def get_inferred_sequences(pairs, genome_dict, add_softclipped_bases=False):
inferred_sequences = []
for read1, read2 in pairs:
if read1.query_name.count('_') == 2:
context_width = int(read1.query_name.split('_')[-2])
name = read1.reference_name + ':' + str(
read1.reference_start +
context_width) + '-' + str(read2.reference_end - context_width)
inferred_sequence = genome_dict[read1.reference_name][
read1.reference_start:read2.reference_end]
if add_softclipped_bases:
inferred_sequence = sctools.left_softclipped_sequence_strict(
read1
) + inferred_sequence + sctools.right_softclipped_sequence_strict(
read2)
inferred_sequence = inferred_sequence[context_width:-context_width]
if read1.query_name.split('_')[-1] == '2':
inferred_sequence = misc.revcomp(inferred_sequence)
contig_edge = False
if sctools.is_left_softclipped_strict(read1) and \
sctools.left_softclipped_position(read1) < 0:
contig_edge = True
elif sctools.is_right_softclipped_strict(read2) and \
sctools.right_softclipped_position(read2) >= len(genome_dict[read2.reference_name]):
contig_edge = True
else:
name = read1.reference_name + ':' + str(
read1.reference_start) + '-' + str(read2.reference_end)
inferred_sequence = genome_dict[read1.reference_name][
read1.reference_start:read2.reference_end]
if add_softclipped_bases:
inferred_sequence = sctools.left_softclipped_sequence_strict(
read1
) + inferred_sequence + sctools.right_softclipped_sequence_strict(
read2)
if read1.query_name.split('_')[-1] == '2':
inferred_sequence = misc.revcomp(inferred_sequence)
contig_edge = False
if sctools.is_left_softclipped_strict(read1) and \
sctools.left_softclipped_position(read1) < 0:
contig_edge = True
elif sctools.is_right_softclipped_strict(read2) and \
sctools.right_softclipped_position(read2) >= len(genome_dict[read2.reference_name]):
contig_edge = True
inferred_sequences.append(
(name, len(inferred_sequence), contig_edge, inferred_sequence))
return inferred_sequences
def get_inferred_sequence(self, forward_read, reverse_read, is_reverse):
contig = forward_read.reference_name
start = forward_read.reference_start
end = reverse_read.reference_end
inferred_sequence = ''.join(self.genome_dict[contig][start:end])
inferred_sequence = sctools.left_softclipped_sequence_strict(forward_read) + \
inferred_sequence + \
sctools.right_softclipped_sequence_strict(reverse_read)
inferred_sequence = inferred_sequence[self.context_width:-self.context_width]
if is_reverse:
inferred_sequence = misc.revcomp(inferred_sequence)
contig_edge = False
if sctools.is_left_softclipped_strict(forward_read) and \
sctools.left_softclipped_position(forward_read) < 0:
contig_edge = True
elif sctools.is_right_softclipped_strict(reverse_read) and \
sctools.right_softclipped_position(reverse_read) >= len(self.genome_dict[contig]):
contig_edge = True
return inferred_sequence, contig_edge
def prefilter_reads(bam, database_dict, min_perc_identity,
max_internal_softclip_prop, max_edge_distance):
keep_reads = defaultdict(lambda: defaultdict(lambda: defaultdict(list)))
for read in bam:
if pysamtools.get_perc_identity(read) < min_perc_identity:
continue
if not read.is_reverse:
if sctools.is_right_softclipped_strict(read) and \
sctools.right_softclipped_position(read) < len(database_dict[read.reference_name]) and \
sctools.right_softclip_proportion(read) > max_internal_softclip_prop:
continue
elif read.reference_start > max_edge_distance:
continue
elif sctools.is_left_softclipped_strict(read) and \
abs(0 - sctools.left_softclip_reference_start(read)) > max_edge_distance:
continue
if read.is_reverse:
if sctools.is_left_softclipped_strict(read) and \
sctools.left_softclipped_position(read) >= 0 and \
sctools.left_softclip_proportion(read) > max_internal_softclip_prop:
continue
elif (len(database_dict[read.reference_name]) -
read.reference_end) > max_edge_distance:
continue
elif sctools.is_right_softclipped_strict(read) and \
abs(0 - (len(database_dict[read.reference_name]) - sctools.right_softclip_reference_end(read))) > max_edge_distance:
continue
pair_id, terminus_id = read.query_name.split('_')
keep_reads[pair_id][read.reference_name][terminus_id].append(read)
return keep_reads
def filter_pairs_max_internal_softclip_prop(self, max_internal_softclip_prop):
keep_pairs = list()
for p in self.pairs:
if sctools.is_left_softclipped_strict(p.forward_read) and \
sctools.get_left_softclip_length(p.forward_read) > 1 and \
sctools.is_right_softclipped_strict(p.reverse_read) and \
sctools.get_right_softclip_length(p.reverse_read) > 1:
continue
if sctools.is_right_softclipped_strict(p.forward_read) and \
p.forward_read.reference_end < p.reverse_read.reference_end and \
sctools.right_softclip_proportion(p.forward_read) > max_internal_softclip_prop:
continue
if sctools.is_left_softclipped_strict(p.reverse_read) and \
p.reverse_read.reference_start > p.forward_read.reference_start and \
sctools.left_softclip_proportion(p.reverse_read) > max_internal_softclip_prop:
continue
keep_pairs.append(p)
self.pairs = keep_pairs
def __prefilter_reads(self):
filtered_bam = []
for read in self.bam:
if pysamtools.get_perc_identity(read) < self.min_perc_identity:
continue
if not read.is_reverse:
if sctools.is_left_softclipped_strict(read) and \
sctools.get_left_softclip_length(read) > 1 and \
sctools.left_softclipped_position(read) >= 0:
continue
if read.is_reverse:
if sctools.is_right_softclipped_strict(read) and \
sctools.get_right_softclip_length(read) > 1 and \
sctools.right_softclipped_position(read) < len(self.genome_dict[read.reference_name]):
continue
filtered_bam.append(read)
self.bam = filtered_bam
