def plot_manhattan_plot(dataframe,
section,
filename_fasta,
map_key2label={},
**kwargs):
plotter = ManhattanPlot(genome_size_file=filename_fasta)
ax = plotter(dataframe, **kwargs)
plt.savefig(E.get_output_file(section))
plt.close()
def plot_mutation_profile_bar_plot(dataframe,
section,
map_key2label={},
**kwargs):
for key, dataframe in dataframe.groupby(by="sample"):
if key == "unique":
continue
if dataframe.empty:
E.warn("no data for {}".format(key))
continue
ax = MutationProfileBarPlot()(dataframe)
label = map_key2label.get(key, key)
plt.savefig(E.get_output_file("-".join((section, label))))
plt.close()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if not argv:
argv = sys.argv
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id",
usage=globals()["__doc__"])
parser.add_option("-u",
"--ucsc-genome",
dest="ucsc_genome",
type="string",
help="UCSC genome identifier [default=%default].")
parser.add_option("-g",
"--genome-file",
dest="genome_file",
type="string",
help="filename with genome [default=%default].")
parser.add_option("--extend",
dest="extension",
type="int",
help="extend tags by this number of bases "
"[default=%default].")
parser.add_option("--shift-size",
dest="shift",
type="int",
help="shift tags by this number of bases "
"[default=%default].")
parser.add_option("--window-size",
dest="window_size",
type="int",
help="window size to be used in the analysis"
"[default=%default].")
parser.add_option("--saturation-iterations",
dest="saturation_iterations",
type="int",
help="iterations for saturation analysis "
"[default=%default].")
parser.add_option("-t",
"--toolset",
dest="toolset",
type="choice",
action="append",
choices=("saturation", "coverage", "enrichment", "dmr",
"rms", "rpm", "all", "convert"),
help="actions to perform [default=%default].")
parser.add_option("-w",
"--bigwig-file",
dest="bigwig",
action="store_true",
help="store wig files as bigwig files - requires a "
"genome file [default=%default]")
parser.add_option("--treatment",
dest="treatment_files",
type="string",
action="append",
help="BAM files for treatment. At least one is required "
"[%default]")
parser.add_option("--control",
dest="control_files",
type="string",
action="append",
help="BAM files for control for differential "
"methylation analysis. Optional [%default].")
parser.add_option("--input",
dest="input_files",
type="string",
action="append",
help="BAM files for input correction. "
"Optional [%default].")
parser.add_option("--is-not-medip",
dest="is_medip",
action="store_false",
help="data is not MeDIP data and is not expected "
"to fit the calibration model. No CpG "
"density normalized rms data is computed"
"[default=%default].")
parser.add_option("--output-rdata",
dest="output_rdata",
action="store_true",
help="in dmr analysis, write R session to file. "
"The file name "
"is given by --ouptut-filename-pattern [%default].")
parser.add_option("--rdata-file",
dest="input_rdata",
type="string",
help="in dmr analysis, read saved R session from "
"file. This can be used to apply different "
"filters [%default]")
parser.add_option("--fdr-threshold",
dest="fdr_threshold",
type="float",
help="FDR threshold to apply for selecting DMR "
"[default=%default].")
parser.add_option("--fdr-method",
dest="fdr_method",
type="choice",
choices=("bonferroni", "BH", "holm", "hochberg",
"hommel", "BY", "fdr", "none"),
help="FDR method to apply for selecting DMR "
"[default=%default].")
parser.add_option("--bwa",
dest="bwa",
action="store_true",
help="alignment generated with bwa"
"[default=%default].")
parser.add_option("--unique",
dest="unique",
type="float",
help="Threshold p-value to determine which read pile\
ups are the result of PCR overamplification"
"[default=%default].")
parser.add_option("--chroms",
dest="chroms",
type="str",
help="Comma delimited list of chromosomes to include"
"[default=%default].")
parser.set_defaults(input_format="bam",
ucsc_genome="Hsapiens.UCSC.hg19",
genome_file=None,
extend=0,
shift=0,
window_size=300,
saturation_iterations=10,
toolset=[],
bigwig=False,
treatment_files=[],
control_files=[],
input_files=[],
output_rdata=False,
input_rdata=None,
is_medip=True,
fdr_threshold=0.1,
fdr_method="BH",
bwa=False,
unique=0.001,
chroms=None)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
if "convert" in options.toolset:
results = []
for line in CSV.DictReader(options.stdin, dialect="excel-tab"):
if line['edgeR.p.value'] == "NA":
continue
# assumes only a single treatment/control
treatment_name = options.treatment_files[0]
control_name = options.control_files[0]
status = "OK"
try:
results.append(
Expression.GeneExpressionResult._make((
"%s:%i-%i" %
(line['chr'], int(line['start']), int(line['stop'])),
treatment_name,
float(line['MSets1.rpkm.mean']),
0,
control_name,
float(line['MSets2.rpkm.mean']),
0,
float(line['edgeR.p.value']),
float(line['edgeR.adj.p.value']),
float(line['edgeR.logFC']),
math.pow(2.0, float(line['edgeR.logFC'])),
float(line['edgeR.logFC']), # no transform
["0", "1"][float(line['edgeR.adj.p.value']) <
options.fdr_threshold],
status)))
except ValueError as msg:
raise ValueError("parsing error %s in line: %s" % (msg, line))
Expression.writeExpressionResults(options.stdout, results)
return
if len(options.treatment_files) < 1:
raise ValueError("please specify a filename with sample data")
if options.bigwig and not options.genome_file:
raise ValueError("please provide a genome file when outputting bigwig")
if options.genome_file:
fasta = IndexedFasta.IndexedFasta(options.genome_file)
contig_sizes = fasta.getContigSizes()
if len(options.toolset) == 0:
options.toolset = ["all"]
do_all = "all" in options.toolset
