def main(argv=None):
settings = process_command_line(argv)
if not os.path.exists(settings.dirout):
os.makedirs(settings.dirout)
if settings.genes_gff:
try:
pos_feat_list, all_features = RILseq.read_gtf(
open(settings.genes_gff), settings.feature, settings.identifier)
except IOError:
settings.genes_gff = None
gcounts = {}
lib_order = []
fastq_1_list = list(RILseq.flat_list(settings.fastq_1))
fastq_2_list = list(RILseq.flat_list(settings.fastq_2))
for i, r1_name in enumerate(RILseq.flat_list(settings.fastq_1)):
try:
r2_name = fastq_2_list[i]
except IndexError:
r2_name = None
outhead = r1_name.rsplit('.', 1)[0]
libname = outhead.rsplit('/',1)[-1]
outhead = '%s_bwa'%libname
bamname = RILseq.run_bwa(
settings.bwa_exec, r1_name, r2_name,
settings.dirout, outhead, settings.allowed_mismatches,
settings.genome_fasta, settings.params_aln, settings.sampe_params,
settings.samse_params, settings.samtools_cmd)
samfile = pysam.Samfile(bamname)
if settings.genes_gff:
lib_order.append(libname)
gcounts[libname] = RILseq.count_features(
pos_feat_list, samfile, settings.overlap,
rev=settings.reverse_complement)
if settings.create_wig:
outwigs = [open("%s/%s_coverage.wig"%(settings.dirout, fastq.split("_cutadapt")[0]), 'w')
for fastq in fastq_1_list]
coverage = RILseq.generate_wig(
samfile, rev=settings.reverse_complement, first_pos=False)
RILseq.print_wiggle(
coverage, "%s_single_fragments_coverage"%libname,
"%s single fragments coverage"%libname, outwigs[i])
# Print the table of counts
if settings.genes_gff:
outtables = [open("%s/%s_counts.txt"%(settings.dirout, fastq.split("_cutadapt")[0]), 'w')
for fastq in fastq_1_list]
for i, r1_name in enumerate(fastq_1_list):
outt = csv.writer(outtables[i], delimiter='\t')
outt.writerow(['Gene name'] + lib_order)
for g in sorted(list(all_features)):
row_out = [g]
for libn in lib_order:
row_out.append(gcounts[libn][g])
outt.writerow(row_out)
return 0 # success
def main(argv=None):
settings = process_command_line(argv)
if not os.path.exists(settings.dirout):
os.makedirs(settings.dirout)
outwig = open("%s/%s_coverage.wig"%(settings.dirout, settings.outhead), 'w')
if settings.genes_gff:
try:
pos_feat_list, all_features = RILseq.read_gtf(
open(settings.genes_gff), settings.feature, settings.identifier)
except IOError:
settings.genes_gff = None
gcounts = {}
lib_order = []
fastq_2_list = list(RILseq.flat_list(settings.fastq_2))
for i, r1_name in enumerate(RILseq.flat_list(settings.fastq_1)):
try:
r2_name = fastq_2_list[i]
except IndexError:
r2_name = None
outhead = r1_name.rsplit('.', 1)[0]
libname = outhead.rsplit('/',1)[-1]
outhead = '%s_bwa'%libname
bamname = RILseq.run_bwa(
settings.bwa_exec, r1_name, r2_name,
settings.dirout, outhead, settings.allowed_mismatches,
settings.genome_fasta, settings.params_aln, settings.sampe_params,
settings.samse_params, settings.samtools_cmd,
processors=settings.processors)
samfile = pysam.Samfile(bamname)
if settings.genes_gff:
lib_order.append(libname)
gcounts[libname] = RILseq.count_features(
pos_feat_list, samfile, settings.overlap,
rev=settings.reverse_complement)
coverage = RILseq.generate_wig(
samfile, rev=settings.reverse_complement, first_pos=False)
RILseq.print_wiggle(
coverage, "%s_single_fragments_coverage"%libname,
