def subsample_ta_se(ta, subsample, non_mito, mito_chr_name, out_dir):
prefix = os.path.join(out_dir,
os.path.basename(strip_ext_ta(ta)))
ta_subsampled = '{}.{}{}tagAlign.gz'.format(
prefix,
'no_chrM.' if non_mito else '',
'{}.'.format(human_readable_number(subsample)) if subsample > 0 else ''
)
# bash-only
cmd = 'zcat -f {} | '
if non_mito:
# cmd += 'awk \'{{if ($1!="'+mito_chr_name+'") print $0}}\' | '
cmd += 'grep -v \'^'+mito_chr_name+'\\b\' | '
if subsample > 0:
cmd += 'shuf -n {} --random-source=<(openssl enc -aes-256-ctr '
cmd += '-pass pass:$(zcat -f {} | wc -c) -nosalt '
cmd += '</dev/zero 2>/dev/null) | '
cmd += 'gzip -nc > {}'
cmd = cmd.format(
ta,
subsample,
ta,
ta_subsampled)
else:
cmd += 'gzip -nc > {}'
cmd = cmd.format(
ta,
ta_subsampled)
run_shell_cmd(cmd)
return ta_subsampled
def spp(ta, ctl_ta, fraglen, cap_num_peak, nth, out_dir):
basename_ta = os.path.basename(strip_ext_ta(ta))
basename_ctl_ta = os.path.basename(strip_ext_ta(ctl_ta))
basename_prefix = '{}_x_{}'.format(basename_ta, basename_ctl_ta)
if len(basename_prefix) > 200: # UNIX cannot have filename > 255
basename_prefix = '{}_x_control'.format(basename_ta)
nth_param = '-p={}'.format(nth) if nth < 2 else ''
prefix = os.path.join(out_dir, basename_prefix)
rpeak = '{}.{}.regionPeak.gz'.format(prefix,
human_readable_number(cap_num_peak))
rpeak_tmp = '{}.tmp'.format(rpeak)
rpeak_tmp_gz = '{}.tmp.gz'.format(rpeak)
cmd0 = 'Rscript --max-ppsize=500000 $(which run_spp.R) -c={} -i={} '
cmd0 += '-npeak={} -odir={} -speak={} -savr={} -rf {}'
cmd0 = cmd0.format(ta, ctl_ta, cap_num_peak, os.path.abspath(out_dir),
fraglen, rpeak_tmp, nth_param)
run_shell_cmd(cmd0)
# if we have scientific representation of chr coord. then convert it to int
cmd1 = 'zcat -f {} | awk \'BEGIN{{OFS="\\t"}}'
cmd1 += '{{if ($2<0) $2=0; '
cmd1 += 'print $1,int($2),int($3),$4,$5,$6,$7,$8,$9,$10;}}\' | '
cmd1 += 'gzip -f -nc > {}'
cmd1 = cmd1.format(rpeak_tmp, rpeak)
run_shell_cmd(cmd1)
rm_f([rpeak_tmp, rpeak_tmp_gz])
return rpeak
def macs2(ta, ctl_ta, chrsz, gensz, pval_thresh, shift, fraglen, cap_num_peak,
ctl_subsample, ctl_paired_end, out_dir):
basename_ta = os.path.basename(strip_ext_ta(ta))
if ctl_ta:
if ctl_subsample:
if ctl_paired_end:
ctl_ta = subsample_ta_pe(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
r1_only=False,
out_dir=out_dir)
else:
ctl_ta = subsample_ta_se(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
out_dir=out_dir)
basename_ctl_ta = os.path.basename(strip_ext_ta(ctl_ta))
basename_prefix = '{}_x_{}'.format(basename_ta, basename_ctl_ta)
if len(basename_prefix) > 200: # UNIX cannot have len(filename) > 255
basename_prefix = '{}_x_control'.format(basename_ta)
else:
basename_prefix = basename_ta
prefix = os.path.join(out_dir, basename_prefix)
npeak = '{}.{}.{}.narrowPeak.gz'.format(
prefix, 'pval{}'.format(pval_thresh),
human_readable_number(cap_num_peak))
npeak_tmp = '{}.tmp'.format(npeak)
npeak_tmp2 = '{}.tmp2'.format(npeak)
temp_files = []
cmd0 = ' macs2 callpeak '
cmd0 += '-t {} {} -f BED -n {} -g {} -p {} '
cmd0 += '--nomodel --shift {} --extsize {} --keep-dup all -B --SPMR'
cmd0 = cmd0.format(ta, '-c {}'.format(ctl_ta) if ctl_ta else '', prefix,
gensz, pval_thresh, 0, fraglen)
run_shell_cmd(cmd0)
cmd1 = 'LC_COLLATE=C sort -k 8gr,8gr "{}"_peaks.narrowPeak | '
cmd1 += 'awk \'BEGIN{{OFS="\\t"}}'
cmd1 += '{{$4="Peak_"NR; if ($2<0) $2=0; if ($3<0) $3=0; if ($10==-1) '
cmd1 += '$10=$2+int(($3-$2+1)/2.0); print $0}}\' > {}'
cmd1 = cmd1.format(prefix, npeak_tmp)
run_shell_cmd(cmd1)
cmd2 = 'head -n {} {} > {}'.format(cap_num_peak, npeak_tmp, npeak_tmp2)
run_shell_cmd(cmd2)