if options.chroms is None:
chrstring = ""
else:
chroms = options.chroms.split(",")
chrstring = ' chr.select=c(\"%s\"), ' % '\",\"'.join(chroms)
# load MEDIPS
R.library('MEDIPS')
genome_file = 'BSgenome.%s' % options.ucsc_genome
R.library(genome_file)
window_size = options.window_size
extend = options.extend
shift = options.shift
saturation_iterations = options.saturation_iterations
uniq = float(options.unique)
if options.bwa is True:
BWA = "TRUE"
else:
BWA = "FALSE"
if "saturation" in options.toolset or do_all:
E.info("saturation analysis")
for fn in options.treatment_files + options.control_files:
paired = isPaired(fn)
R('''sr = MEDIPS.saturation(
file='%(fn)s',
BSgenome='%(genome_file)s',
shift=%(shift)i,
extend=%(extend)i,
window_size=%(window_size)i,
uniq=%(uniq)s,
nit = %(saturation_iterations)i,
paired = %(paired)s,
bwa = %(BWA)s,
%(chrstring)s
nrit = 1)''' % locals())
R.png(E.get_output_file("%s_saturation.png" % fn))
R('''MEDIPS.plotSaturation(sr)''')
R('''dev.off()''')
R('''write.table(sr$estimation, file ='%s', sep='\t')''' %
E.get_output_file("%s_saturation_estimation.tsv" % fn))
outfile = IOTools.open_file(
E.get_output_file("%s_saturation.tsv" % fn), "w")
outfile.write("category\tvalues\n")
outfile.write("estimated_correlation\t%s\n" %
",".join(["%f" % x for x in R('''sr$maxEstCor''')]))
outfile.write("true_correlation\t%s\n" %
",".join(["%f" % x for x in R('''sr$maxTruCor''')]))
outfile.write("nreads\t%s\n" %
",".join(["%i" % x
for x in R('''sr$numberReads''')]))
outfile.close()
if "coverage" in options.toolset or do_all:
E.info("CpG coverage analysis")
for fn in options.treatment_files + options.control_files:
paired = isPaired(fn)
R('''cr = MEDIPS.seqCoverage(
file='%(fn)s',
BSgenome='%(genome_file)s',
pattern='CG',
shift=%(shift)i,
extend=%(extend)i,
paired=%(paired)s,
bwa=%(BWA)s,
%(chrstring)s
uniq=%(uniq)s)''' % locals())
R.png(E.get_output_file("%s_cpg_coverage_pie.png" % fn))
R('''MEDIPS.plotSeqCoverage(seqCoverageObj=cr,
type = "pie", cov.level = c(0, 1, 2, 3, 4, 5))''')
R('''dev.off()''')
R.png(E.get_output_file("%s_cpg_coverage_hist.png" % fn))
R('''MEDIPS.plotSeqCoverage(seqCoverageObj=cr,
type = "hist", t=15)''')
R('''dev.off()''')
# note: this file is large
R('''write.table(cr$cov.res, file=gzfile('%s','w'),
sep='\t')''' %
E.get_output_file("%s_saturation_coveredpos.tsv.gz" % fn))
if 'enrichment' in options.toolset or do_all:
E.info("CpG enrichment analysis")
outfile = IOTools.open_file(E.get_output_file("enrichment.tsv.gz"),
"w")
slotnames = (("regions.CG", "regions_CG",
"%i"), ("regions.C", "regions_C",
"%s"), ("regions.G", "regions_G", "%f"),
("regions.relH", "regions_relH",
"%i"), ("regions.GoGe", "regions_GoGe",
"%i"), ("genome.CG", "genome_CG",
"%s"), ("genome.C", "genome_C", "%s"),
("genome.G", "genome_G", "%i"), ("genome.relH",
"genome_relH", "%i"),
("enrichment.score.relH", "enrichment_relH", "%s"),
("enrichment.score.GoGe", "enrichment_GoGe", "%s"))
outfile.write("\t".join(['sample'] + [x[1] for x in slotnames]) + "\n")
for fn in options.treatment_files + options.control_files:
paired = isPaired(fn)
R('''ce = MEDIPS.CpGenrich(
file='%(fn)s',
BSgenome='%(genome_file)s',
shift=%(shift)i,
extend=%(extend)i,
paired=%(paired)s,
bwa=%(BWA)s,
%(chrstring)s
uniq=%(uniq)s)''' % locals())
outfile.write("%s" % fn)
for slotname, label, pattern in slotnames:
value = tuple(R('''ce$%s''' % slotname))
if len(value) == 0:
value = ""
outfile.write("\t%s" % pattern % value[0])
outfile.write("\n")
outfile.close()
if options.input_rdata:
E.info("reading R session info from '%s'" % options.input_rdata)
R('''load('%s')''' % options.input_rdata)
else:
if "dmr" in options.toolset or "correlation" in options.toolset \
or do_all:
# build four sets
for x, fn in enumerate(options.treatment_files):
paired = isPaired(fn)
E.info("loading '%s'" % fn)
R('''treatment_R%(x)i = MEDIPS.createSet(
file='%(fn)s',
BSgenome='%(genome_file)s',
shift=%(shift)i,
extend=%(extend)i,
window_size=%(window_size)i,
paired=%(paired)s,
bwa=%(BWA)s,
%(chrstring)s
uniq=%(uniq)s)''' % locals())
R('''treatment_set = c(%s)''' % ",".join([
"treatment_R%i" % x
for x in range(len(options.treatment_files))
]))
if options.control_files:
for x, fn in enumerate(options.control_files):
paired = isPaired(fn)
E.info("loading '%s'" % fn)
R('''control_R%(x)i = MEDIPS.createSet(
file='%(fn)s',
BSgenome='%(genome_file)s',
shift=%(shift)i,
extend=%(extend)i,
window_size=%(window_size)i,
paired=%(paired)s,
bwa=%(BWA)s,
%(chrstring)s
uniq=%(uniq)s)''' % locals())
R('''control_set = c(%s)''' % ",".join([
"control_R%i" % x
for x in range(len(options.control_files))
]))
# build coupling vector
R('''CS = MEDIPS.couplingVector(pattern="CG",
refObj = treatment_set[[1]])''')
if "correlation" in options.toolset or do_all:
R('''cor.matrix = MEDIPS.correlation(
c(treatment_set, control_set))''')
R('''write.table(cor.matrix,
file='%s',
sep="\t")''' % E.get_output_file("correlation"))
if "dmr" in options.toolset or do_all:
# Data that does not fit the model causes
# "Error in 1:max_signal_index : argument of length 0"
# The advice is to set MeDIP=FALSE
# See: http://comments.gmane.org/
# gmane.science.biology.informatics.conductor/52319
if options.is_medip:
medip = "TRUE"
else:
medip = "FALSE"
fdr_method = options.fdr_method
E.info("applying test for differential methylation")
R('''meth = MEDIPS.meth(
MSet1 = treatment_set,
MSet2 = control_set,
CSet = CS,
ISet1 = NULL,
ISet2 = NULL,
p.adj = "%(fdr_method)s",
diff.method = "edgeR",
MeDIP = %(medip)s,
CNV = F,
minRowSum = 1)''' % locals())
# Note: several Gb in size
# Output full methylation data table
R('''write.table(meth,
file=gzfile('%s', 'w'),
