def write_output(transcripts,
cluster_shelve_file,
cluster_pair_file,
read_name_file,
output_file,
annotation_source="ensembl"):
# load cluster and read name database files
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
read_name_fh = open(read_name_file, 'r')
# map genome coordinates to transcripts
logging.debug(
"Creating mapping between genome coordinates and transcripts")
transcript_dict, genome_tx_trees = build_genome_transcript_trees(
transcripts)
logging.debug("Writing output")
outfh = open(output_file, "w")
print >> outfh, '#' + '\t'.join(Chimera._fields)
for cluster_pair in parse_discordant_cluster_pair_file(
open(cluster_pair_file)):
c = make_chimera(cluster_pair, cluster_shelve, transcript_dict,
genome_tx_trees, annotation_source)
print >> outfh, str(c)
# cleanup
outfh.close()
read_name_fh.close()
cluster_shelve.close()
return config.JOB_SUCCESS
def process_spanning_alignments(cluster_shelve_file,
cluster_pair_file,
bam_file,
output_sam_file,
output_cluster_pair_file,
local_anchor_length):
# load cluster database file
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
# parse breakpoint alignments and output spanning reads
bamfh = pysam.Samfile(bam_file, "rb")
outsamfh = pysam.Samfile(output_sam_file, "wh", template=bamfh)
outfh = open(output_cluster_pair_file, "w")
cluster_pair_iter = parse_discordant_cluster_pair_file(open(cluster_pair_file))
# get cluster reads from BAM file
num_spanning_reads = 0
for pair_id, cluster_reads in _parse_bam_by_cluster_pair(bamfh):
# synch with cluster pair file
cluster_pair = cluster_pair_iter.next()
while pair_id != cluster_pair.pair_id:
# no spanning reads here
print >>outfh, '\t'.join(map(str, [cluster_pair.pair_id,
cluster_pair.id5p,
cluster_pair.id3p,
','.join(cluster_pair.qnames),
'']))
cluster_pair = cluster_pair_iter.next()
# get spanning read alignments
spanning_reads = nominate_spanning_reads(cluster_pair,
cluster_shelve,
bamfh,
cluster_reads,
local_anchor_length)
spanning_qnames = sorted(set(r5p.qname for r5p,r3p in spanning_reads))
# write new cluster pair file
print >>outfh, '\t'.join(map(str, [cluster_pair.pair_id,
cluster_pair.id5p,
cluster_pair.id3p,
','.join(cluster_pair.qnames),
','.join(spanning_qnames)]))
# write spanning reads to SAM file
for r5p,r3p in spanning_reads:
outsamfh.write(r5p)
outsamfh.write(r3p)
num_spanning_reads += len(spanning_reads)
# finish outputting remaining clusters
for cluster_pair in cluster_pair_iter:
print >>outfh, '\t'.join(map(str, [cluster_pair.pair_id,
cluster_pair.id5p,
cluster_pair.id3p,
','.join(cluster_pair.qnames),
'']))
logging.debug("\tFound %d spanning read alignments" % (num_spanning_reads))
outsamfh.close()
outfh.close()
bamfh.close()
cluster_shelve.close()
return config.JOB_SUCCESS
def process_spanning_alignments(cluster_shelve_file, cluster_pair_file,
bam_file, output_sam_file,
output_cluster_pair_file, local_anchor_length):
# load cluster database file
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
# parse breakpoint alignments and output spanning reads
bamfh = pysam.Samfile(bam_file, "rb")
outsamfh = pysam.Samfile(output_sam_file, "wh", template=bamfh)
outfh = open(output_cluster_pair_file, "w")
cluster_pair_iter = parse_discordant_cluster_pair_file(
open(cluster_pair_file))
# get cluster reads from BAM file
num_spanning_reads = 0
for pair_id, cluster_reads in _parse_bam_by_cluster_pair(bamfh):
# synch with cluster pair file
cluster_pair = cluster_pair_iter.next()
while pair_id != cluster_pair.pair_id:
# no spanning reads here
print >> outfh, '\t'.join(
map(str, [
cluster_pair.pair_id, cluster_pair.id5p, cluster_pair.id3p,
','.join(cluster_pair.qnames), ''
]))
cluster_pair = cluster_pair_iter.next()
# get spanning read alignments
spanning_reads = nominate_spanning_reads(cluster_pair, cluster_shelve,
bamfh, cluster_reads,
local_anchor_length)
spanning_qnames = sorted(set(r5p.qname for r5p, r3p in spanning_reads))
# write new cluster pair file
print >> outfh, '\t'.join(
map(str, [
cluster_pair.pair_id, cluster_pair.id5p, cluster_pair.id3p,
','.join(cluster_pair.qnames), ','.join(spanning_qnames)
]))
# write spanning reads to SAM file
for r5p, r3p in spanning_reads:
outsamfh.write(r5p)
outsamfh.write(r3p)
num_spanning_reads += len(spanning_reads)
# finish outputting remaining clusters
for cluster_pair in cluster_pair_iter:
print >> outfh, '\t'.join(
map(str, [
cluster_pair.pair_id, cluster_pair.id5p, cluster_pair.id3p,
','.join(cluster_pair.qnames), ''
]))
