def test_bam_file_to_fasta_pair_files_region(self):
'''Test bam_file_to_fasta_pair_files with a region'''
tmp1 = 'tmp.to_fasta_1.fa'
tmp2 = 'tmp.to_fasta_2.fa'
mapping.bam_file_to_fasta_pair_files(os.path.join(data_dir, 'mapping_test.smalt.out.sorted.bam'), tmp1, tmp2, chromosome='ref', start=25, end=150)
self.assertTrue(filecmp.cmp(os.path.join(data_dir, 'mapping_test.bam_to_region_1.fa'), tmp1))
self.assertTrue(filecmp.cmp(os.path.join(data_dir, 'mapping_test.bam_to_region_2.fa'), tmp2))
os.unlink(tmp1)
os.unlink(tmp2)
def test_bam_file_to_fasta_pair_files(self):
'''Test bam_file_to_fasta_pair_files'''
tmp1 = 'tmp.to_fasta_1.fa'
tmp2 = 'tmp.to_fasta_2.fa'
mapping.bam_file_to_fasta_pair_files(os.path.join(data_dir, 'mapping_test.smalt.out.bam'), tmp1, tmp2)
self.assertTrue(filecmp.cmp(os.path.join(data_dir, 'mapping_test.reads_1.fasta'), tmp1))
self.assertTrue(filecmp.cmp(os.path.join(data_dir, 'mapping_test.reads_2.fasta'), tmp2))
os.unlink(tmp1)
os.unlink(tmp2)
def _trim_strand_biased_ends(self,
reads_prefix,
out_prefix=None,
tag_as_trimmed=False,
break_contigs=False):
tmpdir = tempfile.mkdtemp(prefix='tmp.trim_strand_biased_ends.',
dir=os.getcwd())
tmp_prefix = os.path.join(tmpdir, 'out')
sorted_bam = tmp_prefix + '.bam'
unsorted_bam = tmp_prefix + '.unsorted.bam'
original_map_minid = self.map_minid
self.map_minid = 0.9
self._map_reads(reads_prefix + '_1.fa',
reads_prefix + '_2.fa',
tmp_prefix,
sort_reads=True)
assert os.path.exists(sorted_bam)
self.map_minid = original_map_minid
new_contigs = []
contigs_to_remove = set()
for ctg in self.contigs:
if break_contigs:
subcontigs = self._subcontigs_from_strand_bias(sorted_bam, ctg)
if len(subcontigs):
new_contigs.extend(subcontigs)
contigs_to_remove.add(ctg)
elif ctg not in self.contigs_trimmed_for_strand_bias:
self._trim_contig_for_strand_bias(sorted_bam, ctg)
# contig could get completely trimmed so nothing left, in which
# case, we need to remove it
if len(self.contigs[ctg]) == 0:
contigs_to_remove.add(ctg)
elif tag_as_trimmed:
self.contigs_trimmed_for_strand_bias.add(ctg)
for ctg in contigs_to_remove:
self._remove_contig(ctg)
for ctg in new_contigs:
self._add_contig(ctg, min_length=0.75 * self.self.seed_stop_length)
if out_prefix is not None:
mapping.bam_file_to_fasta_pair_files(unsorted_bam,
out_prefix + '_1.fa',
out_prefix + '_2.fa',
remove_proper_pairs=True)
shutil.rmtree(tmpdir)
def _trim_strand_biased_ends(self, reads_prefix, out_prefix=None, tag_as_trimmed=False, break_contigs=False):
tmpdir = tempfile.mkdtemp(prefix='tmp.trim_strand_biased_ends.', dir=os.getcwd())
tmp_prefix = os.path.join(tmpdir, 'out')
sorted_bam = tmp_prefix + '.bam'
unsorted_bam = tmp_prefix + '.unsorted.bam'
original_map_minid = self.map_minid
self.map_minid = 0.9
self._map_reads(reads_prefix + '_1.fa', reads_prefix + '_2.fa', tmp_prefix, sort_reads=True)
assert os.path.exists(sorted_bam)
self.map_minid = original_map_minid
new_contigs = []
contigs_to_remove = set()
for ctg in self.contigs:
if break_contigs:
subcontigs = self._subcontigs_from_strand_bias(sorted_bam, ctg)
if len(subcontigs):
new_contigs.extend(subcontigs)
contigs_to_remove.add(ctg)
elif ctg not in self.contigs_trimmed_for_strand_bias:
self._trim_contig_for_strand_bias(sorted_bam, ctg)
# contig could get completely trimmed so nothing left, in which
# case, we need to remove it
if len(self.contigs[ctg]) == 0:
contigs_to_remove.add(ctg)
elif tag_as_trimmed:
self.contigs_trimmed_for_strand_bias.add(ctg)
for ctg in contigs_to_remove:
self._remove_contig(ctg)
for ctg in new_contigs:
self._add_contig(ctg, min_length=0.75 * self.self.seed_stop_length)
if out_prefix is not None:
mapping.bam_file_to_fasta_pair_files(unsorted_bam, out_prefix + '_1.fa', out_prefix + '_2.fa', remove_proper_pairs=True)
shutil.rmtree(tmpdir)
def _get_unmapped_pairs(self, reads1, reads2, out_prefix):
self._map_reads(reads1, reads2, out_prefix, required_flag=12)
mapping.bam_file_to_fasta_pair_files(out_prefix + '.bam',
out_prefix + '_1.fa',
out_prefix + '_2.fa')
os.unlink(out_prefix + '.bam')
def _read_pair_extension_iterations(self,
reads_prefix,
out_prefix,
no_map_contigs=None):
if no_map_contigs is None:
no_map_contigs = set()
assert (len(self.contigs) > len(no_map_contigs))
if self.verbose:
print('{:-^79}'.format(' ' + out_prefix +
