def get_cov(args, bases = 50000, splitsize = 1000):
"""function to get coverages
"""
if not args.out:
if args.bed is None:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
else:
args.out = '.'.join(os.path.basename(args.bed).split('.')[0:-1])
if args.bed is None:
chrs = read_chrom_sizes_from_bam(args.bam)
chunks = ChunkList.convertChromSizes(chrs, splitsize = splitsize)
sets = chunks.split(items = bases/splitsize)
else:
chunks = ChunkList.read(args.bed)
chunks.merge()
sets = chunks.split(bases = bases)
maxQueueSize = max(2,int(2 * bases / np.mean([chunk.length() for chunk in chunks])))
pool1 = mp.Pool(processes = max(1,args.cores-1))
out_handle = open(args.out + '.cov.bedgraph','w')
out_handle.close()
write_queue = mp.JoinableQueue(maxsize = maxQueueSize)
write_process = mp.Process(target = _writeCov, args=(write_queue, args.out))
write_process.start()
for j in sets:
tmp = pool1.map(_covHelper, zip(j,itertools.repeat(args)))
for track in tmp:
write_queue.put(track)
pool1.close()
pool1.join()
write_queue.put('STOP')
write_process.join()
pysam.tabix_compress(args.out + '.cov.bedgraph', args.out + '.cov.bedgraph.gz', force = True)
shell_command('rm ' + args.out + '.cov.bedgraph')
pysam.tabix_index(args.out + '.cov.bedgraph.gz', preset = "bed", force = True)
def make_bias_track(args, bases = 500000, splitsize = 1000):
"""function to compute bias track
"""
if args.out is None:
if args.bed is not None:
args.out = '.'.join(os.path.basename(args.bed).split('.')[0:-1])
else:
args.out = '.'.join(os.path.basename(args.fasta).split('.')[0:-1])
params = _BiasParams(args.fasta, args.pwm)
if args.bed is None:
chunks = ChunkList.convertChromSizes(params.chrs, splitsize = splitsize)
sets = chunks.split(items = bases/splitsize)
else:
chunks = ChunkList.read(args.bed)
chunks.merge()
sets = chunks.split(bases = bases)
maxQueueSize = max(2,int(2 * bases / np.mean([chunk.length() for chunk in chunks])))
pool = mp.Pool(processes = max(1,args.cores-1))
out_handle = open(args.out + '.Scores.bedgraph','w')
out_handle.close()
write_queue = mp.JoinableQueue(maxsize = maxQueueSize)
write_process = mp.Process(target = _writeBias, args=(write_queue, args.out))
write_process.start()
for j in sets:
tmp = pool.map(_biasHelper, zip(j,itertools.repeat(params)))
for track in tmp:
write_queue.put(track)
pool.close()
pool.join()
write_queue.put('STOP')
write_process.join()
pysam.tabix_compress(args.out + '.Scores.bedgraph', args.out + '.Scores.bedgraph.gz', force = True)
shell_command('rm ' + args.out + '.Scores.bedgraph')
pysam.tabix_index(args.out + '.Scores.bedgraph.gz', preset = "bed", force = True)
def get_pwm(args, bases=50000, splitsize=1000):
"""Functiono obtain PWM around ATAC insertion"""
if not args.out:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
chrs = read_chrom_sizes_from_fasta(args.fasta)
if args.bed is None:
chunks = ChunkList.convertChromSizes(chrs,
splitsize=splitsize,
offset=args.flank)
sets = chunks.split(items=bases / splitsize)
else:
chunks = ChunkList.read(args.bed,
chromDict=chrs,
min_offset=args.flank)
sets = chunks.split(bases=bases)
params = _PWMParameters(bam=args.bam,
up=args.flank,
down=args.flank,
fasta=args.fasta,
lower=args.lower,
upper=args.upper,
atac=args.atac,
sym=args.sym)
pool = Pool(processes=args.cores)
tmp = pool.map(_pwmHelper, zip(sets, itertools.repeat(params)))
pool.close()
pool.join()
n = 0.0
result = np.zeros((len(params.nucleotides), params.up + params.down + 1))
for i in tmp:
result += i[0]
n += i[1]
result /= n
if args.bed:
normfreqs = seq.getNucFreqsFromChunkList(chunks, args.fasta,
params.nucleotides)
else:
normfreqs = seq.getNucFreqs(args.fasta, params.nucleotides)
result = result / np.reshape(np.repeat(normfreqs, result.shape[1]),
result.shape)
if args.sym:
#Symmetrize
left = result[:, 0:(args.flank + 1)]
right = result[:, args.flank:]
rightflipped = np.fliplr(np.flipud(right))
combined = (left + rightflipped) / 2
result = np.hstack(
(combined, np.fliplr(np.flipud(combined[:, 0:args.flank]))))
#save
pwm = PWM(result, args.flank, args.flank, params.nucleotides)
pwm.save(args.out + '.PWM.txt')
def make_bias_track(args, bases=500000, splitsize=1000):
"""function to compute bias track