sep="\t",
row.names=F,
quote=F)''' % E.get_output_file("data.tsv.gz"))
# save R session
if options.output_rdata:
R('''save.image(file='%s', safe=FALSE)''' %
E.get_output_file("session.RData"))
# DMR analysis - test for windows and output
if "dmr" in options.toolset:
E.info("selecting differentially methylated windows")
# test windows for differential methylation
fdr_threshold = options.fdr_threshold
R('''tested = MEDIPS.selectSig(meth,
adj=T,
ratio=NULL,
p.value=%(fdr_threshold)f,
bg.counts=NULL,
CNV=F)''' % locals())
R('''write.table(tested,
file=gzfile('%s', 'w'),
sep="\t",
quote=F)''' % E.get_output_file("significant_windows.gz"))
# select gain and merge adjacent windows
try:
R('''gain = tested[which(tested[, grep("logFC", colnames(tested))] > 0),];
gain_merged = MEDIPS.mergeFrames(frames=gain, distance=1)''')
E.info('gain output: %s, merged: %s' %
(str(R('''dim(gain)''')), str(R('''dim(gain_merged)'''))))
R('''of=gzfile('%s', 'w');
write.table(gain_merged,
file=of,
sep="\t",
quote=F,
row.names=FALSE,
col.names=FALSE); close(of)''' % E.get_output_file("gain.bed.gz"))
except rpy2.rinterface.RRuntimeError as msg:
E.warn("could not compute gain windows: msg=%s" % msg)
# select loss and merge adjacent windows
try:
R('''loss = tested[which(tested[, grep("logFC", colnames(tested))] < 0),];
loss_merged = MEDIPS.mergeFrames(frames=loss, distance=1)''')
E.info('loss output: %s, merged: %s' %
(str(R('''dim(loss)''')), str(R('''dim(loss_merged)'''))))
R('''of=gzfile('%s', 'w');
write.table(loss_merged,
file=of,
sep="\t",
quote=F,
row.names=F,
col.names=F); close(of)''' % E.get_output_file("loss.bed.gz"))
except rpy2.rinterface.RRuntimeError as msg:
E.warn("could not compute loss windows: msg=%s" % msg)
# if "rpm" in options.toolset or do_all:
# outputfile = E.get_output_file("rpm.wig")
# R('''MEDIPS.exportWIG(file = '%(outputfile)s',
# data = CONTROL.SET, raw = T, descr = "rpm")''' %
# locals())
# if options.bigwig:
# bigwig(outputfile, contig_sizes)
# else:
# compress(outputfile)
# if "rms" in options.toolset or do_all:
# outputfile = E.get_output_file("rms.wig")
# R('''MEDIPS.exportWIG(file = '%(outputfile)s',
# data = CONTROL.SET, raw = F, descr = "rms")''' %
# locals())
# if options.bigwig:
# bigwig(outputfile, contig_sizes)
# else:
# compress(outputfile)
# write footer and output benchmark information.
E.stop()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if not argv:
argv = sys.argv
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id$",
usage=globals()["__doc__"])
parser.add_option(
"-e",
"--input-bed-file",
dest="input_bed_file",
type="string",
help="input file with intervals. Tab-delimited file of intervals "
"in bed format to restrict analysis to. [%default]")
parser.add_option(
"-m",
"--merge-intervals",
dest="merge_intervals",
action="store_true",
help="merge intervals in bed file. Useful if you have a site bed-file "
"[%default]")
parser.add_option("-f",
"--reference-fasta-file",
dest="reference_fasta_file",
help="reference genomic sequence in fasta format. "
"[%default]")
parser.add_option(
"-c",
"--barcode-fasta-file",
dest="barcode_fasta_file",
help="barcode sequence in fasta format. Variable positions "
"should be marked by N "
"[%default]")
parser.set_defaults(
reference_fasta_file=None,
barcode_fasta_file=None,
merge_intervals=False,
input_bed_file=None,
anchor=5,
)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
if options.stdin != sys.stdin:
bamfile = options.stdin.name
elif args:
if len(args) > 1:
raise ValueError("multiple bam files provided in arguments")
bamfile = args[0]
else:
bamfile = "-"
if options.barcode_fasta_file:
with pysam.FastxFile(options.barcode_fasta_file) as inf:
barcode_sequence = next(inf).sequence
else:
barcode_sequence = None
if not os.path.exists(options.reference_fasta_file):
raise OSError("reference fasta file {} does not exist".format(
options.reference_fasta_file))
if not os.path.exists(options.input_bed_file):
raise OSError("input bed file {} does not exist".format(
options.input_bed_file))
bed_in = pysam.TabixFile(options.input_bed_file)
pysam_in = pysam.AlignmentFile(bamfile)
anchor = options.anchor
for region_idx, vals in enumerate(
iterate_bed(bed_in, options.merge_intervals)):
if region_idx > 0:
raise NotImplementedError(
"output for multiple regions not yet implemented")
contig, region_start, region_end = vals
upstream_anchors, downstream_anchors = [], []
counter = E.Counter()
unaligned_fn = E.get_output_file(
"unaligned_{}.fasta".format(region_idx))
with IOTools.open_file(unaligned_fn, "w") as outf:
for read in pysam_in.fetch(contig, region_start, region_end):
counter.overlapping_reads += 1
try:
pairs = read.get_aligned_pairs(with_seq=True)
except ValueError:
counter.no_md_tag += 1
continue
map_ref2read_pos = dict(
(x[1], x[0]) for x in pairs if x[0] is not None)
map_ref2ref_base = dict(
(x[1], x[2]) for x in pairs if x[0] is not None)
upstream_anchor = "".join(
map_ref2ref_base.get(x, "")
for x in range(region_start - anchor, region_start))
downstream_anchor = "".join(
map_ref2ref_base.get(x, "")
for x in range(region_end, region_end + anchor))
# check if at least one anchor is aligned
upstream_matches = sum([x.isupper() for x in upstream_anchor])
downstream_matches = sum(
[x.isupper() for x in downstream_anchor])
if upstream_matches < anchor and downstream_matches < anchor:
counter.no_anchor += 1
continue
seq = read.query_alignment_sequence
# collect full length anchors
upstream_anchor_start, upstream_anchor_end = region_start - anchor, region_start
downstream_anchor_start, downstream_anchor_end = region_end, region_end + anchor
