def _star_sam2bam(workflow, conf): # SAM -> BAM
"""
convert SAM to BAM and use mapping quality as cutoff
:param workflow: samflow defined class
:param conf: parsed config file
:return: void
"""
import os
for target in conf.sample_targets:
sam2bam = attach_back(
workflow,
ShellCommand("""
ln -s {input[sam]} {output[sam]}
{tool} view -q 255 -bt {param[genome]} {input[sam]} -o {output[bam]}
""",
tool="samtools",
input={"sam": target + "Aligned.out.sam"},
output={
"bam": target + ".bam",
"sam": target + ".sam"
},
param={
"genome":
conf.get(conf.get("basics", "species"),
"chrom_len"),
},
name="star sam2dam"))
workflow.update(param=conf.items("sam2bam"))
#From bwa/dc.py
sam2bamnochrm = attach_back(
workflow, ## use mapping quality 1 defined by samtools official FAQ
ShellCommand(
"""
awk \'BEGIN{{OFS="\\t"}} {{print $1,0,$2}}\' {param[genome]} > {param[chrom_bed]}
grep -v chrM {param[chrom_bed]} > {output[nochrmbed]}
{tool} view -h -b -L {output[nochrmbed]} {input[bam]} > {output[nochrmbam]}
{tool} view -h {output[nochrmbam]} > {output[nochrmsam]}
{tool} view -h {input[bam]} > {output[usam]}
""",
tool="samtools",
input={"bam": target + ".bam"},
output={
"nochrmbed": target + ".nochrM",
"nochrmbam": target + "_nochrM.bam",
"usam":
target + "_u.sam", ## uniquely mapping sam for sampling
"nochrmsam": target + "_nochrM.sam"
},
param={
"tmp_bam": target + ".tmp.bam",
"output_prefix": target,
"chrom_bed": os.path.join(conf.target_dir, "chrom.bed"),
"mapq": 1,
"genome": conf.get(conf.get("basics", "species"),
"chrom_len")
},
name="filtering mapping and convert")
) # Use 5G memory as default
sam2bamnochrm.update(param=conf.items("sam2bam"))
def merge_bams(workflow, conf): ## merge input and chip bam
"""
input multiple input and multiple control to merge into one file separately
:return:
"""
# merge all treatments into one
merge_bams_treat = ShellCommand(
"{tool} merge {output[merged]} {param[bams]}",
tool="samtools",
input=[target + ".bam" for target in conf.treatment_targets],
output={"merged": conf.prefix + "_treatment.bam"})
merge_bams_treat.param = {"bams": " ".join(merge_bams_treat.input)}
merge_bams_treat.allow_fail = True
merge_bams_treat.allow_dangling = True
if len(conf.treatment_targets) > 1:
attach_back(workflow, merge_bams_treat)
else:
# when there's only one treatment sample, use copying instead of merging
attach_back(workflow, make_link_command(merge_bams_treat.input[0], merge_bams_treat.output["merged"]))
# merging step will be skipped if control sample does not exist
# So be careful to check whether there are control samples before using `_control.bam`
if len(conf.control_targets) > 1:
merge_bams_control = merge_bams_treat.clone
merge_bams_control.input = [target + ".bam" for target in conf.control_targets]
merge_bams_control.output = {"merged": conf.prefix + "_control.bam"}
merge_bams_control.param = {"bams": " ".join(merge_bams_control.input)}
attach_back(workflow, merge_bams_control)
elif len(conf.control_targets) == 1:
attach_back(workflow, make_link_command(conf.control_targets[0] + ".bam", conf.prefix + "_control.bam"))
def sample_bam_stat(workflow, conf, tex):
""" sample non chrm bam to 15M for NSC and PBC
sample non chrm bam to 5M for spot
"""
for i, target in enumerate(conf.treatment_targets):
## for PE, use name sorted in order to calculate PBC
input_bam = target + "_name_sorted.bam" if conf.pe else target + "_final_nochrm.bam"
attach_back(workflow, ShellCommand(
"{tool} {input[namesorted]} {param[run_spp]} {output[bamstat]} {output[sppstat]} {param[pe]} {output[pbc]}",
tool = "eap_dnase_stats",
input = {"namesorted": input_bam},
output = {"bamstat": target + "_bam_stat.qc", ## 15M
"sppstat": target + "_spp.qc",
"pbc": target + "_final_nochrm_15M_pbc.qc"},
param = {"pe": "pe" if conf.pe else "se",
"run_spp": conf.get("tool", "spp")}))
if not "macs" in conf.get("tool", "peak_calling"):
attach_back(workflow, ShellCommand(
"{tool} {input[bamwithoutchrm]} {param[genome]} {param[readsize]} {output[spot]} {param[hotspot_dir]} {param[hotspot_output]} {param[hotspot_tmp]} {param[spot_tmp]}",
tool = "dac_spot", ## 5M
input = {"bamwithoutchrm": target + "_final_nochrm.bam"},
output = {"spot": target + "_spot_nochrm_5M.qc"},
param = {"genome": conf.species,
"spot_tmp": conf.hotspot_reps_tmp_prefix[i] + "_final_nochrm.bam.5000000.spot.out",
"readsize": conf.readsize,
"hotspot_dir": conf.get("tool", "peak_calling"),
"hotspot_output": target + "_hotspot",
"hotspot_tmp": target + "_hotspot_tmp"}))
def read_quality(workflow, conf, tex):
if conf.pe:
for raw, target in conf.treatment_pairs_pe:
attach_back(
workflow,
ShellCommand(
"{tool} {input[fastq][0]} {input[fastq][1]} {output[stat][0]} {output[stat][1]}",
tool="dac_pe_read_quality",
input={"fastq": raw},
output={"stat": [i + "_read_quality.qc" for i in target]}))
# attach_back(workflow, PythonCommand(stat_fastqStat,
# input = {"seq": [ [ p + "_100k.seq" for p in target ] for target in conf.treatment_pair_data ]},
# output = {"json": conf.json_prefix + "_seq_quality.json"},
# param = {"samples": conf.treatment_bases, "seq_type": conf.pe}))
# attach_back(workflow, PythonCommand(
# seq_quality_doc,
# input = {"tex": tex, "json": conf.json_prefix + "_seq_quality.json"},
# output = {"seq": conf.latex_prefix + "seq_quality.tex", "len": conf.latex_prefix + "len.tex"},
# param = {"seq_type": conf.seq_type, "reps": len(conf.treatment_pairs),
# "pe_samples": conf.treatment_bases}))
else:
for raw, target in conf.treatment_pairs:
sample_fq = {"stat": target + "_read_quality.qc"}
attach_back(
workflow,
ShellCommand(
"{tool} {input} {output[stat]}",
tool="dac_se_read_quality",
input=raw,
output=sample_fq,
name="100k read sequence quality and sequence length"))
def _bwa(workflow, conf):
"""
incorpate ENCODE ChIP-seq alignment parameters
"""
for raw, target in conf.treatment_pairs:
param = {"threads": conf.threads,
"index":conf.get(conf.species, "genome_index"),
