def test_normByInsertDist(self):
"""test that normalization by insert distribution works as expected"""
isizes = FragmentSizes(lower=100,upper=200, vals = np.array(range(100,200)))
self.biasmat.normByInsertDist(isizes)
a1 = self.biastrack.get(pos = self.biasmat.start -50)
a2 = self.biastrack.get(pos = self.biasmat.start + 50)
correct = np.exp(a1+a2)*isizes.get(size = 101)
self.assertTrue(abs(correct - self.biasmat.mat[1,0])<0.01*correct)
def test_normByInsertDist(self):
"""test that normalization by insert distribution works as expected"""
isizes = FragmentSizes(lower=100,
upper=200,
vals=np.array(range(100, 200)))
self.biasmat.normByInsertDist(isizes)
a1 = self.biastrack.get(pos=self.biasmat.start - 50)
a2 = self.biastrack.get(pos=self.biasmat.start + 50)
correct = np.exp(a1 + a2) * isizes.get(size=101)
self.assertTrue(abs(correct - self.biasmat.mat[1, 0]) < 0.01 * correct)
def get_sizes(args):
"""function to get fragment sizes
"""
if args.out is None:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
sizes = FragmentSizes(lower = args.lower, upper = args.upper, atac = args.atac)
if args.bed:
chunks = ChunkList.read(args.bed)
chunks.merge()
sizes.calculateSizes(args.bam, chunks)
else:
sizes.calculateSizes(args.bam)
sizes.save(args.out+'.fragmentsizes.txt')
if not args.no_plot:
#make figure
fig = plt.figure()
plt.plot(range(sizes.lower,sizes.upper),sizes.get(sizes.lower,sizes.upper),label = args.out)
plt.xlabel("Fragment Size")
plt.ylabel("Frequency")
fig.savefig(args.out+'.fragmentsizes.eps')
plt.close(fig)
def get_sizes(args):
"""function to get fragment sizes
"""
if args.out is None:
args.out = '.'.join(os.path.basename(args.bam).split('.')[0:-1])
sizes = FragmentSizes(lower=args.lower, upper=args.upper, atac=args.atac)
if args.bed:
chunks = ChunkList.read(args.bed)
chunks.merge()
sizes.calculateSizes(args.bam, chunks)
else:
sizes.calculateSizes(args.bam)
sizes.save(args.out + '.fragmentsizes.txt')
if not args.no_plot:
#make figure
fig = plt.figure()
plt.plot(list(range(sizes.lower, sizes.upper)),
sizes.get(sizes.lower, sizes.upper),
label=args.out)
plt.xlabel("Fragment Size")
plt.ylabel("Frequency")
fig.savefig(args.out + '.fragmentsizes.pdf')
plt.close(fig)
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)
class FragmentMixDistribution:
"""Class for modelling insert size distribution"""
def __init__(self, lower = 0, upper =2000):
self.lower = lower
self.upper = upper
def getFragmentSizes(self, bamfile, chunklist = None):
self.fragmentsizes = FragmentSizes(self.lower, self.upper)
self.fragmentsizes.calculateSizes(bamfile, chunks = chunklist)
def modelNFR(self, boundaries = (35,115)):
"""Model NFR distribution with gamma distribution"""
b = np.where(self.fragmentsizes.get(self.lower,boundaries[1]) == max(self.fragmentsizes.get(self.lower,boundaries[1])))[0][0] + self.lower
boundaries = (min(boundaries[0],b), boundaries[1])
x = np.arange(boundaries[0],boundaries[1])
y = self.fragmentsizes.get(boundaries[0],boundaries[1])
def gamma_fit(X,o,p):
k = p[0]
theta = p[1]
a = p[2]
x_mod = X-o
res = np.zeros(len(x_mod))
if k>=1:
nz = x_mod >= 0
else:
nz = x_mod > 0
res[nz] = a * x_mod[nz]**(k-1) * np.exp(-x_mod[nz]/theta) / (theta **k * gamma(k))
return res
res_score = np.ones(boundaries[0]+1)*np.float('inf')
res_param = [0 for i in range(boundaries[0]+1)]
pranges = ((0.01,10),(0.01,150),(0.01,1))
for i in range(15,boundaries[0]+1):
f = lambda p: np.sum((gamma_fit(x,i,p) - y)**2)