# clip peaks between 0-chromSize.
bed_clip(npeak_tmp2, chrsz, npeak)
rm_f([npeak_tmp, npeak_tmp2])
# remove temporary files
temp_files.append("{}_*".format(prefix))
rm_f(temp_files)
return npeak
def macs2(ta, chrsz, gensz, pval_thresh, smooth_win, cap_num_peak, mem_gb,
out_dir):
prefix = os.path.join(out_dir, os.path.basename(strip_ext_ta(ta)))
npeak = '{}.{}.{}.narrowPeak.gz'.format(
prefix, 'pval{}'.format(pval_thresh),
human_readable_number(cap_num_peak))
# temporary files
npeak_tmp = '{}.tmp'.format(npeak)
npeak_tmp2 = '{}.tmp2'.format(npeak)
shiftsize = -int(round(float(smooth_win) / 2.0))
temp_files = []
run_shell_cmd('macs2 callpeak '
'-t {ta} -f BED -n {prefix} -g {gensz} -p {pval_thresh} '
'--shift {shiftsize} --extsize {extsize} '
'--nomodel -B --SPMR --keep-dup all --call-summits'.format(
ta=ta,
prefix=prefix,
gensz=gensz,
pval_thresh=pval_thresh,
shiftsize=shiftsize,
extsize=smooth_win,
))
run_shell_cmd(
'LC_COLLATE=C sort -k 8gr,8gr {sort_param} "{prefix}"_peaks.narrowPeak | '
'awk \'BEGIN{{OFS="\\t"}}'
'{{$4="Peak_"NR; if ($2<0) $2=0; if ($3<0) $3=0; if ($10==-1) '
'$10=$2+int(($3-$2+1)/2.0); print $0}}\' > {npeak_tmp}'.format(
sort_param=get_gnu_sort_param(mem_gb * 1024**3, ratio=0.5),
prefix=prefix,
npeak_tmp=npeak_tmp,
))
run_shell_cmd('head -n {cap_num_peak} {npeak_tmp} > {npeak_tmp2}'.format(
cap_num_peak=cap_num_peak,
npeak_tmp=npeak_tmp,
npeak_tmp2=npeak_tmp2,
))
