def make_output_and(cov, control_cov, out_prefix, window):
# dump raw counts to file
raw_out = open('%s_raw.txt' % out_prefix,'w')
for i in range(-window/2,window/2+1):
print >> raw_out, '%d\t%e\t%e' % (i, cov[window/2+i], control_cov[window/2+i])
raw_out.close()
# make plot data structures
splice_i = ro.IntVector(2*range(-window/2,window/2+1))
cov_r = ro.FloatVector(cov+control_cov)
labels = ro.StrVector(['Main']*len(cov)+['Control']*len(control_cov))
df = ro.DataFrame({'splice_i':splice_i, 'cov':cov_r, 'label':labels})
# construct plot
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='splice_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('Position relative to splice site') + \
ggplot2.scale_y_continuous('Coverage') + \
ggplot2.scale_colour_discrete('')
# plot to file
grdevices.pdf(file='%s.pdf' % out_prefix)
gp.plot()
grdevices.dev_off()
def make_output_and(cov, control_cov, out_prefix, window):
# dump raw counts to file
raw_out = open("%s_raw.txt" % out_prefix, "w")
for i in range(-window / 2, window / 2 + 1):
print >> raw_out, "%d\t%e\t%e" % (i, cov[window / 2 + i], control_cov[window / 2 + i])
raw_out.close()
# make plot data structures
splice_i = ro.IntVector(2 * range(-window / 2, window / 2 + 1))
cov_r = ro.FloatVector(cov + control_cov)
labels = ro.StrVector(["Main"] * len(cov) + ["Control"] * len(control_cov))
df = ro.DataFrame({"splice_i": splice_i, "cov": cov_r, "label": labels})
# construct plot
gp = (
ggplot2.ggplot(df)
+ ggplot2.aes_string(x="splice_i", y="cov", colour="label")
+ ggplot2.geom_point()
+ ggplot2.scale_x_continuous("Position relative to splice site")
+ ggplot2.scale_y_continuous("Coverage")
+ ggplot2.scale_colour_discrete("")
)
# plot to file
grdevices.pdf(file="%s.pdf" % out_prefix)
gp.plot()
grdevices.dev_off()
def make_output_and(te_tss_cov, control_te_tss_cov, out_prefix, upstream,
downstream):
# clean raw counts dir
if os.path.isdir('%s_raw' % out_prefix):
shutil.rmtree('%s_raw' % out_prefix)
os.mkdir('%s_raw' % out_prefix)
# dump raw counts to file
for te in te_tss_cov:
if te[0] in [
'n', '*', 'HERVH-int', 'L2a', 'AluSx', 'AluJb', 'MIRb', 'LTR7'
] and te[1] in [
'n', '*', 'LINE/L1', 'SINE/Alu', 'LTR/ERV1', 'LTR/ERVL-MaLR',
'LINE/L2', 'LTR/ERVL', 'SINE/MIR', 'DNA/hAT-Charlie',
'LTR/ERVK', 'DNA/TcMar-Tigger'
]:
raw_out = open(
'%s_raw/%s_%s.txt' %
(out_prefix, te[0].replace('/', '_'), te[1].replace('/', '_')),
'w')
for i in range(-upstream, downstream + 1):
print >> raw_out, '%d\t%e\t%e' % (i, te_tss_cov[te][
upstream + i], control_te_tss_cov[te][upstream + i])
raw_out.close()
# clean plot dirs
if os.path.isdir('%s_plot' % out_prefix):
shutil.rmtree('%s_plot' % out_prefix)
os.mkdir('%s_plot' % out_prefix)
# make data structures
tss_i = ro.IntVector(2 * range(-upstream, downstream + 1))
labels = ro.StrVector(['Main'] * (upstream + downstream + 1) +
['Control'] * (upstream + downstream + 1))
for te in te_tss_cov:
if te[0] in [
'n', '*', 'HERVH-int', 'L2a', 'AluSx', 'AluJb', 'MIRb', 'LTR7'
] and te[1] in [
'n', '*', 'LINE/L1', 'SINE/Alu', 'LTR/ERV1', 'LTR/ERVL-MaLR',
'LINE/L2', 'LTR/ERVL', 'SINE/MIR', 'DNA/hAT-Charlie',
'LTR/ERVK', 'DNA/TcMar-Tigger'
]:
cov = ro.FloatVector(te_tss_cov[te] + control_te_tss_cov[te])
df = ro.DataFrame({'tss_i': tss_i, 'cov': cov, 'label': labels})
# construct full plot
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index') + \
ggplot2.scale_y_continuous('Coverage') + \
ggplot2.scale_colour_discrete('')
# plot to file
grdevices.pdf(
file='%s_plot/%s_%s.pdf' %
(out_prefix, te[0].replace('/', '_'), te[1].replace('/', '_')))
gp.plot()
grdevices.dev_off()
def main():
usage = 'usage: %prog [options] <raw file>'
parser = OptionParser(usage)
parser.add_option('-d', dest='downstream', default=2000, type='int', help='TSS downstream [Default: %default]')
parser.add_option('-o', dest='out_prefix', default='tss', help='Output prefix [Default: %default]')
parser.add_option('-u', dest='upstream', default=5000, type='int', help='TSS upstream [Default: %default]')
