def makemut(args, bedline, alignopts):
if args.seed is not None: random.seed(int(args.seed))
mutid = '_'.join(map(str, bedline.strip().split()))
try:
bamfile = pysam.Samfile(args.bamFileName, 'rb')
reffile = pysam.Fastafile(args.refFasta)
logfn = '_'.join(map(os.path.basename,
bedline.strip().split())) + ".log"
logfile = open(
'addsv_logs_' + os.path.basename(args.outBamFile) + '/' +
os.path.basename(args.outBamFile) + '_' + logfn, 'w')
exclfile = args.tmpdir + '/' + '.'.join(
(mutid, 'exclude', str(uuid4()), 'txt'))
exclude = open(exclfile, 'w')
# optional CNV file
cnv = None
if (args.cnvfile):
cnv = pysam.Tabixfile(args.cnvfile, 'r')
# temporary file to hold mutated reads
outbam_mutsfile = args.tmpdir + '/' + '.'.join(
(mutid, str(uuid4()), "muts.bam"))
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3:len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
# translocation specific
trn_chrom = None
trn_start = None
trn_end = None
is_transloc = c[3] == 'TRN'
if is_transloc:
start -= 3000
end += 3000
if start < 0: start = 0
trn_chrom = c[4]
trn_start = int(c[5]) - 3000
trn_end = int(c[5]) + 3000
if trn_start < 0: trn_start = 0
actions = map(lambda x: x.strip(), ' '.join(araw).split(','))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom, start, end):
cn = float(cnregion.strip().split()
[3]) # expect chrom,start,end,CN
sys.stdout.write("INFO\t" + now() + "\t" + mutid + "\t" +
' '.join(("copy number in sv region:",
chrom, str(start), str(end),
"=", str(cn))) + "\n")
svfrac = 1.0 / float(cn)
assert svfrac <= 1.0
sys.stdout.write("INFO\t" + now() + "\t" + mutid +
"\tadjusted MAF: " + str(svfrac) + "\n")
print "INFO\t" + now() + "\t" + mutid + "\tinterval:", c
print "INFO\t" + now() + "\t" + mutid + "\tlength:", end - start
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3 * int(args.maxlibsize) # maxctglen is too short
if end - start > maxctglen:
adj = (end - start) - maxctglen
rndpt = random.randint(0, adj)
start = start + rndpt
end = end - (adj - rndpt)
print "INFO\t" + now(
) + "\t" + mutid + "\tnote: interval size too long, adjusted:", chrom, start, end
dfrac = discordant_fraction(args.bamFileName, chrom, start, end)
print "INFO\t" + now() + "\t" + mutid + "\tdiscordant fraction:", dfrac
maxdfrac = 0.1 # FIXME make a parameter
if dfrac > .1:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tdiscordant fraction > " + str(maxdfrac) +
" aborting mutation!\n")
return None, None
contigs = ar.asm(chrom,
start,
end,
args.bamFileName,
reffile,
int(args.kmersize),
args.tmpdir,
args.noref,
args.recycle,
mutid=mutid,
debug=args.debug)
trn_contigs = None
if is_transloc:
trn_contigs = ar.asm(trn_chrom,
trn_start,
trn_end,
args.bamFileName,
reffile,
int(args.kmersize),
args.tmpdir,
args.noref,
args.recycle,
mutid=mutid,
debug=args.debug)
maxcontig = sorted(contigs)[-1]
trn_maxcontig = None
if is_transloc: trn_maxcontig = sorted(trn_contigs)[-1]
# be strict about contig quality
if re.search('N', maxcontig.seq):
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid +
"\tcontig dropped due to ambiguous base (N), aborting mutation.\n"
)
return None, None
if is_transloc and re.search('N', trn_maxcontig.seq):
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid +
"\tcontig dropped due to ambiguous base (N), aborting mutation.\n"
)
return None, None
if maxcontig is None:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tmaxcontig has length 0, aborting mutation!\n")
return None, None
if is_transloc and trn_maxcontig is None:
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid +
"\ttransloc maxcontig has length 0, aborting mutation!\n")
return None, None
print "INFO\t" + now(
) + "\t" + mutid + "\tbest contig length:", sorted(contigs)[-1].len
if is_transloc:
print "INFO\t" + now(
) + "\t" + mutid + "\tbest transloc contig length:", sorted(
trn_contigs)[-1].len
# trim contig to get best ungapped aligned region to ref.
maxcontig, refseq, alignstats, refstart, refend, qrystart, qryend, tgtstart, tgtend = trim_contig(
mutid, chrom, start, end, maxcontig, reffile)
print "INFO\t" + now(
) + "\t" + mutid + "\tstart, end, tgtstart, tgtend, refstart, refend:", start, end, tgtstart, tgtend, refstart, refend
if is_transloc:
trn_maxcontig, trn_refseq, trn_alignstats, trn_refstart, trn_refend, trn_qrystart, trn_qryend, trn_tgtstart, trn_tgtend = trim_contig(
mutid, trn_chrom, trn_start, trn_end, trn_maxcontig, reffile)
print "INFO\t" + now(
) + "\t" + mutid + "\ttrn_start, trn_end, trn_tgtstart, trn_tgtend, trn_refstart, trn_refend:", trn_start, trn_end, trn_tgtstart, trn_tgtend, trn_refstart, trn_refend
# is there anough room to make mutations?
if maxcontig.len < 3 * int(args.maxlibsize):
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tbest contig too short to make mutation!\n")
return None, None
if is_transloc and trn_maxcontig.len < 3 * int(args.maxlibsize):
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid +
"\tbest transloc contig too short to make mutation!\n")
return None, None
# make mutation in the largest contig
mutseq = ms.MutableSeq(maxcontig.seq)
if is_transloc: trn_mutseq = ms.MutableSeq(trn_maxcontig.seq)
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print "INFO\t" + now(
) + "\t" + mutid + "\taction: ", actionstr, action
insseqfile = None
insseq = ''
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == 'INS':
assert len(
a) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not (
os.path.exists(insseqfile) or insseqfile == 'RND'
): # not a file... is it a sequence? (support indel ins.)