"prefix": target + "_raw_sorted",
"qc2": target + "_rawbam_stats.qc"}
if conf.pe:
bwa = attach_back(workflow, ShellCommand(
"{tool} {param[threads]} {param[index]} {input[fastq][0]} {input[fastq][1]} {output[bam]} {output[qc]} {param[prefix]} {param[qc2]}",
tool = "eap_run_bwa_pe",
input = {"fastq": raw},
output = {"bam": target + "_raw_sorted.bam", "qc": target + "_rawbam.qc"},
param = param,
name = "pair end mapping"))
else:
bwa = attach_back(workflow, ShellCommand(
"{tool} {param[threads]} {param[index]} {input[fastq]} {output[bam]} {output[qc]} {param[prefix]} {param[qc2]}",
tool = "eap_run_bwa_se",
input = {"fastq": raw},
output = {"bam": target + "_raw_sorted.bam", "qc": target + "_rawbam.qc"},
param = param,
name = "single end mapping"))
bwa.update(param = conf.items("bwa"))
def filter_bam(workflow, conf, tex):
""" filter bam file by samtools and sample by ucsc app
"""
for target in conf.treatment_targets:
input = {"raw": target + "_raw_sorted.bam"}
if conf.pe:
name = "pair"
tool = "dac_bam_pe_post_filter"
param = {
"mapq": 3,
"namesortedbamprefix": target + "_name_sorted",
"finalprefix": target + "_final",
"qc2": target + "_filter_bam_stats.qc"
}
output = {
"finalbam": target + "_final.bam",
"namesortedbam": target + "_name_sorted.bam",
"bamwithoutchrm": target + "_final_nochrm.bam",
"qc": target + "_filter_bam.qc"
}
attach_back(
workflow,
ShellCommand(
"{tool} {input[raw]} {param[namesortedbamprefix]} {output[namesortedbam]} {param[finalprefix]} {output[finalbam]} {param[mapq]} {output[bamwithoutchrm]} {output[qc]} {param[qc2]}",
tool=tool,
input=input,
output=output,
param=param,
name="%s end filtering" % name))
else:
name = "single"
tool = "dac_bam_se_post_filter"
param = {
"mapq": 3,
"finalprefix": target + "_final",
"qc2": target + "_filter_bam_stats.qc"
}
output = {
"finalbam": target + "_final.bam",
"bamwithoutchrm": target + "_final_nochrm.bam",
"qc": target + "_filter_bam.qc"
}
attach_back(
workflow,
ShellCommand(
"{tool} {input[raw]} {output[finalbam]} {param[mapq]} {output[qc]} {output[bamwithoutchrm]} {param[finalprefix]} {param[qc2]}",
tool=tool,
input=input,
output=output,
param=param,
name="%s end filtering" % name))
def bowtie(workflow, conf): # Mapping
"""
Use bowtie to map reads to genome,
call __bwa_sam2bam to convert sam to bam
:param workflow: samflow defined class
:param conf: parsed config files
:return: void
"""
for target in conf.sample_targets:
bowtie = attach_back(workflow,
ShellCommand(
"{tool} -p {param[NUM_THREADS]} -S -m 1 {param[index]} {input[fastq]} {output[sam]}",
tool = "bowtie",
input = {"fastq": target + ".fastq"},
output = {"sam": target + ".sam"},
param = {"NUM_THREADS": conf.threads,
## judge chosen species from basics section
"index": conf.get_path(conf.get("basics", "species"), "genome_index")},
name = "bowtie aln"))
bowtie.update(param = conf.items("bowtie"))
bowtie.allow_dangling = True
bowtie.allow_fail = True
_bowtie_sam2bam(workflow, conf)
## QC part--NOTE keeping the bwa legacy code!
stat_bwa(workflow, conf)
if conf.long:
tex_bwa(workflow, conf)
def merge_latex(workflow, conf):
## begin and end of the docs
latex_order = [
"_begin.tex",
"_summary_table.tex",
]
if conf.long:
latex_order += [
"_fastqc.tex",
"_fastqc_gc.tex",
"_map.tex",
"_conserv.tex",
# "_macs2.latex", "_macs2_on_sample.latex",
# "_phan.tex",
"_motif.tex",
"_contam.tex",
"_frip.tex",
]
latex_order.append("_end.tex")
latex_list = [conf.latex_prefix + i for i in latex_order]
merge_cmd = attach_back(
workflow,
ShellCommand("cat {param[tex]} > {output}",
output=conf.prefix + "_report.tex"))
merge_cmd.allow_fail = True
merge_cmd.param = {"tex": " ".join(latex_list)}
def replicates_peaks_overlap(workflow, conf): # peaks bed from each replicate
"""
:param workflow: class from samflow
:param conf: external parsed config file
:return: workflow through attach_back
"""
for i in range(len(conf.treatment_targets)):
for j in range(i + 1, len(conf.treatment_targets)):
replicates_overlap = attach_back(
workflow,
ShellCommand(
"{tool} -f {param[p]} -a {input[0]} -b {input[1]} | wc -l > {output}",
tool="intersectBed",
input=[
conf.treatment_targets[i] + "_sort_peaks.narrowPeak"
if conf.get("macs2", "type").lower()
in ["both", "narrow"] else conf.treatment_targets[i] +
"_b_sort_peaks.broadPeak", conf.treatment_targets[j] +
"_sort_peaks.narrowPeak" if conf.get(
"macs2", "type").lower() in ["both", "narrow"] else
conf.treatment_targets[j] + "_b_sort_peaks.broadPeak"
],
output=conf.prefix + "_%s_%s.overlap" % (i, j),
param={"p": 0.3},
name="Replicates peaks overlap QC"))
replicates_overlap.allow_fail = True # in case 0 peak in macs2
replicates_overlap.allow_dangling = True
## generate a barplot for meta distribution
replicates_overlap.update(param=conf.items("replicates"))
return workflow
def PBC(workflow, conf): # PBC1
"""
Introduce ENCODE II library complexity assessment methods
N1 / Nd, N1 is the location with exact one read, Nd is distinct location number
:param workflow: samflow class
:param conf: parsed config
:return: void
"""
for t in conf.sample_targets:
pbc1 = attach_back(
workflow,
ShellCommand(
"""
bamToBed -i {input[bam]} | {tool} \'{{l[$1"\\t"$2"\\t"$3"\\t"$6]+=1}} END {{for(i in l) print l[i]}}\' \\
| awk \'{{n[$1]+=1}} END {{for (i in n) print i"\\t"n[i]}}\' \\
| sort -k1n - > {output[hist]}
awk '{{
if (NR==1) {{N1=$2}}
Nd+=$2
}} END {{print N1,Nd,N1/Nd}}' {output[hist]} > {output[pbc]}
""",
tool="awk",
input={"bam": t + "_4000000.bam" if conf.down else t + ".bam"},
output={
"pbc": t + ".pbc",
"hist": t + ".hist"
},
name="PBC"))
pbc1.allow_fail = True
pbc1.allow_dangling = True
## QC part
stat_pbc(workflow, conf)
def r_exec(jinja_template_r):
ShellCommand("{tool} {input}",
tool = "Rscript",
name = 'Rscript',
input=jinja_template_r.param["render_dump"],
param={},
output=jinja_template_r.param["pdf"]).invoke()
def Phan(workflow, conf): # NSC, RSC, Qtag
"""
for calculating NSC, RSC score at 4M level
