def test_pcdhit():
import os
import lilbio
import pcdhit
from lilbio.funcs import uppercase_only
source = os.path.join(tests_dir(), alignment_file)
records = lilbio.parse(source, 'stockholm', func=uppercase_only)
filtered_records = pcdhit.filter(records, 0.7)
assert len(list(filtered_records)) == 61
def filter(records, threshold):
"""Filter non-redundant records via cd-hit.
cdhit: http://weizhongli-lab.org/cd-hit/
cdhit will cluster sequences that meet a similarity threshold and return a
representative record for each cluster:
cdhit -i <fin> -c <threshold> -o <fout>
Parameters
----------
records : iterable
Iterable of (header, sequence) tuples.
threshold : float, optional (0.9)
Sequence identity threshold (cd-hit '-c <thr>' option).
Yields
------
(header, sequence) : tuple (str, str)
For each non-redundant record, a tuple containing header and sequence.
"""
# check for cd-hit on path
cdhit_exe = is_command(['cd-hit', 'cdhit'])
logger.debug('cd-hit executable: %r', cdhit_exe)
if cdhit_exe is None:
raise CdhitNotFoundError
if not 0.7 <= threshold <= 1.0:
raise IdentityThresholdError
# open tmp files
with opentf() as fin, opentf() as fout:
print_input_fasta(records, fin)
call_cdhit(cdhit_exe, fin, fout, threshold)
for rec in lilbio.parse(fout, fmt='fasta'):
head, seq = rec[0].split('@')
yield head, seq
def test_stockholm():
fname = os.path.join(tests_dir(), '1.sto')
a = list(lilbio.parse(fname, 'stockholm'))
assert repr(a) == RECORDS
def test_fasta():
fname = os.path.join(tests_dir(), '1.fa')
a = list(lilbio.parse(fname, 'fasta'))
assert repr(a) == RECORDS