def make_call_from_reads(queue,
idx,
calls,
refrfile,
ksize=31,
delta=50,
seedsize=31,
maxdiff=None,
match=1,
mismatch=2,
gapopen=5,
gapextend=0,
min_ikmers=None,
refrseqs=None,
logstream=sys.stderr):
while True:
if queue.empty():
sleep(3)
continue
reads = queue.get()
ccmatch = re.search(r'kvcc=(\d+)', reads[0].name)
cc = ccmatch.group(1) if ccmatch else None
message = '[kevlar::alac::make_call_from_reads'
message += ' (thread={:d})]'.format(idx)
message += ' grabbed partition={} from queue,'.format(cc)
message += ' queue size now {:d}'.format(queue.qsize())
print(message, file=sys.stderr)
# Assemble partitioned reads into contig(s)
contigs = list(assemble_fml_asm(reads, logstream=logstream))
if min_ikmers is not None:
# Apply min ikmer filter if it's set
contigs = [c for c in contigs if len(c.annotations) >= min_ikmers]
if len(contigs) == 0:
queue.task_done()
continue
# Identify the genomic region(s) associated with each contig
localizer = localize(contigs,
refrfile,
seedsize,
delta=delta,
maxdiff=maxdiff,
refrseqs=refrseqs,
logstream=logstream)
targets = list(localizer)
if len(targets) == 0:
queue.task_done()
continue
# Align contigs to genomic targets to make variant calls
caller = call(targets, contigs, match, mismatch, gapopen, gapextend,
ksize, refrfile)
for varcall in caller:
if cc:
varcall.annotate('PART', cc)
calls.append(varcall)
queue.task_done()
def test_localize_no_match(capsys):
refr_file = data_file('fiveparts-refr.fa.gz')
contig_file = data_file('wasp-pass.contig.augfasta')
contigstream = kevlar.parse_augmented_fastx(kevlar.open(contig_file, 'r'))
pstream = kevlar.parse_partitioned_reads(contigstream)
localizer = localize(pstream, refr_file, seedsize=41, debug=True)
cutoutdata = list(localizer)
assert cutoutdata == []
out, err = capsys.readouterr()
assert 'WARNING: no reference matches' in err
def test_localize_new_single_partition(partid, testdeflines):
refr_file = data_file('fiveparts-refr.fa.gz')
contig_file = data_file('fiveparts.contigs.augfasta.gz')
contigstream = kevlar.parse_augmented_fastx(kevlar.open(contig_file, 'r'))
pstream = kevlar.parse_single_partition(contigstream, partid)
localizer = localize(pstream, refr_file, seedsize=51)
cutoutdata = list(localizer)
partids = [partid for partid, gdna in cutoutdata]
gdnas = [gdna for partid, gdna in cutoutdata]
deflines = sorted([g.defline for g in gdnas])
assert deflines == testdeflines
def alac(pstream, refrfile, ksize=31, delta=25, maxdiff=10000, match=1,
mismatch=2, gapopen=5, gapextend=0, greedy=False,
logstream=sys.stderr):
assembler = assemble_greedy if greedy else assemble_fml_asm
for partition in pstream:
contigs = [c for c in assembler(partition, logstream=logstream)]
targets = [t for t in localize(contigs, refrfile, ksize, delta=delta)]
caller = call(
targets, contigs, match, mismatch, gapopen, gapextend, ksize
)
for varcall in caller:
yield varcall
def test_localize_new():
refr_file = data_file('fiveparts-refr.fa.gz')
contig_file = data_file('fiveparts.contigs.augfasta.gz')
contigstream = kevlar.parse_augmented_fastx(kevlar.open(contig_file, 'r'))
pstream = kevlar.parse_partitioned_reads(contigstream)
localizer = localize(pstream, refr_file, seedsize=51, debug=True)
cutoutdata = list(localizer)
partids = [partid for partid, gdna in cutoutdata]
gdnas = [gdna for partid, gdna in cutoutdata]
deflines = [g.defline for g in gdnas]
assert partids == ['1', '1', '2', '3', '4', '5']
assert sorted(deflines) == sorted([
'seq1_284663-284950', 'seq1_1924681-1925049', 'seq1_1660589-1660884',
'seq1_2315741-2316037', 'seq1_2321099-2321322', 'seq1_593102-593389'
])