def classify_mapped_reads_new(bamfile,
settings=get_setting('CHIMERAS_SETTINGS'),
file_format='fastq',
mate_length_range=[2000, 4000],
out_format=SEQITEM):
#settings. Include in function properties with default values
max_coincidences = settings['MAX_COINCIDENCES']
max_mapq_difference = settings['MAX_MAPQ_DIFFERENCE']
limit = settings['MAX_DISTANCE_TO_END']
max_clipping = settings['MAX_CLIPPING']
max_pe_len = settings['MAX_PE_LEN']
min_mp_len = settings['MIN_MP_LEN']
#It tries to find out the kind of each pair of sequences
for grouped_mates in _group_alignments_by_reads(bamfile):
mates_alignments = _split_mates(grouped_mates)
if _mates_are_not_chimeric(mates_alignments, max_clipping,
mate_length_range, bamfile,
max_coincidences, max_mapq_difference,
limit):
kind = NON_CHIMERIC
elif _mates_are_chimeric(mates_alignments, bamfile, max_clipping,
max_pe_len):
kind = CHIMERA
else:
kind = UNKNOWN
if out_format == SEQITEM:
pair = [alignedread_to_seqitem(read[0], file_format) for read in mates_alignments]
elif out_format == 'aligned_read':
pair = mates_alignments
yield [pair, kind]
def classify_mapped_reads(bam_fhand, mate_distance,
settings=get_setting('CHIMERAS_SETTINGS')):
'''It classifies sequences from bam file in chimeric, unknown and
non chimeric, according to its distance and orientation in the reference
sequence'''
bamfile = Samfile(bam_fhand.name)
# settings. Include in function properties with default values
max_clipping = settings['MAX_CLIPPING']
max_pe_len = settings['MAX_PE_LEN']
variation = settings['MATE_DISTANCE_VARIATION']
mate_length_range = [mate_distance - variation, mate_distance + variation]
reference_lengths = _get_ref_lengths(bamfile)
# It tries to find out the kind of each pair of sequences
for grouped_mates in _group_alignments_reads_by_qname(bamfile):
mates_alignments = _split_mates(grouped_mates)
if _mates_are_not_chimeric(mates_alignments, max_clipping,
mate_length_range, bamfile,
reference_lengths):
kind = NON_CHIMERIC
elif _mates_are_chimeric(mates_alignments, bamfile, max_clipping,
max_pe_len, reference_lengths):
kind = CHIMERA
else:
kind = UNKNOWN
pair = [alignedread_to_seqitem(_get_primary_alignment(mates))
for mates in mates_alignments]
if None not in pair:
yield pair, kind
def classify_mapped_reads(bam_fhand,
mate_distance,
settings=get_setting('CHIMERAS_SETTINGS')):
'''It classifies sequences from bam file in chimeric, unknown and
non chimeric, according to its distance and orientation in the reference
sequence'''
bamfile = AlignmentFile(bam_fhand.name)
# settings. Include in function properties with default values
max_clipping = settings['MAX_CLIPPING']
max_pe_len = settings['MAX_PE_LEN']
variation = settings['MATE_DISTANCE_VARIATION']
mate_length_range = [mate_distance - variation, mate_distance + variation]
reference_lengths = _get_ref_lengths(bamfile)
# It tries to find out the kind of each pair of sequences
for grouped_mates in _group_alignments_reads_by_qname(bamfile):
mates_alignments = _split_mates(grouped_mates)
if _mates_are_not_chimeric(mates_alignments, max_clipping,
mate_length_range, bamfile,
reference_lengths):
kind = NON_CHIMERIC
elif _mates_are_chimeric(mates_alignments, bamfile, max_clipping,
max_pe_len, reference_lengths):
kind = CHIMERA
else:
kind = UNKNOWN
pair = [
alignedread_to_seqitem(_get_primary_alignment(mates))
for mates in mates_alignments
]
if None not in pair:
yield pair, kind
def _do_trim(self, aligned_reads):
max_clipping = self.max_clipping
primary_alignment = _get_primary_alignment(aligned_reads)
_5end = _get_longest_5end_alinged_read(aligned_reads, max_clipping)
seq = alignedread_to_seqitem(primary_alignment)
segments = None
if _5end is not None:
if not _read_is_totally_mapped([_5end], max_clipping):
if not _5end.is_reverse:
qend = _get_qend(_5end)
else:
qend = get_length(seq) - _get_qstart(_5end)
segments = [(qend, get_length(seq) - 1)]
if segments is not None:
_add_trim_segments(segments, seq, kind=OTHER)
return seq
def _do_trim(self, aligned_reads):
max_clipping = self.max_clipping
