def runBiSeq_liver(infiles, outfile):
job_options = "-l mem_free=10G -pe dedicated 10"
RRBS.runBiSeq(infiles,
outfile,
"Liver",
submit=True,
job_options=job_options)
def runM3D(infile, outfile, root, design):
job_options = "-l mem_free=4G -pe dedicated 1"
groups = [x for x in itertools.combinations(EXPERIMENTS, 2)]
# **code repeated - refactor**
for pair in groups:
pair = [re.sub("-agg", "", str(x)) for x in pair]
pair1, pair2 = pair
pair1_split = pair1.split("-")
pair2_split = pair2.split("-")
# only want pairs with one difference
# e.g treatment or tissue but not both
if not (pair1_split[0] != pair2_split[0]
and pair1_split[1] != pair2_split[1]):
outfile = ("%(root)s%(pair1)s_vs_%(pair2)s.tsv" % locals())
if pair1_split[0] != pair2_split[0]:
groups = [pair1_split[0], pair2_split[0]]
elif pair1_split[1] != pair2_split[1]:
groups = [pair1_split[1], pair2_split[1]]
else:
E.error(
"This pair does not contain any comparisons: %(pair)s" %
locals())
RRBS.calculateM3DStat(infile,
outfile,
design,
pair=pair,
groups=groups,
submit=True,
job_options=job_options)
def runBiSeq_germline(infiles, outfile):
job_options = "-l mem_free=10G -pe dedicated 10"
RRBS.runBiSeq(infiles,
outfile,
"Germline",
submit=True,
job_options=job_options)
def clusterSpikeInsPowerAnalysis(infiles, outfile):
job_options = "-l mem_free=23G"
RRBS.spikeInClustersAnalysis(infiles,
outfile,
submit=True,
job_options=job_options)
def runM3DSpikeClusters(infiles, outfile):
job_options = "-l mem_free=4G -pe dedicated 1"
infile, design = infiles
RRBS.calculateM3DStat(infile,
outfile,
design,
submit=True,
job_options=job_options)
def findCpGs(outfile):
genome_infile = PARAMS["methylation_summary_genome_fasta"]
job_options = "-l mem_free=2G"
RRBS.fasta2CpG(genome_infile,
outfile,
submit=True,
job_options=job_options)
def addTreatmentMeans(infile, outfile):
job_options = "-l mem_free=48G"
RRBS.addTreatmentMean(infile,
outfile,
submit=True,
job_options=job_options)
def extractRepeatCpGs(outfile):
'''extract repeats sequences and identify CpG locations'''
RRBS.findRepeatCpGs(outfile,
PARAMS["methylation_summary_genome_fasta"],
PARAMS["annotation_repeats_gff"],
submit=True,
job_memory="4G")
def categorisePromoterCpGs(outfile):
'''extract promoter sequences and categorise them by CpG density'''
RRBS.categorisePromoterCpGs(outfile,
PARAMS["methylation_summary_genome_fasta"],
PARAMS['annotation_database'],
submit=True,
job_memory="4G")
def subsetCpGsToCovered(infile, outfile):
job_options = "-l mem_free=48G"
RRBS.subsetToCovered(infile,
outfile,
cov_threshold=10,
submit=True,
job_options=job_options)
def calculateM3DSpikeClustersPvalue(infiles, outfile):
job_options = "-l mem_free=4G -pe dedicated 1"
design = infiles[-1]
infiles = infiles[:-1]
RRBS.calculateM3DSpikepvalue(infiles,
outfile,
design,
submit=True,
job_options=job_options)
P.touch(outfile)
def splitClustersDataframe(infile, outfiles):
outprefix = "subframes.dir/cluster_subframe_"
suffix = ".tsv"
job_options = "-l mem_free=8G -pe dedicated 1"
RRBS.splitDataframeClusters(infile,
outprefix,
suffix,
submit=True,
job_options=job_options)
def extractDMRCpGs(outfile):
'''extract sequences for Highly conserved non-coding element and
identify CpG locations'''
RRBS.findCpGsFromBed(outfile,
PARAMS["methylation_summary_genome_fasta"],
PARAMS["annotation_dmr"],
"DMR",
both_strands=True,
submit=True,
job_memory="4G")
def plotReadBias(infile, outfile):
job_options = "-l mem_free=1G"
m_bias_infile = P.snip(infile, ".bismark.cov") + ".M-bias.txt"
print(m_bias_infile)
RRBS.plotReadBias(m_bias_infile,
outfile,
submit=True,
job_options=job_options)
def mergeCoverage(infiles, outfile):
cpgs_infile = infiles[-1]
coverage_infiles = infiles[:-1]
# this should be replaced with a non-pandas based solution
# very memory intensive! - find out why and re-code
job_options = "-l mem_free=48G"
job_threads = 2
RRBS.mergeAndDrop(cpgs_infile,
coverage_infiles,
outfile,
submit=True,
job_options=job_options)
def calculateM3DClustersPvalue(infiles, outfile, pair1, pair2):
job_options = "-l mem_free=4G -pe dedicated 1"
infiles = infiles[:-1]
print("pair1: %s" % pair1)
print("pair2: %s" % pair2)
pair = [pair1, pair2]
print(infiles, outfile, pair)
RRBS.calculateM3Dpvalue(infiles,
outfile,
pair,
submit=True,
job_options=job_options)
def mergeCpGAnnotations(infiles, outfile):
'''merge together the CpG annotations for plotting'''
meth_inf, prom_inf, repeat_inf, hcne_inf, dmr_inf = infiles
RRBS.mergeCpGAnnotations(meth_inf,
prom_inf,
repeat_inf,
hcne_inf,
dmr_inf,
outfile,
submit=True,
job_memory="4G")
def addCpGIs(infiles, outfile):
infile, CpGI = infiles
# TS: still memory intensive even after supplying data types
# for all columns!
# this should be replaced with a non-pandas based solution
job_memory = "40G"
job_threads = 1
RRBS.pandasMerge(infile,
CpGI,
outfile,
merge_type="left",
left=['contig', 'position'],
right=['contig', 'position'],
submit=True,
job_memory=job_memory)
def plotCoverage(infile, outfiles):
RRBS.plotCoverage(infile, outfiles, submit=True, job_memory="6G")
def makeSummaryPlots(infile, outfile):
job_options = "-l mem_free=48G"
RRBS.summaryPlots(infile, outfile, submit=True, job_options=job_options)
P.touch(outfile)
def plotMethylationFrequency(infile, outfile):
RRBS.plotMethFrequency(infile, outfile, job_memory="2G", submit=True)
def plotCpGAnnotations(infile, outfiles):
''' make histogram and boxplots for the CpGs facetted per annotation'''
outfile_hist, outfile_box = outfiles
RRBS.plotCpGAnnotations(infile, outfile_hist, outfile_box)
def calculateCoverage(infile, outfile):
RRBS.calculateCoverage(infile, outfile, submit=True, job_memory="2G")
def summariseM3D(infile, outfile):
''' summarise the number of cluster passing threshold'''
# adjusted p-value threshold
threshold = 0.05
print(infile, outfile, threshold)
RRBS.summariseM3D(infile, outfile, threshold, submit=True)