def get_reads_unpaired(sample, datahub):
logger.info("Loading more reads...")
cur_reads = []
search_regions = datahub.variant.search_regions(sample.search_distance)
for region in search_regions:
chrom, start, end = region.chrom, region.start, region.end
chrom = misc.match_chrom_format(chrom, sample.bam.references)
for read in sample.bam.fetch(chrom, start, end):
# if read.query_name != "m150105_192231_42177R_c100761782550000001823161607221526_s1_p0/138972/39862_46995":
# continue
if read.is_supplementary or read.is_duplicate or read.is_secondary:
continue
if datahub.args.min_mapq and read.mapq < datahub.args.min_mapq:
continue
cur_reads.append(alignment.Alignment(read))
if datahub.args.batch_size is not None and len(
cur_reads) >= datahub.args.batch_size:
yield cur_reads
logger.info("Loading more reads...")
cur_reads = []
yield cur_reads
def get_seq(self, chrom, start, end, strand):
chrom = misc.match_chrom_format(chrom, list(self.fasta.keys()))
seq = self.fasta[chrom][start:end + 1]
if strand == "-":
seq = misc.reverse_comp(seq)
return seq
def blacklist(self, blacklist_loci):
self._blacklist = []
for locus in blacklist_loci:
cur_chrom = misc.match_chrom_format(locus.chrom, list(self.keys()))
self._blacklist.append(
intervals.Locus(cur_chrom, locus.start, locus.end,
locus.strand))
def do_homology_search(datahub):
logger.info(
"Finding homologous genomic regions (segmental duplications)...")
variant = datahub.variant
segment_loci = []
for part in variant.chrom_parts("ref"):
for segment in part.segments:
cur_chrom = misc.match_chrom_format(segment.chrom,
datahub.genome.keys())
segment = intervals.Locus(cur_chrom, max(0, segment.start - 100),
segment.end + 100, "+")
segment_loci.append(segment)
for part in variant.chrom_parts("ref"):
breakpoints = numpy.cumsum([len(segment)
for segment in part.segments])[:-1]
seq = part.get_seq()
homologous_regions = find_homologous_regions(seq, datahub.genome,
segment_loci)
plot_homologous_regions(seq, homologous_regions, datahub.genome,
part.id, breakpoints)
# also look for any reasonably-sized segments that are unique to the alt allele
ref_ids = set(segment.id for part in variant.chrom_parts("ref")
for segment in part.segments)
alt_ids = set(segment.id for part in variant.chrom_parts("alt")
for segment in part.segments)
alt_only_ids = alt_ids - ref_ids
alt_only_segments = [
segment for part in variant.chrom_parts("alt")
for segment in part.segments if segment.id in alt_only_ids
]
for segment in alt_only_segments:
if len(segment) >= 50:
seq = datahub.variant.sources[segment.chrom].get_seq(
segment.chrom, segment.start, segment.end, segment.strand)
homologous_regions = find_homologous_regions(
seq, datahub.genome, [])
plot_homologous_regions(seq,
homologous_regions,
datahub.genome,
label="segment_{}".format(segment))