def test_make_canon_kmers(self):
"""Test that make kmers works."""
seq = 'ATCGGTA'
expected_kmers = set(['ATCG', 'CCGA', 'ACCG', 'GGTA'])
actual_kmers = make_kmers(seq, 4, canon=True)
for kmer in actual_kmers:
self.assertIn(kmer, expected_kmers)
def test_kmer_stats(self):
kmers = make_kmers(ECOLI.longest_contig()[:1000], 31, canon=True)
radial_cover = GreedyRadialCover(hamming_distance, 2)
for kmer in kmers:
radial_cover.add(kmer)
radial_cover.stats()
for kmer in kmers[:10]:
radial_cover.search(kmer, 1)
def kmer_entropy(seqs, k=31):
tbl = {}
for seq in seqs:
for kmer in make_kmers(seq, k, canon=True):
tbl[kmer] = 1 + tbl.get(kmer, 0)
H, total = 0, sum(tbl.values())
for count in tbl.values():
p = count / total
H += p * log2(p)
return -H
def kmer_search(cls, seq, eps=0.5):
uuids = {}
kmers = make_kmers(seq, MAX_K)
for kmer in kmers:
for uuid in cls.query.filter_by(kmer=kmer).all():
uuids[uuid] = uuids.get(uuid, 0) + 1
seqs = []
for uuid, count in uuids.items():
if count < (eps * len(kmers)):
continue
seqs.append(Contig.from_uuid(uuid))
return seqs
def sub_k_dist(row):
q_ks = make_kmers(row['query'], sub_k)
t_ks = make_kmers(row['target'], sub_k)
return len(q_ks & t_ks) / len(q_ks)
def num_unique_subwords(seq, n):
out = set()
for kmer in make_kmers(seq, n, canon=False):
out.add(kmer)
return len(out)
def add_contig(cls, contig):
kmers = set(make_kmers(contig.seq, MAX_K))
for kmer in kmers:
db.session.add(cls(contig.uuid, kmer))
db.session.commit()
def test_add_to_cover(self):
kmers = make_kmers(ECOLI.longest_contig()[:1000], 31, canon=True)
radial_cover = GreedyRadialCover(hamming_distance, 2)
for kmer in kmers:
radial_cover.add(kmer)
def test_make_kmers(self):
"""Idiot check myself."""
make_kmers(ECOLI.longest_contig()[:1000], 31, canon=True)