def load_from_inputs(args):
#Read in the VCF file
sys.stderr.write("Reading in the VCF file\n")
alleles = {}
#with open(args.phased_VCF) as inf:
with open(args.inputs[1]) as inf:
for line in inf:
vcf = VCF(line)
if not vcf.is_snp(): continue
g = vcf.get_phased_genotype()
if not g: continue
if vcf.value('chrom') not in alleles:
alleles[vcf.value('chrom')] = {}
if vcf.value('pos') in alleles[vcf.value('chrom')]:
sys.stderr.write("WARNING: seeing the same position twice.\n" +
line.rstrip() + "\n")
alleles[vcf.value('chrom')][vcf.value(
'pos')] = g # set our left and right
sys.stderr.write("Reading in the reference genome\n")
#ref = read_fasta_into_hash(args.reference_genome)
ref = read_fasta_into_hash(args.inputs[0])
res1 = []
res2 = []
p = None
sys.stderr.write("Introducing VCF changes to reference sequences\n")
# Pretty memory intesnive to so don't go with all possible threads
if args.threads > 1: p = Pool(processes=max(1, int(args.threads / 4)))
for chrom in ref:
# handle the case where there is no allele information
if chrom not in alleles:
r1q = Queue()
r1q.put([0, chrom, ref[chrom]])
res1.append(r1q)
r2q = Queue()
r2q.put([0, chrom, ref[chrom]])
res2.append(r2q)
elif args.threads > 1:
res1.append(
p.apply_async(adjust_reference_genome,
args=(alleles[chrom], ref[chrom], 0, chrom)))
res2.append(
p.apply_async(adjust_reference_genome,
args=(alleles[chrom], ref[chrom], 1, chrom)))
else:
r1q = Queue()
r1q.put(
adjust_reference_genome(alleles[chrom], ref[chrom], 0, chrom))
res1.append(r1q)
r2q = Queue()
r2q.put(
adjust_reference_genome(alleles[chrom], ref[chrom], 1, chrom))
res2.append(r2q)
if args.threads > 1:
p.close()
p.join()
# now we can fill reference 1 with all our new sequences
ref1 = {}
c1 = 0
for i in range(0, len(res1)):
res = res1[i].get()
c1 += res[0]
ref1[res[1]] = res[2]
# now we can fill reference 2 with all our new sequences
ref2 = {}
c2 = 0
for i in range(0, len(res2)):
res = res2[i].get()
c2 += res[0]
ref2[res[1]] = res[2]
sys.stderr.write("Made " + str(c1) + "|" + str(c2) +
" changes to the reference\n")
# Now ref1 and ref2 have are the diploid sources of the transcriptome
gpdnames = {}
txn1 = Transcriptome()
txn2 = Transcriptome()
txn1.set_reference_genome_dictionary(ref1)
txn2.set_reference_genome_dictionary(ref2)
#with open(args.transcripts_genepred) as inf:
with open(args.inputs[2]) as inf:
for line in inf:
if line[0] == '#': continue
txn1.add_genepred_line(line.rstrip())
txn2.add_genepred_line(line.rstrip())
gpd = GenePredEntry(line.rstrip())
gpdnames[gpd.value('name')] = gpd.value('gene_name')
# The transcriptomes are set but we dont' really need the references anymore
# Empty our big memory things
txn1.ref_hash = None
txn2.ref_hash = None
for chrom in ref1.keys():
del ref1[chrom]
for chrom in ref2.keys():
del ref2[chrom]
for chrom in ref.keys():
del ref[chrom]
if not args.locus_by_gene_name:
#[locus2name,name2locus] = get_loci(args.transcripts_genepred)
[locus2name, name2locus] = get_loci(args.inputs[2])
else: # set locus by gene name
sys.stderr.write("Organizing loci by gene name\n")
locus2name = {}
name2locus = {}
numname = {}
m = 0
for name in sorted(gpdnames):
gene = gpdnames[name]
if gene not in numname:
m += 1
numname[gene] = m
num = numname[gene]
if num not in locus2name:
locus2name[num] = set()
locus2name[num].add(name)
name2locus[name] = num
sys.stderr.write("Ended with " + str(len(locus2name.keys())) +
" loci\n")
if args.isoform_expression:
sys.stderr.write("Reading expression from a TSV\n")
with open(args.isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn1.add_expression(f[0], float(f[1]))
txn2.add_expression(f[0], float(f[1]))
