def test_18_filter_reads(self):
if ONLY and ONLY != '18':
return
if CHKTIME:
t0 = time()
for ali in ['map', 'sam']:
seed(1)
if 13436 == int(random()*100000):
same_seed = True
genome = generate_random_ali(ali)
genome_bis = parse_fasta('test.fa~', verbose=False)
self.assertEqual(genome, genome_bis)
else:
same_seed = False
genome = parse_fasta('test.fa~')
# PARSE SAM
if ali == 'map':
from pytadbit.parsers.map_parser import parse_map as parser
else:
try:
from pytadbit.parsers.sam_parser import parse_sam as parser
except ImportError:
print 'ERROR: PYSAM not found, skipping test\n'
continue
parser(['test_read1.%s~' % (ali)], ['test_read2.%s~' % (ali)],
'./lala1-%s~' % (ali), './lala2-%s~' % (ali), genome,
re_name='DPNII', mapper='GEM')
# GET INTERSECTION
from pytadbit.mapping import get_intersection
get_intersection('lala1-%s~' % (ali), 'lala2-%s~' % (ali),
'lala-%s~' % (ali))
# FILTER
masked = filter_reads('lala-%s~' % (ali), verbose=False,
fast=(ali=='map'))
self.assertEqual(masked[1]['reads'], 1000)
self.assertEqual(masked[2]['reads'], 1000)
self.assertEqual(masked[3]['reads'], 1000)
self.assertEqual(masked[4]['reads'], 1000)
if same_seed:
self.assertEqual(masked[5]['reads'], 1110)
self.assertEqual(masked[6]['reads'], 2332)
self.assertEqual(masked[7]['reads'], 0)
self.assertEqual(masked[8]['reads'], 141)
self.assertEqual(masked[10]['reads'], 1)
else:
self.assertTrue (masked[5]['reads'] > 1000)
self.assertEqual(masked[9]['reads'], 1000)
apply_filter('lala-map~', 'lala-map-filt~', masked, filters=[1],
reverse=True, verbose=False)
self.assertEqual(len([True for l in open('lala-map-filt~')
if not l.startswith('#')]), 1000)
d = plot_iterative_mapping('lala1-map~', 'lala2-map~')
self.assertEqual(d[0][1], 6000)
if CHKTIME:
self.assertEqual(True, True)
print '18', time() - t0
def run(opts):
check_options(opts)
launch_time = time.localtime()
fname1, fname2 = load_parameters_fromdb(opts)
param_hash = digest_parameters(opts)
reads = path.join(opts.workdir, '03_filtered_reads',
'all_r1-r2_intersection_%s.tsv' % param_hash)
mreads = path.join(opts.workdir, '03_filtered_reads',
'valid_r1-r2_intersection_%s.tsv' % param_hash)
if not opts.resume:
mkdir(path.join(opts.workdir, '03_filtered_reads'))
# compute the intersection of the two read ends
print 'Getting intersection between read 1 and read 2'
count, multiples = get_intersection(fname1, fname2, reads)
# compute insert size
print 'Get insert size...'
hist_path = path.join(opts.workdir,
'histogram_fragment_sizes_%s.pdf' % param_hash)
median, max_f, mad = insert_sizes(
reads, nreads=1000000, stats=('median', 'first_decay', 'MAD'),
savefig=hist_path)
print ' - median insert size =', median
print ' - double median absolution of insert size =', mad
print ' - max insert size (when a gap in continuity of > 10 bp is found in fragment lengths) =', max_f
max_mole = max_f # pseudo DEs
min_dist = max_f + mad # random breaks
print (' Using the maximum continuous fragment size'
'(%d bp) to check '
'for pseudo-dangling ends') % max_mole
print (' Using maximum continuous fragment size plus the MAD '
'(%d bp) to check for random breaks') % min_dist
print "identify pairs to filter..."