http://code.google.com/p/phantompeakqualtools/
(1) Determine strand cross-correlation peak / predominant fragment length OR print out quality measures
Rscript run_spp.R -c=<tagAlign/BAMfile> -savp -out=<outFile>
"""
# peaks calling by SPP needs control, for phantomqc, we do both treat and control independently
for t in conf.sample_targets:
if conf.down: ## default, this option
ibam = t + "_4000000.bam"
# elif conf.unsc: ## --total --unsc
# ibam = t + "_rmdup.bam"
else: ## --total
ibam = t + ".bam"
attach_back(
workflow,
ShellCommand(
"{tool} {param[script]} -c={input[chip]} -rf -savp -out={output[spp]} -odir={param[dir]}",
tool="Rscript",
input={"chip": ibam},
output={
"spp": t + ".spp",
"pdf": t + "_4000000.pdf" if conf.down else t + ".pdf"
},
param={
"script": conf.get("tool", "spp"),
"dir": os.path.dirname(t + ".spp")
},
name="SPP"))
stat_phan(workflow, conf)
if conf.long:
tex_phan(workflow, conf)
def DHS(workflow, conf): # DHS overlap percentage
"""
get peaks overlapping percentage with union DHS
:param workflow: uniform pipeline workflow from samflow
:param conf: parsed config files
:return: workflow
"""
peaks = conf.prefix + "_sort_peaks.narrowPeak" if conf.get(
"macs2", "type") in ["both", "narrow"
] else conf.prefix + "_b_sort_peaks.broadPeak"
DHS = attach_back(
workflow,
ShellCommand("""
n=$(head -n {param[p]} {input[MACS2_bed]} | wc -l)
dhs=$(head -n {param[p]} {input[MACS2_bed]} | {tool} -wa -u -a - -b {input[DHS_peaks_bed]}|wc -l)
##dhs=$(echo \"scale=5;$dhs/$n\" | bc)
echo $n,$dhs > {output}
""",
tool="intersectBed",
input={
"MACS2_bed":
peaks,
"DHS_peaks_bed":
conf.get(conf.get("basics", "species"), "dhs")
},
output=conf.prefix + ".dhs",
param={"p": 5000},
name="intersect DHS"))
DHS.allow_dangling = True
DHS.allow_fail = True
def star(workflow, conf): # Mapping
"""
Use star to map reads to genome,
call __bwa_sam2bam to convert sam to bam
:param workflow: samflow defined class
:param conf: parsed config files
:return: void
"""
for target in conf.sample_targets:
star = attach_back(
workflow,
ShellCommand(
"{tool} --genomeDir {param[index]} --runThreadN {param[NUM_THREADS]} --readFilesIn {input[fastq]} --outFileNamePrefix {param[prefix]}",
tool="STAR",
input={"fastq": target + ".fastq"},
output={"sam": target + "Aligned.out.sam"},
param={
"NUM_THREADS":
conf.threads,
"prefix":
target,
## judge chosen species from basics section
"index":
conf.get_path(conf.get("basics", "species"),
"genome_index")
},
name="star aln"))
star.update(param=conf.items("bowtie"))
_star_sam2bam(workflow, conf)
## QC part--NOTE keeping the bwa legacy code!
stat_bwa(workflow, conf)
if conf.long:
tex_bwa(workflow, conf)
def fastqc(workflow, conf):
"""
fastqc to extract gc contents(not yet) and median sequence quality
:param workflow:
:param conf:
:return:
"""
for raw, target in conf.sample_pairs:
if conf.pe:
fastqc_run = attach_back(
workflow,
ShellCommand(
"{tool} {input} --extract -t {param[threads]} -o {output[target_dir]}",
## only check one pair
input=target[0] + "_100k.fastq",
output={
"target_dir":
conf.target_dir,
"fastqc_summary":
target[0] + "_100k_fastqc/fastqc_data.txt"
},
tool="fastqc",
param={"threads": conf.threads},
name="fastqc"))
else:
fastqc_run = attach_back(
workflow,
ShellCommand(
"{tool} {input} --extract -t {param[threads]} -o {output[target_dir]}",
input=target + "_100k.fastq",
output={
"target_dir": conf.target_dir,
"fastqc_summary":
target + "_100k_fastqc/fastqc_data.txt"
},
tool="fastqc",
param={"threads": conf.threads},
name="fastqc"))
fastqc_run.update(param=conf.items("fastqc"))
fastqc.allow_fail = True
fastqc.allow_dangling = True
## QC part of chilin
## use conf property conf.long = True
stat_fastqc(workflow, conf)
if conf.long:
tex_fastqc(workflow, conf)
def hotspotv4(workflow, conf, tex):
for target in conf.treatment_targets:
hotspot=attach_back(workflow,
ShellCommand(
"{tool} {param[hotspot_dir]} {param[genome]} {input[bam]} {param[readsize]} {output[narrowbb]} {output[broadbb]} {output[bigwig]} {param[tmp]} {output[hotspot_output]} {input[narrowas]} {input[broadas]} {param[chromsize]} {output[narrow]} {output[broad]}",
tool = "eap_run_hotspot",
input = {"bam": target + "_final_nochrm.bam",
"narrowas": narrow,
"broadas": broad},
output = {"narrowbb": target + ".narrowPeak.bigBed",
"broadbb": target + ".broadPeak.bigBed",
"narrow": target + ".narrowPeak",
# "qc1": target + ".narrowPeak.qc",
# "qc2": target + ".broadPeak.qc",
"broad": target + ".broadPeak",
"bigwig": target + ".bigWig",
"hotspot_output": target + "_hotspot"},
param = {"hotspot_dir": conf.get("tool", "peak_calling"),
"genome": conf.species,
"chromsize": conf.get(conf.species, "chrom_len"),
"tmp": target + "_hotspot_peak_call_tmp",
"readsize": 36}))
have_treat_reps = len(conf.treatment_pairs) >= 2 ## replicates
if have_treat_reps:
eval_reps(workflow, conf, tex)
catsam = attach_back(workflow, ShellCommand(
"{tool} cat {param[bams]} > {output[bam]}",
tool = "samtools",
input ={"bams": [ target + "_final.bam" for target in conf.treatment_targets]},
output = {"bam": conf.prefix + "_pool.bam"}))
catsam.param.update(bams=' '.join(catsam.input["bams"]))
hotspot_merge = hotspot.clone
hotspot_merge.param.update(tmp=conf.prefix+"_hotspot_peak_call_tmp")
hotspot_merge.input.update(bam = conf.prefix + "_pool.bam")
hotspot_merge.output ={"narrowbb": conf.prefix + ".narrowPeak.bigBed",
"broadbb": conf.prefix + ".broadPeak.bigBed",
"narrow": conf.prefix + ".narrowPeak",
# "qc1": conf.prefix + ".narrowPeak.qc",
# "qc2": conf.prefix + ".broadPeak.qc",
"broad": conf.prefix + ".broadPeak",
"bigwig": conf.prefix + ".bigWig",
"hotspot_output": conf.prefix + "_hotspot"}
attach_back(workflow, hotspot_merge)
def make_link_command(orig, dest):
"""
link original input to destination files
:param orig: input
:param dest: link symbol
:return: ShellCommand Class
ln has machine type problem
""" ## not use symbol link
return ShellCommand("cp -fr {input} {output}",