' start extension subiteration 0001 '),
flush=True)
bases_added = self._extend_with_reads(reads_prefix, out_prefix + '.1',
no_map_contigs)
current_reads_prefix = reads_prefix
if bases_added == 0:
return True
try_contig_trim = False
i = 1
while self._worth_extending() or try_contig_trim:
i += 1
if self.verbose:
print('{:-^79}'.format(' ' + out_prefix +
' start extension subiteration ' +
str(i).zfill(4) + ' '),
flush=True)
if i % 5 == 0:
tmpdir = tempfile.mkdtemp(prefix='tmp.filter_reads.',
dir=os.getcwd())
tmp_prefix = os.path.join(tmpdir, 'out')
bam = tmp_prefix + '.bam'
original_map_minid = self.map_minid
self.map_minid = 0.9
self._map_reads(current_reads_prefix + '_1.fa',
current_reads_prefix + '_2.fa', tmp_prefix)
self.map_minid = original_map_minid
filter_prefix = reads_prefix + '.subiter.' + str(i) + '.reads'
mapping.bam_file_to_fasta_pair_files(bam,
filter_prefix + '_1.fa',
filter_prefix + '_2.fa',
remove_proper_pairs=True)
if current_reads_prefix != reads_prefix:
os.unlink(current_reads_prefix + '_1.fa')
os.unlink(current_reads_prefix + '_2.fa')
current_reads_prefix = filter_prefix
shutil.rmtree(tmpdir)
iter_prefix = out_prefix + '.' + str(i)
bases_added = self._extend_with_reads(current_reads_prefix,
iter_prefix, no_map_contigs)
if bases_added == 0:
if not try_contig_trim:
if self.verbose:
print(' No bases added. Try trimming contigs')
self._trim_strand_biased_ends(reads_prefix,
tag_as_trimmed=False)
if len(self.contigs) <= len(no_map_contigs):
if self.verbose:
print(
' lost contigs during trimming. No more iterations'
)
return False
self.trim_contigs(self.contig_iter_trim)
try_contig_trim = True
else:
if self.verbose:
print(
' No bases added after trimming. No more iterations'
)
break
else:
try_contig_trim = False
if current_reads_prefix != reads_prefix:
os.unlink(current_reads_prefix + '_1.fa')
os.unlink(current_reads_prefix + '_2.fa')
return True
def _get_unmapped_pairs(self, reads1, reads2, out_prefix):
self._map_reads(reads1, reads2, out_prefix, required_flag=12)
mapping.bam_file_to_fasta_pair_files(out_prefix + '.bam', out_prefix + '_1.fa', out_prefix + '_2.fa')
os.unlink(out_prefix + '.bam')
def _read_pair_extension_iterations(self, reads_prefix, out_prefix, no_map_contigs=None):
if no_map_contigs is None:
no_map_contigs = set()
assert(len(self.contigs) > len(no_map_contigs))
if self.verbose:
print('{:-^79}'.format(' ' + out_prefix + ' start extension subiteration 0001 '), flush=True)
bases_added = self._extend_with_reads(reads_prefix, out_prefix + '.1', no_map_contigs)
current_reads_prefix = reads_prefix
if bases_added == 0:
return True
try_contig_trim = False
i = 1
while self._worth_extending() or try_contig_trim:
i += 1
if self.verbose:
print('{:-^79}'.format(' ' + out_prefix + ' start extension subiteration ' + str(i).zfill(4) + ' '), flush=True)
if i % 5 == 0:
tmpdir = tempfile.mkdtemp(prefix='tmp.filter_reads.', dir=os.getcwd())
tmp_prefix = os.path.join(tmpdir, 'out')
bam = tmp_prefix + '.bam'
original_map_minid = self.map_minid
self.map_minid = 0.9
self._map_reads(current_reads_prefix + '_1.fa', current_reads_prefix + '_2.fa', tmp_prefix)
self.map_minid = original_map_minid
filter_prefix = reads_prefix + '.subiter.' + str(i) + '.reads'
mapping.bam_file_to_fasta_pair_files(bam, filter_prefix + '_1.fa', filter_prefix + '_2.fa', remove_proper_pairs=True)
if current_reads_prefix != reads_prefix:
os.unlink(current_reads_prefix + '_1.fa')
os.unlink(current_reads_prefix + '_2.fa')
current_reads_prefix = filter_prefix
shutil.rmtree(tmpdir)
iter_prefix = out_prefix + '.' + str(i)
bases_added = self._extend_with_reads(current_reads_prefix, iter_prefix, no_map_contigs)
if bases_added == 0:
if not try_contig_trim:
if self.verbose:
print(' No bases added. Try trimming contigs')
self._trim_strand_biased_ends(reads_prefix, tag_as_trimmed=False)
if len(self.contigs) <= len(no_map_contigs):
if self.verbose:
print(' lost contigs during trimming. No more iterations')
return False
self.trim_contigs(self.contig_iter_trim)
try_contig_trim = True
else:
if self.verbose:
print(' No bases added after trimming. No more iterations')
break
else:
try_contig_trim = False
if current_reads_prefix != reads_prefix:
os.unlink(current_reads_prefix + '_1.fa')
os.unlink(current_reads_prefix + '_2.fa')
return True