# clip peaks between 0-chromSize.
bed_clip(npeak_tmp2, chrsz, npeak)
rm_f([npeak_tmp, npeak_tmp2])
# remove temporary files
temp_files.append("{prefix}_*".format(prefix=prefix))
rm_f(temp_files)
return npeak
def spp(ta, ctl_ta, chrsz, fraglen, cap_num_peak, fdr_thresh, ctl_subsample,
ctl_paired_end, nth, out_dir):
basename_ta = os.path.basename(strip_ext_ta(ta))
if ctl_subsample:
if ctl_paired_end:
ctl_ta = subsample_ta_pe(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
r1_only=False,
out_dir=out_dir)
else:
ctl_ta = subsample_ta_se(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
out_dir=out_dir)
basename_ctl_ta = os.path.basename(strip_ext_ta(ctl_ta))
basename_prefix = '{}_x_{}'.format(basename_ta, basename_ctl_ta)
if len(basename_prefix) > 200: # UNIX cannot have filename > 255
basename_prefix = '{}_x_control'.format(basename_ta)
nth_param = '-p={}'.format(nth) if nth >= 2 else ''
prefix = os.path.join(out_dir, basename_prefix)
rpeak = '{}.{}.regionPeak.gz'.format(prefix,
human_readable_number(cap_num_peak))
rpeak_tmp_prefix = '{}.tmp'.format(rpeak)
rpeak_tmp_gz = '{}.tmp.gz'.format(rpeak)
rpeak_tmp2 = '{}.tmp2'.format(rpeak)
cmd0 = 'Rscript --max-ppsize=500000 $(which run_spp.R) -c={} -i={} '
cmd0 += '-npeak={} -odir={} -speak={} -savr={} -fdr={} -rf {}'
cmd0 = cmd0.format(ta, ctl_ta, cap_num_peak, os.path.abspath(out_dir),
fraglen, rpeak_tmp_prefix, fdr_thresh, nth_param)
run_shell_cmd(cmd0)
# if we have scientific representation of chr coord. then convert it to int
cmd1 = 'zcat -f {} | awk \'BEGIN{{OFS="\\t"}}'
cmd1 += '{{if ($2<0) $2=0; '
cmd1 += 'print $1,int($2),int($3),$4,$5,$6,$7,$8,$9,$10;}}\' > {}'
cmd1 = cmd1.format(rpeak_tmp_gz, rpeak_tmp2)
run_shell_cmd(cmd1)
rm_f(rpeak_tmp_gz)
# clip peaks between 0-chromSize.
bed_clip(rpeak_tmp2, chrsz, rpeak)
rm_f(rpeak_tmp2)
return rpeak
def subsample_ta_pe(ta, subsample, non_mito, mito_chr_name, r1_only, out_dir):
prefix = os.path.join(out_dir,
os.path.basename(strip_ext_ta(ta)))
ta_subsampled = '{}.{}{}{}tagAlign.gz'.format(
prefix,
'no_chrM.' if non_mito else '',
'R1.' if r1_only else '',
'{}.'.format(human_readable_number(subsample)) if subsample > 0 else ''
)
ta_tmp = '{}.tagAlign.tmp'.format(prefix)
cmd0 = 'zcat -f {} | '
if non_mito:
# cmd0 += 'awk \'{{if ($1!="'+mito_chr_name+'") print $0}}\' | '
cmd0 += 'grep -v \'^'+mito_chr_name+'\\b\' | '
cmd0 += 'sed \'N;s/\\n/\\t/\' '
if subsample > 0:
cmd0 += '| shuf -n {} --random-source=<(openssl enc -aes-256-ctr '
cmd0 += '-pass pass:$(zcat -f {} | wc -c) -nosalt '
cmd0 += '</dev/zero 2>/dev/null) > {}'
cmd0 = cmd0.format(
ta,
subsample,
ta,
ta_tmp)
else:
cmd0 += '> {}'
cmd0 = cmd0.format(
ta,
ta_tmp)
run_shell_cmd(cmd0)
cmd = 'cat {} | '
cmd += 'awk \'BEGIN{{OFS="\\t"}} '
if r1_only:
cmd += '{{printf "%s\\t%s\\t%s\\t%s\\t%s\\t%s\\n'
cmd += '",$1,$2,$3,$4,$5,$6}}\' | '
else:
cmd += '{{printf "%s\\t%s\\t%s\\t%s\\t%s\\t%s\\n'
cmd += '%s\\t%s\\t%s\\t%s\\t%s\\t%s\\n",'
cmd += '$1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12}}\' | '
cmd += 'gzip -nc > {}'
cmd = cmd.format(
ta_tmp,
ta_subsampled)
run_shell_cmd(cmd)
rm_f(ta_tmp)
return ta_subsampled
def macs2(ta, chrsz, gensz, pval_thresh, smooth_win, cap_num_peak, out_dir):
prefix = os.path.join(out_dir, os.path.basename(strip_ext_ta(ta)))
npeak = '{}.{}.{}.narrowPeak.gz'.format(
prefix, 'pval{}'.format(pval_thresh),
human_readable_number(cap_num_peak))
# temporary files
npeak_tmp = '{}.tmp'.format(npeak)
npeak_tmp2 = '{}.tmp2'.format(npeak)
shiftsize = -int(round(float(smooth_win) / 2.0))
temp_files = []
cmd0 = 'macs2 callpeak '
cmd0 += '-t {} -f BED -n {} -g {} -p {} '
cmd0 += '--shift {} --extsize {} '
cmd0 += '--nomodel -B --SPMR '
cmd0 += '--keep-dup all --call-summits '
cmd0 = cmd0.format(ta, prefix, gensz, pval_thresh, shiftsize, smooth_win)
run_shell_cmd(cmd0)
cmd1 = 'LC_COLLATE=C sort -k 8gr,8gr "{}"_peaks.narrowPeak | '
cmd1 += 'awk \'BEGIN{{OFS="\\t"}}'
cmd1 += '{{$4="Peak_"NR; if ($2<0) $2=0; if ($3<0) $3=0; if ($10==-1) '
cmd1 += '$10=$2+int(($3-$2+1)/2.0); print $0}}\' > {}'
cmd1 = cmd1.format(prefix, npeak_tmp)
run_shell_cmd(cmd1)
cmd2 = 'head -n {} {} > {}'.format(cap_num_peak, npeak_tmp, npeak_tmp2)
run_shell_cmd(cmd2)