"%s single fragments coverage"%libname, outwig)
# Print the table of counts
if settings.genes_gff:
outtable = open(
"%s/%s_counts.txt"%(settings.dirout, settings.outhead), 'w')
outt = csv.writer(outtable, delimiter='\t')
outt.writerow(['Gene name'] + lib_order)
for g in sorted(list(all_features)):
row_out = [g]
for libn in lib_order:
row_out.append(gcounts[libn][g])
outt.writerow(row_out)
return 0 # success
def main(argv=None):
settings = process_command_line(argv)
if settings.ribozero and settings.bc_dir:
try:
# sys.stderr.write("{}\n".format(','.join(settings.rrna_list)))
uid_pos,_,_,_,_,rRNAs = RILseq.ecocyc_parser.read_genes_data(
settings.bc_dir, rRNA_prod=settings.rrna_list)
except IOError:
raise
rr_pos = []
chr_dict = dict(zip(
settings.BC_chrlist.split(',')[1::2],
settings.BC_chrlist.split(',')[0::2]))
for rrgene in rRNAs:
# Pad the position of the rRNA gene with the alignment length
rr_pos.append([chr_dict[uid_pos[rrgene][0]]] +\
[uid_pos[rrgene][1]-settings.length] +\
[uid_pos[rrgene][2]+settings.length] +\
[uid_pos[rrgene][3]])
# sys.stderr.write('\n'.join(rRNAs))
# print uid_pos.items()[:10]
# sys.stderr.write('\n'.join(' '.join(map(str,sl)) for sl in rr_pos))
else:
rr_pos = None
region_interactions, region_ints_as1, region_ints_as2, total_interactions=\
RILseq.read_reads_table(
open(settings.reads_in), settings.seglen, rr_pos, settings.only_singles)
# If all interactions are desired, skip the tests and report all
if settings.all_interactions:
interacting_regions = []
for reg1, r1data in region_interactions.items():
for reg2, clist in r1data.items():
interacting_regions.append(
(1, len(clist), 0, reg1[0], reg1[0]+settings.seglen,
reg1[1], reg1[2], reg2[0], reg2[0]+settings.seglen,
reg2[1], reg2[2], 0, 0, 0))
else:
# Now run the test for each pair of interacting regions
found_in_interaction = defaultdict(bool)
interacting_regions = []
# Start with the regions with the most interactions
pairs_num = {}
for reg1 in list(region_interactions.keys()):
if region_ints_as1[reg1] < settings.min_int:
continue
for reg2 in list(region_interactions[reg1].keys()):
if len(region_interactions[reg1][reg2]) >= settings.min_int:
pairs_num[(reg1, reg2)] = len(region_interactions[reg1][reg2])
# Iterate the list of regions from the pairs with many interactions down
for (reg1, reg2) in sorted(pairs_num, key=pairs_num.get, reverse=True):
pv, ints, odds, r1_from, r1_to, r2_from, r2_to, mat_b, mat_c,mat_d=\
RILseq.minpv_regions(
reg1, reg2, region_interactions, region_ints_as1,
region_ints_as2, total_interactions, found_in_interaction,
settings.seglen, settings.maxsegs, settings.min_odds_ratio)
pv *= len(pairs_num)
if pv <= settings.max_pv:
# Mark as participating
for r1 in range(r1_from, r1_to, settings.seglen):
for r2 in range(r2_from, r2_to, settings.seglen):
found_in_interaction[
(r1, reg1[1], reg1[2], r2, reg2[1], reg2[2])] = True
# Report this interaction
interacting_regions.append(
(pv, ints, odds, r1_from, r1_to, reg1[1], reg1[2], r2_from,
r2_to, reg2[1], reg2[2], mat_b, mat_c, mat_d))
# Read the number of total RNAs in each region if the bam file is given
sum_reads=0
if settings.total_RNA:
# prepare a dictionary of features
feat_dict = defaultdict(lambda: defaultdict(set))
for region1 in region_ints_as1:
for i in range(settings.seglen):
feat_dict[region1[2]+region1[1]][region1[0]+i].add(region1)
for region2 in region_ints_as2:
for i in range(settings.seglen):
feat_dict[region2[2]+region2[1]][region2[0]+i].add(region2)
feat_list = {}
for chrom, data in feat_dict.items():
maxpos = max(data.keys())
list_of_sets = []
for k in range(maxpos+1):
list_of_sets.append(list(data[k]))
feat_list[chrom] = list_of_sets
totRNA_counts = defaultdict(int)
sum_reads = 0
for bamfile in settings.total_RNA.split(','):
saminf = pysam.Samfile(bamfile)
totcounts, sum_of_counts_lib = RILseq.count_features(
feat_list, saminf, 5,
rev=settings.total_reverse, get_sum=True)
for k, v in totcounts.items():
totRNA_counts[k] += v
sum_reads += sum_of_counts_lib
# Collect all the ratios between IP and total then choose the 90%
# percentile to avoid liers
max_IP_div_total_as1 = []
max_IP_div_total_as2 = []
for reg, counts in totRNA_counts.items():
if counts > settings.min_total_counts:
counts = float(counts+1)
# div_prod = max(region_ints_as1[reg]/counts, region_ints_as2[reg]/counts)
max_IP_div_total_as1.append(region_ints_as1[reg]/counts)
max_IP_div_total_as2.append(region_ints_as2[reg]/counts)
# mm1_sorted = sorted(max_IP_div_total_as1)
mm_sorted = sorted(max_IP_div_total_as2+max_IP_div_total_as1)
max_IP_div_total = mm_sorted[
int(len(mm_sorted)*settings.norm_percentile)]
sys.stderr.write("%f\n"%(max_IP_div_total))
else:
totRNA_counts = defaultdict(int)
max_IP_div_total = 0
if (settings.shuffles ==0 and settings.run_RNAup):
settings.shuffles=-1
# Read the additional data to decorate the results with
RILseq.report_interactions(
region_interactions, sys.stdout, interacting_regions, settings.seglen,
settings.bc_dir, settings.genome, settings.BC_chrlist, settings.refseq_dir,
settings.targets_file, settings.rep_table, settings.single_counts,
settings.shuffles, settings.RNAup_cmd, settings.servers,
settings.length, settings.est_utr_lens, settings.pad_seqs,
totRNA_counts, max_IP_div_total, total_interactions, sum_reads, settings.linear_chromosome_list)
return 0 # success
def main(argv=None):
settings = process_command_line(argv)
if settings.ribozero:
try:
uid_pos,_,_,_,_,rRNAs = RILseq.ecocyc_parser.read_genes_data(
settings.ec_dir)
except IOError:
raise
rr_pos = []
chr_dict = dict(zip(
settings.EC_chrlist.split(',')[1::2],
settings.EC_chrlist.split(',')[0::2]))
for rrgene in rRNAs:
# Pad the position of the rRNA gene with the alignment lenght
rr_pos.append([chr_dict[uid_pos[rrgene][0]]] +\
[uid_pos[rrgene][1]-settings.length] +\
[uid_pos[rrgene][2]+settings.length] +\
[uid_pos[rrgene][3]])
else:
rr_pos = None
region_interactions, region_ints_as1, region_ints_as2, total_interactions=\
RILseq.read_reads_table(
open(settings.reads_in), settings.seglen, rr_pos, settings.only_singles)
# If all interactions are desired, skip the tests and report all
if settings.all_interactions:
interacting_regions = []
for reg1, r1data in region_interactions.items():
for reg2, clist in r1data.items():
interacting_regions.append(
(1, len(clist), 0, reg1[0], reg1[0]+settings.seglen,
reg1[1], reg1[2], reg2[0], reg2[0]+settings.seglen,
reg2[1], reg2[2], 0, 0, 0))
else:
# Now run the test for each pair of interacting regions
found_in_interaction = defaultdict(bool)