if upstream_anchor_start in map_ref2read_pos and upstream_anchor_end in map_ref2read_pos:
upstream_anchors.append(
seq[map_ref2read_pos[upstream_anchor_start]:
map_ref2read_pos[upstream_anchor_end]])
if downstream_anchor_start in map_ref2read_pos and downstream_anchor_end in map_ref2read_pos:
downstream_anchors.append(
seq[map_ref2read_pos[downstream_anchor_start]:
map_ref2read_pos[downstream_anchor_end]])
# get region to align
read_start = min(
(map_ref2read_pos.get(x, len(seq))
for x in range(region_start - anchor, region_start)))
if read_start == len(seq):
read_start = 0
read_end = max(
(map_ref2read_pos.get(x, 0) + 1
for x in range(region_end, region_end + anchor)))
if read_end == 1:
read_end = len(seq)
counter.collected_reads += 1
outf.write(">{}/{}-{}\n{}\n".format(read.query_name,
read_start, read_end,
seq[read_start:read_end]))
counter.downstream_anchors = len(downstream_anchors)
counter.upstream_anchors = len(upstream_anchors)
E.info(counter)
if counter.overlapping_reads == 0:
E.warn("no sequences overlapping region")
continue
if counter.downstream_anchors == 0 or counter.upstream_anchors == 0:
E.warn("at least one anchor undefined")
continue
if counter.collected_reads == 1:
E.warn("only single sequence, multiple aligment skipped")
with IOTools.open_file(unaligned_fn) as inf:
stdout = inf.read()
else:
# G-INS-i -> global alignment algorithm
E.info("starting mafft multiple alignment")
stdout = E.run(
"mafft --globalpair --maxiterate 100 --quiet --op 2 --ep 0.5 {}"
.format(unaligned_fn),
return_stdout=True)
aligned_fn = E.get_output_file("aligned_{}.fasta".format(region_idx))
with IOTools.open_file(aligned_fn, "w") as outf:
outf.write(stdout)
mali = stdout.splitlines()
identifiers = [mali[x] for x in range(0, len(mali), 2)]
sequences = [mali[x].upper() for x in range(1, len(mali), 2)]
consensus = get_consensus(sequences)
E.info("after alignment: consensus={}".format(consensus))
# gap filtering -> remove highly gappy columns
consensus = get_consensus(sequences, min_gap_proportion=0.9)
E.info("after anchor trimming: consensus={}".format(consensus))
take = [idx for idx, x in enumerate(consensus) if x != "-"]
sequences = ["".join([s[x] for x in take]) for s in sequences]
consensus = get_consensus(sequences, min_gap_proportion=0.9)
E.info("after gap filtering: consensus={}".format(consensus))
# get anchor consensus and chop it off
consensus = get_consensus(sequences, ignore_gaps=True)
upstream_anchor = get_anchor_consensus(upstream_anchors)
downstream_anchor = get_anchor_consensus(downstream_anchors)
upstream_anchor_start = consensus.find(upstream_anchor)
downstream_anchor_start = consensus.rfind(downstream_anchor)
E.info(
"anchor consensus (no gaps)={}, upstream={}, downstream={}, upstream_idx={}, downstream_idx={}"
.format(consensus, upstream_anchor, downstream_anchor,
upstream_anchor_start, downstream_anchor_start))
if upstream_anchor_start < 0 or downstream_anchor_start < 0:
E.warn("can't locate anchor, no output produced")
continue
upstream_anchor_end = upstream_anchor_start + len(upstream_anchor)
if upstream_anchor_end >= downstream_anchor_start:
E.warn("anchor not in correct order, no output produced")
continue
sequences = [
x[upstream_anchor_end:downstream_anchor_start] for x in sequences
]
consensus = get_consensus(sequences)
E.info("after anchor trimming: consensus={}".format(consensus))
truncated_fn = E.get_output_file(
"aligned_truncated_{}.fasta".format(region_idx))
with IOTools.open_file(truncated_fn, "w") as outf:
outf.write("\n".join("{}\n{}\n".format(x, y)
for x, y in zip(identifiers, sequences)))
positions = list(zip(*sequences))
bases = ["A", "C", "G", "T"]
df = pandas.DataFrame([collections.Counter(x)
for x in positions]).fillna(0)
for missing_base in [x for x in bases if x not in df.columns]:
df[missing_base] = 0
df["gapped_depth"] = df.sum(axis=1)
df["depth"] = df[bases].sum(axis=1)
df["consensus"] = df[bases].idxmax(axis=1)
df["consensus_counts"] = df.lookup(df.index, df.consensus)
df["consensus_support"] = df.consensus_counts / df.depth
df["offconsensus_counts"] = df.depth - df.consensus_counts
df.loc[df.consensus_counts == 0, "consensus"] = "N"
df["region_id"] = region_idx
# replace "gap" consensus positions with + character
alignment = global_align(re.sub("-", "+", consensus), barcode_sequence)
E.info("alignment: consensus {}".format(alignment[0]))
E.info("alignment: barcode {}".format(alignment[1]))
barcode_idx = 0
deleted_barcode_bases = []
rows = []
for c, b in zip(*alignment):
if c == "-":
deleted_barcode_bases.append(barcode_idx)
barcode_idx += 1
elif b == "N":
rows.append((barcode_idx, "variable"))
barcode_idx += 1
elif b == "-":
rows.append(("", "insertion"))
elif b == c:
rows.append((barcode_idx, "fixed-match"))
barcode_idx += 1
else:
rows.append((barcode_idx, "fixed-mismatch"))
barcode_idx += 1
alignment_df = pandas.DataFrame.from_records(
rows, columns=["barcode_pos", "barcode_class"])
assert len(alignment_df) == len(df)
df = pandas.concat([df, alignment_df], axis=1)
with E.open_output_file("pileup") as outf:
df.to_csv(outf, sep="\t", index=True, index_label="position")
observed_barcode_sequence = "".join(
df[df.barcode_class == "variable"].consensus)
headers = df.consensus_support.describe().index
eval_df = df.loc[df.barcode_class.isin(
("variable", "fixed-match", "fixed-mismatch")), ]
median_consensus_depth = eval_df.consensus_counts.median()
# zero stuff out if depth is low
if median_consensus_depth <= 2:
deleted_barcode_bases = []
outf = options.stdout
# modules to recover partial bar-codes
outf.write("\t".join(
map(str, [
"barcode", "ndeleted_barcode_bases", "deleted_barcode_bases"
] + ["support_{}".format(x)