"""
if args.out is None:
if args.bed is not None:
args.out = '.'.join(os.path.basename(args.bed).split('.')[0:-1])
else:
args.out = '.'.join(os.path.basename(args.fasta).split('.')[0:-1])
params = _BiasParams(args.fasta, args.pwm)
if args.bed is None:
chunks = ChunkList.convertChromSizes(params.chrs, splitsize=splitsize)
sets = chunks.split(items=bases // splitsize)
else:
chunks = ChunkList.read(args.bed)
chunks.checkChroms(list(params.chrs.keys()))
chunks.merge()
sets = chunks.split(bases=bases)
maxQueueSize = max(
2, int(2 * bases / np.mean([chunk.length() for chunk in chunks])))
pool = mp.Pool(processes=max(1, args.cores - 1))
out_handle = open(args.out + '.Scores.bedgraph', 'w')
out_handle.close()
write_queue = mp.JoinableQueue(maxsize=maxQueueSize)
write_process = mp.Process(target=_writeBias, args=(write_queue, args.out))
write_process.start()
for j in sets:
tmp = pool.map(_biasHelper, list(zip(j, itertools.repeat(params))))
for track in tmp:
write_queue.put(track)
pool.close()
pool.join()
write_queue.put('STOP')
write_process.join()
pysam.tabix_compress(args.out + '.Scores.bedgraph',
args.out + '.Scores.bedgraph.gz',
force=True)
shell_command('rm ' + args.out + '.Scores.bedgraph')
pysam.tabix_index(args.out + '.Scores.bedgraph.gz',
preset="bed",
force=True)
def get_ins(args, bases=50000, splitsize=1000):
"""function to get insertions
"""
if not args.out:
if args.bed is None:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
else:
args.out = '.'.join(os.path.basename(args.bed).split('.')[0:-1])
if args.bed is None:
chrs = read_chrom_sizes_from_bam(args.bam)
chunks = ChunkList.convertChromSizes(chrs, splitsize=splitsize)
sets = chunks.split(items=bases / splitsize)
else:
chunks = ChunkList.read(args.bed)
chunks.merge()
sets = chunks.split(bases=bases)
maxQueueSize = max(
2, int(2 * bases / np.mean([chunk.length() for chunk in chunks])))
pool1 = mp.Pool(processes=max(1, args.cores - 1))
out_handle = open(args.out + '.ins.bedgraph', 'w')
out_handle.close()
write_queue = mp.JoinableQueue(maxsize=maxQueueSize)
write_process = mp.Process(target=_writeIns, args=(write_queue, args.out))
write_process.start()
for j in sets:
if args.smooth:
tmp = pool1.map(_insHelperSmooth,
list(zip(j, itertools.repeat(args))))
else:
tmp = pool1.map(_insHelper, list(zip(j, itertools.repeat(args))))
for track in tmp:
write_queue.put(track)
pool1.close()
pool1.join()
write_queue.put('STOP')
write_process.join()
pysam.tabix_compress(args.out + '.ins.bedgraph',
args.out + '.ins.bedgraph.gz',
force=True)
shell_command('rm ' + args.out + '.ins.bedgraph')
pysam.tabix_index(args.out + '.ins.bedgraph.gz', preset="bed", force=True)
def get_pwm(args, bases = 50000, splitsize = 1000):
"""Functiono obtain PWM around ATAC insertion"""
if not args.out:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
chrs = read_chrom_sizes_from_fasta(args.fasta)
if args.bed is None:
chunks = ChunkList.convertChromSizes(chrs, splitsize = splitsize, offset = args.flank)
sets = chunks.split(items = bases/splitsize)
else:
chunks = ChunkList.read(args.bed, chromDict = chrs, min_offset = args.flank)
sets = chunks.split(bases = bases)
params = _PWMParameters(bam = args.bam, up = args.flank, down = args.flank, fasta = args.fasta,
lower = args.lower, upper = args.upper, atac = args.atac, sym = args.sym)
pool = Pool(processes = args.cores)
tmp = pool.map(_pwmHelper, zip(sets,itertools.repeat(params)))
pool.close()
pool.join()
n = 0.0
result = np.zeros((len(params.nucleotides), params.up + params.down + 1))
for i in tmp:
result += i[0]
n += i[1]
result /= n
if args.bed:
normfreqs = seq.getNucFreqsFromChunkList(chunks, args.fasta, params.nucleotides)
else:
normfreqs = seq.getNucFreqs(args.fasta, params.nucleotides)
result = result / np.reshape(np.repeat(normfreqs,result.shape[1]),result.shape)
if args.sym:
#Symmetrize
left = result[:,0:(args.flank + 1)]
right = result[:,args.flank:]
rightflipped = np.fliplr(np.flipud(right))
combined = (left + rightflipped) / 2
result = np.hstack((combined, np.fliplr(np.flipud(combined[:,0:args.flank]))))
#save
pwm = PWM(result, args.flank, args.flank, params.nucleotides)
pwm.save(args.out + '.PWM.txt')