interacting_regions = []
# Start with the regions with the most interactions
pairs_num = {}
for reg1 in list(region_interactions.keys()):
if region_ints_as1[reg1] < settings.min_int:
continue
for reg2 in list(region_interactions[reg1].keys()):
if len(region_interactions[reg1][reg2]) >= settings.min_int:
pairs_num[(reg1, reg2)] = len(region_interactions[reg1][reg2])
# Iterate the list of regions from the pairs with many interactions down
for (reg1, reg2) in sorted(pairs_num, key=pairs_num.get, reverse=True):
pv, ints, odds, r1_from, r1_to, r2_from, r2_to, mat_b, mat_c,mat_d=\
RILseq.minpv_regions(
reg1, reg2, region_interactions, region_ints_as1,
region_ints_as2, total_interactions, found_in_interaction,
settings.seglen, settings.maxsegs, settings.min_odds_ratio)
pv *= len(pairs_num)
if pv <= settings.max_pv:
# Mark as participating
for r1 in range(r1_from, r1_to, settings.seglen):
for r2 in range(r2_from, r2_to, settings.seglen):
found_in_interaction[
(r1, reg1[1], reg1[2], r2, reg2[1], reg2[2])] = True
# Report this interaction
interacting_regions.append(
(pv, ints, odds, r1_from, r1_to, reg1[1], reg1[2], r2_from,
r2_to, reg2[1], reg2[2], mat_b, mat_c, mat_d))
# Read the number of total RNAs in each region if the bam file is given
if settings.total_RNA:
# prepare a dictionary of features
feat_dict = defaultdict(lambda: defaultdict(set))
for region1 in region_ints_as1:
for i in range(settings.seglen):
feat_dict[region1[2]+region1[1]][region1[0]+i].add(region1)
for region2 in region_ints_as2:
for i in range(settings.seglen):
feat_dict[region2[2]+region2[1]][region2[0]+i].add(region2)
feat_list = {}
for chrom, data in feat_dict.items():
maxpos = max(data.keys())
list_of_sets = []
for k in range(maxpos+1):
list_of_sets.append(list(data[k]))
feat_list[chrom] = list_of_sets
totRNA_counts = defaultdict(int)
for bamfile in settings.total_RNA.split(','):
totcounts = RILseq.count_features(
feat_list, pysam.Samfile(bamfile), 5,
rev=settings.total_reverse)
for k, v in totcounts.items():
totRNA_counts[k] += v
# Collect all the ratios between IP and total then choose the 90%
# percentile to avoid liers
max_IP_div_total_as1 = []
max_IP_div_total_as2 = []
for reg, counts in totRNA_counts.items():
if counts > settings.min_total_counts:
counts = float(counts+1)
# div_prod = max(region_ints_as1[reg]/counts, region_ints_as2[reg]/counts)
max_IP_div_total_as1.append(region_ints_as1[reg]/counts)
max_IP_div_total_as2.append(region_ints_as2[reg]/counts)
# mm1_sorted = sorted(max_IP_div_total_as1)
mm_sorted = sorted(max_IP_div_total_as2+max_IP_div_total_as1)
max_IP_div_total = mm_sorted[
int(len(mm_sorted)*settings.norm_percentile)]
sys.stderr.write("%f\n"%(max_IP_div_total))
else:
totRNA_counts = defaultdict(int)
max_IP_div_total = 0
# Read the additional data to decorate the results with
RILseq.report_interactions(
region_interactions, sys.stdout, interacting_regions, settings.seglen,
settings.ec_dir, settings.EC_chrlist, settings.refseq_dir,
settings.targets_file, settings.rep_table, settings.single_counts,
settings.shuffles, settings.RNAup_cmd, settings.servers,
settings.length, settings.est_utr_lens, settings.pad_seqs,
totRNA_counts, max_IP_div_total)
return 0 # success