for x in headers] + ["counts_{}".format(x) for x in headers] +
["offcounts_{}".format(x) for x in headers])) + "\n")
outf.write("\t".join(
map(str, [
observed_barcode_sequence,
len(deleted_barcode_bases), ",".join(
map(str, deleted_barcode_bases))
] + eval_df.consensus_support.describe().tolist() +
eval_df.consensus_counts.describe().tolist() +
eval_df.offconsensus_counts.describe().tolist())) + "\n")
E.stop()
def main(argv=sys.argv):
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id$",
usage=globals()["__doc__"])
parser.add_option("-s",
"--session",
dest="session",
type="string",
help="load session before creating plots "
"[%default]")
parser.add_option("-d",
"--snapshot-dir",
dest="snapshotdir",
type="string",
help="directory to save snapshots in [%default]")
parser.add_option("-f",
"--format",
dest="format",
type="choice",
choices=("png", "eps", "svg"),
help="output file format [%default]")
parser.add_option("-o",
"--host",
dest="host",
type="string",
help="host that IGV is running on [%default]")
parser.add_option("-p",
"--port",
dest="port",
type="int",
help="port that IGV listens at [%default]")
parser.add_option("-e",
"--extend",
dest="extend",
type="int",
help="extend each interval by a number of bases "
"[%default]")
parser.add_option("-x",
"--expand",
dest="expand",
type="float",
help="expand each region by a certain factor "
"[%default]")
parser.add_option("--session-only",
dest="session_only",
action="store_true",
help="plot session after opening, "
"ignore intervals "
"[%default]")
parser.add_option("-n",
"--name",
dest="name",
type="choice",
choices=("bed-name", "increment"),
help="name to use for snapshot "
"[%default]")
parser.set_defaults(
command="igv.sh",
host='127.0.0.1',
port=61111,
snapshotdir=os.getcwd(),
extend=0,
format="png",
expand=1.0,
session=None,
session_only=False,
keep_open=False,
name="bed-name",
)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
igv_process = None
if options.new_instance:
E.info("starting new IGV process")
igv_process = IGV.startIGV(command=options.command, port=options.port)
E.info("new IGV process started")
E.info("connection to process on %s:%s" % (options.host, options.port))
E.info("saving images in %s" % options.snapshotdir)
igv = IGV(host=options.host,
port=options.port,
snapshot_dir=os.path.abspath(options.snapshotdir))
if options.session:
E.info('loading session from %s' % options.session)
igv.load(options.session)
E.info('loaded session')
if options.session_only:
E.info('plotting session only ignoring any intervals')
fn = "%s.%s" % (os.path.basename(options.session), options.format)
E.info("writing snapshot to '%s'" %
os.path.join(options.snapshotdir, fn))
igv.save(fn)
else:
c = E.Counter()
for bed in pysam.tabix_iterator(options.stdin, parser=pysam.asBed()):
c.input += 1
# IGV can not deal with white-space in filenames
if options.name == "bed-name":
name = re.sub("\s", "_", bed.name)
elif options.name == "increment":
name = str(c.input)
E.info("going to %s:%i-%i for %s" %
(bed.contig, bed.start, bed.end, name))
start, end = bed.start, bed.end
extend = options.extend
if options.expand:
d = end - start
extend = max(extend, (options.expand * d - d) // 2)
start -= extend
end += extend
igv.go("%s:%i-%i" % (bed.contig, start, end))
fn = E.get_output_file("%s.%s" % (name, options.format))
E.info("writing snapshot to '%s'" % fn)
igv.save(fn)
c.snapshots += 1
E.info(c)
if igv_process is not None and not options.keep_open:
E.info('shutting down IGV')
igv_process.send_signal(signal.SIGKILL)
E.stop()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if not argv:
argv = sys.argv
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id$",
usage=globals()["__doc__"])
parser.add_option("-m",
"--method",
dest="methods",
type="choice",
action="append",
choices=(
"geneprofile",
"tssprofile",
"utrprofile",
"intervalprofile",
"midpointprofile",
"geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
"separateexonprofile",
"separateexonprofilewithintrons",
),
help='counters to use. Counters describe the '
'meta-gene structure to use. '
'Note using geneprofilewithintrons, or '
'geneprofileabsolutedistancefromthreeprimeend will '
'automatically turn on the --use-base-accuracy option'
'[%default].')
parser.add_option("-b",
"--bam-file",
"--bedfile",
"--bigwigfile",
dest="infiles",
metavar="BAM",
type="string",
action="append",
help="BAM/bed/bigwig files to use. Do not mix "
"different types [%default]")
parser.add_option("-c",
"--control-bam-file",
dest="controlfiles",
metavar="BAM",
type="string",
action="append",
help="control/input to use. Should be of the same "
"type as the bam/bed/bigwig file"
" [%default]")
parser.add_option("-g",
"--gtf-file",
dest="gtffile",
type="string",
metavar="GTF",
help="GTF file to use. "
"[%default]")
parser.add_option("--normalize-transcript",
dest="transcript_normalization",
type="choice",
choices=("none", "max", "sum", "total-max", "total-sum"),
help="normalization to apply on each transcript "
"profile before adding to meta-gene profile. "
"[%default]")
parser.add_option("--normalize-profile",
dest="profile_normalizations",
type="choice",
action="append",
choices=("all", "none", "area", "counts", "background"),
help="normalization to apply on meta-gene "
"profile normalization. "
"[%default]")
parser.add_option(
"-r",
"--reporter",
dest="reporter",
type="choice",
choices=("gene", "transcript"),
help="report results for genes or transcripts."
" When 'genes` is chosen, exons across all transcripts for"
" a gene are merged. When 'transcript' is chosen, counts are"
" computed for each transcript separately with each transcript"
" contributing equally to the meta-gene profile."
