def filter_mappings(self):
num_unmapped = 0
num_entirely_genomic = 0
num_nonunique = 0
num_unique = 0
nongenomic_lengths = Counter()
sam_file = pysam.Samfile(self.file_names['accepted_hits'])
region_fetcher = genomes.build_region_fetcher(self.file_names['genome'],
load_references=True,
sam_file=sam_file,
)
extended_sorter = sam.AlignmentSorter(sam_file.references,
sam_file.lengths,
self.file_names['extended'],
)
filtered_sorter = sam.AlignmentSorter(sam_file.references,
sam_file.lengths,
self.file_names['extended_filtered'],
)
extended_mappings = (trim.extend_polyA_end(mapping, region_fetcher) for mapping in sam_file)
mapping_groups = utilities.group_by(extended_mappings, lambda m: m.qname)
with extended_sorter, filtered_sorter:
for qname, group in mapping_groups:
for m in group:
extended_sorter.write(m)
min_nongenomic_length = min(trim.get_nongenomic_length(m) for m in group)
nongenomic_lengths[min_nongenomic_length] += 1
if min_nongenomic_length == 0:
num_entirely_genomic += 1
continue
nonunique = len(group) > 1 or any(m.mapq < 40 for m in group)
if nonunique:
num_nonunique += 1
continue
num_unique += 1
for m in group:
filtered_sorter.write(m)
self.summary.extend(
[('Mapped with no non-genomic A\'s', num_entirely_genomic),
('Nonunique', num_nonunique),
('Unique', num_unique),
],
)
nongenomic_lengths = utilities.counts_to_array(nongenomic_lengths)
self.write_file('nongenomic_lengths', nongenomic_lengths)
def filter_mappings(self):
num_unmapped = 0
num_entirely_genomic = 0
num_nonunique = 0
num_unique = 0
nongenomic_lengths = Counter()
sam_file = pysam.Samfile(self.file_names['accepted_hits'])
region_fetcher = genomes.build_region_fetcher(
self.file_names['genome'],
load_references=True,
sam_file=sam_file,
)
extended_sorter = sam.AlignmentSorter(
sam_file.references,
sam_file.lengths,
self.file_names['extended'],
)
filtered_sorter = sam.AlignmentSorter(
sam_file.references,
sam_file.lengths,
self.file_names['extended_filtered'],
)
extended_mappings = (trim.extend_polyA_end(mapping, region_fetcher)
for mapping in sam_file)
mapping_groups = utilities.group_by(extended_mappings,
lambda m: m.qname)
with extended_sorter, filtered_sorter:
for qname, group in mapping_groups:
for m in group:
extended_sorter.write(m)
min_nongenomic_length = min(
trim.get_nongenomic_length(m) for m in group)
nongenomic_lengths[min_nongenomic_length] += 1
if min_nongenomic_length == 0:
num_entirely_genomic += 1
continue
nonunique = len(group) > 1 or any(m.mapq < 40 for m in group)
if nonunique:
num_nonunique += 1
continue
num_unique += 1
for m in group:
filtered_sorter.write(m)
self.summary.extend([
('Mapped with no non-genomic A\'s', num_entirely_genomic),
('Nonunique', num_nonunique),
('Unique', num_unique),
], )
nongenomic_lengths = utilities.counts_to_array(nongenomic_lengths)
self.write_file('nongenomic_lengths', nongenomic_lengths)
def get_Transcript_position_counts(clean_bam_fn,
transcripts,
relevant_lengths,
left_buffer=left_buffer,
right_buffer=right_buffer,
):
gene_infos = {}
bam_file = pysam.Samfile(clean_bam_fn)
max_nongenomic_length = 5
for transcript in transcripts:
transcript.build_coordinate_maps(left_buffer, right_buffer)
nonunique = 0
alternatively_spliced = 0
landmarks = {'start': 0,
'start_codon': transcript.transcript_start_codon,
'stop_codon': transcript.transcript_stop_codon,
'end': transcript.transcript_length,
}
five_prime_positions = {l: PositionCounts(landmarks, left_buffer, right_buffer)
for l in relevant_lengths + ['all', 'all_nonunique']}
three_prime_positions = {l: PositionCounts(landmarks, left_buffer, right_buffer)
for l in range(max_nongenomic_length + 1) + ['all', 'all_nonunique']}
transcript_sequence = transcript.get_transcript_sequence(left_buffer, right_buffer)
