def read_bam_from_bed(infile, bedfile, position_threshold):
chrregions = {}
chrends = {}
regions = read_bed(bedfile)
regions = sort_regions(regions)
regions = merge_regions(regions, position_threshold)
regions = expand_regions_from_bed(regions, infile)
newregions = []
for contig in regions:
for a, b, c in regions[contig]:
newregions.append((contig, a, b, c))
with pysam.AlignmentFile(infile, 'rb') as f:
chrs = get_chromosome_list_from_bam(f)
for contig, start, end, name in newregions:
if contig in chrs:
if contig not in chrregions:
chrregions[contig] = {}
chrregions[contig][start] = count_umis_in_region(
f, contig, start, end)
if contig not in chrends:
chrends[contig] = {}
chrends[contig][start] = end
regions = remove_singleton_regions(chrregions, 2)
return (regions, chrends)
def main(bamfilename, bedfilename):
with open('/home/xsteto/umierrorcorrect/umi.pickle', 'rb') as f:
umis = pickle.load(f)
fasta = pysam.FastaFile('/medstore/External_References/hg19/Homo_sapiens_sequence_hg19.fasta')
ref_seq = get_reference_sequence(fasta, '17', 7577495, 7577800)
contig = '17'
start = 7577495
end = 7577800
position_matrix, singleton_matrix = get_cons_dict(bamfilename, umis, contig, start, end, True)
consensus_seq = get_all_consensus(position_matrix, umis, contig)
cons = get_cons_info(consensus_seq, singleton_matrix)
regions = read_bed(bedfilename)
regions = sort_regions(regions)
regions = merge_regions(regions, 0)
annotation = regions[contig]
with open('out/cons.out', 'w') as f:
write_consensus(f, cons, ref_seq, start, contig, annotation, False)
def run_umi_errorcorrect(args):
'''Run the umi clustering and consensus read generation (error correction)'''
logging.info("Starting UMI clustering")
args.output_path = check_output_directory(args.output_path)
if args.regions_from_bed:
group_method = 'fromBed'
elif args.regions_from_tag:
group_method = 'fromTag'
else:
group_method = 'automatic'
logging.info('Group by position method: {}'.format(group_method))
if not args.sample_name:
args.sample_name = get_sample_name(args.bam_file)
if group_method == 'fromTag':
regions, ends, starts = cluster_umis_on_position(args.bam_file, args.position_threshold,
group_method, args.bed_file)
else:
regions, ends = cluster_umis_on_position(args.bam_file, args.position_threshold,
group_method, args.bed_file)
#print(regions)
nregions = 0
for chrx in regions:
nregions += len(regions[chrx])
logging.info("Number of regions, {}".format(nregions))
edit_distance_threshold = args.edit_distance_threshold
if args.num_threads:
num_cpus = int(args.num_threads)
else:
num_cpus = int(cpu_count())
logging.info("Starting Consensus sequence generation")
logging.info("Starting {} threads".format(num_cpus))
fasta = args.reference_file
if args.bed_file:
bedregions = read_bed(args.bed_file)
bedregions = sort_regions(bedregions)
if group_method=='fromBed':
bedregions = merge_regions(bedregions, 0)
else:
bedregions = []
if group_method=='fromTag':
bamfilelist = cluster_umis_all_regions(regions, ends, edit_distance_threshold,
args.sample_name, args.bam_file, args.output_path,
args.include_singletons, fasta, bedregions,
num_cpus, args.indel_frequency_threshold,
args.consensus_frequency_threshold,
args.regions_from_tag, starts)
else:
bamfilelist = cluster_umis_all_regions(regions, ends, edit_distance_threshold,
args.sample_name, args.bam_file, args.output_path,
args.include_singletons, fasta, bedregions,
num_cpus, args.indel_frequency_threshold,
args.consensus_frequency_threshold)
merge_bams(args.output_path, bamfilelist, args.sample_name)
index_bam_file(args.output_path + '/' + args.sample_name + '_consensus_reads.bam',
num_cpus)
consfilelist = [x.rstrip('.bam') + '.cons' for x in bamfilelist]
merge_cons(args.output_path, consfilelist, args.sample_name)
cons_file = args.output_path + '/' + args.sample_name + '_cons.tsv'
if args.remove_large_files:
os.remove(args.output_path+'/' +args.bam_file)
statfilelist = [x.rstrip('.bam') + '.hist' for x in bamfilelist]
merge_stat(args.output_path, statfilelist, args.sample_name)
duppos = check_duplicate_positions(cons_file)
if any(duppos):
merge_duplicate_positions_all_chromosomes(duppos,cons_file,num_cpus)
merge_duplicate_stat(args.output_path,args.sample_name)
logging.info("Consensus generation complete, output written to {}, {}".format(args.output_path +
'/' + args.sample_name + '_consensus_reads.bam',
args.output_path + '/' + args.sample_name + '_cons.tsv'))