" [%default]")
parser.add_option("-i",
"--shift-size",
dest="shifts",
type="int",
action="append",
help="shift reads in :term:`bam` formatted file "
"before computing densities (ChIP-Seq). "
"[%default]")
parser.add_option("-a",
"--merge-pairs",
dest="merge_pairs",
action="store_true",
help="merge pairs in :term:`bam` formatted "
"file before computing "
"densities (ChIP-Seq). "
"[%default]")
parser.add_option("-u",
"--use-base-accuracy",
dest="base_accuracy",
action="store_true",
help="compute densities with base accuracy. The default "
"is to only use the start and end of the aligned region "
"(RNA-Seq) "
"[%default]")
parser.add_option("-e",
"--extend",
dest="extends",
type="int",
action="append",
help="extend reads in :term:`bam` formatted file "
"(ChIP-Seq). "
"[%default]")
parser.add_option("--resolution-upstream",
dest="resolution_upstream",
type="int",
help="resolution of upstream region in bp "
"[%default]")
parser.add_option("--resolution-downstream",
dest="resolution_downstream",
type="int",
help="resolution of downstream region in bp "
"[%default]")
parser.add_option("--resolution-upstream-utr",
dest="resolution_upstream_utr",
type="int",
help="resolution of upstream UTR region in bp "
"[%default]")
parser.add_option("--resolution-downstream-utr",
dest="resolution_downstream_utr",
type="int",
help="resolution of downstream UTR region in bp "
"[%default]")
parser.add_option("--resolution-cds",
dest="resolution_cds",
type="int",
help="resolution of cds region in bp "
"[%default]")
parser.add_option("--resolution-first-exon",
dest="resolution_first",
type="int",
help="resolution of first exon in gene, in bp"
"[%default]")
parser.add_option("--resolution-last-exon",
dest="resolution_last",
type="int",
help="resolution of last exon in gene, in bp"
"[%default]")
parser.add_option("--resolution-introns",
dest="resolution_introns",
type="int",
help="resolution of introns region in bp "
"[%default]")
parser.add_option("--resolution-exons-absolute-distance-topolya",
dest="resolution_exons_absolute_distance_topolya",
type="int",
help="resolution of exons absolute distance "
"topolya in bp "
"[%default]")
parser.add_option("--resolution-introns-absolute-distance-topolya",
dest="resolution_introns_absolute_distance_topolya",
type="int",
help="resolution of introns absolute distance "
"topolya in bp "
"[%default]")
parser.add_option("--extension-exons-absolute-distance-topolya",
dest="extension_exons_absolute_distance_topolya",
type="int",
help="extension for exons from the absolute "
"distance from the topolya in bp "
"[%default]")
parser.add_option(
"--extension-introns-absolute-distance-topolya",
dest="extension_introns_absolute_distance_topolya",
type="int",
help="extension for introns from the absolute distance from "
"the topolya in bp [%default]")
parser.add_option("--extension-upstream",
dest="extension_upstream",
type="int",
help="extension upstream from the first exon in bp"
"[%default]")
parser.add_option("--extension-downstream",
dest="extension_downstream",
type="int",
help="extension downstream from the last exon in bp"
"[%default]")
parser.add_option("--extension-inward",
dest="extension_inward",
type="int",
help="extension inward from a TSS start site in bp"
"[%default]")
parser.add_option("--extension-outward",
dest="extension_outward",
type="int",
help="extension outward from a TSS start site in bp"
"[%default]")
parser.add_option("--scale-flank-length",
dest="scale_flanks",
type="int",
help="scale flanks to (integer multiples of) gene length"
"[%default]")
parser.add_option(
"--control-factor",
dest="control_factor",
type="float",
help="factor for normalizing control and foreground data. "
"Computed from data if not set. "
"[%default]")
parser.add_option("--output-all-profiles",
dest="output_all_profiles",
action="store_true",
help="keep individual profiles for each "
"transcript and output. "
"[%default]")
parser.add_option("--counts-tsv-file",
dest="input_filename_counts",
type="string",
help="filename with count data for each transcript. "
"Use this instead "
"of recomputing the profile. Useful for plotting the "
"meta-gene profile "
"from previously computed counts "
"[%default]")
parser.add_option(
"--background-region-bins",
dest="background_region_bins",
type="int",
help="number of bins on either end of the profile "
"to be considered for background meta-gene normalization "
"[%default]")
parser.set_defaults(
remove_rna=False,
ignore_pairs=False,
force_output=False,
bin_size=10,
extends=[],
shifts=[],
sort=[],
reporter="transcript",
resolution_cds=1000,
resolution_introns=1000,
# 3kb is a good balance of seeing long enough 3 prime bias and not omit
# too many genes. Tim 31th Aug 2013
resolution_exons_absolute_distance_topolya=3000,
# introns is only for assess the noise level, thus do ont need a long
# region, a long region has the side effect of omit more genes. Tim
# 31th Aug 2013
resolution_introns_absolute_distance_topolya=500,
# extension can simply just be the same as resolution
extension_exons_absolute_distance_topolya=3000,
extension_introns_absolute_distance_topolya=500,
resolution_upstream_utr=1000,
resolution_downstream_utr=1000,
resolution_upstream=1000,
resolution_downstream=1000,
resolution_first=1000,
resolution_last=1000,
# mean length of transcripts: about 2.5 kb
extension_upstream=2500,
extension_downstream=2500,
extension_inward=3000,
extension_outward=3000,
plot=True,
methods=[],
infiles=[],
controlfiles=[],
gtffile=None,
profile_normalizations=[],
transcript_normalization=None,
scale_flanks=0,
merge_pairs=False,
min_insert_size=0,
max_insert_size=1000,
base_accuracy=False,
matrix_format="single",
control_factor=None,
output_all_profiles=False,
background_region_bins=10,
input_filename_counts=None,
)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
# Keep for backwards compatability
if len(args) == 2:
infile, gtf = args
options.infiles.append(infile)
options.gtffile = gtf
if not options.gtffile:
raise ValueError("no GTF file specified")
if options.gtffile == "-":
options.gtffile = options.stdin
else:
options.gtffile = IOTools.open_file(options.gtffile)
if len(options.infiles) == 0:
raise ValueError("no bam/wig/bed files specified")
for methodsRequiresBaseAccuracy in [
"geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
]:
# If you implemented any methods that you do not want the
# spliced out introns or exons appear to be covered by
# non-existent reads, it is better you let those methods imply
# --base-accurarcy by add them here.
if methodsRequiresBaseAccuracy in options.methods:
options.base_accuracy = True
if options.reporter == "gene":
gtf_iterator = GTF.flat_gene_iterator(GTF.iterator(options.gtffile))
elif options.reporter == "transcript":
gtf_iterator = GTF.transcript_iterator(GTF.iterator(options.gtffile))
# Select rangecounter based on file type
if len(options.infiles) > 0:
if options.infiles[0].endswith(".bam"):
bamfiles = [pysam.AlignmentFile(x, "rb") for x in options.infiles]
if options.controlfiles:
controlfiles = [
pysam.AlignmentFile(x, "rb") for x in options.controlfiles
]
else:
controlfiles = None
format = "bam"
if options.merge_pairs:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
shifts=options.shifts,
extends=options.extends,
merge_pairs=options.merge_pairs,
min_insert_size=options.min_insert_size,
max_insert_size=options.max_insert_size,
controfiles=controlfiles,
control_factor=options.control_factor)
elif options.shifts or options.extends:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
shifts=options.shifts,
extends=options.extends,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.base_accuracy:
range_counter = geneprofile.RangeCounterBAMBaseAccuracy(
bamfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
else:
range_counter = geneprofile.RangeCounterBAM(
bamfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.infiles[0].endswith(".bed.gz"):
bedfiles = [pysam.Tabixfile(x) for x in options.infiles]
if options.controlfiles:
controlfiles = [
pysam.Tabixfile(x) for x in options.controlfiles
]
else:
controlfiles = None
range_counter = geneprofile.RangeCounterBed(
bedfiles,
controlfiles=controlfiles,
control_factor=options.control_factor)
elif options.infiles[0].endswith(".bw"):
wigfiles = [BigWigFile(file=open(x)) for x in options.infiles]
range_counter = geneprofile.RangeCounterBigWig(wigfiles)
else:
raise NotImplementedError("can't determine file type for %s" %
str(options.infiles))
counters = []
for method in options.methods:
if method == "utrprofile":
counters.append(
geneprofile.UTRCounter(
range_counter,
options.resolution_upstream,
options.resolution_upstream_utr,
options.resolution_cds,
options.resolution_downstream_utr,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream,
))
elif method == "geneprofile":
counters.append(
geneprofile.GeneCounter(
range_counter, options.resolution_upstream,
options.resolution_cds, options.resolution_downstream,
options.extension_upstream, options.extension_downstream,
options.scale_flanks))
elif method == "geneprofilewithintrons":
counters.append(
geneprofile.GeneCounterWithIntrons(
range_counter, options.resolution_upstream,
options.resolution_cds, options.resolution_introns,
options.resolution_downstream, options.extension_upstream,
options.extension_downstream, options.scale_flanks))
elif method == "geneprofileabsolutedistancefromthreeprimeend":
# options.extension_exons_absolute_distance_tostartsite,
# options.extension_introns_absolute_distance_tostartsite,
# Tim 31th Aug 2013: a possible feature for future, if five prime
# bias is of your interest.
# (you need to create another class). It is not very difficult to
# derive from this class, but is not implemented yet
# This future feature is slightly different the TSS profile
# already implemented, because in this future feature introns are
# skipped,
counters.append(
geneprofile.GeneCounterAbsoluteDistanceFromThreePrimeEnd(
range_counter, options.resolution_upstream,
options.resolution_downstream,
options.resolution_exons_absolute_distance_topolya,
options.resolution_introns_absolute_distance_topolya,
options.extension_upstream, options.extension_downstream,
options.extension_exons_absolute_distance_topolya,
options.extension_introns_absolute_distance_topolya,
options.scale_flanks))
elif method == "tssprofile":
counters.append(
geneprofile.TSSCounter(range_counter,
options.extension_outward,
options.extension_inward))
elif method == "intervalprofile":
counters.append(
geneprofile.RegionCounter(range_counter,
options.resolution_upstream,
options.resolution_cds,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream))
elif method == "midpointprofile":
counters.append(
geneprofile.MidpointCounter(range_counter,
options.resolution_upstream,
options.resolution_downstream,
options.extension_upstream,
options.extension_downstream))
# add new method to split 1st and last exons out
# requires a representative transcript for reach gene
# gtf should be sorted gene-position
elif method == "separateexonprofile":
counters.append(
geneprofile.SeparateExonCounter(
range_counter, options.resolution_upstream,
options.resolution_first, options.resolution_last,
options.resolution_cds, options.resolution_downstream,
options.extension_upstream, options.extension_downstream))
elif method == "separateexonprofilewithintrons":
counters.append(
geneprofile.SeparateExonWithIntronCounter(
range_counter, options.resolution_upstream,
options.resolution_first, options.resolution_last,
options.resolution_cds, options.resolution_introns,
options.resolution_downstream, options.extension_upstream,
options.extension_downstream))
# set normalization
for c in counters:
c.setNormalization(options.transcript_normalization)
if options.output_all_profiles:
c.setOutputProfiles(
IOTools.open_file(
E.get_output_file(c.name) + ".profiles.tsv.gz", "w"))
if options.input_filename_counts:
# read counts from file
E.info("reading counts from %s" % options.input_filename_counts)
all_counts = pandas.read_csv(IOTools.open_file(
options.input_filename_counts),
sep='\t',
header=0,
index_col=0)
if len(counters) != 1:
raise NotImplementedError(
'counting from matrix only implemented for 1 counter.')
# build counter based on reference counter
counter = geneprofile.UnsegmentedCounter(counters[0])
counters = [counter]
geneprofile.countFromCounts(counters, all_counts)
else:
E.info("starting counting with %i counters" % len(counters))
feature_names = geneprofile.countFromGTF(counters, gtf_iterator)
# output matrices
if not options.profile_normalizations:
options.profile_normalizations.append("none")
elif "all" in options.profile_normalizations:
options.profile_normalizations = [
"none", "area", "counts", "background"
]
for method, counter in zip(options.methods, counters):
profiles = []
for norm in options.profile_normalizations:
# build matrix, apply normalization
profile = counter.getProfile(
normalize=norm,
background_region_bins=options.background_region_bins)
profiles.append(profile)
for x in range(1, len(profiles)):
assert profiles[0].shape == profiles[x].shape
# build a single matrix of all profiles for output
matrix = numpy.concatenate(profiles)