assert re.search('^[ATGCatgc]*$',
insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == 'DUP':
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == 'DEL':
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
if action == 'TRN':
pass
logfile.write(">" + chrom + ":" + str(refstart) + "-" +
str(refend) + " BEFORE\n" + str(mutseq) + "\n")
if action == 'INS':
if insseqfile: # seq in file
if insseqfile == 'RND':
assert args.inslib is not None # insertion library needs to exist
insseqfile = random.choice(args.inslib.keys())
print "INFO\t" + now(
) + "\t" + mutid + "\tchose sequence from insertion library: " + insseqfile
mutseq.insertion(mutseq.length() / 2,
args.inslib[insseqfile], tsdlen)
else:
mutseq.insertion(mutseq.length() / 2,
singleseqfa(insseqfile, mutid=mutid),
tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length() / 2, insseq, tsdlen)
logfile.write("\t".join(
('ins', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(mutseq.length() / 2),
str(insseqfile), str(tsdlen))) + "\n")
elif action == 'INV':
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart, invend)
logfile.write("\t".join(
('inv', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(invstart), str(invend))) + "\n")
elif action == 'DEL':
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend - delstart) * dsize) + 0.5)
dadj = delend - delstart - dlen
if dadj < 0:
dadj = 0
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\twarning: deletion of length 0\n")
delstart += dadj / 2
delend -= dadj / 2
mutseq.deletion(delstart, delend)
logfile.write("\t".join(('del', chrom, str(refstart),
str(refend), action,
str(mutseq.length()), str(delstart),
str(delend), str(dlen))) + "\n")
elif action == 'DUP':
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart, dupend, ndups)
logfile.write("\t".join(('dup', chrom, str(refstart),
str(refend), action,
str(mutseq.length()), str(dupstart),
str(dupend), str(ndups))) + "\n")
elif action == 'TRN':
mutseq.fusion(mutseq.length() / 2, trn_mutseq,
trn_mutseq.length() / 2)
logfile.write("\t".join(
('trn', chrom, str(refstart), str(refend), action,
str(mutseq.length()), trn_chrom, str(trn_refstart),
str(trn_refend), str(trn_mutseq.length()))) + "\n")
else:
raise ValueError(
"ERROR\t" + now() + "\t" + mutid +
"\t: mutation not one of: INS,INV,DEL,DUP,TRN\n")
logfile.write(">" + chrom + ":" + str(refstart) + "-" +
str(refend) + " AFTER\n" + str(mutseq) + "\n")
pemean, pesd = float(args.ismean), float(args.issd)
print "INFO\t" + now(
) + "\t" + mutid + "\tset paired end mean distance: " + str(
args.ismean)
print "INFO\t" + now(
) + "\t" + mutid + "\tset paired end distance stddev: " + str(
args.issd)
# simulate reads
(fq1, fq2) = runwgsim(maxcontig,
mutseq.seq,
svfrac,
actions,
exclude,
pemean,
pesd,
args.tmpdir,
mutid=mutid,
seed=args.seed,
trn_contig=trn_maxcontig)
outreads = aligners.remap_fastq(args.aligner,
fq1,
fq2,
args.refFasta,
outbam_mutsfile,
alignopts,
mutid=mutid,
threads=1)
if outreads == 0:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\toutbam " +
outbam_mutsfile + " has no mapped reads!\n")
return None, None
print "INFO\t" + now(
) + "\t" + mutid + "\ttemporary bam: " + outbam_mutsfile
exclude.close()
bamfile.close()
return outbam_mutsfile, exclfile
except Exception, e:
sys.stderr.write("*" * 60 +
"\nencountered error in mutation spikein: " +
bedline + "\n")
traceback.print_exc(file=sys.stderr)
sys.stderr.write("*" * 60 + "\n")
return None, None
def makemut(args, bedline, alignopts):
if args.seed is not None:
random.seed(int(args.seed))
mutid = "_".join(map(str, bedline.strip().split()))
try:
bamfile = pysam.Samfile(args.bamFileName, "rb")
reffile = pysam.Fastafile(args.refFasta)
logfn = "_".join(map(os.path.basename, bedline.strip().split())) + ".log"
logfile = open(
"addsv_logs_" + os.path.basename(args.outBamFile) + "/" + os.path.basename(args.outBamFile) + "_" + logfn,
"w",
)
exclfile = args.tmpdir + "/" + ".".join((mutid, "exclude", str(uuid4()), "txt"))
exclude = open(exclfile, "w")
# optional CNV file
cnv = None
if args.cnvfile:
cnv = pysam.Tabixfile(args.cnvfile, "r")
# temporary file to hold mutated reads
outbam_mutsfile = args.tmpdir + "/" + ".".join((mutid, str(uuid4()), "muts.bam"))
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3 : len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
# translocation specific
trn_chrom = None
trn_start = None
trn_end = None
is_transloc = c[3] == "TRN"
if is_transloc:
start -= 3000
end += 3000
if start < 0:
start = 0