masked = filter_reads(reads, max_molecule_length=max_mole,
over_represented=opts.over_represented,
max_frag_size=opts.max_frag_size,
min_frag_size=opts.min_frag_size,
re_proximity=opts.re_proximity,
min_dist_to_re=min_dist, fast=True)
n_valid_pairs = apply_filter(reads, mreads, masked,
filters=opts.apply)
finish_time = time.localtime()
print median, max_f, mad
# save all job information to sqlite DB
save_to_db(opts, count, multiples, reads, mreads, n_valid_pairs, masked,
hist_path, median, max_f, mad, launch_time, finish_time)
def test_18_filter_reads(self):
if ONLY and ONLY != "18":
return
if CHKTIME:
t0 = time()
for ali in ["map", "sam"]:
seed(1)
if 13436 == int(random() * 100000):
same_seed = True
genome = generate_random_ali(ali)
genome_bis = parse_fasta("test.fa~", verbose=False)
self.assertEqual(genome, genome_bis)
else:
same_seed = False
genome = parse_fasta("test.fa~")
# PARSE SAM
if ali == "map":
from pytadbit.parsers.map_parser import parse_map as parser
else:
try:
from pytadbit.parsers.sam_parser import parse_sam as parser
except ImportError:
print "ERROR: PYSAM not found, skipping test\n"
continue
parser(
["test_read1.%s~" % (ali)],
["test_read2.%s~" % (ali)],
"./lala1-%s~" % (ali),
"./lala2-%s~" % (ali),
genome,
re_name="DPNII",
mapper="GEM",
)
# GET INTERSECTION
from pytadbit.mapping import get_intersection
get_intersection("lala1-%s~" % (ali), "lala2-%s~" % (ali), "lala-%s~" % (ali))
# FILTER
masked = filter_reads("lala-%s~" % (ali), verbose=False, fast=(ali == "map"))
self.assertEqual(masked[1]["reads"], 1000)
self.assertEqual(masked[2]["reads"], 1000)
self.assertEqual(masked[3]["reads"], 1000)
self.assertEqual(masked[4]["reads"], 1000)
if same_seed:
self.assertEqual(masked[5]["reads"], 1110)
self.assertEqual(masked[6]["reads"], 2332)
self.assertEqual(masked[7]["reads"], 0)
self.assertEqual(masked[8]["reads"], 141)
self.assertEqual(masked[10]["reads"], 1)
else:
self.assertTrue(masked[5]["reads"] > 1000)
self.assertEqual(masked[9]["reads"], 1000)
apply_filter("lala-map~", "lala-map-filt~", masked, filters=[1], reverse=True, verbose=False)
self.assertEqual(len([True for l in open("lala-map-filt~") if not l.startswith("#")]), 1000)
d = plot_iterative_mapping("lala1-map~", "lala2-map~")
self.assertEqual(d[0][1], 6000)
if CHKTIME:
self.assertEqual(True, True)
print "18", time() - t0
def run(opts):
check_options(opts)
launch_time = time.localtime()
fname1, fname2 = load_parameters_fromdb(opts)
param_hash = digest_parameters(opts)
reads = path.join(opts.workdir, '03_filtered_reads',
'all_r1-r2_intersection_%s.tsv' % param_hash)
mreads = path.join(opts.workdir, '03_filtered_reads',
'valid_r1-r2_intersection_%s.tsv' % param_hash)
if not opts.resume:
mkdir(path.join(opts.workdir, '03_filtered_reads'))
# compute the intersection of the two read ends
print 'Getting intersection between read 1 and read 2'
count, multiples = get_intersection(fname1, fname2, reads)
# compute insert size
print 'Get insert size...'
hist_path = path.join(opts.workdir,
'histogram_fragment_sizes_%s.pdf' % param_hash)
median, max_f, mad = fragment_size(reads,
nreads=1000000,
stats=('median', 'first_decay',
'MAD'),
savefig=hist_path)
print ' - median insert size =', median
print ' - double median absolution of insert size =', mad
print ' - max insert size (when a gap in continuity of > 10 bp is found in fragment lengths) =', max_f
max_mole = max_f # pseudo DEs
min_dist = max_f + mad # random breaks
print(
' Using the maximum continuous fragment size'
'(%d bp) to check '
'for pseudo-dangling ends') % max_mole
print(
' Using maximum continuous fragment size plus the MAD '
'(%d bp) to check for random breaks') % min_dist
print "identify pairs to filter..."