matrix.shape = len(profiles), len(profiles[0])
matrix = matrix.transpose()
with IOTools.open_file(
E.get_output_file(counter.name) + ".matrix.tsv.gz",
"w") as outfile:
outfile.write("bin\tregion\tregion_bin\t%s\n" %
"\t".join(options.profile_normalizations))
fields = []
bins = []
for field, nbins in zip(counter.fields, counter.nbins):
fields.extend([field] * nbins)
bins.extend(list(range(nbins)))
for row, cols in enumerate(zip(fields, bins, matrix)):
outfile.write("%i\t%s\t" %
(row, "\t".join([str(x) for x in cols[:-1]])))
outfile.write("%s\n" % ("\t".join([str(x) for x in cols[-1]])))
with IOTools.open_file(
E.get_output_file(counter.name) + ".lengths.tsv.gz",
"w") as outfile:
counter.writeLengthStats(outfile)
if options.output_all_profiles:
counter.closeOutputProfiles()
if options.plot:
import matplotlib
# avoid Tk or any X
matplotlib.use("Agg")
import matplotlib.pyplot as plt
for method, counter in zip(options.methods, counters):
if method in ("geneprofile", "geneprofilewithintrons",
"geneprofileabsolutedistancefromthreeprimeend",
"utrprofile", "intervalprofile",
"separateexonprofile",
"separateexonprofilewithintrons"):
plt.figure()
plt.subplots_adjust(wspace=0.05)
max_scale = max([max(x) for x in counter.aggregate_counts])
for x, counts in enumerate(counter.aggregate_counts):
plt.subplot(6, 1, x + 1)
plt.plot(list(range(len(counts))), counts)
plt.title(counter.fields[x])
plt.ylim(0, max_scale)
figname = counter.name + ".full"
fn = E.get_output_file(figname) + ".png"
plt.savefig(os.path.expanduser(fn))
plt.figure()
points = []
cuts = []
for x, counts in enumerate(counter.aggregate_counts):
points.extend(counts)
cuts.append(len(counts))
plt.plot(list(range(len(points))), points)
xx, xxx = 0, []
for x in cuts:
xxx.append(xx + x // 2)
xx += x
plt.axvline(xx, color="r", ls="--")
plt.xticks(xxx, counter.fields)
figname = counter.name + ".detail"
fn = E.get_output_file(figname) + ".png"
plt.savefig(os.path.expanduser(fn))
elif method == "tssprofile":
plt.figure()
plt.subplot(1, 3, 1)
plt.plot(
list(
range(-options.extension_outward,
options.extension_inward)),
counter.aggregate_counts[0])
plt.title(counter.fields[0])
plt.subplot(1, 3, 2)
plt.plot(
list(
range(-options.extension_inward,
options.extension_outward)),
counter.aggregate_counts[1])
plt.title(counter.fields[1])
plt.subplot(1, 3, 3)
plt.title("combined")
plt.plot(
list(
range(-options.extension_outward,
options.extension_inward)),
counter.aggregate_counts[0])
plt.plot(
list(
range(-options.extension_inward,
options.extension_outward)),
counter.aggregate_counts[1])
plt.legend(counter.fields[:2])
fn = E.get_output_file(counter.name) + ".png"
plt.savefig(os.path.expanduser(fn))
elif method == "midpointprofile":
plt.figure()
plt.plot(numpy.arange(-options.resolution_upstream, 0),
counter.aggregate_counts[0])
plt.plot(numpy.arange(0, options.resolution_downstream),
counter.aggregate_counts[1])
fn = E.get_output_file(counter.name) + ".png"
plt.savefig(os.path.expanduser(fn))
# write footer and output benchmark information.
E.stop()
def plot_depth_profile_plot(dataframe, section, map_key2label={}, **kwargs):
ax = DepthProfilePlot()(dataframe, map_sample2label={})
plt.savefig(E.get_output_file(section))
plt.close()
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if not argv:
argv = sys.argv
# setup command line parser
parser = E.OptionParser(version="%prog version: $Id$",
usage=globals()["__doc__"])
parser.add_option(
"-e",
"--exclusive-overlap",
dest="exclusive",
action="store_true",
help="Intervals reported will be merged across the "
"positive set and do not overlap any interval in any of the "
"other sets [default=%default].")
parser.add_option("-p",
"--pattern-identifier",
dest="pattern_id",
type="string",
help="pattern to convert a filename "
"to an id [default=%default].")
parser.add_option("-m",
"--method",
dest="method",
type="choice",
choices=("merged-combinations", "unmerged-combinations"),
help="method to perform [default=%default]")
parser.set_defaults(
pattern_id="(.*).bed.gz",
exclusive=False,
method="merged-combinations",
)
# add common options (-h/--help, ...) and parse command line
(options, args) = E.start(parser, argv=argv, add_output_options=True)
if len(args) < 2:
raise ValueError("at least two arguments required")
tags, bedfiles = [], []
for infile in args:
bedfiles.append(pysam.Tabixfile(infile, "r"))
tags.append(re.search(options.pattern_id, infile).groups()[0])
indices = list(range(len(bedfiles)))
is_exclusive = options.exclusive
if options.method == "merged-combinations":
if is_exclusive:
start = 1
else:
start = 2
options.stdout.write("combination\twithout\tcounts\n")
for ncombinants in range(start, len(bedfiles) + 1):
for combination in itertools.combinations(indices, ncombinants):
other = [x for x in indices if x not in combination]
tag = ":".join([tags[x] for x in combination])
E.debug("combination %s started" % tag)
E.debug("other: %s" % ":".join([tags[x] for x in other]))
other_bed = [bedfiles[x] for x in other]
outf = IOTools.open_file(E.get_output_file(tag),
"w",
create_dir=True)
c = E.Counter()
for contig, start, end in combineMergedIntervals(
[bedfiles[x] for x in combination]):
c.found += 1
if is_exclusive and isContainedInOne(
contig, start, end, other_bed):
c.removed += 1
continue
c.output += 1
outf.write("%s\t%i\t%i\n" % (contig, start, end))
outf.close()
E.info("combination %s finished: %s" % (tag, c))
options.stdout.write("%s\t%s\t%i\n" %
(":".join([tags[x] for x in combination]),
":".join([tags[x]
for x in other]), c.output))
elif options.method == "unmerged-combinations":
options.stdout.write("track\tcombination\twithout\tcounts\n")
for foreground in indices:
start = 0
background = [x for x in indices if x != foreground]
for ncombinants in range(0, len(background) + 1):
for combination in itertools.combinations(
background, ncombinants):
other = [x for x in background if x not in combination]
combination_bed = [bedfiles[x] for x in combination]
other_bed = [bedfiles[x] for x in other]
tag = ":".join([tags[foreground]] +
[tags[x] for x in combination])
E.debug("fg=%i, combination=%s, other=%s" %
(foreground, combination, other))
E.debug("combination %s started" % tag)
E.debug("other: %s" % ":".join([tags[x] for x in other]))
outf = IOTools.open_file(E.get_output_file(tag),
"w",
create_dir=True)
c = E.Counter()
for bed in combineUnmergedIntervals(
bedfiles[foreground], combination_bed):
c.found += 1
if is_exclusive and isContainedInOne(
bed.contig, bed.start, bed.end, other_bed):
c.removed += 1
continue
c.output += 1
outf.write("%s\n" % str(bed))
outf.close()
E.info("combination %s finished: %s" % (tag, c))
options.stdout.write(
"%s\t%s\t%s\t%i\n" % (tags[foreground], ":".join([
tags[x] for x in combination
]), ":".join([tags[x] for x in other]), c.output))
E.stop()