input=orig,
output=dest,
name="copy")
def merge_bams(workflow, conf): ## merge input and chip bam
"""
input multiple input and multiple control to merge into one file separately
:return:
"""
# merge all treatments into one
merge_bams_treat = ShellCommand(
"{tool} merge {output[merged]} {param[bams]}",
tool="samtools",
input=[target + ".bam" for target in conf.treatment_targets],
output={"merged": conf.prefix + "_treatment.bam"})
merge_bams_treat.param = {"bams": " ".join(merge_bams_treat.input)}
merge_bams_treat.allow_fail = True
merge_bams_treat.allow_dangling = True
if len(conf.treatment_targets) > 1:
attach_back(workflow, merge_bams_treat)
else:
# when there's only one treatment sample, use copying instead of merging
attach_back(
workflow,
make_link_command(merge_bams_treat.input[0],
merge_bams_treat.output["merged"]))
# merging step will be skipped if control sample does not exist
# So be careful to check whether there are control samples before using `_control.bam`
if len(conf.control_targets) > 1:
merge_bams_control = merge_bams_treat.clone
merge_bams_control.input = [
target + ".bam" for target in conf.control_targets
]
merge_bams_control.output = {"merged": conf.prefix + "_control.bam"}
merge_bams_control.param = {"bams": " ".join(merge_bams_control.input)}
attach_back(workflow, merge_bams_control)
elif len(conf.control_targets) == 1:
attach_back(
workflow,
make_link_command(conf.control_targets[0] + ".bam",
conf.prefix + "_control.bam"))
def sampling(orig, dest, rand, format, conf): # call fastq_sampling
"""
prepare sampling fastq files for library contamination and fastqc
rand: the number of random selected fastq reads
use lh3's https://github.com/lh3/seqtk/ to sample fastq and fastq.gz
"""
if format == "fastq":
#return PythonCommand(fastq_sampling,
# input=orig,
# output=dest,
# param={"random_number": rand})
## faster and support fastq.gz
## if paired end, we must use same -s
return ShellCommand("{tool} sample -s 11 {input[fastq]} {param[rand]} > {output[fastq_sample]}",
tool = "seqtk",
input = orig,
output = dest,
param = {"rand": 100000})
elif format == "sam":
## samtools sampling
## add judge condition
return ShellCommand("""
count=$({tool} view -Sc {input[sam]})
## judge mapped reads number less than sampling number
if [ $count -le {param[random_number]} ]
then
ln -f {input[sam]} {input[sam]}.{param[random_number]}
{tool} view -bS {input[sam]}.{param[random_number]} > {output[samp]}
else
sampling_pe_sam.py {input[sam]} {param[random_number]} {param[pair]}
{tool} view -bS {input[sam]}.{param[random_number]} > {output[samp]}
fi
""",
tool = "samtools",
input={"sam": orig},
output={"samp": dest},
param={"random_number": rand, "pair": str(conf.pe)},
name = "sampling bam")
def fragment(workflow, conf):
## this is done after FRiP
if conf.get("tool", "macs2"):
macs2_bin = conf.get("tool", "macs2")
else:
macs2_bin = "macs2"
for target in conf.treatment_targets:
fragment_size = attach_back(
workflow,
ShellCommand(
"{tool} predictd -i {input[bam]} --rfile {param[prefix]} -g {param[species]}",
tool=macs2_bin,
input={"bam": target + ".bam"},
output={"R": target + "_model.R"},
param={
"prefix": target + "_model.R",
"species": 'hs'
}))
fragment_size.update(param=conf.items("macs2"))
## except too few peaks for modeling
fragment_size.allow_fail = True
fragment_size.allow_dangling = True
## extract standard deviation from MACS2 model.R,
## use m, p, and pileup value for standard deviation; mean fragment size is provided (choose the one with highest correlation)
frag_qc = attach_back(
workflow,
PythonCommand(
stat_frag_std,
input={
"r":
[target + "_model.R" for target in conf.treatment_targets]
},
output={
"json": conf.json_prefix + "_frag.json",
"r":
[target + "_frag_sd.R" for target in conf.treatment_targets]
},
param={
"samples": conf.treatment_bases,
"frag_tool": "BAMSE"
},
name="macs2 model R script parser"))
frag_qc.allow_fail = True
frag_qc.allow_dangling = True
def bedAnnotate_ceas(workflow, conf):
"""
Calls bedAnnotate to get the genome distribution of the summits
"""
import os
summits = conf.prefix + "_sort_summits.bed" if conf.get(
"macs2", "type") in ["both", "narrow"
] else conf.prefix + "_b_sort_peaks.broadPeak"
ceas = attach_back(
workflow,
ShellCommand(
"""{tool} -g {param[geneTable]} -b {input} -e {output[exon]} -t {output[gene]}> {output[meta]}
meta_info.sh {output[gene]} {output[exon]} 2000 {param[chrominfo]}
""",
tool="bedAnnotate.py",
input=summits,
output={
"meta": conf.prefix + ".meta",
"gene": os.path.join(conf.target_dir, "gene.bed"),
"exon": os.path.join(conf.target_dir, "exon.bed"),
"promoter": os.path.join(conf.target_dir, "gene.bed_promoter"),
"exon": os.path.join(conf.target_dir, "gene.bed_exon")
},
param={
"geneTable":
conf.get_path(conf.get("basics", "species"), "geneTable"),
"chrominfo":
conf.get_path(conf.get("basics", "species"), "chrom_len")
},
name="bedAnnotate (ceas)"))
try:
has_velcro = conf.get(conf.get("basics", "species"), "velcro")
has_dhs = conf.get(conf.get("basics", "species"), "dhs")
except:
has_velcro = ""
has_dhs = ""
ceas.allow_fail = True
ceas.allow_dangling = True
if has_dhs:
DHS(workflow, conf)
if has_velcro:
velcro(workflow, conf)
stat_bedAnnotate(workflow, conf, has_dhs, has_velcro)
def FRiP(workflow, conf): # FRiP
"""
Fraction of Reads in Peaks regions at 4M reads level
For example: 2 treat, 2 control
modify: without down sampling read peaks calling, use merged peaks for comparison
"""
## use merged peaks for evaluation after removing chrM reads
for t in conf.sample_targets:
if conf.frip: ## sampling 5M reads
reads = t + "_5000000_nochrM.bam"
else:
reads = t + "_4000000_nochrM.bam"
frip = attach_back(
workflow,
ShellCommand("""
fr=$(bedtools intersect -f {param[p]} -wa -u -abam {input[reads]} -b {input[peaks]} -bed | wc -l)
total=$(samtools flagstat {input[reads]} | head -1 | cut -d" " -f1)
echo $fr,$total > {output[frip]}
""",
tool="intersectBed",