elif args.cufflinks_isoform_expression:
sys.stderr.write("Using cufflinks expression\n")
cuffz = 0
with open(args.cufflinks_isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
cuffz += 1
sys.stderr.write(str(cuffz) + " cufflinks entries processed\r")
f = line.rstrip().split("\t")
txn1.add_expression_no_update(f[0], float(f[9]))
txn2.add_expression_no_update(f[0], float(f[9]))
txn1.update_expression()
txn2.update_expression()
sys.stderr.write("\n")
elif args.uniform_expression:
sys.stderr.write("Using uniform expression model\n")
else:
sys.stderr.write(
"Warning isoform expression not sepcified, using uniform expression model.\n"
)
# Now we have the transcriptomes set
rhos = {} # The ASE of allele 1 (the left side)
randos = {}
if args.seed:
random.seed(args.seed)
for z in locus2name:
randos[z] = random.random()
sys.stderr.write("Setting rho for each transcript\n")
# Lets set rho for ASE for each transcript
for tname in sorted(txn1.transcripts):
if args.ASE_identical or args.ASE_identical == 0:
rhos[tname] = float(args.ASE_identical)
elif args.ASE_isoform_random:
rhos[tname] = random.random()
else: # we must be on locus random
rhos[tname] = randos[name2locus[tname]]
#Now our dataset is set up
rbe = SimulationBasics.RandomBiallelicTranscriptomeEmitter(txn1, txn2)
rbe.gene_names = gpdnames
rbe.name2locus = name2locus
rbe.set_transcriptome1_rho(rhos)
return rbe
def main():
parser = argparse.ArgumentParser(description="Create a simulated RNA-seq dataset")
parser.add_argument('reference_genome',help="The reference genome.")
parser.add_argument('transcripts_genepred',help="A genepred file describing the transcripts. Each transcript name must be unique.")
group = parser.add_mutually_exclusive_group(required=True)
group.add_argument('--uniform_expression',action='store_true',help="Uniform distribution of transcript expression")
group.add_argument('--isoform_expression',help="The transcript expression in TSV format <Transcript name> tab <Expression>")
group.add_argument('--cufflinks_isoform_expression',help="The expression of the isoforms or - for a uniform distribution of transcript expression")
group2 = parser.add_mutually_exclusive_group()
group2.add_argument('--long_reads_only',action='store_true')
group2.add_argument('--short_reads_only',action='store_true')
group2.add_argument('--output',help="Directory name for output")
parser.add_argument('--short_read_count',type=int,default=10000,help="INT number of short reads")
parser.add_argument('--short_read_length',type=int,default=101,help="INT length of the short reads")
parser.add_argument('--long_read_count',type=int,default=4000,help="INT default number of long reads")
parser.add_argument('--no_errors',action='store_true')
parser.add_argument('--threads',type=int,default=1)
args = parser.parse_args()
if args.output:
args.output = args.output.rstrip('/')
fq_prof_pacbio_ccs95 = None
fq_prof_pacbio_subreads = None
fq_prof_illumina = None
if not args.no_errors:
fq_prof_pacbio_ccs95 = default_pacbio_ccs95()
fq_prof_pacbio_subreads = default_pacbio_subreads()
fq_prof_illumina = default_illumina()
ref = read_fasta_into_hash(args.reference_genome)
txn = Transcriptome()
txn.set_reference_genome_dictionary(ref)
with open(args.transcripts_genepred) as inf:
for line in inf:
if line[0]=='#': continue
txn.add_genepred_line(line.rstrip())
if args.isoform_expression:
sys.stderr.write("Reading expression from a TSV\n")
with open(args.isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn.add_expression(f[0],float(f[1]))
elif args.uniform_expression:
sys.stderr.write("Using uniform expression model\n")
elif args.cufflinks_isoform_expression:
sys.stderr.write("Using cufflinks expression\n")
with open(args.cufflinks_isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn.add_expression(f[0],float(f[9]))
sys.stderr.write("have transcriptome\n")
for n in txn.ref_hash.keys(): del txn.ref_hash[n]
rbe = SimulationBasics.RandomTranscriptomeEmitter(txn)
# Now we have the transcriptomes set
#Now our dataset is set up
if args.short_reads_only:
rbe.set_gaussian_fragmentation_default_hiseq()
for zi in range(0,args.short_read_count):
[name,seq] = rbe.emit_short_read(args.short_read_length)
if args.no_errors:
print "@SRSIM"+str(zi+1)
print seq
print "+"
print 'I'*len(seq)
else:
l1perm = fq_prof_illumina.create_fastq_and_permute_sequence(seq)
print "@SRSIM"+str(zi+1)
print l1perm['seq']
print "+"
print l1perm['qual']
return
if args.long_reads_only:
rbe.set_gaussian_fragmentation_default_pacbio()
for zi in range(0,args.long_read_count):
[name,seq] = rbe.emit_long_read()
if args.no_errors:
g = 'm150101_010101_11111_c111111111111111111_s1_p0/'+str(zi+1)+'/ccs'
print "@"+g
print seq
print "+"
print 'I'*len(seq)
else:
g = 'm150101_010101_11111_c111111111111111111_s1_p0/'+str(zi+1)+'/ccs'
seqperm = fq_prof_pacbio_ccs95.create_fastq_and_permute_sequence(seq)
print "@"+g
print seqperm['seq']
print "+"
print seqperm['qual']
return
if not os.path.exists(args.output):
os.makedirs(args.output)
rbe.set_gaussian_fragmentation_default_hiseq()
# Lets prepare to output now
sys.stderr.write("Sequencing short reads\n")
global left_handle
global right_handle
left_handle = gzip.open(args.output+"/SR_1.fq.gz",'wb')
right_handle = gzip.open(args.output+"/SR_2.fq.gz",'wb')
buffer_size = 10000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
z = 0
for i in range(0,args.short_read_count):
z = i+1
if z %1000==0: sys.stderr.write(str(z)+"\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_short_read_buffer(buffer[:],rbe,args,fq_prof_illumina)
do_short(v)
else:
p.apply_async(process_short_read_buffer,args=(buffer[:],rbe,args,fq_prof_illumina),callback=do_short)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_short_read_buffer(buffer[:],rbe,args,fq_prof_illumina)
do_short(v)
else:
p.apply_async(process_short_read_buffer,args=(buffer[:],rbe,args,fq_prof_illumina),callback=do_short)
buffer = []
if args.threads > 1:
p.close()
p.join()
global greport
of = open(args.output+"/SR_report.txt",'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]])+"\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing short reads\n")
left_handle.close()
right_handle.close()
# Now lets create the long read set
rbe.set_gaussian_fragmentation_default_pacbio()
sys.stderr.write("Sequencing ccs long reads\n")
global long_handle
long_handle = gzip.open(args.output+"/LR_ccs.fq.gz",'wb')
buffer_size = 1000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
z = 0
for i in range(0,args.long_read_count):
z = i+1
if z %100==0: sys.stderr.write(str(z)+"\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_long_reads(buffer[:],rbe,args,fq_prof_pacbio_ccs95,'ccs')
do_long(v)
else:
p.apply_async(process_long_reads,args=(buffer[:],rbe,args,fq_prof_pacbio_ccs95,'ccs'),callback=do_long)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_long_reads(buffer[:],rbe,args,fq_prof_pacbio_ccs95,'ccs')
do_long(v)
else:
p.apply_async(process_long_reads,args=(buffer[:],rbe,args,fq_prof_pacbio_ccs95,'ccs'),callback=do_long)
buffer = []
if args.threads > 1:
p.close()
p.join()
long_handle.close()
of = open(args.output+"/LR_ccs_report.txt",'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]])+"\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing ccs long reads\n")