# clip peaks between 0-chromSize.
bed_clip(npeak_tmp2, chrsz, npeak)
rm_f([npeak_tmp, npeak_tmp2])
# remove temporary files
temp_files.append("{}_*".format(prefix))
rm_f(temp_files)
return npeak
def macs2(ta, ctl_ta, chrsz, gensz, pval_thresh, shift, fraglen, cap_num_peak,
ctl_subsample, ctl_paired_end, mem_gb, out_dir):
basename_ta = os.path.basename(strip_ext_ta(ta))
if ctl_ta:
if ctl_subsample:
if ctl_paired_end:
ctl_ta = subsample_ta_pe(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
r1_only=False,
out_dir=out_dir)
else:
ctl_ta = subsample_ta_se(ctl_ta,
ctl_subsample,
non_mito=False,
mito_chr_name=None,
out_dir=out_dir)
basename_ctl_ta = os.path.basename(strip_ext_ta(ctl_ta))
basename_prefix = '{}_x_{}'.format(basename_ta, basename_ctl_ta)
if len(basename_prefix) > 200: # UNIX cannot have len(filename) > 255
basename_prefix = '{}_x_control'.format(basename_ta)
else:
basename_prefix = basename_ta
prefix = os.path.join(out_dir, basename_prefix)
npeak = '{}.{}.{}.narrowPeak.gz'.format(
prefix, 'pval{}'.format(pval_thresh),
human_readable_number(cap_num_peak))
npeak_tmp = '{}.tmp'.format(npeak)
npeak_tmp2 = '{}.tmp2'.format(npeak)
temp_files = []
run_shell_cmd(
' macs2 callpeak '
'-t {ta} {ctl_param} -f BED -n {prefix} -g {gensz} -p {pval_thresh} '
'--nomodel --shift {shiftsize} --extsize {extsize} --keep-dup all -B --SPMR'
.format(
ta=ta,
ctl_param='-c {ctl_ta}'.format(ctl_ta=ctl_ta) if ctl_ta else '',
prefix=prefix,
gensz=gensz,
pval_thresh=pval_thresh,
shiftsize=0,
extsize=fraglen,
))
run_shell_cmd(
'LC_COLLATE=C sort -k 8gr,8gr {sort_param} "{prefix}_peaks.narrowPeak" | '
'awk \'BEGIN{{OFS="\\t"}}'
'{{$4="Peak_"NR; if ($2<0) $2=0; if ($3<0) $3=0; if ($10==-1) '
'$10=$2+int(($3-$2+1)/2.0); print $0}}\' > {npeak_tmp}'.format(
sort_param=get_gnu_sort_param(mem_gb * 1024**3, ratio=0.5),
prefix=prefix,
npeak_tmp=npeak_tmp,
))
run_shell_cmd('head -n {cap_num_peak} {npeak_tmp} > {npeak_tmp2}'.format(
cap_num_peak=cap_num_peak,
npeak_tmp=npeak_tmp,
npeak_tmp2=npeak_tmp2,
))
# clip peaks between 0-chromSize.
bed_clip(npeak_tmp2, chrsz, npeak)
rm_f([npeak_tmp, npeak_tmp2])
# remove temporary files
temp_files.append("{prefix}_*".format(prefix=prefix))
rm_f(temp_files)
return npeak