trn_chrom = c[4]
trn_start = int(c[5]) - 3000
trn_end = int(c[5]) + 3000
if trn_start < 0:
trn_start = 0
actions = map(lambda x: x.strip(), " ".join(araw).split(","))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom, start, end):
cn = float(cnregion.strip().split()[3]) # expect chrom,start,end,CN
sys.stdout.write(
"INFO\t"
+ now()
+ "\t"
+ mutid
+ "\t"
+ " ".join(("copy number in sv region:", chrom, str(start), str(end), "=", str(cn)))
+ "\n"
)
svfrac = 1.0 / float(cn)
assert svfrac <= 1.0
sys.stdout.write("INFO\t" + now() + "\t" + mutid + "\tadjusted MAF: " + str(svfrac) + "\n")
print "INFO\t" + now() + "\t" + mutid + "\tinterval:", c
print "INFO\t" + now() + "\t" + mutid + "\tlength:", end - start
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3 * int(args.maxlibsize) # maxctglen is too short
if end - start > maxctglen:
adj = (end - start) - maxctglen
rndpt = random.randint(0, adj)
start = start + rndpt
end = end - (adj - rndpt)
print "INFO\t" + now() + "\t" + mutid + "\tnote: interval size too long, adjusted:", chrom, start, end
dfrac = discordant_fraction(args.bamFileName, chrom, start, end)
print "INFO\t" + now() + "\t" + mutid + "\tdiscordant fraction:", dfrac
maxdfrac = 0.1 # FIXME make a parameter
if dfrac > 0.1:
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid + "\tdiscordant fraction > " + str(maxdfrac) + " aborting mutation!\n"
)
return None, None
contigs = ar.asm(
chrom,
start,
end,
args.bamFileName,
reffile,
int(args.kmersize),
args.tmpdir,
args.noref,
args.recycle,
mutid=mutid,
debug=args.debug,
)
trn_contigs = None
if is_transloc:
trn_contigs = ar.asm(
trn_chrom,
trn_start,
trn_end,
args.bamFileName,
reffile,
int(args.kmersize),
args.tmpdir,
args.noref,
args.recycle,
mutid=mutid,
debug=args.debug,
)
maxcontig = sorted(contigs)[-1]
trn_maxcontig = None
if is_transloc:
trn_maxcontig = sorted(trn_contigs)[-1]
# be strict about contig quality
if re.search("N", maxcontig.seq):
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid + "\tcontig dropped due to ambiguous base (N), aborting mutation.\n"
)
return None, None
if is_transloc and re.search("N", trn_maxcontig.seq):
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid + "\tcontig dropped due to ambiguous base (N), aborting mutation.\n"
)
return None, None
if maxcontig is None:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tmaxcontig has length 0, aborting mutation!\n")
return None, None
if is_transloc and trn_maxcontig is None:
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid + "\ttransloc maxcontig has length 0, aborting mutation!\n"
)
return None, None
print "INFO\t" + now() + "\t" + mutid + "\tbest contig length:", sorted(contigs)[-1].len
if is_transloc:
print "INFO\t" + now() + "\t" + mutid + "\tbest transloc contig length:", sorted(trn_contigs)[-1].len
# trim contig to get best ungapped aligned region to ref.
maxcontig, refseq, alignstats, refstart, refend, qrystart, qryend, tgtstart, tgtend = trim_contig(
mutid, chrom, start, end, maxcontig, reffile
)
print "INFO\t" + now() + "\t" + mutid + "\tstart, end, tgtstart, tgtend, refstart, refend:", start, end, tgtstart, tgtend, refstart, refend
if is_transloc:
trn_maxcontig, trn_refseq, trn_alignstats, trn_refstart, trn_refend, trn_qrystart, trn_qryend, trn_tgtstart, trn_tgtend = trim_contig(
mutid, trn_chrom, trn_start, trn_end, trn_maxcontig, reffile
)
print "INFO\t" + now() + "\t" + mutid + "\ttrn_start, trn_end, trn_tgtstart, trn_tgtend, trn_refstart, trn_refend:", trn_start, trn_end, trn_tgtstart, trn_tgtend, trn_refstart, trn_refend
# is there anough room to make mutations?
if maxcontig.len < 3 * int(args.maxlibsize):
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tbest contig too short to make mutation!\n")
return None, None
if is_transloc and trn_maxcontig.len < 3 * int(args.maxlibsize):
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tbest transloc contig too short to make mutation!\n")
return None, None
# make mutation in the largest contig
mutseq = ms.MutableSeq(maxcontig.seq)
if is_transloc:
trn_mutseq = ms.MutableSeq(trn_maxcontig.seq)
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print "INFO\t" + now() + "\t" + mutid + "\taction: ", actionstr, action
insseqfile = None
insseq = ""
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == "INS":
assert len(a) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not (
os.path.exists(insseqfile) or insseqfile == "RND"
): # not a file... is it a sequence? (support indel ins.)
