def run_diff(args, bases=500000):
"""run differential occupancy calling
"""
chrs = read_chrom_sizes_from_bam(args.bam)
pwm = PWM.open(args.pwm)
chunks = ChunkList.read(args.bed,
chromDict=chrs,
min_offset=args.flank + args.upper / 2 +
max(pwm.up, pwm.down))
chunks.merge()
maxQueueSize = max(
2, int(100 * bases / np.mean([chunk.length() for chunk in chunks])))
#get fragmentsizes
fragment_dist1 = FragmentMixDistribution(0, upper=args.upper)
fragment_dist1.fragmentsizes = FragmentSizes(
0, args.upper, vals=FragmentSizes.open(args.sizes1).get(0, args.upper))
fragment_dist1.modelNFR()
fragment_dist2 = FragmentMixDistribution(0, upper=args.upper)
fragment_dist2.fragmentsizes = FragmentSizes(
0, args.upper, vals=FragmentSizes.open(args.sizes2).get(0, args.upper))
fragment_dist2.modelNFR()
params = OccupancyParameters(fragment_dist,
args.upper,
args.fasta,
args.pwm,
sep=args.nuc_sep,
min_occ=args.min_occ,
flank=args.flank,
bam=args.bam,
ci=args.confidence_interval)
sets = chunks.split(bases=bases)
pool1 = mp.Pool(processes=max(1, args.cores - 1))
diff_handle = open(args.out + '.occdiff.bed', 'w')
diff_handle.close()
diff_queue = mp.JoinableQueue()
diff_process = mp.Process(target=_writeDiff, args=(diff_queue, args.out))
diff_process.start()
nuc_dist = np.zeros(args.upper)
for j in sets:
tmp = pool1.map(_occHelper, zip(j, itertools.repeat(params)))
for result in tmp:
diff_queue.put(result[1])
pool1.close()
pool1.join()
diff_queue.put('STOP')
diff_process.join()
pysam.tabix_compress(args.out + '.occdiff.bed',
args.out + '.occdiff.bed.gz',
force=True)
shell_command('rm ' + args.out + '.occdiff.bed')
pysam.tabix_index(args.out + '.occdiff.bed.gz', preset="bed", force=True)
def setUp(self):
"""setup Test_occupancy class by establishing parameters"""
self.fragment_dist = FragmentMixDistribution(0, 3)
self.fragment_dist.nfr_fit = FragmentSizes(0,
3,
vals=np.array(
[0.5, 0.49, 0.01]))
self.fragment_dist.nuc_fit = FragmentSizes(0,
3,
vals=np.array(
[0.01, 0.49, 0.5]))
self.params = OccupancyCalcParams(0, 3, self.fragment_dist)
def run_diff(args, bases = 500000):
"""run differential occupancy calling
"""
chrs = read_chrom_sizes_from_bam(args.bam)
pwm = PWM.open(args.pwm)
chunks = ChunkList.read(args.bed, chromDict = chrs, min_offset = args.flank + args.upper/2 + max(pwm.up,pwm.down))
chunks.merge()
maxQueueSize = max(2,int(100 * bases / np.mean([chunk.length() for chunk in chunks])))
#get fragmentsizes
fragment_dist1 = FragmentMixDistribution(0, upper = args.upper)
fragment_dist1.fragmentsizes = FragmentSizes(0, args.upper, vals = FragmentSizes.open(args.sizes1).get(0,args.upper))
fragment_dist1.modelNFR()
fragment_dist2 = FragmentMixDistribution(0, upper = args.upper)
fragment_dist2.fragmentsizes = FragmentSizes(0, args.upper, vals = FragmentSizes.open(args.sizes2).get(0,args.upper))
fragment_dist2.modelNFR()
params = OccupancyParameters(fragment_dist, args.upper, args.fasta, args.pwm, sep = args.nuc_sep, min_occ = args.min_occ,
flank = args.flank, bam = args.bam, ci = args.confidence_interval)
sets = chunks.split(bases = bases)
pool1 = mp.Pool(processes = max(1,args.cores-1))
diff_handle = open(args.out + '.occdiff.bed','w')
diff_handle.close()
diff_queue = mp.JoinableQueue()
diff_process = mp.Process(target = _writeDiff, args=(diff_queue, args.out))
diff_process.start()
nuc_dist = np.zeros(args.upper)