input={
"reads":
reads if conf.down else t + "_nochrM.bam",
"peaks":
conf.prefix + "_sort_peaks.narrowPeak" if
conf.get("macs2", "type") in ["both", "narrow"]
else conf.prefix + "_b_sort_peaks.broadPeak"
},
output={"frip": t + ".frip"},
param={"p": "1E-9"},
name="FRiP score"))
## in case that peaks calling on 4M reads may be very poor,
## no peaks generated, allow fail and dangling
frip.allow_fail = True
frip.allow_dangling = True
frip.update(param=conf.items("bedtools"))
## QC part
stat_frip(workflow, conf)
if conf.long:
tex_frip(workflow, conf)
def bowtie(workflow, conf, target, output, index): # Mapping
"""
Use bowtie to map reads to genome,
"""
bowtie = attach_back(
workflow,
ShellCommand(
"{tool} -p {param[NUM_THREADS]} -S -m 1 {param[index]} {input[fastq]} {output[sam]}",
tool="bowtie",
input={"fastq": target + "_100k.fastq"},
output={"sam": output},
param={
"NUM_THREADS": conf.threads,
## judge chosen species from basics section
"index": index
},
name="bowtie aln"))
bowtie.update(param=conf.items("bowtie"))
bowtie.allow_fail = True
bowtie.allow_dangling = True
return workflow
def eval_reps(workflow, conf, tex):
peaks = [ target + ".narrowPeak" for target in conf.treatment_targets ]
attach_back(workflow, ShellCommand(
"""
cat {param[narrowPeaks]} | sort -k1,1 -k2,2n - | bedtools merge -i - > {output[mergedPeak]}
bedToBigBed {output[mergedPeak]} {param[chromsize]} {output[mergedPeakbb]}
bigWigCorrelate -restrict={output[mergedPeakbb]} {param[bigwigs]} 1>{output[qc1]}
{tool} {param[narrowPeaksbb]} {output[qc2]}
""",
tool = "edwComparePeaks",
input = {"narrowPeaks": peaks,
"bigwigs": [ target + ".bigWig" for target in conf.treatment_targets ],
"narrowPeakbbs": [ target + ".narrowPeak.bigBed" for target in conf.treatment_targets ]},
output = {"mergedPeak": conf.prefix + "_merge.bed",
"mergedPeakbb": conf.prefix + "_merged.bigBed",
"qc1": conf.prefix + "_cor.qc",
"qc2": conf.prefix + "_overlap.qc"},
param = {"narrowPeaksbb": " ".join([ target + ".narrowPeak.bigBed" for target in conf.treatment_targets ]),
"narrowPeaks": " ".join([ target + ".narrowPeak" for target in conf.treatment_targets ]),
"bigwigs": " ".join([ target + ".bigWig" for target in conf.treatment_targets ]),
"chromsize": conf.get(conf.species, "chrom_len")}))
def star(workflow, conf, target, output, index): # Mapping
"""
Use star to map reads to genome,
"""
star = attach_back(
workflow,
ShellCommand(
"{tool} --genomeDir {param[index]} --runThreadN {param[NUM_THREADS]} --readFilesIn {input[fastq]} --outFileNamePrefix {param[prefix]}",
tool="STAR",
input={"fastq": target + "_100k.fastq"},
output={"sam": output},
param={
"NUM_THREADS": conf.threads,
"prefix": target,
## judge chosen species from basics section
"index": index
},
name="star aln"))
star.update(param=conf.items("bowtie"))
star.allow_dangling = True
star.allow_fail = True
return workflow
def replicates_bw_correlation(workflow,
conf): ## correlation among different replicates
"""
Use UCSC binary bigWigCorrelate to calculate reads density correlation
collections in json files from qc
:param workflow: samflow class
:param conf: parsed config files
:return: void
"""
replicates_correlation = attach_back(
workflow,
ShellCommand(
"{tool} {param[input_list]} > {output}",
tool="wigCorrelate",
input=[target + "_treat.bw" for target in conf.treatment_targets],
output=conf.prefix + ".cor",
param={"input_list": []},
name="correlation between bigwiggle"))
replicates_correlation.update(
param={"input_list": " ".join(replicates_correlation.input)})
replicates_correlation.allow_fail = True # in case 0 peak in macs2
replicates_correlation.allow_dangling = True
def render_pdf(workflow, conf, long=True):
latex_environ(workflow, conf)
summary_table_latex(workflow, conf)
merge_latex(workflow, conf)
render = attach_back(
workflow,
ShellCommand(
# Somehow the pdflatex has to be invoked twice..
"{tool} -output-directory {output[dir]} -jobname={param[name]} {input} \
&& {tool} -output-directory {output[dir]} -jobname={param[name]} {input}",
tool="pdflatex",
input=conf.prefix + "_report.tex",
# output[pdf] should use "conf.prefix" to have the absolute path
output={
"dir": conf.target_dir,
"pdf": conf.prefix + "_report.pdf"
},
# param[name] should use "conf.id" to avoid using absolute path
param={"name": conf.id + "_report"},
name="report"))
render.allow_fail = True
def stat_bwa(workflow, conf): ## use samtools to parse mappable reads from bwa
"""
bam files are filtered by samtools -q 1, so mapped reads are considered to be unique
"""
for t in conf.sample_targets:
stat = attach_back(workflow, ShellCommand(
"""
{tool} view -Sc {input[sam]} > {output[total]}
{tool} flagstat {input[bam]} > {output[stat]}
""",
tool = "samtools",
input = {"bam": t + ".bam",
"sam": t + ".sam"},
output = {"stat": t + "_mapped.bwa",
"total": t + "_total.bwa"}))
stat.allow_fail = True
stat.allow_dangling = True
collect = attach_back(workflow, PythonCommand(json_bwa,
input={"bwa_mapped": [ t + "_mapped.bwa" for t in conf.sample_targets ],
"bwa_total": [ t + "_total.bwa" for t in conf.sample_targets ]},
output={"json": conf.json_prefix+"_map.json"},
param={"sample":conf.sample_bases},
name="bwa qc"))
collect.allow_dangling = True
collect.allow_fail = True
if conf.long:
long_collect = attach_back(workflow, PythonCommand(bwa_figures,
input = {"dbaccessor": resource_filename("chilin2.modules.dbaccessor", "ChiLinQC.db"),
"json": conf.json_prefix + "_map.json",
"template": resource_filename("chilin2.modules.summary", "R_culmulative_plot.R")},
output = {"pdf": conf.prefix + "_bwa_compare.pdf", "R": conf.prefix+"_bwa_compare.R"},
param = {"sample": conf.sample_bases}))
long_collect.allow_fail = True
long_collect.allow_fail = True
def velcro(workflow, conf):
vel = attach_back(
workflow,
ShellCommand(
"""
n=$(head -n {param[p]} {input[MACS2_bed]} | wc -l)
velcro=$(head -n {param[p]} {input[MACS2_bed]} | {tool} -wa -u -a - -b {input[velcro_peaks_bed]} | wc -l)