sys.stderr.write("Sequencing long sub reads\n")
long_handle = gzip.open(args.output+"/LR_sub.fq.gz",'wb')
buffer_size = 1000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
for i in range(z,z+args.long_read_count):
z = i+1
if z %100==0: sys.stderr.write(str(z)+"\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_long_reads(buffer[:],rbe,args,fq_prof_pacbio_subreads,'sub')
do_long(v)
else:
p.apply_async(process_long_reads,args=(buffer[:],rbe,args,fq_prof_pacbio_subreads,'sub'),callback=do_long)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_long_reads(buffer[:],rbe,args,fq_prof_pacbio_subreads,'sub')
do_long(v)
else:
p.apply_async(process_long_reads,args=(buffer[:],rbe,args,fq_prof_pacbio_subreads,'sub'),callback=do_long)
buffer = []
if args.threads > 1:
p.close()
p.join()
long_handle.close()
of = open(args.output+"/LR_sub_report.txt",'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]])+"\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing long sub reads\n")
combo = {}
with open(args.output+"/SR_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name,express,left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
with open(args.output+"/LR_ccs_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name,express,left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
with open(args.output+"/LR_sub_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name,express,left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
of = open(args.output+"/LR_SR_combo_report.txt",'w')
for name in sorted(combo):
of.write(name+"\t"+combo[name]['express']+"\t"+str(combo[name]['left'])+"\n")
of.close()
def main():
parser = argparse.ArgumentParser(
description="Create a simulated RNA-seq dataset")
parser.add_argument('reference_genome', help="The reference genome.")
parser.add_argument(
'transcripts_genepred',
help=
"A genepred file describing the transcripts. Each transcript name must be unique."
)
group = parser.add_mutually_exclusive_group(required=True)
group.add_argument('--uniform_expression',
action='store_true',
help="Uniform distribution of transcript expression")
group.add_argument(
'--isoform_expression',
help=
"The transcript expression in TSV format <Transcript name> tab <Expression>"
)
group.add_argument(
'--cufflinks_isoform_expression',
help=
"The expression of the isoforms or - for a uniform distribution of transcript expression"
)
group2 = parser.add_mutually_exclusive_group()
group2.add_argument('--long_reads_only', action='store_true')
group2.add_argument('--short_reads_only', action='store_true')
group2.add_argument('--output', help="Directory name for output")
parser.add_argument('--short_read_count',
type=int,
default=10000,
help="INT number of short reads")
parser.add_argument('--short_read_length',
type=int,
default=101,
help="INT length of the short reads")
parser.add_argument('--long_read_count',
type=int,
default=4000,
help="INT default number of long reads")
parser.add_argument('--no_errors', action='store_true')
parser.add_argument('--threads', type=int, default=1)
args = parser.parse_args()
if args.output:
args.output = args.output.rstrip('/')
fq_prof_pacbio_ccs95 = None
fq_prof_pacbio_subreads = None
fq_prof_illumina = None
if not args.no_errors:
fq_prof_pacbio_ccs95 = default_pacbio_ccs95()
fq_prof_pacbio_subreads = default_pacbio_subreads()
fq_prof_illumina = default_illumina()
ref = read_fasta_into_hash(args.reference_genome)
txn = Transcriptome()
txn.set_reference_genome_dictionary(ref)
with open(args.transcripts_genepred) as inf:
for line in inf:
if line[0] == '#': continue
txn.add_genepred_line(line.rstrip())
if args.isoform_expression:
sys.stderr.write("Reading expression from a TSV\n")
with open(args.isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn.add_expression(f[0], float(f[1]))
elif args.uniform_expression:
sys.stderr.write("Using uniform expression model\n")
elif args.cufflinks_isoform_expression:
sys.stderr.write("Using cufflinks expression\n")
with open(args.cufflinks_isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn.add_expression(f[0], float(f[9]))
sys.stderr.write("have transcriptome\n")
for n in txn.ref_hash.keys():
del txn.ref_hash[n]
rbe = SimulationBasics.RandomTranscriptomeEmitter(txn)
# Now we have the transcriptomes set
#Now our dataset is set up
if args.short_reads_only:
rbe.set_gaussian_fragmentation_default_hiseq()
for zi in range(0, args.short_read_count):
[name, seq] = rbe.emit_short_read(args.short_read_length)
if args.no_errors:
print "@SRSIM" + str(zi + 1)
print seq
print "+"
print 'I' * len(seq)
else:
l1perm = fq_prof_illumina.create_fastq_and_permute_sequence(
seq)
print "@SRSIM" + str(zi + 1)
print l1perm['seq']
print "+"
print l1perm['qual']
return
if args.long_reads_only:
rbe.set_gaussian_fragmentation_default_pacbio()
for zi in range(0, args.long_read_count):
[name, seq] = rbe.emit_long_read()
if args.no_errors:
g = 'm150101_010101_11111_c111111111111111111_s1_p0/' + str(
zi + 1) + '/ccs'
print "@" + g
print seq
print "+"
print 'I' * len(seq)
else:
g = 'm150101_010101_11111_c111111111111111111_s1_p0/' + str(
zi + 1) + '/ccs'
seqperm = fq_prof_pacbio_ccs95.create_fastq_and_permute_sequence(
seq)
print "@" + g
print seqperm['seq']
print "+"
print seqperm['qual']
return
if not os.path.exists(args.output):
os.makedirs(args.output)
rbe.set_gaussian_fragmentation_default_hiseq()
# Lets prepare to output now
sys.stderr.write("Sequencing short reads\n")
global left_handle
global right_handle
left_handle = gzip.open(args.output + "/SR_1.fq.gz", 'wb')
right_handle = gzip.open(args.output + "/SR_2.fq.gz", 'wb')
buffer_size = 10000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
z = 0
for i in range(0, args.short_read_count):
z = i + 1
if z % 1000 == 0: sys.stderr.write(str(z) + "\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_short_read_buffer(buffer[:], rbe, args,
fq_prof_illumina)
do_short(v)
else:
p.apply_async(process_short_read_buffer,
args=(buffer[:], rbe, args, fq_prof_illumina),
callback=do_short)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_short_read_buffer(buffer[:], rbe, args,
fq_prof_illumina)
do_short(v)
else:
p.apply_async(process_short_read_buffer,
args=(buffer[:], rbe, args, fq_prof_illumina),
callback=do_short)
buffer = []
if args.threads > 1:
p.close()
p.join()
global greport
of = open(args.output + "/SR_report.txt", 'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]]) + "\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing short reads\n")
left_handle.close()
right_handle.close()
# Now lets create the long read set
rbe.set_gaussian_fragmentation_default_pacbio()
sys.stderr.write("Sequencing ccs long reads\n")
global long_handle
long_handle = gzip.open(args.output + "/LR_ccs.fq.gz", 'wb')
buffer_size = 1000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
z = 0
for i in range(0, args.long_read_count):
z = i + 1
if z % 100 == 0: sys.stderr.write(str(z) + "\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_long_reads(buffer[:], rbe, args,
fq_prof_pacbio_ccs95, 'ccs')
do_long(v)
else:
p.apply_async(process_long_reads,
args=(buffer[:], rbe, args, fq_prof_pacbio_ccs95,
'ccs'),
callback=do_long)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_long_reads(buffer[:], rbe, args, fq_prof_pacbio_ccs95,
'ccs')