parser.add_option('--ymax', dest='ymax', default=None, type='float', help='Y-coordinate limit [Default: %default]')
(options,args) = parser.parse_args()
if len(args) != 1:
parser.error('Must provide raw file')
else:
raw_file = args[0]
# collect data
coords = []
main_cov = []
control_cov = []
for line in open(raw_file):
a = line.split()
coords.append(int(a[0]))
main_cov.append(float(a[1]))
control_cov.append(float(a[2]))
# data structures
tss_i = ro.IntVector(range(-options.upstream,options.downstream+1))
labels = ro.StrVector(['Main']*(options.upstream+options.downstream+1)+['Control']*(options.upstream+options.downstream+1))
cov = ro.FloatVector(main_cov + control_cov)
df = ro.DataFrame({'tss_i':tss_i, 'cov':cov, 'label':labels})
# plot
'''
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index') + \
ggplot2.scale_colour_discrete('')
'''
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_smooth(method='loess', size=1, span=0.2, se=False) + \
ggplot2.scale_x_continuous('TSS Position') + \
ggplot2.scale_colour_discrete('') + \
ggplot2.theme_bw()
if options.ymax == None:
gp += ggplot2.scale_y_continuous('Coverage')
else:
gp += ggplot2.scale_y_continuous('Coverage', limits=ro.FloatVector([0,options.ymax]))
# save to file
grdevices.pdf(file='%s_and.pdf' % options.out_prefix)
gp.plot()
grdevices.dev_off()
def make_output_and(tss_cov, control_tss_cov, out_prefix, upstream, downstream):
# dump raw counts to file
raw_out = open('%s_raw.txt' % out_prefix,'w')
for i in range(-upstream,downstream+1):
print >> raw_out, '%d\t%e\t%e' % (i, tss_cov[upstream+i], control_tss_cov[upstream+i])
raw_out.close()
# make plot data structures
tss_i = ro.IntVector(2*range(-upstream,downstream+1))
cov = ro.FloatVector(tss_cov+control_tss_cov)
labels = ro.StrVector(['Main']*len(tss_cov)+['Control']*len(control_tss_cov))
df = ro.DataFrame({'tss_i':tss_i, 'cov':cov, 'label':labels})
# construct full plot
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index') + \
ggplot2.scale_y_continuous('Coverage') + \
ggplot2.scale_colour_discrete('')
# plot to file
grdevices.pdf(file='%s_full.pdf' % out_prefix)
gp.plot()
grdevices.dev_off()
# construct zoomed plot
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index',limits=ro.IntVector([-1000,1000])) + \
ggplot2.scale_y_continuous('Coverage') + \
ggplot2.scale_colour_discrete('')
# plot to file
grdevices.pdf(file='%s_zoom.pdf' % out_prefix)
gp.plot()
grdevices.dev_off()
def make_output_and(te_tss_cov, control_te_tss_cov, out_prefix, upstream, downstream):
# clean raw counts dir
if os.path.isdir('%s_raw' % out_prefix):
shutil.rmtree('%s_raw' % out_prefix)
os.mkdir('%s_raw' % out_prefix)
# dump raw counts to file
for te in te_tss_cov:
if te[0] in ['n','*','HERVH-int','L2a','AluSx','AluJb','MIRb','LTR7'] and te[1] in ['n','*','LINE/L1','SINE/Alu','LTR/ERV1','LTR/ERVL-MaLR','LINE/L2','LTR/ERVL','SINE/MIR','DNA/hAT-Charlie','LTR/ERVK','DNA/TcMar-Tigger']:
raw_out = open('%s_raw/%s_%s.txt' % (out_prefix,te[0].replace('/','_'),te[1].replace('/','_')),'w')
for i in range(-upstream,downstream+1):
print >> raw_out, '%d\t%e\t%e' % (i, te_tss_cov[te][upstream+i], control_te_tss_cov[te][upstream+i])
raw_out.close()
# clean plot dirs
if os.path.isdir('%s_plot' % out_prefix):
shutil.rmtree('%s_plot' % out_prefix)
os.mkdir('%s_plot' % out_prefix)
# make data structures
tss_i = ro.IntVector(2*range(-upstream,downstream+1))
labels = ro.StrVector(['Main']*(upstream+downstream+1)+['Control']*(upstream+downstream+1))
for te in te_tss_cov:
if te[0] in ['n','*','HERVH-int','L2a','AluSx','AluJb','MIRb','LTR7'] and te[1] in ['n','*','LINE/L1','SINE/Alu','LTR/ERV1','LTR/ERVL-MaLR','LINE/L2','LTR/ERVL','SINE/MIR','DNA/hAT-Charlie','LTR/ERVK','DNA/TcMar-Tigger']:
cov = ro.FloatVector(te_tss_cov[te] + control_te_tss_cov[te])
df = ro.DataFrame({'tss_i':tss_i, 'cov':cov, 'label':labels})
# construct full plot
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index') + \
ggplot2.scale_y_continuous('Coverage') + \
ggplot2.scale_colour_discrete('')