assert re.search("^[ATGCatgc]*$", insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == "DUP":
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == "DEL":
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
if action == "TRN":
pass
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) + " BEFORE\n" + str(mutseq) + "\n")
if action == "INS":
if insseqfile: # seq in file
if insseqfile == "RND":
assert args.inslib is not None # insertion library needs to exist
insseqfile = random.choice(args.inslib.keys())
print "INFO\t" + now() + "\t" + mutid + "\tchose sequence from insertion library: " + insseqfile
mutseq.insertion(mutseq.length() / 2, args.inslib[insseqfile], tsdlen)
else:
mutseq.insertion(mutseq.length() / 2, singleseqfa(insseqfile, mutid=mutid), tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length() / 2, insseq, tsdlen)
logfile.write(
"\t".join(
(
"ins",
chrom,
str(refstart),
str(refend),
action,
str(mutseq.length()),
str(mutseq.length() / 2),
str(insseqfile),
str(tsdlen),
)
)
+ "\n"
)
elif action == "INV":
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart, invend)
logfile.write(
"\t".join(
(
"inv",
chrom,
str(refstart),
str(refend),
action,
str(mutseq.length()),
str(invstart),
str(invend),
)
)
+ "\n"
)
elif action == "DEL":
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend - delstart) * dsize) + 0.5)
dadj = delend - delstart - dlen
if dadj < 0:
dadj = 0
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\twarning: deletion of length 0\n")
delstart += dadj / 2
delend -= dadj / 2
mutseq.deletion(delstart, delend)
logfile.write(
"\t".join(
(
"del",
chrom,
str(refstart),
str(refend),
action,
str(mutseq.length()),
str(delstart),
str(delend),
str(dlen),
)
)
+ "\n"
)
elif action == "DUP":
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart, dupend, ndups)
logfile.write(
"\t".join(
(
"dup",
chrom,
str(refstart),
str(refend),
action,
str(mutseq.length()),
str(dupstart),
str(dupend),
str(ndups),
)
)
+ "\n"
)
elif action == "TRN":
mutseq.fusion(mutseq.length() / 2, trn_mutseq, trn_mutseq.length() / 2)
logfile.write(
"\t".join(
(
"trn",
chrom,
str(refstart),
str(refend),
action,
str(mutseq.length()),
trn_chrom,
str(trn_refstart),
str(trn_refend),
str(trn_mutseq.length()),
)
)
+ "\n"
)
else:
raise ValueError("ERROR\t" + now() + "\t" + mutid + "\t: mutation not one of: INS,INV,DEL,DUP,TRN\n")
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) + " AFTER\n" + str(mutseq) + "\n")
pemean, pesd = float(args.ismean), float(args.issd)
print "INFO\t" + now() + "\t" + mutid + "\tset paired end mean distance: " + str(args.ismean)
print "INFO\t" + now() + "\t" + mutid + "\tset paired end distance stddev: " + str(args.issd)
# simulate reads
(fq1, fq2) = runwgsim(
maxcontig,
mutseq.seq,
svfrac,
actions,
exclude,
pemean,
pesd,
args.tmpdir,
mutid=mutid,
seed=args.seed,
trn_contig=trn_maxcontig,
)
outreads = aligners.remap_fastq(
args.aligner, fq1, fq2, args.refFasta, outbam_mutsfile, alignopts, mutid=mutid, threads=1
)
if outreads == 0:
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid + "\toutbam " + outbam_mutsfile + " has no mapped reads!\n"
)
return None, None
print "INFO\t" + now() + "\t" + mutid + "\ttemporary bam: " + outbam_mutsfile
exclude.close()
bamfile.close()
return outbam_mutsfile, exclfile
except Exception, e:
sys.stderr.write("*" * 60 + "\nencountered error in mutation spikein: " + bedline + "\n")
traceback.print_exc(file=sys.stderr)
sys.stderr.write("*" * 60 + "\n")
return None, None
def main(args):
""" needs refactoring
"""
varfile = open(args.varFileName, 'r')
bamfile = pysam.Samfile(args.bamFileName, 'rb')
reffile = pysam.Fastafile(args.refFasta)
logfile = open(args.outBamFile + ".log", 'w')
exclude = open(args.exclfile, 'w')
# optional CNV file
cnv = None
if (args.cnvfile):
cnv = pysam.Tabixfile(args.cnvfile, 'r')
# temporary file to hold mutated reads
outbam_mutsfile = "tmp." + str(random.random()) + ".muts.bam"
nmuts = 0
for bedline in varfile:
if re.search('^#',bedline):
continue
if args.maxmuts and nmuts >= int(args.maxmuts):
break
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3:len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
actions = map(lambda x: x.strip(),' '.join(araw).split(','))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom,start,end):
cn = float(cnregion.strip().split()[3]) # expect chrom,start,end,CN
sys.stderr.write(' '.join(("copy number in snp region:",chrom,str(start),str(end),"=",str(cn))) + "\n")
svfrac = 1.0/float(cn)
sys.stderr.write("adjusted MAF: " + str(svfrac) + "\n")
print "interval:",c
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3*int(args.maxlibsize) # maxctglen is too short
if end-start > maxctglen:
adj = (end-start) - maxctglen
rndpt = random.randint(0,adj)
start = start + rndpt
end = end - (adj-rndpt)
print "note: interval size too long, adjusted:",chrom,start,end
contigs = ar.asm(chrom, start, end, args.bamFileName, reffile, int(args.kmersize), args.noref, args.recycle)
# find the largest contig
maxlen = 0
maxcontig = None
for contig in contigs:
if contig.len > maxlen:
maxlen = contig.len
maxcontig = contig
# is there anough room to make mutations?
if maxlen > 3*int(args.maxlibsize):
# make mutation in the largest contig
mutseq = ms.MutableSeq(maxcontig.seq)
# if we're this far along, we're making a mutation
nmuts += 1
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print actionstr,action
insseqfile = None
insseq = ''
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == 'INS':
assert len(a) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not os.path.exists(insseqfile): # not a file... is it a sequence? (support indel ins.)
assert re.search('^[ATGCatgc]*$',insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == 'DUP':
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == 'DEL':
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
print "BEFORE:",mutseq
if action == 'INS':
if insseqfile: # seq in file
mutseq.insertion(mutseq.length()/2,singleseqfa(insseqfile),tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length()/2,insseq,tsdlen)
logfile.write("\t".join(('ins',chrom,str(start),str(end),action,str(mutseq.length()),str(mutseq.length()/2),str(insseqfile),str(tsdlen))) + "\n")
elif action == 'INV':
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart,invend)
logfile.write("\t".join(('inv',chrom,str(start),str(end),action,str(mutseq.length()),str(invstart),str(invend))) + "\n")
elif action == 'DEL':
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend-delstart) * dsize)+0.5)
dadj = delend-delstart-dlen
if dadj < 0:
dadj = 0
print "warning: deletion of length 0"
delstart += dadj/2
delend -= dadj/2
mutseq.deletion(delstart,delend)
logfile.write("\t".join(('del',chrom,str(start),str(end),action,str(mutseq.length()),str(delstart),str(delend),str(dlen))) + "\n")
elif action == 'DUP':
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart,dupend,ndups)
logfile.write("\t".join(('dup',chrom,str(start),str(end),action,str(mutseq.length()),str(dupstart),str(dupend),str(ndups))) + "\n")
else:
raise ValueError(bedline.strip() + ": mutation not one of: INS,INV,DEL,DUP")
print "AFTER:",mutseq
# simulate reads
(fq1, fq2) = runwgsim(maxcontig, mutseq.seq, svfrac, exclude)
# remap reads
remap(fq1, fq2, 4, args.refFasta, outbam_mutsfile)
else:
print "best contig too short to make mutation: ",bedline.strip()
print "addsv.py finished, made", nmuts, "mutations."