do_long(v)
else:
p.apply_async(process_long_reads,
args=(buffer[:], rbe, args, fq_prof_pacbio_ccs95,
'ccs'),
callback=do_long)
buffer = []
if args.threads > 1:
p.close()
p.join()
long_handle.close()
of = open(args.output + "/LR_ccs_report.txt", 'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]]) + "\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing ccs long reads\n")
sys.stderr.write("Sequencing long sub reads\n")
long_handle = gzip.open(args.output + "/LR_sub.fq.gz", 'wb')
buffer_size = 1000
buffer = []
if args.threads > 1:
p = Pool(processes=args.threads)
for i in range(z, z + args.long_read_count):
z = i + 1
if z % 100 == 0: sys.stderr.write(str(z) + "\r")
buffer.append(z)
if len(buffer) >= buffer_size:
if args.threads <= 1:
v = process_long_reads(buffer[:], rbe, args,
fq_prof_pacbio_subreads, 'sub')
do_long(v)
else:
p.apply_async(process_long_reads,
args=(buffer[:], rbe, args,
fq_prof_pacbio_subreads, 'sub'),
callback=do_long)
buffer = []
if len(buffer) > 0:
if args.threads <= 1:
v = process_long_reads(buffer[:], rbe, args,
fq_prof_pacbio_subreads, 'sub')
do_long(v)
else:
p.apply_async(process_long_reads,
args=(buffer[:], rbe, args, fq_prof_pacbio_subreads,
'sub'),
callback=do_long)
buffer = []
if args.threads > 1:
p.close()
p.join()
long_handle.close()
of = open(args.output + "/LR_sub_report.txt", 'w')
for name in greport:
of.write("\t".join([str(x) for x in greport[name]]) + "\n")
of.close()
greport = {}
sys.stderr.write("\nFinished sequencing long sub reads\n")
combo = {}
with open(args.output + "/SR_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name, express, left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
with open(args.output + "/LR_ccs_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name, express, left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
with open(args.output + "/LR_sub_report.txt") as inf:
for line in inf:
f = line.rstrip().split("\t")
[name, express, left] = f
if name not in combo:
combo[name] = {}
combo[name]['express'] = express
combo[name]['left'] = 0
combo[name]['left'] += int(left)
of = open(args.output + "/LR_SR_combo_report.txt", 'w')
for name in sorted(combo):
of.write(name + "\t" + combo[name]['express'] + "\t" +
str(combo[name]['left']) + "\n")
of.close()
def load_from_inputs(args):
#Read in the VCF file
sys.stderr.write("Reading in the VCF file\n")
alleles = {}
#with open(args.phased_VCF) as inf:
with open(args.inputs[1]) as inf:
for line in inf:
vcf = VCF(line)
if not vcf.is_snp(): continue
g = vcf.get_phased_genotype()
if not g: continue
if vcf.value('chrom') not in alleles:
alleles[vcf.value('chrom')] = {}
if vcf.value('pos') in alleles[vcf.value('chrom')]:
sys.stderr.write("WARNING: seeing the same position twice.\n"+line.rstrip()+"\n")
alleles[vcf.value('chrom')][vcf.value('pos')] = g # set our left and right
sys.stderr.write("Reading in the reference genome\n")
#ref = read_fasta_into_hash(args.reference_genome)
ref = read_fasta_into_hash(args.inputs[0])
res1 = []
res2 = []
p = None
sys.stderr.write("Introducing VCF changes to reference sequences\n")
# Pretty memory intesnive to so don't go with all possible threads
if args.threads > 1: p = Pool(processes=max(1,int(args.threads/4)))
for chrom in ref:
# handle the case where there is no allele information
if chrom not in alleles:
r1q = Queue()
r1q.put([0,chrom,ref[chrom]])
res1.append(r1q)
r2q = Queue()
r2q.put([0,chrom,ref[chrom]])
res2.append(r2q)
elif args.threads > 1:
res1.append(p.apply_async(adjust_reference_genome,args=(alleles[chrom],ref[chrom],0,chrom)))
res2.append(p.apply_async(adjust_reference_genome,args=(alleles[chrom],ref[chrom],1,chrom)))
