACCCC\n"
"123444567890123456\n"
">TEST.grm.1 small fasta for testing\n"
"AAAAGGCCAAAACCCC\n"
"1234447890123456\n"))
somseq = DipSeq(refseq.seqid + '.som',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
with captured_output() as (out, err):
expected_msg = MSG_SKIP_MUT % {'allele': 1, 'POS': 5}
somseq.transform(grmseq, self.vcfsom)
self.assertEqual(err.getvalue(), expected_msg)
out = StringIO()
somseq.write(out)
# somatic insertion at 4 and deletion at 5 both applied to allele 0
# somatic deletion at 5 isn't applied to allele 1
self.assertEqual(out.getvalue(),
(">TEST.som.0 small fasta for testing\n"
"AAAATTGGCAAAACCCC\n"
"12344444590123456\n"
">TEST.som.1 small fasta for testing\n"
"AAAATTGGCCAAAACCCC\n"
"123444447890123456\n"))
def test_transform_0(self):
"""germline het & hom snps"""
refseq = self.fa0
mutseq = DipSeq(refseq.seqid + '.mut',
refseq.description,
size=refseq.seqA.shape[0] * 2,
fold=refseq.fold)
mutseq.transform(refseq, self.vcf0)
for i in range(refseq.stopA):
POS = i + 1 # VCF coords
if POS == 5:
self.assertEqual(base(mutseq.seqA[i]), base(refseq.seqA[i]))
self.assertEqual(base(mutseq.seqB[i]), "G")
elif POS == 9:
self.assertEqual(base(mutseq.seqA[i]), "G")
self.assertEqual(base(mutseq.seqB[i]), base(refseq.seqA[i]))
elif POS == 13:
self.assertEqual(base(mutseq.seqA[i]), "T")
self.assertEqual(base(mutseq.seqA[i]), base(mutseq.seqB[i]))
self.assertNotEqual(base(mutseq.seqA[i]), base(refseq.seqA[i]))
else:
self.assertEqual(base(mutseq.seqA[i]), base(refseq.seqA[i]))
self.assertEqual(base(mutseq.seqB[i]), base(refseq.seqA[i]))