exclude.close()
varfile.close()
bamfile.close()
logfile.close()
print "merging mutations into", args.bamFileName, "-->", args.outBamFile
replace(args.bamFileName, outbam_mutsfile, args.outBamFile, args.exclfile)
# cleanup
os.remove(outbam_mutsfile)
def makemut(args, bedline, alignopts):
mutid = '_'.join(map(str, bedline.strip().split()))
try:
bamfile = pysam.Samfile(args.bamFileName, 'rb')
reffile = pysam.Fastafile(args.refFasta)
logfn = '_'.join(map(os.path.basename, bedline.strip().split())) + ".log"
logfile = open('addsv_logs_' + os.path.basename(args.outBamFile) + '/' + os.path.basename(args.outBamFile) + '_' + logfn, 'w')
exclfile = args.tmpdir + '/' + '.'.join((mutid, 'exclude', str(uuid4()), 'txt'))
exclude = open(exclfile, 'w')
# optional CNV file
cnv = None
if (args.cnvfile):
cnv = pysam.Tabixfile(args.cnvfile, 'r')
# temporary file to hold mutated reads
outbam_mutsfile = args.tmpdir + '/' + '.'.join((mutid, str(uuid4()), "muts.bam"))
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3:len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
actions = map(lambda x: x.strip(),' '.join(araw).split(','))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom,start,end):
cn = float(cnregion.strip().split()[3]) # expect chrom,start,end,CN
sys.stdout.write("INFO\t" + now() + "\t" + mutid + "\t" + ' '.join(("copy number in sv region:",chrom,str(start),str(end),"=",str(cn))) + "\n")
svfrac = 1.0/float(cn)
assert svfrac <= 1.0
sys.stdout.write("INFO\t" + now() + "\t" + mutid + "\tadjusted MAF: " + str(svfrac) + "\n")
print "INFO\t" + now() + "\t" + mutid + "\tinterval:", c
print "INFO\t" + now() + "\t" + mutid + "\tlength:", end-start
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3*int(args.maxlibsize) # maxctglen is too short
if end-start > maxctglen:
adj = (end-start) - maxctglen
rndpt = random.randint(0,adj)
start = start + rndpt
end = end - (adj-rndpt)
print "INFO\t" + now() + "\t" + mutid + "\tnote: interval size too long, adjusted:",chrom,start,end
dfrac = discordant_fraction(args.bamFileName, chrom, start, end)
print "INFO\t" + now() + "\t" + mutid + "\tdiscordant fraction:", dfrac
maxdfrac = 0.1 # FIXME make a parameter
if dfrac > .1:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tdiscordant fraction > " + str(maxdfrac) + " aborting mutation!\n")
return None, None
contigs = ar.asm(chrom, start, end, args.bamFileName, reffile, int(args.kmersize), args.tmpdir, args.noref, args.recycle, mutid=mutid, debug=args.debug)
# find the largest contig
maxlen = 0
maxcontig = None
for contig in contigs:
if contig.len > maxlen:
maxlen = contig.len
maxcontig = contig
# be strict about contig quality
if re.search('N', maxcontig.seq):
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tcontig dropped due to ambiguous base (N), aborting mutation.\n")
return None, None
if maxcontig is None:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tmaxcontig has length 0, aborting mutation!\n")
return None, None
# trim contig to get best ungapped aligned region to ref.
refseq = reffile.fetch(chrom,start,end)
alignstats = align(maxcontig.seq, refseq)
if len(alignstats) < 6:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\talignstats:" + str(alignstats) + "\n")
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tNo good alignment between mutated contig and original, aborting mutation!\n")
return None, None
qrystart, qryend = map(int, alignstats[2:4])
tgtstart, tgtend = map(int, alignstats[4:6])
refseq = refseq[tgtstart:tgtend]
print "INFO\t" + now() + "\t" + mutid + "\tbest contig length:", maxlen
print "INFO\t" + now() + "\t" + mutid + "\talignment result:", alignstats
maxcontig.trimseq(qrystart, qryend)
print "INFO\t" + now() + "\t" + mutid + "\ttrimmed contig length:", maxcontig.len
refstart = start + tgtstart
refend = start + tgtend
if refstart > refend:
refstart, refend = refend, refstart
print "INFO\t" + now() + "\t" + mutid + "\tstart, end, tgtstart, tgtend, refstart, refend:", start, end, tgtstart, tgtend, refstart, refend
# is there anough room to make mutations?
if maxcontig.len > 3*int(args.maxlibsize):
# make mutation in the largest contig
mutseq = ms.MutableSeq(maxcontig.seq)
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print "INFO\t" + now() + "\t" + mutid + "\taction: ", actionstr, action
insseqfile = None
insseq = ''
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == 'INS':
assert len(a) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not (os.path.exists(insseqfile) or insseqfile == 'RND'): # not a file... is it a sequence? (support indel ins.)
