def extract_boundary_sequences(self):
read_pairs = self.get_read_pairs()
trimmed_read_pairs = self.trim_barcodes(read_pairs)
total_reads = 0
well_formed = 0
long_enough = 0
counters = {'positions': {orientation: Counter() for orientation in orientations},
'control_ids': Counter(),
'polyA_lengths': Counter(),
'left_ids': Counter(),
'right_ids': Counter(),
'joint_lengths': Counter(),
}
with open(self.file_names['five_prime_boundaries'], 'w') as fives_fh, \
open(self.file_names['three_prime_boundaries'], 'w') as threes_fh:
for R1, R2 in trimmed_read_pairs:
total_reads += 1
five_payload_read, three_payload_read = TIF_seq_structure.find_boundary_sequences(R1, R2, counters)
if five_payload_read and three_payload_read:
well_formed += 1
if len(five_payload_read.seq) >= self.min_payload_length and \
len(three_payload_read.seq) >= self.min_payload_length:
long_enough += 1
fives_fh.write(fastq.make_record(*five_payload_read))
threes_fh.write(fastq.make_record(*three_payload_read))
# Pop off of counters so that what is left at the end can be written
# directly to the id_counts file.
position_counts = counters.pop('positions')
for orientation in orientations:
key = '{0}_{1}'.format(orientation, 'positions')
array = counts_to_array(position_counts[orientation])
self.write_file(key, array)
polyA_lengths = counts_to_array(counters.pop('polyA_lengths'))
self.write_file('polyA_lengths', polyA_lengths)
joint_lengths = counts_to_array(counters.pop('joint_lengths'), dim=2)
self.write_file('joint_lengths', joint_lengths)
self.write_file('id_counts', counters)
self.summary.extend(
[('Total read pairs', total_reads),
('Well-formed', well_formed),
('Long enough', long_enough),
],
)
def preprocess(self):
''' tophat can't handle named pipes, so need to make a file. '''
reads = self.get_reads()
total_reads = 0
with open(self.file_names['preprocessed_reads'], 'w') as preprocessed_file:
for read in reads:
total_reads += 1
record = fastq.make_record(*read)
preprocessed_file.write(record)
self.summary.extend(
[('Total reads', total_reads),
],
)
def preprocess(self):
''' tophat can't handle named pipes, so need to make a file. '''
reads = self.get_reads()
total_reads = 0
with open(self.file_names['preprocessed_reads'],
'w') as preprocessed_file:
for read in reads:
total_reads += 1
record = fastq.make_record(*read)
preprocessed_file.write(record)
self.summary.extend([
('Total reads', total_reads),
], )
def trim_reads(self, read_pairs):
total_reads = 0
long_enough_reads = 0
trimmed_lengths = Counter()
barcodes = Counter()
truncated_in_R1 = self.adapter_in_R1[1:]
truncated_in_R2 = self.adapter_in_R2[1:]
for R1, R2 in read_pairs:
total_reads += 1
barcodes[R2.seq[:len(self.barcode)]] += 1
# Check for weird thing where expected overhang base doesn't
# exist in primer dimers.
R1_dimer_distance = adapters.adapter_hamming_distance(
R1.seq,
truncated_in_R1,
len(R1.seq),
len(truncated_in_R1),
len(self.barcode),
)
R2_dimer_distance = adapters.adapter_hamming_distance(
R2.seq,
truncated_in_R2,
len(R2.seq),
len(truncated_in_R2),
len(self.barcode),
)
if R1_dimer_distance <= 3 and R2_dimer_distance <= 3:
position = len(self.barcode)
else:
position = adapters.consistent_paired_position(
R1.seq,
R2.seq,
self.adapter_in_R1,
self.adapter_in_R2,
19,
3,
)
if position != None:
trimmed_lengths[position] += 1
if position - len(self.barcode) < 12:
continue
else:
position = len(R1.seq)
long_enough_reads += 1
payload_slice = slice(len(self.barcode), position)
processed_R1 = fastq.Read(R1.name, R1.seq[payload_slice],
R1.qual[payload_slice])
processed_R2 = fastq.make_record(R2.name, R2.seq[payload_slice],
R2.qual[payload_slice])
yield processed_R1, processed_R2
trimmed_lengths = utilities.counts_to_array(trimmed_lengths)
self.write_file('trimmed_lengths', trimmed_lengths)
self.write_file('barcodes', barcodes)
self.summary.extend([
('Total read pairs', total_reads),
('Long enough', long_enough_reads),
])
def trim_reads(self, read_pairs):
total_reads = 0
long_enough_reads = 0
trimmed_lengths = Counter()
barcodes = Counter()
truncated_in_R1 = self.adapter_in_R1[1:]
truncated_in_R2 = self.adapter_in_R2[1:]
for R1, R2 in read_pairs:
total_reads += 1
barcodes[R2.seq[:len(self.barcode)]] += 1
# Check for weird thing where expected overhang base doesn't
# exist in primer dimers.
R1_dimer_distance = adapters.adapter_hamming_distance(R1.seq,
truncated_in_R1,
len(R1.seq),
len(truncated_in_R1),
len(self.barcode),
)
R2_dimer_distance = adapters.adapter_hamming_distance(R2.seq,
truncated_in_R2,
len(R2.seq),
len(truncated_in_R2),
len(self.barcode),
)
if R1_dimer_distance <= 3 and R2_dimer_distance <= 3:
position = len(self.barcode)
else:
position = adapters.consistent_paired_position(R1.seq,
R2.seq,
self.adapter_in_R1,
self.adapter_in_R2,
19,
3,
)
if position != None:
trimmed_lengths[position] += 1
if position - len(self.barcode) < 12:
continue
else:
position = len(R1.seq)
long_enough_reads += 1
payload_slice = slice(len(self.barcode), position)
processed_R1 = fastq.Read(R1.name, R1.seq[payload_slice], R1.qual[payload_slice])
processed_R2 = fastq.make_record(R2.name, R2.seq[payload_slice], R2.qual[payload_slice])
yield processed_R1, processed_R2
trimmed_lengths = utilities.counts_to_array(trimmed_lengths)
self.write_file('trimmed_lengths', trimmed_lengths)
self.write_file('barcodes', barcodes)
self.summary.extend(
[('Total read pairs', total_reads),
('Long enough', long_enough_reads),
]
)