for j in sets:
tmp = pool1.map(_occHelper, zip(j,itertools.repeat(params)))
for result in tmp:
diff_queue.put(result[1])
pool1.close()
pool1.join()
diff_queue.put('STOP')
diff_process.join()
pysam.tabix_compress(args.out + '.occdiff.bed', args.out + '.occdiff.bed.gz',force = True)
shell_command('rm ' + args.out + '.occdiff.bed')
pysam.tabix_index(args.out + '.occdiff.bed.gz', preset = "bed", force = True)
def run_occ(args):
"""run occupancy calling
"""
if args.fasta:
chrs = read_chrom_sizes_from_fasta(args.fasta)
else:
chrs = read_chrom_sizes_from_bam(args.bam)
pwm = PWM.open(args.pwm)
chunks = ChunkList.read(args.bed, chromDict = chrs, min_offset = args.flank + args.upper/2 + max(pwm.up,pwm.down) + args.nuc_sep/2)
chunks.slop(chrs, up = args.nuc_sep/2, down = args.nuc_sep/2)
chunks.merge()
maxQueueSize = args.cores*10
fragment_dist = FragmentMixDistribution(0, upper = args.upper)
if args.sizes is not None:
tmp = FragmentSizes.open(args.sizes)
fragment_dist.fragmentsizes = FragmentSizes(0, args.upper, vals = tmp.get(0,args.upper))
else:
fragment_dist.getFragmentSizes(args.bam, chunks)
fragment_dist.modelNFR()
fragment_dist.plotFits(args.out + '.occ_fit.eps')
fragment_dist.fragmentsizes.save(args.out + '.fragmentsizes.txt')
params = OccupancyParameters(fragment_dist, args.upper, args.fasta, args.pwm, sep = args.nuc_sep, min_occ = args.min_occ,
flank = args.flank, bam = args.bam, ci = args.confidence_interval, step = args.step)
sets = chunks.split(items = args.cores * 5)
pool1 = mp.Pool(processes = max(1,args.cores-1))
out_handle1 = open(args.out + '.occ.bedgraph','w')
out_handle1.close()
out_handle2 = open(args.out + '.occ.lower_bound.bedgraph','w')
out_handle2.close()
out_handle3 = open(args.out + '.occ.upper_bound.bedgraph','w')
out_handle3.close()
write_queue = mp.JoinableQueue(maxsize = maxQueueSize)
write_process = mp.Process(target = _writeOcc, args=(write_queue, args.out))
write_process.start()
peaks_handle = open(args.out + '.occpeaks.bed','w')
peaks_handle.close()
peaks_queue = mp.JoinableQueue()
peaks_process = mp.Process(target = _writePeaks, args=(peaks_queue, args.out))
peaks_process.start()
nuc_dist = np.zeros(args.upper)
for j in sets:
tmp = pool1.map(_occHelper, zip(j,itertools.repeat(params)))
for result in tmp:
nuc_dist += result[0]
write_queue.put(result[1])
peaks_queue.put(result[2])
pool1.close()
pool1.join()
write_queue.put('STOP')
peaks_queue.put('STOP')
write_process.join()
peaks_process.join()
pysam.tabix_compress(args.out + '.occpeaks.bed', args.out + '.occpeaks.bed.gz',force = True)
shell_command('rm ' + args.out + '.occpeaks.bed')
pysam.tabix_index(args.out + '.occpeaks.bed.gz', preset = "bed", force = True)
for i in ('occ','occ.lower_bound','occ.upper_bound'):
pysam.tabix_compress(args.out + '.' + i + '.bedgraph', args.out + '.'+i+'.bedgraph.gz',force = True)
shell_command('rm ' + args.out + '.' + i + '.bedgraph')
pysam.tabix_index(args.out + '.' + i + '.bedgraph.gz', preset = "bed", force = True)
dist_out = FragmentSizes(0, args.upper, vals = nuc_dist)
dist_out.save(args.out + '.nuc_dist.txt')
print "Making figure"
#make figure
fig = plt.figure()
plt.plot(range(0,args.upper),dist_out.get(0,args.upper),label = "Nucleosome Distribution")
plt.xlabel("Fragment Size")
plt.ylabel("Frequency")
fig.savefig(args.out+'.nuc_dist.eps')
plt.close(fig)