tmpres = optimize.brute(f, pranges, full_output=True,
finish=optimize.fmin)
res_score[i] = tmpres[1]
res_param[i] = tmpres[0]
whichres = np.argmin(res_score)
res = res_param[whichres]
self.nfr_fit0 = FragmentSizes(self.lower,self.upper, vals = gamma_fit(np.arange(self.lower,self.upper),whichres,res_param[whichres]))
nfr = np.concatenate((self.fragmentsizes.get(self.lower,boundaries[1]), self.nfr_fit0.get(boundaries[1],self.upper)))
nfr[nfr==0] = min(nfr[nfr!=0])*0.01
self.nfr_fit = FragmentSizes(self.lower,self.upper, vals = nfr)
nuc = np.concatenate((np.zeros(boundaries[1]-self.lower),
self.fragmentsizes.get(boundaries[1],self.upper) -
self.nfr_fit.get(boundaries[1],self.upper)))
nuc[nuc<=0]=min(min(nfr)*0.1,min(nuc[nuc>0])*0.001)
self.nuc_fit = FragmentSizes(self.lower, self.upper, vals = nuc)
def plotFits(self,filename=None):
"""plot the Fits"""
fig = plt.figure()
plt.plot(range(self.lower,self.upper),self.fragmentsizes.get(),
label = "Observed")
plt.plot(range(self.lower,self.upper),self.nfr_fit0.get(), label = "NFR Fit")
plt.plot(range(self.lower,self.upper),self.nuc_fit.get(), label = "Nucleosome Model")
plt.plot(range(self.lower,self.upper),self.nfr_fit.get(), label = "NFR Model")
plt.legend()
plt.xlabel("Fragment size")
plt.ylabel("Relative Frequency")
if filename:
fig.savefig(filename)
plt.close(fig)
#Also save text output!
filename2 = ".".join(filename.split(".")[:-1]+['txt'])
out = np.vstack((self.fragmentsizes.get(), #self.smoothed.get(),
self.nuc_fit.get(), self.nfr_fit.get()))
np.savetxt(filename2,out,delimiter="\t")
else:
fig.show()
class FragmentMixDistribution:
"""Class for modelling insert size distribution"""
def __init__(self, lower=0, upper=2000):
self.lower = lower
self.upper = upper
def getFragmentSizes(self, bamfile, chunklist=None):
self.fragmentsizes = FragmentSizes(self.lower, self.upper)
self.fragmentsizes.calculateSizes(bamfile, chunks=chunklist)
def modelNFR(self, boundary=115):
"""Model NFR distribution with exponential distribution"""
b = np.where(
self.fragmentsizes.get(self.lower, boundary) == max(
self.fragmentsizes.get(self.lower,
boundary)))[0][0] + 10 + self.lower
def exp_pdf(x, *p): #defines the PDF
k = p[0]
a = p[1]
x = x - b
return a * k * np.exp(-k * x)
x = np.array(range(b, boundary))
p0 = (.1, 1)
coeff, var_matrix = optimize.curve_fit(exp_pdf,
x,
self.fragmentsizes.get(
b, boundary),
p0=p0)
nfr = np.concatenate((self.fragmentsizes.get(self.lower, boundary),
exp_pdf(np.array(range(boundary, self.upper)),
*coeff)))
nfr[nfr == 0] = min(nfr[nfr != 0]) * 0.01
self.nfr_fit = FragmentSizes(self.lower, self.upper, vals=nfr)
nuc = np.concatenate((np.zeros(boundary - self.lower),
self.fragmentsizes.get(boundary, self.upper) -
self.nfr_fit.get(boundary, self.upper)))
nuc[nuc <= 0] = min(min(nfr) * 0.1, min(nuc[nuc > 0]) * 0.001)
self.nuc_fit = FragmentSizes(self.lower, self.upper, vals=nuc)
def plotFits(self, filename=None):
"""plot the Fits"""
fig = plt.figure()
plt.plot(range(self.lower, self.upper),
self.fragmentsizes.get(),
label="Observed")
#plt.plot(range(self.lower,self.upper),self.smoothed.get(), label = "Smoothed")
plt.plot(range(self.lower, self.upper),
self.nuc_fit.get(),
label="Nucleosome Fit")
plt.plot(range(self.lower, self.upper),
self.nfr_fit.get(),
label="NFR Fit")
plt.legend()
plt.xlabel("Fragment size")
plt.ylabel("Relative Frequency")
if filename:
fig.savefig(filename)
plt.close(fig)
#Also save text output!