assert re.search('^[ATGCatgc]*$',insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == 'DUP':
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == 'DEL':
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) + " BEFORE\n" + str(mutseq) + "\n")
if action == 'INS':
if insseqfile: # seq in file
if insseqfile == 'RND':
assert args.inslib is not None # insertion library needs to exist
mutseq.insertion(mutseq.length()/2,pickseq(args.inslib, mutid=mutid),tsdlen)
else:
mutseq.insertion(mutseq.length()/2,singleseqfa(insseqfile, mutid=mutid),tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length()/2,insseq,tsdlen)
logfile.write("\t".join(('ins',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(mutseq.length()/2),str(insseqfile),str(tsdlen))) + "\n")
elif action == 'INV':
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart,invend)
logfile.write("\t".join(('inv',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(invstart),str(invend))) + "\n")
elif action == 'DEL':
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend-delstart) * dsize)+0.5)
dadj = delend-delstart-dlen
if dadj < 0:
dadj = 0
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\twarning: deletion of length 0\n")
delstart += dadj/2
delend -= dadj/2
mutseq.deletion(delstart,delend)
logfile.write("\t".join(('del',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(delstart),str(delend),str(dlen))) + "\n")
elif action == 'DUP':
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart,dupend,ndups)
logfile.write("\t".join(('dup',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(dupstart),str(dupend),str(ndups))) + "\n")
else:
raise ValueError("ERROR\t" + now() + "\t" + mutid + "\t: mutation not one of: INS,INV,DEL,DUP\n")
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) +" AFTER\n" + str(mutseq) + "\n")
pemean, pesd = float(args.ismean), float(args.issd)
print "INFO\t" + now() + "\t" + mutid + "\tset paired end mean distance: " + str(args.ismean)
print "INFO\t" + now() + "\t" + mutid + "\tset paired end distance stddev: " + str(args.issd)
# simulate reads
(fq1, fq2) = runwgsim(maxcontig, mutseq.seq, svfrac, actions, exclude, pemean, pesd, args.tmpdir, mutid=mutid)
outreads = aligners.remap_fastq(args.aligner, fq1, fq2, args.refFasta, outbam_mutsfile, alignopts, mutid=mutid, threads=1)
if outreads == 0:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\toutbam " + outbam_mutsfile + " has no mapped reads!\n")
return None, None
else:
sys.stderr.write("WARN\t" + now() + "\t" + mutid + "\tbest contig too short to make mutation!\n")
return None, None
print "INFO\t" + now() + "\t" + mutid + "\ttemporary bam: " + outbam_mutsfile
exclude.close()
bamfile.close()
return outbam_mutsfile, exclfile
except Exception, e:
sys.stderr.write("*"*60 + "\nencountered error in mutation spikein: " + bedline + "\n")
traceback.print_exc(file=sys.stderr)
sys.stderr.write("*"*60 + "\n")
return None, None
def makemut(args, bedline):
mutid = ':'.join(map(str, bedline.strip().split()))
try:
bamfile = pysam.Samfile(args.bamFileName, 'rb')
reffile = pysam.Fastafile(args.refFasta)
logfn = '_'.join(map(os.path.basename,
bedline.strip().split())) + ".log"
logfile = open(
'addsv_logs_' + os.path.basename(args.outBamFile) + '/' +
os.path.basename(args.outBamFile) + '_' + logfn, 'w')
exclfile = 'exclude.' + str(random.random()) + '.txt'
exclude = open(exclfile, 'w')
# optional CNV file
cnv = None
if (args.cnvfile):
cnv = pysam.Tabixfile(args.cnvfile, 'r')
# temporary file to hold mutated reads
outbam_mutsfile = "tmp." + str(random.random()) + ".muts.bam"
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3:len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
actions = map(lambda x: x.strip(), ' '.join(araw).split(','))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom, start, end):
cn = float(cnregion.strip().split()
[3]) # expect chrom,start,end,CN
sys.stdout.write("INFO\t" + now() + "\t" + mutid + "\t" +
' '.join(("copy number in snp region:",
chrom, str(start), str(end),
"=", str(cn))) + "\n")
svfrac = 1.0 / float(cn)
assert svfrac < 1.0
sys.stdout.write("INFO\t" + now() + "\t" + mutid +
"\tadjusted MAF: " + str(svfrac) + "\n")
print "INFO\t" + now() + "\t" + mutid + "\tinterval:", c
print "INFO\t" + now() + "\t" + mutid + "\tlength:", end - start
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3 * int(args.maxlibsize) # maxctglen is too short
if end - start > maxctglen:
adj = (end - start) - maxctglen
rndpt = random.randint(0, adj)
start = start + rndpt
end = end - (adj - rndpt)
print "INFO\t" + now(
) + "\t" + mutid + "\tnote: interval size too long, adjusted:", chrom, start, end
dfrac = discordant_fraction(args.bamFileName, chrom, start, end)
print "INFO\t" + now() + "\t" + mutid + "\tdiscordant fraction:", dfrac
maxdfrac = 0.1 # FIXME make a parameter
if dfrac > .1:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tdiscordant fraction > " + str(maxdfrac) +
" aborting mutation!\n")
return None, None
contigs = ar.asm(chrom,
start,
end,
args.bamFileName,
reffile,
int(args.kmersize),
args.noref,
args.recycle,
mutid=mutid)
# find the largest contig
maxlen = 0
maxcontig = None
for contig in contigs:
if contig.len > maxlen:
maxlen = contig.len
maxcontig = contig
if maxcontig is None:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tmaxcontig has length 0, aborting mutation!\n")
return None, None
# trim contig to get best ungapped aligned region to ref.
refseq = reffile.fetch(chrom, start, end)
alignstats = align(maxcontig.seq, refseq)
if len(alignstats) < 6:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\talignstats:" + str(alignstats) + "\n")
sys.stderr.write(
"WARN\t" + now() + "\t" + mutid +
"\tNo good alignment between mutated contig and original, aborting mutation!\n"
)
return None, None
qrystart, qryend = map(int, alignstats[2:4])
tgtstart, tgtend = map(int, alignstats[4:6])
refseq = refseq[tgtstart:tgtend]
print "INFO\t" + now() + "\t" + mutid + "\tbest contig length:", maxlen
print "INFO\t" + now(
) + "\t" + mutid + "\talignment result:", alignstats
maxcontig.trimseq(qrystart, qryend)
print "INFO\t" + now(
) + "\t" + mutid + "\ttrimmed contig length:", maxcontig.len
refstart = start + tgtstart
refend = start + tgtend
if refstart > refend:
refstart, refend = refend, refstart
print "INFO\t" + now(
) + "\t" + mutid + "\tstart, end, tgtstart, tgtend, refstart, refend:", start, end, tgtstart, tgtend, refstart, refend
# is there anough room to make mutations?
if maxcontig.len > 3 * int(args.maxlibsize):
# make mutation in the largest contig
mutseq = ms.MutableSeq(maxcontig.seq)
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print "INFO\t" + now(
) + "\t" + mutid + "\taction: ", actionstr, action
insseqfile = None
insseq = ''
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == 'INS':
assert len(
a
) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not (
os.path.exists(insseqfile) or insseqfile == 'RND'
): # not a file... is it a sequence? (support indel ins.)