# fetch raises a ValueError if given a negative start, but it doesn't
# care if the end is valid.
left_edge = max(0, transcript.start - left_buffer)
right_edge = transcript.end + right_buffer
overlapping_reads = bam_file.fetch(transcript.seqname, left_edge, right_edge)
for read in overlapping_reads:
if any(transcript.is_spliced_out(position) for position in read.positions):
alternatively_spliced += 1
continue
if read.mapq != 50:
nonunique += 1
is_unique = False
else:
is_unique = True
read_strand = '-' if read.is_reverse else '+'
if read_strand != transcript.strand:
continue
left_edge = read.pos
right_edge = read.aend - 1
if read_strand == '+':
five_prime_position = left_edge
three_prime_position = right_edge
elif read_strand == '-':
five_prime_position = right_edge
three_prime_position = left_edge
if five_prime_position in transcript.genomic_to_transcript:
transcript_coord = transcript.genomic_to_transcript[five_prime_position]
if is_unique:
five_prime_positions['all']['start', transcript_coord] += 1
if read.qlen in relevant_lengths:
five_prime_positions[read.qlen]['start', transcript_coord] += 1
elif not read.is_secondary:
five_prime_positions['all_nonunique']['start', transcript_coord] += 1
if three_prime_position in transcript.genomic_to_transcript:
transcript_coord = transcript.genomic_to_transcript[three_prime_position]
if is_unique:
three_prime_positions['all']['start', transcript_coord] += 1
nongenomic_length = trim.get_nongenomic_length(read)
if nongenomic_length <= max_nongenomic_length:
three_prime_positions[nongenomic_length]['start', transcript_coord] += 1
elif not read.is_secondary:
three_prime_positions['all_nonunique']['start', transcript_coord] += 1
gene_infos[transcript.name] = {'CDS_length': transcript.CDS_length,
'five_prime_positions': five_prime_positions,
'three_prime_positions': three_prime_positions,
'nonunique': nonunique,
'alternatively_spliced': alternatively_spliced,
'sequence': transcript_sequence,
}
transcript.delete_coordinate_maps()
return gene_infos
def get_Transcript_position_counts(
clean_bam_fn,
transcripts,
relevant_lengths,
left_buffer=left_buffer,
right_buffer=right_buffer,
):
gene_infos = {}
bam_file = pysam.Samfile(clean_bam_fn)
max_nongenomic_length = 5
for transcript in transcripts:
transcript.build_coordinate_maps(left_buffer, right_buffer)
nonunique = 0
alternatively_spliced = 0
landmarks = {
'start': 0,
'start_codon': transcript.transcript_start_codon,
'stop_codon': transcript.transcript_stop_codon,
'end': transcript.transcript_length,
}
five_prime_positions = {
l: PositionCounts(landmarks, left_buffer, right_buffer)
for l in relevant_lengths + ['all', 'all_nonunique']
}
three_prime_positions = {
l: PositionCounts(landmarks, left_buffer, right_buffer)
for l in range(max_nongenomic_length + 1) +
['all', 'all_nonunique']
}
transcript_sequence = transcript.get_transcript_sequence(
left_buffer, right_buffer)
# fetch raises a ValueError if given a negative start, but it doesn't
# care if the end is valid.
left_edge = max(0, transcript.start - left_buffer)
right_edge = transcript.end + right_buffer
overlapping_reads = bam_file.fetch(transcript.seqname, left_edge,
right_edge)
for read in overlapping_reads:
if any(
transcript.is_spliced_out(position)
for position in read.positions):
alternatively_spliced += 1
continue
if read.mapq != 50:
nonunique += 1
is_unique = False
else:
is_unique = True
read_strand = '-' if read.is_reverse else '+'
if read_strand != transcript.strand:
continue
left_edge = read.pos
right_edge = read.aend - 1
if read_strand == '+':
five_prime_position = left_edge
three_prime_position = right_edge
elif read_strand == '-':
five_prime_position = right_edge
three_prime_position = left_edge
if five_prime_position in transcript.genomic_to_transcript:
transcript_coord = transcript.genomic_to_transcript[
five_prime_position]
if is_unique:
five_prime_positions['all']['start', transcript_coord] += 1
if read.qlen in relevant_lengths:
five_prime_positions[read.qlen]['start',
transcript_coord] += 1
elif not read.is_secondary:
five_prime_positions['all_nonunique'][
'start', transcript_coord] += 1
if three_prime_position in transcript.genomic_to_transcript:
transcript_coord = transcript.genomic_to_transcript[
three_prime_position]
if is_unique:
three_prime_positions['all']['start',
transcript_coord] += 1
nongenomic_length = trim.get_nongenomic_length(read)
if nongenomic_length <= max_nongenomic_length:
three_prime_positions[nongenomic_length][
'start', transcript_coord] += 1
elif not read.is_secondary:
three_prime_positions['all_nonunique'][
'start', transcript_coord] += 1
gene_infos[transcript.name] = {
'CDS_length': transcript.CDS_length,
'five_prime_positions': five_prime_positions,
'three_prime_positions': three_prime_positions,
'nonunique': nonunique,
'alternatively_spliced': alternatively_spliced,
'sequence': transcript_sequence,
}
transcript.delete_coordinate_maps()
return gene_infos