else:
r1q = Queue()
r1q.put(adjust_reference_genome(alleles[chrom],ref[chrom],0,chrom))
res1.append(r1q)
r2q = Queue()
r2q.put(adjust_reference_genome(alleles[chrom],ref[chrom],1,chrom))
res2.append(r2q)
if args.threads > 1:
p.close()
p.join()
# now we can fill reference 1 with all our new sequences
ref1 = {}
c1 = 0
for i in range(0,len(res1)):
res = res1[i].get()
c1 += res[0]
ref1[res[1]]=res[2]
# now we can fill reference 2 with all our new sequences
ref2 = {}
c2 = 0
for i in range(0,len(res2)):
res = res2[i].get()
c2 += res[0]
ref2[res[1]]=res[2]
sys.stderr.write("Made "+str(c1)+"|"+str(c2)+" changes to the reference\n")
# Now ref1 and ref2 have are the diploid sources of the transcriptome
gpdnames = {}
txn1 = Transcriptome()
txn2 = Transcriptome()
txn1.set_reference_genome_dictionary(ref1)
txn2.set_reference_genome_dictionary(ref2)
#with open(args.transcripts_genepred) as inf:
with open(args.inputs[2]) as inf:
for line in inf:
if line[0]=='#': continue
txn1.add_genepred_line(line.rstrip())
txn2.add_genepred_line(line.rstrip())
gpd = GenePredEntry(line.rstrip())
gpdnames[gpd.value('name')] = gpd.value('gene_name')
# The transcriptomes are set but we dont' really need the references anymore
# Empty our big memory things
txn1.ref_hash = None
txn2.ref_hash = None
for chrom in ref1.keys(): del ref1[chrom]
for chrom in ref2.keys(): del ref2[chrom]
for chrom in ref.keys(): del ref[chrom]
if not args.locus_by_gene_name:
#[locus2name,name2locus] = get_loci(args.transcripts_genepred)
[locus2name,name2locus] = get_loci(args.inputs[2])
else: # set locus by gene name
sys.stderr.write("Organizing loci by gene name\n")
locus2name = {}
name2locus = {}
numname = {}
m = 0
for name in sorted(gpdnames):
gene = gpdnames[name]
if gene not in numname:
m+=1
numname[gene] = m
num = numname[gene]
if num not in locus2name:
locus2name[num] = set()
locus2name[num].add(name)
name2locus[name] = num
sys.stderr.write("Ended with "+str(len(locus2name.keys()))+" loci\n")
if args.isoform_expression:
sys.stderr.write("Reading expression from a TSV\n")
with open(args.isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
f = line.rstrip().split("\t")
txn1.add_expression(f[0],float(f[1]))
txn2.add_expression(f[0],float(f[1]))
elif args.cufflinks_isoform_expression:
sys.stderr.write("Using cufflinks expression\n")
cuffz = 0
with open(args.cufflinks_isoform_expression) as inf:
line1 = inf.readline()
for line in inf:
cuffz +=1
sys.stderr.write(str(cuffz)+" cufflinks entries processed\r")
f = line.rstrip().split("\t")
txn1.add_expression_no_update(f[0],float(f[9]))
txn2.add_expression_no_update(f[0],float(f[9]))
txn1.update_expression()
txn2.update_expression()
sys.stderr.write("\n")
elif args.uniform_expression:
sys.stderr.write("Using uniform expression model\n")
else:
sys.stderr.write("Warning isoform expression not sepcified, using uniform expression model.\n")
# Now we have the transcriptomes set
rhos = {} # The ASE of allele 1 (the left side)
randos = {}
if args.seed:
random.seed(args.seed)
for z in locus2name: randos[z] = random.random()
sys.stderr.write("Setting rho for each transcript\n")
# Lets set rho for ASE for each transcript
for tname in sorted(txn1.transcripts):
if args.ASE_identical or args.ASE_identical == 0:
rhos[tname] = float(args.ASE_identical)
elif args.ASE_isoform_random:
rhos[tname] = random.random()
else: # we must be on locus random
rhos[tname] = randos[name2locus[tname]]
#Now our dataset is set up
rbe = SimulationBasics.RandomBiallelicTranscriptomeEmitter(txn1,txn2)
rbe.gene_names = gpdnames
rbe.name2locus = name2locus
rbe.set_transcriptome1_rho(rhos)
return rbe
