def process_full_length_mappings(self):
clean_bam = pysam.Samfile(self.file_names['clean_bam'])
type_shape = (self.max_read_length + 1,
self.max_read_length,
fastq.MAX_EXPECTED_QUAL + 1,
6,
6,
)
type_counts = np.zeros(type_shape, int)
# To avoid counting mismatches in non-unique mappings multiple times,
# a dummy secondary_type_counts array is passed to
# trim_mismatches_from_start for secondary mappings.
secondary_type_counts = np.zeros(type_shape, int)
clean_trimmed_length_counts = Counter()
region_fetcher = genomes.build_region_fetcher(self.file_names['genome'],
load_references=True,
sam_file=clean_bam,
)
for mapping in clean_bam:
if mapping.is_secondary:
counts_array = secondary_type_counts
else:
counts_array = type_counts
trimmed_from_start = trim.trim_mismatches_from_start(mapping,
region_fetcher,
counts_array,
)
trimmed_from_end = trim.trim_nongenomic_polyA_from_end(trimmed_from_start,
region_fetcher,
)
if not trimmed_from_end.is_unmapped and not trimmed_from_end.is_secondary:
clean_trimmed_length_counts[trimmed_from_end.qlen] += 1
yield trimmed_from_end
self.write_file('mismatches', type_counts)
clean_trimmed_lengths = self.zero_padded_array(clean_trimmed_length_counts)
self.write_file('lengths', {'clean_trimmed': clean_trimmed_lengths})
def process_remapped(self):
clean_bam = pysam.Samfile(self.file_names['remapped_accepted_hits'])
type_shape = (self.max_read_length + 1,
self.max_read_length,
fastq.MAX_EXPECTED_QUAL + 1,
6,
6,
)
type_counts = np.zeros(type_shape, int)
remapped_length_counts = Counter()
region_fetcher = genomes.build_region_fetcher(self.file_names['genome'],
load_references=True,
sam_file=clean_bam,
)
for mapping in clean_bam:
trimmed_from_start = trim.trim_mismatches_from_start(mapping,
region_fetcher,
type_counts,
)
# Add back any genomic A's that were trimmed as part of mappings and
# any remaining A's from the first non-genomic onward as soft clipped
# bases for visualization in IGV.
extended = trim.extend_polyA_end(trimmed_from_start,
region_fetcher,
trimmed_twice=True,
)
if not extended.is_unmapped and not extended.is_secondary:
remapped_length_counts[extended.qlen] += 1
yield extended
remapped_lengths = self.zero_padded_array(remapped_length_counts)
self.write_file('lengths', {'remapped': remapped_lengths})
