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 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 entropy(options):
'''
Calculate entropy for each cluster
'''
# parse options
folder = check_dir(options['<rampagedir>'])
link_f = check_bed(os.path.join(folder, 'rampage_link.bed'),
return_handle=False)
threads = int(options['--thread'])
with open(os.path.join(folder, 'rampage_peaks.txt'), 'r') as peak:
result = Parallel(n_jobs=threads)(delayed(cal_entropy)(line, link_f)
for line in peak)
with open(os.path.join(folder, 'rampage_entropy.txt'), 'w') as out:
for r in result:
out.write(r)
def main():
# parse options
options = docopt(__doc__, version=__version__)
if options['--seq']:
if not os.path.isfile(options['--seq']):
sys.exit('Error: wrong seq file!')
seq = os.path.abspath(options['--seq'])
seq_flag = True
else:
seq = None
seq_flag = False
fa = check_fasta(options['--genome'])
chrom = options['--chrom']
site = int(options['--site'])
strand = '+' if options['--strand'] == '1' else '-'
rlen = int(options['--read-length'])
alen = int(options['--region-length'])
clen = int(options['--check-region-length'])
thread = options['--thread']
skip_flag = options['--skip-alignment']
# check output directory
if not skip_flag: # not skip alignment
out_dir = create_dir(options['<out_dir>'])
else: # skip alignment
out_dir = check_dir(options['<out_dir>'])
# build index for sgRNA
index_path, offset = build_index(fa, chrom, site, strand, rlen, thread,
out_dir, seq, seq_flag)
if not skip_flag: # not skip alignment
# deal with reads file
reads = tempfile.NamedTemporaryFile(mode='w+')
if options['-R']:
fq_lst = options['-R'].split(',')
convert_read(reads, single=fq_lst)
else:
fq1_lst = options['-1'].split(',')
fq2_lst = options['-2'].split(',')
convert_read(reads, fq1=fq1_lst, fq2=fq2_lst)
reads.seek(0)
read_path = reads.name
# mapped reads with bowtie2
bam = bowtie2_align(index_path, read_path, thread, out_dir)
# remove tempfile
reads.close()
else:
bam = os.path.join(out_dir, 'cs.bam')
# fetch cleavage site reads
fetch_reads(index_path, offset, alen, clen, bam, out_dir)
def assign_peak(options):
'''
Call rampage peaks
'''
# parse options
if options['--ref']:
db = options['--ref']
ref_flag = True
elif options['--db']:
import gffutils
db = gffutils.FeatureDB(options['--db'])
ref_flag = False
else:
import gffutils
gtf_f = options['--gtf']
prefix = os.path.splitext(os.path.basename(gtf_f))[0]
db = gffutils.create_db(gtf_f,
prefix + '.db',
force=True,
disable_infer_transcripts=True)
ref_flag = False
folder = check_dir(options['<rampagepeak>'])
rampage = check_bed(os.path.join(folder, 'rampage_link.bed'),
return_handle=False)
rampage_peak = check_bed(os.path.join(folder, 'rampage_peaks.txt'),
return_handle=False)
prom = int(options['--promoter'])
# align and filter candidate peak
p = Pool(int(options['--thread']))
results = []
for gene_info, gpromoter in parse_gene(db, ref_flag, prom):
results.append(
p.apply_async(assign_peak_to_gene,
args=(rampage, rampage_peak, gene_info, gpromoter,
prom)))
p.close()
p.join()
# output results
with open(os.path.join(folder, 'rampage_assigned_peaks.txt'), 'w') as outf:
for r in results:
gene_info, peak = r.get()
if gene_info:
for p in peak:
outf.write('%s\t%s\n' % (p, gene_info))
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 dbloci(options):
'''
Fetch bidirectionally transcribed loci
'''
# parse options
folder = check_dir(options['<rampagedir>'])
size = int(options['-l'])
filter_flag = options['--filter']
if filter_flag == 'rpm':
filter = float(options['--rpm'])
elif filter_flag == 'height':
filter = int(options['--height'])
# fetch rampage pairs
peak_f = os.path.join(folder, 'rampage_peaks.txt')
peak_bed = check_bed(peak_f)
up_cluster = {}
down_cluster = {}
pairs = defaultdict(list)
with open(peak_f, 'r') as peak:
for ucluster in peak: # parse upstream cluster
chrom, _, _, _, _, ustrand, upos = ucluster.split()[:7]
if ustrand == '+': # upstream cluster should be minus
continue
if filter_cluster(ucluster, filter_flag, filter):
continue
uheight = int(ucluster.split()[7])
up_id = '\t'.join([chrom, upos, ustrand])
up_cluster[up_id] = uheight
# parse downstream cluster
start = int(upos)
end = start + size * 2
for dcluster in peak_bed.fetch(chrom, start, end):
dstrand, dpos = dcluster.split()[5:7]
if dstrand == '-': # downstream cluster should be plus
continue
if filter_cluster(dcluster, filter_flag, filter):
continue
dheight = int(dcluster.split()[7])
down_id = '\t'.join([chrom, dpos, dstrand])
down_cluster[down_id] = dheight
# construct pairs
pairs[up_id].append(down_id)
pairs[down_id].append(up_id)
# output enhancers
outf = os.path.join(folder, 'enhancers.txt')
with open(outf, 'w') as out:
for pair_set in fetch_pair(pairs):
up_site, down_site = 0, 0
up_height, down_height = 0, 0
for site_id in pair_set:
chrom, site, strand = site_id.split()
if strand == '-': # upstream
height = up_cluster[site_id]
if height > up_height:
up_site = int(site)
up_height = height
else: # downstream
height = down_cluster[site_id]
if height > down_height:
down_site = int(site)
down_height = height
middle_site = int((up_site + down_site) / 2)
forward_plus = cal_density(folder, chrom, middle_site,
middle_site + size, 'plus')
forward_minus = cal_density(folder, chrom, middle_site,
middle_site + size, 'minus')
reverse_plus = cal_density(folder, chrom, middle_site - size,
middle_site, 'plus')
reverse_minus = cal_density(folder, chrom, middle_site - size,
middle_site, 'minus')
if forward_minus >= forward_plus or reverse_plus >= reverse_minus:
continue
else:
forward = forward_plus
reverse = reverse_minus
forward_dis = fetch_dis(folder, chrom, middle_site,
middle_site + size, '+')
reverse_dis = fetch_dis(folder, chrom, middle_site - size,
middle_site, '-')
fold = (forward - reverse) * 1.0 / (forward + reverse)
start = middle_site - size
end = middle_site + size
out_format = '%s\t%d\t%d\tenhancer\t0\t+\t%d\t%d\t%d\t%d\t%d\t%f\n'
out.write(out_format %
(chrom, start, end, middle_site, reverse_dis,
forward_dis, reverse, forward, fold))