def build_index(fa, chrom, site, strand, rlen, thread, out_dir, seq, seq_flag):
print('Build index...')
if strand == '+':
start = site - (rlen - 10)
end = site + (rlen - 20)
offset = rlen - 10
else:
start = site - (rlen - 20)
end = site + (rlen - 10)
offset = rlen - 20
index_path = os.path.join(out_dir, 'sgRNA.fa')
if seq_flag:
os.symlink(seq, index_path)
else:
# fetch sgRNA region sequence
with open(index_path, 'w') as out:
out.write('>sgRNA_region\n')
out.write(
dna_to_rna(fa.fetch(chrom, start, end), strand=strand) + '\n')
# build index
if which('bowtie2-build'):
command = 'bowtie2-build -q --threads %s %s %s'
command = command % (thread, index_path, index_path)
run_command(command, 'Error: cannot build index for sgRNA!')
else:
sys.exit('Error: no bowtie2-build installed!')
return index_path, offset
def main():
# parse options
options = docopt(__doc__, version=__version__)
# check output_dir
if options['-o'] == './':
dir = os.getcwd()
else:
dir = check_dir(options['-o'])
# fetch junction bed file
junc_f = fetch_juncfile(options['<bam>'],
url=options['--url'],
dir=dir,
uniq=options['--uniq'],
stranded=options['--stranded'],
min=int(options['--min-reads']))
# create junction bigbed file in case
if options['--bb'] and which('bedToBigBed') is not None:
prefix = os.path.splitext(os.path.split(options['<bam>'])[-1])[0]
bamf = pysam.AlignmentFile(options['<bam>'], 'rb')
with tempfile.NamedTemporaryFile() as chrom_size:
for seq in bamf.header['SQ']:
chrom_size.write('%s\t%s\n' % (seq['SN'], seq['LN']))
chrom_size.seek(0)
bb_path = os.path.join(dir, prefix + '_junc.bb')
return_code = os.system('bedToBigBed -type=bed12 %s %s %s' %
(junc_f, chrom_size.name, bb_path)) >> 8
if return_code:
sys.exit('Error: cannot convert bed to BigBed!')
bamf.close()
def fseq(options):
'''
Call peaks using F-seq
'''
# parse options
if not which('fseq'):
sys.exit('Error: No F-seq installed!')
folder = check_dir(options['<rampagedir>'])
flength = options['-l']
wig_flag = options['--wig']
percent = float(options['-p'])
with open(os.path.join(folder, 'total_counts.txt'), 'r') as f:
total = int(f.read().rstrip())
# run F-seq
flist = {'+': 'rampage_plus_5end.bed', '-': 'rampage_minus_5end.bed'}
all_peak_f = os.path.join(folder, 'rampage_peaks.txt')
with open(all_peak_f, 'w') as out:
for strand in flist:
peak_f = run_fseq(folder, flist[strand], strand, flength, wig_flag,
percent)
with open(peak_f, 'r') as f:
for line in f:
if total: # calculate RPM
reads = int(line.rstrip().split()[9])
rpm = reads * 1000000.0 / total
out.write(line.rstrip() + '\t%f\n' % rpm)
else:
out.write(line)
def assemble(options):
'''
Assemble RNA-seq with StringTie
'''
# parse options
if not which('stringtie'):
sys.exit('Error: No StringTie installed!')
gtf = options['--gtf']
thread = options['--thread']
bamf = options['<bam>']
dir_path = options['--dir']
if len(bamf) == 1: # no replicates
# run StringTie
out_gtf = run_stringtie(bamf[0], gtf, dir_path, thread)
# convert GTF to GenePred
convert_gtf(out_gtf)
else: # have replicates
# run StringTie
gtf_list = []
for f in bamf:
gtf_list.append(run_stringtie(f, gtf, dir_path, thread))
merged_prefix = options['--prefix']
out_gtf = merge_stringtie(gtf_list, gtf, dir_path, merged_prefix,
thread)
# convert GTF to GenePred
convert_gtf(out_gtf, merge=True)
def bowtie2_align(index, read, thread, out_dir):
print('Align reads...')
if which('bowtie2'):
bam = os.path.join(out_dir, 'cs.bam')
sam = tempfile.NamedTemporaryFile('w+')
command = 'bowtie2 --quiet --end-to-end -p %s -x %s -U %s -S %s'
command = command % (thread, index, read, sam.name)
run_command(command, 'Error in bowtie2 alignment!')
sam.seek(0)
with pysam.AlignmentFile(sam.name, 'r') as sam_f:
with pysam.AlignmentFile(bam, 'wb', template=sam_f) as bam_f:
for read in sam_f:
if not read.is_unmapped:
bam_f.write(read)
sam.close()
return bam
else:
sys.exit('Error: no bowtie2 installed!')
def fseq(options):
'''
Call peaks using F-seq
'''
# parse options
if not which('fseq'):
sys.exit('Error: No F-seq installed!')
folder = check_dir(options['<rampagedir>'])
flength = options['-l']
percent = [0.95, 0.9, 0.85, 0.8, 0.75, 0.7]
with open(os.path.join(folder, 'total_counts.txt'), 'r') as f:
total = int(f.read().rstrip())
# run F-seq
flist = {'+': 'rampage_plus_5end.bed', '-': 'rampage_minus_5end.bed'}
all_peak_f = options['-o']
with open(all_peak_f, 'w') as out:
for strand in flist:
peak_f = run_fseq(folder, flist[strand], strand, flength, percent)
with open(peak_f, 'r') as f:
for line in f:
if total:
out.write(line.rstrip() + '\t%d\n' % total)
else:
out.write(line)
def check_pattern(options):
# parse options
strand_flag = options['--stranded']
remote_flag = options['--remote']
thread = int(options['--thread'])
# check bigwig
if options['--bam']:
if which('bedGraphToBigWig') is not None:
if strand_flag:
plus_bg, minus_bg = bam_to_bedgraph(options['--bam'],
url=remote_flag,
stranded=True)
plus_bw = bg_to_bw(plus_bg, options['--chrom-size'])
minus_bw = bg_to_bw(minus_bg, options['--chrom-size'])
else:
bg = bam_to_bedgraph(options['--bam'], url=remote_flag)
bw = bg_to_bw(bg, options['--chrom-size'])
else:
sys.exit('Could not find bedGraphToBigWig!')
else:
if os.path.isfile(options['--bigwig']):
bw = options['--bigwig']
# parse junction file
if os.path.isfile(options['--junc']):
junc_path = options['--junc']
else:
sys.exit('Wrong junc file: %s' % options['--junc'])
# check pattern
junc_info = defaultdict(list)
p = Pool(thread)
result = []
with open(junc_path, 'r') as junc_f:
for junc in junc_f:
info = junc.split()
chrom = info[1]
strand = info[4]
rs_site = info[6].split('|')[0]
rs_info = '\t'.join([chrom, rs_site, strand])
if rs_info not in junc_info:
if strand_flag:
if strand == '+':
result.append(
p.apply_async(cal_ratio,
args=(
plus_bw,
rs_info,
options,
)))
else:
result.append(
p.apply_async(cal_ratio,
args=(
minus_bw,
rs_info,
options,
)))
else:
result.append(
p.apply_async(cal_ratio, args=(
bw,
rs_info,
options,
)))
junc_info[rs_info].append(info)
p.close()
p.join()
with open(options['<rs_pattern>'], 'w') as out:
for r in result:
pvalue, fold, rs_info = r.get()
if pvalue is not None:
for junc in junc_info[rs_info]:
out.write('\t'.join(junc))
out.write('\t%f\t%f\n' % (fold, pvalue))