masked = filter_reads(reads,
max_molecule_length=max_mole,
over_represented=opts.over_represented,
max_frag_size=opts.max_frag_size,
min_frag_size=opts.min_frag_size,
re_proximity=opts.re_proximity,
min_dist_to_re=min_dist,
fast=True)
n_valid_pairs = apply_filter(reads, mreads, masked, filters=opts.apply)
outbam = path.join(opts.workdir, '03_filtered_reads',
'intersection_%s' % param_hash)
if opts.valid:
infile = mreads
else:
infile = reads
bed2D_to_BAMhic(infile,
opts.valid,
opts.cpus,
outbam,
opts.format,
masked,
samtools=opts.samtools)
finish_time = time.localtime()
print median, max_f, mad
# save all job information to sqlite DB
save_to_db(opts, count, multiples, reads, mreads, n_valid_pairs, masked,
outbam + '.bam', hist_path, median, max_f, mad, launch_time,
finish_time)
def test_18_filter_reads(self):
if ONLY and not "18" in ONLY:
return
if CHKTIME:
t0 = time()
for ali in ["map", "sam"]:
seed(1)
if 13436 == int(random() * 100000):
same_seed = True
genome = generate_random_ali(ali)
genome_bis = parse_fasta("test.fa~",
verbose=False,
save_cache=False)
self.assertEqual(genome, genome_bis)
else:
same_seed = False
genome = parse_fasta("test.fa~", save_cache=False)
# PARSE SAM
if ali == "map":
from pytadbit.parsers.map_parser import parse_map as parser
else:
try:
from pytadbit.parsers.sam_parser import parse_sam as parser
except ImportError:
print("ERROR: PYSAM not found, skipping test\n")
continue
parser(["test_read1.%s~" % (ali)], ["test_read2.%s~" % (ali)],
"./lala1-%s~" % (ali),
"./lala2-%s~" % (ali),
genome,
re_name="DPNII",
mapper="GEM")
# GET INTERSECTION
from pytadbit.mapping import get_intersection
get_intersection("lala1-%s~" % (ali), "lala2-%s~" % (ali),
"lala-%s~" % (ali))
# FILTER
masked = filter_reads("lala-%s~" % (ali),
verbose=False,
fast=(ali == "map"))
self.assertEqual(masked[1]["reads"], 1000)
self.assertEqual(masked[2]["reads"], 1000)
self.assertEqual(masked[3]["reads"], 1000)
self.assertEqual(masked[4]["reads"], 1000)
if same_seed:
self.assertEqual(masked[5]["reads"], 1091)
self.assertEqual(masked[6]["reads"], 2230)
self.assertEqual(masked[7]["reads"], 0)
self.assertEqual(masked[8]["reads"], 100)
self.assertEqual(masked[10]["reads"], 5)
else:
self.assertTrue(masked[5]["reads"] > 1000)
self.assertEqual(masked[9]["reads"], 1001)
apply_filter("lala-map~",
"lala-map-filt~",
masked,
filters=[1],
reverse=True,
verbose=False)
with open("lala-map-filt~") as f_lala_filt:
self.assertEqual(
len([True for l in f_lala_filt if not l.startswith("#")]),
1000)
d = plot_iterative_mapping("lala1-map~", "lala2-map~")
self.assertEqual(d[0][1], 6000)
if CHKTIME:
self.assertEqual(True, True)
print("18", time() - t0)
from pytadbit.parsers.map_parser import parse_map
from pytadbit.parsers.genome_parser import parse_fasta
from pytadbit.mapping import get_intersection
# Load the genome
genome_seq = parse_fasta(fasta)
# Output directory
RESULTS = '%s/results/%s/processed_reads' % (SAMPLE, version)
if not os.path.exists(RESULTS):
os.makedirs(RESULTS)
infiles = []
outfiles = []
for infile in [paired1, paired2]:
bname = infile.split("/")[-1].replace(".fastq.gz", "")
maps = glob.glob('%s/%s/*' % (MAP_DIR, bname))
infiles.append(maps)
outfiles.append('%s/%s_map.tsv' % (RESULTS, bname))
parse_map(infiles[0],
infiles[1],
outfiles[0],
outfiles[1],
genome_seq,
restriction_enzyme,
verbose=True,
ncpus=slots)
final_output = outfiles[0].replace('read1', 'both')
get_intersection(outfiles[0], outfiles[1], final_output, verbose=True)
def tb_parse_mapping_iter(self, genome_seq, enzyme_name, window1_1,
window1_2, window1_3, window1_4, window2_1,
window2_2, window2_3, window2_4, reads):
"""
Function to map the aligned reads and return the matching pairs
Parameters
----------
genome_seq : dict
Object containing the sequence of each of the chromosomes
enzyme_name : str
Name of the enzyme used to digest the genome