primary_alignment = _get_primary_alignment(aligned_reads)
_5end = _get_longest_5end_alinged_read(aligned_reads, max_clipping)
seq = alignedread_to_seqitem(primary_alignment)
segments = None
if _5end is not None:
if not _read_is_totally_mapped([_5end], max_clipping):
if not _5end.is_reverse:
qend = _get_qend(_5end)
else:
qend = get_length(seq) - _get_qstart(_5end)
segments = [(qend, get_length(seq) - 1)]
if segments is not None:
_add_trim_segments(segments, seq, kind=OTHER)
return seq
def classify_mapped_reads(bamfile, settings=get_setting('CHIMERAS_SETTINGS'),
paired_result=True, file_format='fastq'):
#settings. Include in function properties with default values
max_coincidences = settings['MAX_COINCIDENCES']
max_mapq_difference = settings['MAX_MAPQ_DIFFERENCE']
limit = settings['MAX_DISTANCE_TO_END']
max_clipping = settings['MAX_CLIPPING']
max_pe_len = settings['MAX_PE_LEN']
min_mp_len = settings['MIN_MP_LEN']
#parameters for statistics
len_mp_non_chimeric = 0
n_non_chimeric_mates = 0
len_pe_sum = 0
len_mp_sum = 0
typic_chimeras = 0
pe_like_chimeras = 0
#It tries to find out the kind of each pair of sequences
for grouped_mates in _group_alignments_by_reads(bamfile):
mates_alignments = _split_mates(grouped_mates)
i = 0
pair = []
for alignments_group in mates_alignments:
i += 1
mate = _get_mate(i, mates_alignments)
primary_mate = _get_primary_alignment(mate)
primary_alignment = _get_primary_alignment(alignments_group)
if _read_is_totally_mapped(alignments_group, max_clipping):
if primary_alignment.mate_is_unmapped:
kind = UNKNOWN
else:
if primary_alignment.rname != primary_alignment.rnext:
kind = NON_CHIMERIC
else:
mates = [primary_mate, primary_alignment]
distance = _find_distance(mates)
if _mapped_read_is_chimeric(primary_alignment, mate,
max_mapq_difference,
max_pe_len):
kind = CHIMERA
elif _mates_are_outies(mates) and distance > min_mp_len:
len_mp_non_chimeric += abs(distance)
n_non_chimeric_mates += 1
kind = NON_CHIMERIC
else:
kind = UNKNOWN
else:
fragment = _find_secondary_fragment(alignments_group,
max_coincidences,
max_mapq_difference=100)
#For fragmented reads. likely to be chimeric
if fragment is not None:
fragments = [primary_alignment, fragment]
if _alignments_in_same_ref(fragments):
if _read_is_typical_chimera(fragments, max_pe_len,
min_mp_len, mate):
len_pe, len_mp = _find_distances_with_mate(fragments, mate)
len_pe_sum += len_pe
len_mp_sum += len_mp
typic_chimeras += 1
#typic chimera
kind = CHIMERA
else:
if (_alignment_at_ends_of_reference(fragments[0],
limit, bamfile)
and _alignment_at_ends_of_reference(fragments[1],
limit, bamfile)):
kind = _guess_kind_fragments_at_ends(fragments,
bamfile, limit)
else:
kind = CHIMERA
else:
#Find PE-like chimeras in partially mapping reads
if primary_alignment.is_unmapped:
kind = UNKNOWN
elif not _3end_mapped(primary_alignment, max_clipping):
kind = UNKNOWN
else:
if primary_alignment.rname == primary_alignment.rnext:
if _mapped_read_is_chimeric(primary_alignment,
mate, max_mapq_difference,
max_pe_len):
pe_like_chimeras += 1
kind = CHIMERA
else:
kind = UNKNOWN
else:
kind = UNKNOWN
read = [alignedread_to_seqitem(alignments_group[0], file_format),
kind]
#read = [alignments_group, kind]
if paired_result == False:
yield [read[0]], read[1]
else:
pair.append(read)
if paired_result:
kinds = [read[1] for read in pair]
reads = [read[0] for read in pair]
if CHIMERA in kinds:
yield [reads, CHIMERA]
elif UNKNOWN in kinds:
yield [reads, UNKNOWN]
else:
yield [reads, NON_CHIMERIC]
try:
mean_mp_non_chimeric = len_mp_non_chimeric / float(n_non_chimeric_mates)
mean_pe_len = len_pe_sum / float(typic_chimeras)
mean_mp_len = len_mp_sum / float(typic_chimeras)
print 'MP mean length in non chimeric reads: ', mean_mp_non_chimeric
print 'Typical chimera number: ', typic_chimeras
print 'PE-like chimera number: ', pe_like_chimeras
print 'PE mean length in chimeric reads: ', mean_pe_len
print 'MP mean length in chimeric reads: ', mean_mp_len
except ZeroDivisionError:
print 'typic chimeras or non_chimeric_mates not found'