assert re.search(
'^[ATGCatgc]*$',
insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == 'DUP':
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == 'DEL':
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
logfile.write(">" + chrom + ":" + str(refstart) + "-" +
str(refend) + " BEFORE\n" + str(mutseq) + "\n")
if action == 'INS':
if insseqfile: # seq in file
if insseqfile == 'RND':
assert args.inslib is not None # insertion library needs to exist
mutseq.insertion(mutseq.length() / 2,
pickseq(args.inslib, mutid=mutid),
tsdlen)
else:
mutseq.insertion(
mutseq.length() / 2,
singleseqfa(insseqfile, mutid=mutid), tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length() / 2, insseq, tsdlen)
logfile.write("\t".join(
('ins', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(mutseq.length() / 2),
str(insseqfile), str(tsdlen))) + "\n")
elif action == 'INV':
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart, invend)
logfile.write("\t".join(
('inv', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(invstart), str(invend))) +
"\n")
elif action == 'DEL':
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend - delstart) * dsize) + 0.5)
dadj = delend - delstart - dlen
if dadj < 0:
dadj = 0
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\twarning: deletion of length 0\n")
delstart += dadj / 2
delend -= dadj / 2
mutseq.deletion(delstart, delend)
logfile.write("\t".join(
('del', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(delstart), str(delend),
str(dlen))) + "\n")
elif action == 'DUP':
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart, dupend, ndups)
logfile.write("\t".join(
('dup', chrom, str(refstart), str(refend), action,
str(mutseq.length()), str(dupstart), str(dupend),
str(ndups))) + "\n")
else:
raise ValueError(
"ERROR\t" + now() + "\t" + mutid +
"\t: mutation not one of: INS,INV,DEL,DUP\n")
logfile.write(">" + chrom + ":" + str(refstart) + "-" +
str(refend) + " AFTER\n" + str(mutseq) + "\n")
pemean, pesd = float(args.ismean), float(args.issd)
print "INFO\t" + now(
) + "\t" + mutid + "\tset paired end mean distance: " + str(
args.ismean)
print "INFO\t" + now(
) + "\t" + mutid + "\tset paired end distance stddev: " + str(
args.issd)
# simulate reads
(fq1, fq2) = runwgsim(maxcontig,
mutseq.seq,
svfrac,
exclude,
pemean,
pesd,
mutid=mutid)
# remap reads
if args.bwamem:
outreads = remap_bwamem(fq1,
fq2,
4,
args.refFasta,
outbam_mutsfile,
mutid=mutid)
else:
outreads = remap(fq1,
fq2,
4,
args.refFasta,
outbam_mutsfile,
mutid=mutid)
if outreads == 0:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\toutbam " + outbam_mutsfile +
" has no mapped reads!\n")
return None, None
else:
sys.stderr.write("WARN\t" + now() + "\t" + mutid +
"\tbest contig too short to make mutation!\n")
return None, None
print "INFO\t" + now(
) + "\t" + mutid + "\ttemporary bam: " + outbam_mutsfile
exclude.close()
bamfile.close()
return outbam_mutsfile, exclfile
except Exception, e:
sys.stderr.write("*" * 60 +
"\nencountered error in mutation spikein: " +
bedline + "\n")
traceback.print_exc(file=sys.stderr)
sys.stderr.write("*" * 60 + "\n")
return None, None
def makemut(args, bedline):
try:
bamfile = pysam.Samfile(args.bamFileName, 'rb')
reffile = pysam.Fastafile(args.refFasta)
logfn = '_'.join(map(os.path.basename, bedline.strip().split())) + ".log"
logfile = open('addsv_logs_' + os.path.basename(args.outBamFile) + '/' + os.path.basename(args.outBamFile) + '_' + logfn, 'w')
exclfile = 'exclude.' + str(random.random()) + '.txt'
exclude = open(exclfile, 'w')
# optional CNV file
cnv = None
if (args.cnvfile):
cnv = pysam.Tabixfile(args.cnvfile, 'r')
# temporary file to hold mutated reads
outbam_mutsfile = "tmp." + str(random.random()) + ".muts.bam"
c = bedline.strip().split()
chrom = c[0]
start = int(c[1])
end = int(c[2])
araw = c[3:len(c)] # INV, DEL, INS seqfile.fa TSDlength, DUP
actions = map(lambda x: x.strip(),' '.join(araw).split(','))
svfrac = float(args.svfrac) # default, can be overridden by cnv file
if cnv: # CNV file is present
if chrom in cnv.contigs:
for cnregion in cnv.fetch(chrom,start,end):
cn = float(cnregion.strip().split()[3]) # expect chrom,start,end,CN
sys.stderr.write(' '.join(("copy number in snp region:",chrom,str(start),str(end),"=",str(cn))) + "\n")
svfrac = 1.0/float(cn)
sys.stderr.write("adjusted MAF: " + str(svfrac) + "\n")
print "interval:", c
print "length:", end-start
# modify start and end if interval is too long
maxctglen = int(args.maxctglen)
assert maxctglen > 3*int(args.maxlibsize) # maxctglen is too short
if end-start > maxctglen:
adj = (end-start) - maxctglen
rndpt = random.randint(0,adj)
start = start + rndpt
end = end - (adj-rndpt)
print "note: interval size too long, adjusted:",chrom,start,end
dfrac = discordant_fraction(args.bamFileName, chrom, start, end)
print "discordant fraction:",dfrac
maxdfrac = 0.1 # FIXME make a parameter
if dfrac > .1:
print "discordant fraction >", maxdfrac, "aborting mutation!"