window1_1 : str
Location of the first window index file
window1_2 : str
Location of the second window index file
window1_3 : str
Location of the third window index file
window1_4 : str
Location of the fourth window index file
window2_1 : str
Location of the first window index file
window2_2 : str
Location of the second window index file
window2_3 : str
Location of the third window index file
window2_4 : str
Location of the fourth window index file
reads : str
Location of the reads thats that has a matching location at both
ends of the paired reads
Returns
-------
reads : str
Location of the intersection of mapped reads that have matching
reads in both pair end files
"""
reads1 = reads + '_reads_1.tsv'
reads2 = reads + '_reads_2.tsv'
reads_both = reads + '_reads_both.tsv'
parse_map(
[window1_1, window1_2, window1_3, window1_4],
[window2_1, window2_2, window2_3, window2_4],
out_file1=reads1,
out_file2=reads2,
genome_seq=genome_seq,
re_name=enzyme_name,
verbose=True,
# ncpus=32
)
get_intersection(reads1, reads2, reads_both, verbose=True)
with open(reads, "wb") as f_out:
with open(reads_both, "rb") as f_in:
f_out.write(f_in.read())
return True
def run(opts):
check_options(opts)
launch_time = time.localtime()
fname1, fname2 = load_parameters_fromdb(opts)
param_hash = digest_parameters(opts)
reads = path.join(opts.workdir, '03_filtered_reads',
'all_r1-r2_intersection_%s.tsv' % param_hash)
mreads = path.join(opts.workdir, '03_filtered_reads',
'valid_r1-r2_intersection_%s.tsv' % param_hash)
if not opts.resume:
mkdir(path.join(opts.workdir, '03_filtered_reads'))
if opts.fast_fragment:
reads = fname1
counts_multis = [
'#' in line.split('\t')[0] for line in open(reads)
]
count = len(counts_multis)
multiples = {}
multiples[1] = sum(
[count_mult for count_mult in counts_multis if count_mult])
del counts_multis
else:
# compute the intersection of the two read ends
print('Getting intersection between read 1 and read 2')
count, multiples = get_intersection(fname1,
fname2,
reads,
compress=opts.compress_input)
# compute insert size
print('Get insert size...')
hist_path = path.join(opts.workdir,
'histogram_fragment_sizes_%s.pdf' % param_hash)
try:
median, max_f, mad = fragment_size(reads,
nreads=1000000,
stats=('median', 'first_decay',
'MAD'),
savefig=hist_path)
except ZeroDivisionError:
warn('WARNING: cannot compute fragment length, too few '
'dangling-ends. Setting median length to 400 nt.')
median, max_f, mad = 400, 100, 1000
print(' - median insert size =', median)
print(' - double median absolution of insert size =', mad)
print(
' - max insert size (when a gap in continuity of > 10 bp is found in fragment lengths) =',
max_f)
max_mole = max_f # pseudo DEs
min_dist = max_f + mad # random breaks
print(' Using the maximum continuous fragment size'
'(%d bp) to check '
'for pseudo-dangling ends' % max_mole)
print(' Using maximum continuous fragment size plus the MAD '
'(%d bp) to check for random breaks' % min_dist)
print("identify pairs to filter...")
masked = filter_reads(reads,
max_molecule_length=max_mole,
over_represented=opts.over_represented,
max_frag_size=opts.max_frag_size,
min_frag_size=opts.min_frag_size,
re_proximity=opts.re_proximity,
strict_duplicates=opts.strict_duplicates,
min_dist_to_re=min_dist,
fast=True)
n_valid_pairs = apply_filter(reads, mreads, masked, filters=opts.apply)
outbam = path.join(opts.workdir, '03_filtered_reads',
'intersection_%s' % param_hash)
if opts.valid:
infile = mreads
else:
infile = reads
bed2D_to_BAMhic(infile,
opts.valid,
opts.cpus,
outbam,
opts.format,
masked,
samtools=opts.samtools)
finish_time = time.localtime()
print(median, max_f, mad)
# save all job information to sqlite DB
save_to_db(opts, count, multiples, reads, mreads, n_valid_pairs, masked,
outbam + '.bam', hist_path, median, max_f, mad, launch_time,
finish_time)