velcro=$(echo \"scale=5;$velcro/$n\" | bc)
echo $velcro > {output}
""",
tool="intersectBed",
input={
"MACS2_bed":
conf.prefix + "_sort_peaks.narrowPeak" if conf.get(
"macs2", "type") in ["both", "narrow"] else conf.prefix +
"_b_sort_peaks.broadPeak",
"velcro_peaks_bed":
conf.get(conf.get("basics", "species"), "velcro")
},
output=conf.prefix + ".velcro",
param={"p": 5000},
name="velcro overlap"))
vel.allow_fail = True
vel.allow_dangling = True
def _macs2(workflow, conf):
# merge all treatments into one
merge_bams_treat = ShellCommand(
"{tool} merge {output[merged]} {param[bams]}",
tool="samtools",
input=[target + ".bam" for target in conf.treatment_targets],
output={"merged": conf.prefix + "_treatment.bam"})
merge_bams_treat.param = {"bams": " ".join(merge_bams_treat.input)}
if len(conf.treatment_targets) > 1:
attach_back(workflow, merge_bams_treat)
else:
# when there's only one treatment sample, use copying instead of merging
attach_back(workflow, make_copy_command(merge_bams_treat.input[0], merge_bams_treat.output["merged"]))
# merging step will be skipped if control sample does not exist
# So be careful to check whether there are control samples before using `_control.bam`
if len(conf.control_targets) > 1:
merge_bams_control = merge_bams_treat.clone
merge_bams_control.input = [target + ".bam" for target in conf.control_targets]
merge_bams_control.output = {"merged": conf.prefix + "_control.bam"}
merge_bams_control.param = {"bams": " ".join(merge_bams_control.input)}
attach_back(workflow, merge_bams_control)
elif len(conf.control_targets) == 1:
attach_back(workflow, make_copy_command(conf.control_targets[0] + ".bam", conf.prefix + "_control.bam"))
macs2_on_merged = attach_back(workflow, ShellCommand(
"{tool} callpeak -B -q 0.01 --keep-dup {param[keep_dup]} --shiftsize={param[shiftsize]} --nomodel \
{param[treat_opt]} {param[control_opt]} -n {param[description]}",
tool="macs2",
input={"treat": conf.prefix + "_treatment.bam"},
output={"peaks": conf.prefix + "_peaks.bed",
"summit": conf.prefix + "_summits.bed",
"treat_bdg": conf.prefix + "_treat_pileup.bdg",
"ENCODE": conf.prefix + "_peaks.encodePeak",
"peaks_xls": conf.prefix + "_peaks.xls",
"control_bdg": conf.prefix + "_control_lambda.bdg"},
param={"description": conf.prefix,
"keep_dup": 1,
"shiftsize": 73},
name="macs2_callpeak_merged"))
macs2_on_merged.param["treat_opt"] = "-t " + macs2_on_merged.input["treat"]
# control option is skipped if control samples does not exist
if len(conf.control_targets) >= 1:
macs2_on_merged.input["control"] = conf.prefix + "_control.bam"
macs2_on_merged.param["control_opt"] = "-c " + macs2_on_merged.input["control"]
else:
macs2_on_merged.param["control_opt"] = ""
macs2_on_merged.update(param=conf.items("macs2"))
# For bedGraphToBigwiggle bugs, we need to remove coordinates over-border coordinates
# As _control_lambda.bdg always exist. There are no need to check whether there are control samples.
bdg_trim_control = attach_back(workflow,
ShellCommand(
'{tool} intersect -a {input[bdg]} -b {input[chrom_bed]} -wa -f 1.00 > {output}',
tool="bedtools",
input={"bdg": conf.prefix + "_control_lambda.bdg",
'chrom_bed': conf.get_path("lib", "chrom_bed")},
output=conf.prefix + "_control_lambda.bdg.tmp",
name="bedGraph filtering"))
bdg_trim_treat = bdg_trim_control.clone
bdg_trim_treat.input["bdg"] = conf.prefix + "_treat_pileup.bdg"
bdg_trim_treat.output = conf.prefix + "_treat_pileup.bdg.tmp"
attach_back(workflow, bdg_trim_treat)
bdg2bw_treat = attach_back(workflow,
ShellCommand(
"{tool} {input[bdg]} {input[chrom_len]} {output[bw]}",
tool="bedGraphToBigWig",
input={"bdg": conf.prefix + "_control_lambda.bdg.tmp",
"chrom_len": conf.get("lib", "chrom_len")},
output={"bw": conf.prefix + "_control.bw"},
name="bdg_to_bw"))
# prototype used here to do the similar thing on treatment file
bdg2bw_control = bdg2bw_treat.clone
bdg2bw_control.input["bdg"] = conf.prefix + "_treat_pileup.bdg.tmp"
bdg2bw_control.output["bw"] = conf.prefix + "_treat.bw"
attach_back(workflow, bdg2bw_control)
attach_back(workflow, PythonCommand(
stat_macs2,
input={"macs2_peaks_xls": conf.prefix + "_peaks.xls",
"db": ChiLinQC_db,
"template": rlang_template},
output={"json": conf.json_prefix + "_macs2.json",
"R": conf.prefix + "_macs2.R",
"pdf": conf.prefix + "_macs2.pdf"},
param={"id": conf.id},
name="MACS2 summary"))
def test_invoke_collect_output(self):
echo_cmd = ShellCommand("echo test_collect").set_stdout_collecting()
self.assertTrue(echo_cmd.invoke())
self.assertEqual(echo_cmd.result, "test_collect\n")
def test_invoke_non_exist_input(self):
non_exist_input_cmd = ShellCommand("cat < {input}", input="non_exist_file")
self.assertFalse(non_exist_input_cmd.invoke())
def macs2_rep(workflow, conf):
# Though macs command already exists, I choose not to use prototype here
# Because the prototype definition and usage might be far from each other, making codes not readable
if conf.get("tool", "macs2"):
macs2_bin = conf.get("tool", "macs2")
else:
macs2_bin = "macs2"
format = " -f BAMPE " if conf.pe else " "
for target in conf.treatment_targets:
## DNase, H3K4, H2AZ, all acetyl marks, or TF
if conf.get("macs2", "type").lower() in ["both", "narrow"]: ## for DNase, H3K4, H2AZ, all acetyl marks, or TF
macs2_on_rep_narrow = attach_back(workflow,
ShellCommand(
"""
{tool} callpeak --SPMR -B -q {param[fdr]} --keep-dup {param[keep_dup]} --extsize={param[extsize]} --nomodel -g {param[species]} {param[format]} {param[treat_opt]} {param[control_opt]} -n {param[description]} && cut -f1,2,3,4,9 {output[peaks]} > {output[bedtmp]}
## remove weird path characters
cp {output[peaks]} {output[peakstmp]}
cp {output[summits]} {output[summitstmp]}
awk \'{{OFS="\\t";n+=1;$4="peak"n;print $0}}\' {output[peakstmp]} > {output[peaks]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[bedtmp]} > {output[bed]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[summitstmp]} > {output[summits]}
