def process_remapped_unmapped(self):
unmapped_lengths = np.zeros(self.max_read_length + 1, int)
unmapped_seq_counts = Counter()
long_polyA_lengths = np.zeros(self.max_read_length + 1, int)
long_polyA_counts = Counter()
unmapped_reads = sam.bam_to_fastq(self.file_names['remapped_unmapped_bam'])
unretrimmed_reads = trim.untrim_reads(unmapped_reads, second_time=True)
synthetic_filtered_reads = self.filter_synthetic_sequences(unretrimmed_reads)
def record_common(reads):
for read in reads:
if predominantly_A(read.seq):
long_polyA_lengths[len(read.seq)] += 1
long_polyA_counts[read.seq] += 1
else:
unmapped_lengths[len(read.seq)] += 1
unmapped_seq_counts[read.seq] += 1
yield read
synthetic_filtered_reads = record_common(synthetic_filtered_reads)
seq_info_pairs = ((read.seq, False) for read in synthetic_filtered_reads)
all_array, _ = composition.length_stratified_composition(seq_info_pairs, self.max_read_length)
self.write_file('unmapped_composition', all_array)
self.write_file('lengths', {'unmapped': unmapped_lengths,
'long_polyA': long_polyA_lengths,
},
)
non_long_polyA = Counter(dict(unmapped_seq_counts.most_common(100)))
long_polyA = Counter(dict(long_polyA_counts.most_common(100)))
common_unmapped = {'non_long_polyA': non_long_polyA,
'long_polyA': long_polyA,
}
self.write_file('common_unmapped', common_unmapped)
def compute_base_composition(self):
seq_info_pairs = composition.get_seq_info_pairs(self.file_names['clean_bam'])
all_array, perfect_array = composition.length_stratified_composition(seq_info_pairs, self.max_read_length)
self.write_file('clean_composition', all_array)
self.write_file('clean_composition_perfect', perfect_array)