# plot to file
grdevices.pdf(file='%s_plot/%s_%s.pdf' % (out_prefix,te[0].replace('/','_'),te[1].replace('/','_')))
gp.plot()
grdevices.dev_off()
def main():
usage = 'usage: %prog [options] <raw file>'
parser = OptionParser(usage)
parser.add_option('-d',
dest='downstream',
default=2000,
type='int',
help='TSS downstream [Default: %default]')
parser.add_option('-o',
dest='out_prefix',
default='tss',
help='Output prefix [Default: %default]')
parser.add_option('-u',
dest='upstream',
default=5000,
type='int',
help='TSS upstream [Default: %default]')
parser.add_option('--ymax',
dest='ymax',
default=None,
type='float',
help='Y-coordinate limit [Default: %default]')
(options, args) = parser.parse_args()
if len(args) != 1:
parser.error('Must provide raw file')
else:
raw_file = args[0]
# collect data
coords = []
main_cov = []
control_cov = []
for line in open(raw_file):
a = line.split()
coords.append(int(a[0]))
main_cov.append(float(a[1]))
control_cov.append(float(a[2]))
# data structures
tss_i = ro.IntVector(range(-options.upstream, options.downstream + 1))
labels = ro.StrVector(['Main'] *
(options.upstream + options.downstream + 1) +
['Control'] *
(options.upstream + options.downstream + 1))
cov = ro.FloatVector(main_cov + control_cov)
df = ro.DataFrame({'tss_i': tss_i, 'cov': cov, 'label': labels})
# plot
'''
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_point() + \
ggplot2.scale_x_continuous('TSS index') + \
ggplot2.scale_colour_discrete('')
'''
gp = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='tss_i', y='cov', colour='label') + \
ggplot2.geom_smooth(method='loess', size=1, span=0.2, se=False) + \
ggplot2.scale_x_continuous('TSS Position') + \
ggplot2.scale_colour_discrete('') + \
ggplot2.theme_bw()
if options.ymax == None:
gp += ggplot2.scale_y_continuous('Coverage')
else:
gp += ggplot2.scale_y_continuous('Coverage',
limits=ro.FloatVector(
[0, options.ymax]))
# save to file
grdevices.pdf(file='%s_and.pdf' % options.out_prefix)
gp.plot()
grdevices.dev_off()
samples = pd.concat((samplebs, samplebg))
# re-index to avoid duplicate row.names in Rdf
samples.index = npy.arange(len(samples))
samplesgrouped = samples.groupby(['model'])
variances = samplesgrouped['Zweighted'].aggregate(npy.var)
print variances
print variances['BG'] / variances['BS']
print estimatesum(samples)
print samplesgrouped['Zweighted'].aggregate(estimatesum)
print trueZnsum
# grdevices.png(file="sampled-Z.png", width=4, height=3, units="in", res=300)
rsamples = com.convert_to_r_dataframe(samples)
pp = ggplot2.ggplot(rsamples) + \
ggplot2.aes_string(x='Z', color='factor(model)') + \
ggplot2.scale_colour_discrete(name="model") + \
ggplot2.geom_density() + \
ggplot2.scale_x_log10()
# ggplot2.scale_x_continuous(limits=FloatVector((0, 1)))
pp.plot()
# grdevices.dev_off()
def makeestimate(sampler, numsamples, **kwargs):
samples = sample(sampler, numsamples, **kwargs)
return estimatesum(samples['Zweighted'])
def makeestimates(sampler, numsamples, numestimates, **kwargs):
estimates = [
makeestimate(sampler, numsamples, **kwargs)
logging.debug('True sum: %s', trueZnsum)
emdf = pd.DataFrame({
'BSdists' : distsbs,
'BGdists' : distsbg,
'truesums' : truesums,
'varratios' : varratios,
})
# Plot sampled Z
logging.info('Plotting sampled Zn')
grdevices.png(file="sampled-Z.png", width=4, height=3, units="in", res=300)
rsamples = com.convert_to_r_dataframe(samples)
pp = ggplot2.ggplot(rsamples) + \
ggplot2.aes_string(x='Z', color='factor(model)') + \
ggplot2.scale_colour_discrete(name="model") + \
ggplot2.geom_density() + \
ggplot2.scale_x_log10()
# ggplot2.scale_x_continuous(limits=FloatVector((0, 1)))
pp.plot()
grdevices.dev_off()
# Plot likelihood ratios
logging.info('Plotting likelihood ratios from binding site samples')
grdevices.png(file="sampled-ratios.png",
width=4, height=3, units="in", res=300)
rsamplesbs = com.convert_to_r_dataframe(samples[samples['model'] == 'BS'])
pp = ggplot2.ggplot(rsamplesbs) + \
ggplot2.aes_string(x='ir') + \
ggplot2.geom_density() + \
ggplot2.scale_x_log10()