""",
tool=macs2_bin,
input={"treat": target + ".bam"},
output={"peaks": target + "_peaks.narrowPeak",
"peakstmp": target + "_peaks.narrowPeak.tmp",
"summits": target + "_summits.bed",
"summitstmp": target + "_summits.bed.tmp",
"bed": target + "_peaks.bed",
"bedtmp": target + "_peaks.bed.tmp",
"treat_bdg": target + "_treat_pileup.bdg",
"peaks_xls": target + "_peaks.xls",
"control_bdg": target + "_control_lambda.bdg"},
param={"description": target, "keep_dup": 1, "extsize": 73*2, "species": "hs", "fdr":0.01, "format": format},
name="macs2_callpeak_rep"))
macs2_on_rep_narrow.param["treat_opt"] = "-t " + macs2_on_rep_narrow.input["treat"]
sort = attach_back(workflow,
ShellCommand(
"{tool} -r -g -k 9 {input} > {output}",
tool = "sort",
input = target + "_peaks.narrowPeak",
output = target + "_sort_peaks.narrowPeak"))
# control option is skipped if control samples does not exist
if len(conf.control_targets) >= 1:
macs2_on_rep_narrow.input["control"] = conf.prefix + "_control.bam"
macs2_on_rep_narrow.param["control_opt"] = "-c " + macs2_on_rep_narrow.input["control"]
else:
macs2_on_rep_narrow.param["control_opt"] = ""
macs2_on_rep_narrow.update(param=conf.items("macs2"))
macs2_on_rep_narrow.allow_dangling = True
macs2_on_rep_narrow.allow_fail = True
if conf.get("macs2", "type").lower() in ["both", "broad"]: # K9, K36, K79 and K27 methylation, both for chromatin regulator, all other histone marks
macs2_on_rep_broad = attach_back(workflow,
ShellCommand(
"""
{tool} callpeak --SPMR -B -q {param[fdr]} {param[treat_opt]} {param[control_opt]} --keep-dup {param[keep_dup]} --broad --broad-cutoff {param[fdr]} -g {param[species]} {param[format]} -n {param[description]} && cut -f1,2,3,4,9 {output[peaks]} > {output[bedtmp]}
## remove weird path characters
cp {output[peaks]} {output[peakstmp]}
awk \'{{OFS="\\t";n+=1;$4="peak"n;print $0}}\' {output[peakstmp]} > {output[peaks]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[bedtmp]} > {output[bed]}
""",
tool=macs2_bin,
input = {"treat": target + ".bam"},
output = {"peaks": target + "_b_peaks.broadPeak",
"peakstmp": target + "_b_peaks.broadPeak.tmp",
"bed": target + "_b_peaks.bed",
"bedtmp": target + "_b_peaks.bed.tmp",
"treat_bdg": target + "_b_treat_pileup.bdg",
"peaks_xls": target + "_b_peaks.xls",
"control_bdg": target + "_b_control_lambda.bdg"},
param = {"description": target + "_b",
"species": "hs",
"format": format,
"fdr": 0.01},
name = " macs2 broad peaks"))
macs2_on_rep_broad.param["treat_opt"] = " -t " + macs2_on_rep_broad.input["treat"]
macs2_on_rep_broad.update(param=conf.items("macs2"))
macs2_on_rep_broad.allow_dangling = True
macs2_on_rep_broad.allow_fail=True
if len(conf.control_targets) >= 1:
macs2_on_rep_broad.input["control"] = conf.prefix + "_control.bam"
macs2_on_rep_broad.param["control_opt"] = "-c " + macs2_on_rep_broad.input["control"]
else:
macs2_on_rep_broad.param["control_opt"] = ""
## some broad peaks cannot be called
macs2_on_rep_broad.update(param=conf.items("macs2"))
sort = attach_back(workflow,
ShellCommand(
"{tool} -r -g -k 9 {input} > {output}",
tool = "sort",
input = target + "_b_peaks.broadPeak",
output = target + "_b_sort_peaks.broadPeak",
name = "sort broad peaks"))
sort.allow_dangling = True
sort.allow_fail=True
## For bedGraphToBigwiggle bugs, we need to remove coordinates outlier
if conf.get("macs2", "type").lower() in ["both", "broad"]:
cont_bdg = target + "_b_control_lambda.bdg"
treat_bdg = target + "_b_treat_pileup.bdg"
if conf.get("macs2", "type").lower() in ["both", "narrow"]:
cont_bdg = target + "_control_lambda.bdg"
treat_bdg = target + "_treat_pileup.bdg"
import os
bdg_trim_controlrep = attach_back(workflow,
ShellCommand(
'{tool} intersect -a {input} -b {param[chrom_bed]} -wa -f 1.00 > {output}',
tool="bedtools",
input=cont_bdg,
output=cont_bdg + ".tmp",
param={"chrom_bed": os.path.join(conf.target_dir, "chrom.bed")},
name="bedGraph control replicate filtering"))
bdg_trim_controlrep.allow_dangling = True
bdg_trim_controlrep.allow_fail=True
bdg_trim_treatrep = bdg_trim_controlrep.clone
bdg_trim_treatrep.input = treat_bdg
bdg_trim_treatrep.output = treat_bdg + ".tmp"
bdg_trim_treatrep.allow_dangling = True
bdg_trim_treatrep.allow_fail=True
attach_back(workflow, bdg_trim_treatrep)
bdg2bw_treatrep = attach_back(workflow,
ShellCommand(
"{tool} {input} {param[chrom_len]} {output}",
tool="bedGraphToBigWig",
input=treat_bdg+".tmp",
output=target + "_treat.bw",
param={"chrom_len": conf.get_path(conf.get("basics", "species"), "chrom_len")},
name="bdg_to_bw treat"))
## in case broad peaks calling failed
bdg2bw_treatrep.allow_dangling = True
bdg2bw_treatrep.allow_fail=True
# prototype used here to do the similar thing on treatment file
bdg2bw_controlrep = bdg2bw_treatrep.clone
bdg2bw_controlrep.input = cont_bdg + ".tmp"
bdg2bw_controlrep.output = target + "_control.bw"
attach_back(workflow, bdg2bw_controlrep)
## in case broad peaks calling failed
bdg2bw_controlrep.allow_dangling = True
bdg2bw_controlrep.allow_fail=True
stat_macs2_on_rep(workflow, conf)
def macs2(workflow, conf):
if conf.get("tool", "macs2"):
macs2_bin = conf.get("tool", "macs2")
else:
macs2_bin = "macs2"
format = "-f BAMPE" if conf.pe else " "
if conf.get("macs2", "type").lower() in ["both", "narrow"]: ## for DNase, H3K4, H2AZ, all acetyl marks, or TF
macs2_on_merged_narrow = attach_back(workflow, ShellCommand(
"""
{tool} callpeak --SPMR -B -q {param[fdr]} --keep-dup {param[keep_dup]} --extsize={param[extsize]} --nomodel -g {param[species]} {param[format]} {param[treat_opt]} {param[control_opt]} -n {param[description]} && cut -f1,2,3,4,9 {output[peaks]} > {output[bedtmp]}
## remove weird path characters
cp {output[peaks]} {output[peakstmp]}
cp {output[summits]} {output[summitstmp]}
awk \'{{OFS="\\t";n+=1;$4="peak"n;print $0}}\' {output[peakstmp]} > {output[peaks]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[bedtmp]} > {output[bed]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[summitstmp]} > {output[summits]}