return None, None
contigs = ar.asm(chrom, start, end, args.bamFileName, reffile, int(args.kmersize), args.noref, args.recycle)
# find the largest contig
maxlen = 0
maxcontig = None
for contig in contigs:
if contig.len > maxlen:
maxlen = contig.len
maxcontig = contig
if maxcontig is None:
print "maxcontig has length 0, aborting mutation!"
return None, None
# trim contig to get best ungapped aligned region to ref.
refseq = reffile.fetch(chrom,start,end)
alignstats = align(maxcontig.seq, refseq)
qrystart, qryend = map(int, alignstats[2:4])
tgtstart, tgtend = map(int, alignstats[4:6])
refseq = refseq[tgtstart:tgtend]
print "best contig length:", maxlen
print "alignment result:", alignstats
maxcontig.trimseq(qrystart, qryend)
print "trimmed contig length:", maxcontig.len
refstart = start + tgtstart
refend = start + tgtend
if refstart > refend:
refstart, refend = refend, refstart
print 'start, end, tgtstart, tgtend, refstart, refend:', start, end, tgtstart, tgtend, refstart, refend
#fixedseq = check_asmvariants(args.bamFileName, maxcontig.seq, reffile, chrom, refstart, refend)
fixedseq = maxcontig.seq # FIXME
# is there anough room to make mutations?
if maxcontig.len > 3*int(args.maxlibsize):
# make mutation in the largest contig
mutseq = ms.MutableSeq(fixedseq)
# support for multiple mutations
for actionstr in actions:
a = actionstr.split()
action = a[0]
print actionstr,action
insseqfile = None
insseq = ''
tsdlen = 0 # target site duplication length
ndups = 0 # number of tandem dups
dsize = 0.0 # deletion size fraction
dlen = 0
if action == 'INS':
assert len(a) > 1 # insertion syntax: INS <file.fa> [optional TSDlen]
insseqfile = a[1]
if not os.path.exists(insseqfile): # not a file... is it a sequence? (support indel ins.)
assert re.search('^[ATGCatgc]*$',insseqfile) # make sure it's a sequence
insseq = insseqfile.upper()
insseqfile = None
if len(a) > 2:
tsdlen = int(a[2])
if action == 'DUP':
if len(a) > 1:
ndups = int(a[1])
else:
ndups = 1
if action == 'DEL':
if len(a) > 1:
dsize = float(a[1])
if dsize >= 1.0: # if DEL size is not a fraction, interpret as bp
# since DEL 1 is default, if DEL 1 is specified, interpret as 1 bp deletion
dlen = int(dsize)
dsize = 1.0
else:
dsize = 1.0
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) + " BEFORE\n" + str(mutseq) + "\n")
if action == 'INS':
if insseqfile: # seq in file
mutseq.insertion(mutseq.length()/2,singleseqfa(insseqfile),tsdlen)
else: # seq is input
mutseq.insertion(mutseq.length()/2,insseq,tsdlen)
logfile.write("\t".join(('ins',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(mutseq.length()/2),str(insseqfile),str(tsdlen))) + "\n")
elif action == 'INV':
invstart = int(args.maxlibsize)
invend = mutseq.length() - invstart
mutseq.inversion(invstart,invend)
logfile.write("\t".join(('inv',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(invstart),str(invend))) + "\n")
elif action == 'DEL':
delstart = int(args.maxlibsize)
delend = mutseq.length() - delstart
if dlen == 0: # bp size not specified, delete fraction of contig
dlen = int((float(delend-delstart) * dsize)+0.5)
dadj = delend-delstart-dlen
if dadj < 0:
dadj = 0
print "warning: deletion of length 0"
delstart += dadj/2
delend -= dadj/2
mutseq.deletion(delstart,delend)
logfile.write("\t".join(('del',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(delstart),str(delend),str(dlen))) + "\n")
elif action == 'DUP':
dupstart = int(args.maxlibsize)
dupend = mutseq.length() - dupstart
mutseq.duplication(dupstart,dupend,ndups)
logfile.write("\t".join(('dup',chrom,str(refstart),str(refend),action,str(mutseq.length()),str(dupstart),str(dupend),str(ndups))) + "\n")
else:
raise ValueError(bedline.strip() + ": mutation not one of: INS,INV,DEL,DUP")
logfile.write(">" + chrom + ":" + str(refstart) + "-" + str(refend) +" AFTER\n" + str(mutseq) + "\n")
# estimate paired-end distribution
print "estimating paired-end insert size mean, stdev..."
pemean, pesd = estimate_pedist(bamfile, chrom, start, end, window=10000, setmean=args.ismean, setsd=args.issd)
# simulate reads
(fq1, fq2) = runwgsim(maxcontig, mutseq.seq, svfrac, exclude, pemean, pesd)
# remap reads
outreads = remap(fq1, fq2, 4, args.refFasta, outbam_mutsfile)
if outreads == 0:
print "outbam", outbam_mutsfile, "has no mapped reads!"
return None, None
else:
print "best contig too short to make mutation: ",bedline.strip()
return None, None
sys.stderr.write("temporary bam: " + outbam_mutsfile + "\n")
exclude.close()
bamfile.close()
return outbam_mutsfile, exclfile
except Exception, e:
sys.stderr.write("*"*60 + "\nencountered error in mutation spikein: " + bedline + "\n")
traceback.print_exc(file=sys.stdout)
sys.stderr.write("*"*60 + "\n")
return None, None