logging.debug("\tFound %d spanning read alignments" % (num_spanning_reads))
outsamfh.close()
outfh.close()
bamfh.close()
cluster_shelve.close()
return config.JOB_SUCCESS
def realign_across_breakpoints(index_dir,
discordant_bam_file,
unpaired_bam_file,
cluster_shelve_file,
cluster_pair_file,
breakpoint_bam_file,
log_dir,
tmp_dir,
num_processors,
local_anchor_length,
local_multihits):
# load cluster database file
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
# open discordant reads file
discordant_bamfh = pysam.Samfile(discordant_bam_file, "rb")
unpaired_bamfh = pysam.Samfile(unpaired_bam_file, "rb")
# create tmp dir if it does not exist
fastq_file = os.path.join(tmp_dir, config.BREAKPOINT_FASTQ_FILE)
fastq_fh = open(fastq_file, 'w')
# iterate through cluster pairs and get breakpoint reads
logging.debug("Extracting breakpoint spanning sequences")
num_seqs = 0
for cluster_pair in parse_discordant_cluster_pair_file(open(cluster_pair_file)):
for fastq_line in _get_cluster_breakpoint_fastq(cluster_pair,
cluster_shelve,
discordant_bamfh,
unpaired_bamfh):
print >>fastq_fh, fastq_line
num_seqs += 1
fastq_fh.close()
discordant_bamfh.close()
unpaired_bamfh.close()
logging.debug("\tFound %d putative breakpoint spanning sequences" % (num_seqs))
# use bowtie2 local alignment to find spanning reads
transcriptome_index = os.path.join(index_dir, config.TRANSCRIPTOME_INDEX)
genome_index = os.path.join(index_dir, config.GENOME_INDEX)
transcript_file = os.path.join(index_dir, config.TRANSCRIPT_FEATURE_FILE)
log_file = os.path.join(log_dir, config.BREAKPOINT_LOG_FILE)
logging.debug("Realigning breakpoint spanning sequences")
bowtie2_align_local(transcriptome_index,
genome_index,
transcript_file,
fastq_file,
breakpoint_bam_file,
log_file,
local_anchor_length=local_anchor_length,
local_multihits=local_multihits,
num_processors=num_processors)
cluster_shelve.close()
return config.JOB_SUCCESS
def realign_across_breakpoints(index_dir, discordant_bam_file,
unpaired_bam_file, cluster_shelve_file,
cluster_pair_file, breakpoint_bam_file, log_dir,
tmp_dir, num_processors, local_anchor_length,
local_multihits):
# load cluster database file
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
# open discordant reads file
discordant_bamfh = pysam.Samfile(discordant_bam_file, "rb")
unpaired_bamfh = pysam.Samfile(unpaired_bam_file, "rb")
# create tmp dir if it does not exist
fastq_file = os.path.join(tmp_dir, config.BREAKPOINT_FASTQ_FILE)
fastq_fh = open(fastq_file, 'w')
# iterate through cluster pairs and get breakpoint reads
logging.debug("Extracting breakpoint spanning sequences")
num_seqs = 0
for cluster_pair in parse_discordant_cluster_pair_file(
open(cluster_pair_file)):
for fastq_line in _get_cluster_breakpoint_fastq(
cluster_pair, cluster_shelve, discordant_bamfh,
unpaired_bamfh):
print >> fastq_fh, fastq_line
num_seqs += 1
fastq_fh.close()
discordant_bamfh.close()
unpaired_bamfh.close()
logging.debug("\tFound %d putative breakpoint spanning sequences" %
(num_seqs))
# use bowtie2 local alignment to find spanning reads
transcriptome_index = os.path.join(index_dir, config.TRANSCRIPTOME_INDEX)
genome_index = os.path.join(index_dir, config.GENOME_INDEX)
transcript_file = os.path.join(index_dir, config.TRANSCRIPT_FEATURE_FILE)
log_file = os.path.join(log_dir, config.BREAKPOINT_LOG_FILE)
logging.debug("Realigning breakpoint spanning sequences")
bowtie2_align_local(transcriptome_index,
genome_index,
transcript_file,
fastq_file,
breakpoint_bam_file,
log_file,
local_anchor_length=local_anchor_length,
local_multihits=local_multihits,
num_processors=num_processors)
cluster_shelve.close()
return config.JOB_SUCCESS
def write_output(transcripts, cluster_shelve_file, cluster_pair_file,
read_name_file, output_file,
annotation_source="ensembl"):
# load cluster and read name database files
cluster_shelve = shelve.open(cluster_shelve_file, 'r')
read_name_fh = open(read_name_file, 'r')
# map genome coordinates to transcripts
logging.debug("Creating mapping between genome coordinates and transcripts")
transcript_dict, genome_tx_trees = build_genome_transcript_trees(transcripts)
logging.debug("Writing output")
outfh = open(output_file, "w")
print >>outfh, '#' + '\t'.join(Chimera._fields)
for cluster_pair in parse_discordant_cluster_pair_file(open(cluster_pair_file)):
c = make_chimera(cluster_pair, cluster_shelve, transcript_dict,
genome_tx_trees, annotation_source)
print >>outfh, str(c)
# cleanup
outfh.close()
read_name_fh.close()
cluster_shelve.close()
return config.JOB_SUCCESS