""",
tool=macs2_bin,
input={"treat": conf.prefix + "_treatment.bam"},
output={"peaks": conf.prefix + "_peaks.narrowPeak",
"peakstmp": conf.prefix + "_peaks.narrowPeak.tmp",
"bed": conf.prefix + "_peaks.bed",
"bedtmp": conf.prefix + "_peaks.bed.tmp",
"summits": conf.prefix + "_summits.bed",
"summitstmp": conf.prefix + "_summits.bed.tmp",
"treat_bdg": conf.prefix + "_treat_pileup.bdg",
"peaks_xls": conf.prefix + "_peaks.xls",
"control_bdg": conf.prefix + "_control_lambda.bdg"},
param={"description": conf.prefix,
"keep_dup": 1,
"format": format,
"extsize": 73 * 2, # extsize=2*shiftsize
"fdr": 0.01,
"species": "hs"},
name="macs2_callpeak_merged"))
macs2_on_merged_narrow.param["treat_opt"] = "-t " + macs2_on_merged_narrow.input["treat"]
# control option is skipped if control samples does not exist
if len(conf.control_targets) >= 1:
macs2_on_merged_narrow.input["control"] = conf.prefix + "_control.bam"
macs2_on_merged_narrow.param["control_opt"] = "-c " + macs2_on_merged_narrow.input["control"]
else:
macs2_on_merged_narrow.param["control_opt"] = ""
macs2_on_merged_narrow.update(param=conf.items("macs2"))
sort = attach_back(workflow,
ShellCommand(
"""{tool} -r -g -k 9 {input[peaks]} > {output[p]}
{tool} -r -g -k 5 {input[summits]} > {output[s]}
""",
tool = "sort",
input = {"peaks": conf.prefix + "_peaks.narrowPeak",
"summits": conf.prefix + "_summits.bed"},
output = {"p":conf.prefix + "_sort_peaks.narrowPeak",
"s":conf.prefix + "_sort_summits.bed"},
name = "sort peaks"))
macs2_on_merged_narrow.allow_fail = True
macs2_on_merged_narrow.allow_dangling = True
sort.allow_fail = True
sort.allow_dangling = True
if conf.get("macs2", "type").lower() in ["both", "broad"]: # K9, K36, K79 and K27 methylation, both for chromatin regulator, all other histone marks
macs2_on_merged_broad = attach_back(workflow,
ShellCommand(
"""
{tool} callpeak --SPMR -B -q {param[fdr]} {param[treat_opt]} {param[control_opt]} --keep-dup {param[keep_dup]} --broad --broad-cutoff {param[fdr]} -g {param[species]} {param[format]} -n {param[description]} && cut -f1,2,3,4,9 {output[peaks]} > {output[bedtmp]}
## remove weird path characters
cp {output[peaks]} {output[peakstmp]}
awk \'{{OFS="\\t";n+=1;$4="peak"n;print $0}}\' {output[peakstmp]} > {output[peaks]}
awk \'{{OFS="\\t";n+=1;$4=n;print $1,$2,$3,"peak"$4,$5}}\' {output[bedtmp]} > {output[bed]}
""",
tool=macs2_bin,
input = {"treat": conf.prefix + "_treatment.bam"},
output = {"peaks": conf.prefix + "_b_peaks.broadPeak",
"peakstmp": conf.prefix + "_b_peaks.broadPeak.tmp",
"bed": conf.prefix + "_b_peaks.bed",
"bedtmp": conf.prefix + "_b_peaks.bed.tmp",
"treat_bdg": conf.prefix + "_b_treat_pileup.bdg",
"peaks_xls": conf.prefix + "_b_peaks.xls",
"control_bdg": conf.prefix + "_b_control_lambda.bdg"},
param = {"description": conf.prefix + "_b",
"species": "hs",
"format": format,
"keep_dup": 1, "fdr": 0.01},
name = "broad peaks calling"))
macs2_on_merged_broad.param["treat_opt"] = " -t " + macs2_on_merged_broad.input["treat"]
macs2_on_merged_broad.allow_fail = True
macs2_on_merged_broad.allow_dangling = True
if len(conf.control_targets) >= 1:
macs2_on_merged_broad.input["control"] = conf.prefix + "_control.bam"
macs2_on_merged_broad.param["control_opt"] = "-c " + macs2_on_merged_broad.input["control"]
else:
macs2_on_merged_broad.param["control_opt"] = ""
macs2_on_merged_broad.update(param=conf.items("macs2"))
sort = attach_back(workflow,
ShellCommand(
"""{tool} -r -g -k 9 {input[peaks]} > {output[p]}
""",
tool = "sort",
input = {"peaks": conf.prefix + "_b_peaks.broadPeak"},
output = {"p": conf.prefix + "_b_sort_peaks.broadPeak"},
name = "sort peaks files"))
sort.allow_dangling = True
sort.allow_fail = True
# For bedGraphToBigwiggle bugs, we need to remove coordinates over-border coordinates
# As _control_lambda.bdg always exist. There are no need to check whether there are control samples.
if conf.get("macs2", "type").lower() in ["both", "broad"]:
cont_bdg = conf.prefix + "_b_control_lambda.bdg"
treat_bdg = conf.prefix + "_b_treat_pileup.bdg"
if conf.get("macs2", "type").lower() in ["both", "narrow"]:
cont_bdg = conf.prefix + "_control_lambda.bdg"
treat_bdg = conf.prefix + "_treat_pileup.bdg"
import os
bdg_trim_control = attach_back(workflow,
ShellCommand(
'{tool} intersect -a {input[bdg]} -b {param[chrom_bed]} -wa -f 1.00 > {output}',
tool="bedtools",
input={"bdg": cont_bdg},
param = {"chrom_bed": os.path.join(conf.target_dir, "chrom.bed")},
output=cont_bdg+".tmp",
name="bedGraph filtering control"))
bdg_trim_control.fail = True
bdg_trim_control.allow_dangling = True
bdg_trim_treat = bdg_trim_control.clone
bdg_trim_treat.input["bdg"] = treat_bdg
bdg_trim_treat.output = treat_bdg + ".tmp"
bdg_trim_treat.fail = True
bdg_trim_treat.allow_dangling = True
attach_back(workflow, bdg_trim_treat)
bdg2bw_treat = attach_back(workflow,
ShellCommand(
"{tool} {input[bdg]} {input[chrom_len]} {output[bw]}",
tool="bedGraphToBigWig",
input={"bdg": cont_bdg+".tmp",
"chrom_len": conf.get_path(conf.get("basics", "species"), "chrom_len")},
output={"bw": conf.prefix + "_control.bw"},
name="bdg_to_bw control"))
## in case broad peaks failed
bdg2bw_treat.allow_fail = True
bdg2bw_treat.allow_dangling = True
# prototype used here to do the similar thing on treatment file
bdg2bw_control = bdg2bw_treat.clone
bdg2bw_control.input["bdg"] = treat_bdg+".tmp"
bdg2bw_control.output["bw"] = conf.prefix + "_treat.bw"
## in case broad peaks failed
bdg2bw_control.allow_fail = True
bdg2bw_control.allow_dangling = True
attach_back(workflow, bdg2bw_control)
stat_macs2(workflow, conf)
def test_invoke_non_exist_output(self):
non_exist_output_cmd = ShellCommand("echo tempfile3", output="tempfile3")
self.assertFalse(non_exist_output_cmd.invoke())
def test_invoke_dangling_tool(self):
dangling_tool_command = ShellCommand("{tool} fun", tool="wolfyp")
self.assertFalse(dangling_tool_command.invoke())