filename2 = ".".join(filename.split(".")[:-1] + ['txt'])
out = np.vstack((
self.fragmentsizes.get(), #self.smoothed.get(),
self.nuc_fit.get(),
self.nfr_fit.get()))
np.savetxt(filename2, out, delimiter="\t")
else:
fig.show()
class FragmentMixDistribution:
"""Class for modelling insert size distribution"""
def __init__(self, lower = 0, upper =2000):
self.lower = lower
self.upper = upper
def getFragmentSizes(self, bamfile, chunklist = None):
self.fragmentsizes = FragmentSizes(self.lower, self.upper)
self.fragmentsizes.calculateSizes(bamfile, chunks = chunklist)
def modelNFR(self, boundary = 115):
"""Model NFR distribution with exponential distribution"""
b = np.where(self.fragmentsizes.get(self.lower,boundary) == max(self.fragmentsizes.get(self.lower,boundary)))[0][0]+10 + self.lower
def exp_pdf(x,*p): #defines the PDF
k=p[0]
a=p[1]
x=x-b
return a*k*np.exp(-k*x)
x = np.array(range(b,boundary))
p0 = (.1,1)
coeff, var_matrix = optimize.curve_fit(exp_pdf,x, self.fragmentsizes.get(b,boundary),
p0=p0)
nfr = np.concatenate((self.fragmentsizes.get(self.lower,boundary), exp_pdf(np.array(range(boundary,self.upper)),*coeff)))
nfr[nfr==0] = min(nfr[nfr!=0])*0.01
self.nfr_fit = FragmentSizes(self.lower,self.upper, vals = nfr)
nuc = np.concatenate((np.zeros(boundary-self.lower),
self.fragmentsizes.get(boundary,self.upper) -
self.nfr_fit.get(boundary,self.upper)))
nuc[nuc<=0]=min(min(nfr)*0.1,min(nuc[nuc>0])*0.001)
self.nuc_fit = FragmentSizes(self.lower, self.upper, vals = nuc)
def plotFits(self,filename=None):
"""plot the Fits"""
fig = plt.figure()
plt.plot(range(self.lower,self.upper),self.fragmentsizes.get(),
label = "Observed")
#plt.plot(range(self.lower,self.upper),self.smoothed.get(), label = "Smoothed")
plt.plot(range(self.lower,self.upper),self.nuc_fit.get(), label = "Nucleosome Fit")
plt.plot(range(self.lower,self.upper),self.nfr_fit.get(), label = "NFR Fit")
plt.legend()
plt.xlabel("Fragment size")
plt.ylabel("Relative Frequency")
if filename:
fig.savefig(filename)
plt.close(fig)
#Also save text output!
filename2 = ".".join(filename.split(".")[:-1]+['txt'])
out = np.vstack((self.fragmentsizes.get(), #self.smoothed.get(),
self.nuc_fit.get(), self.nfr_fit.get()))
np.savetxt(filename2,out,delimiter="\t")
else:
fig.show()
