def _plt_distr(dat,
col,
title='',
splitBy_pfill=True,
pfill='label',
independentpdf=False,
fname='xdistr.pdf'):
df = dat[dat[pfill] != 'NA'] ## remove invalid pairs
n = len(df)
df = {
col: robjects.FloatVector(list(df[col])),
pfill: robjects.StrVector(list(df[pfill]))
}
df = robjects.DataFrame(df)
pp = ggplot2.ggplot(df) + \
ggplot2.ggtitle('%s [Total = %s]' % (title, n))
## Plot1: counts
if splitBy_pfill:
p1 = pp + ggplot2.aes_string(x=col, fill=pfill)
else:
p1 = pp + ggplot2.aes_string(x=col)
## Plot2: density
if splitBy_pfill:
p2 = pp + ggplot2.aes_string(x=col, fill=pfill, y='..density..')
else:
p2 = pp + ggplot2.aes_string(x=col, y='..density..')
p2 = p2 + ggplot2.geom_density(alpha=.5, origin=-500)
if col == 'distance':
p1 = p1 + \
ggplot2.geom_histogram(binwidth=1000, alpha=.5, position='identity', origin=-500) + \
ggplot2.xlim(-1000, 51000)
p2 = p2 + \
ggplot2.geom_histogram(binwidth=1000, alpha=.33, position='identity', origin=-500) + \
ggplot2.xlim(-1000, 51000)
else:
p1 = p1 + \
ggplot2.geom_histogram(alpha=.5, position='identity')
p2 = p2 + \
ggplot2.geom_histogram(alpha=.33, position='identity')
if col == 'correlation':
p1 = p1 + ggplot2.xlim(-1.1, 1.1)
p2 = p2 + ggplot2.xlim(-1.1, 1.1)
if independentpdf:
grdevices = importr('grDevices')
grdevices.pdf(file=fname)
p1.plot()
p2.plot()
grdevices.dev_off()
else:
p1.plot()
p2.plot()
return
def makeDistancePlot( alldata, figurename, feature="distance") :
alldata["distance"] = alldata.het + alldata.hom
r_dataframe = com.convert_to_r_dataframe(alldata)
p = ggplot2.ggplot(r_dataframe) + \
ggplot2.aes_string(x=feature ) + \
ggplot2.geom_density(ggplot2.aes_string(fill="factor(continent)")) + \
ggplot2.ggtitle("Distance from Reference by Continent") + \
ggplot2.theme(**mytheme) #+ \
#ggplot2.theme(**{'axis.text.x': ggplot2.element_text(angle = 45)}) + \
#ggplot2.facet_grid( robjects.Formula('RVIS_type ~ .') )
grdevices.png(figurename)
p.plot()
grdevices.dev_off()
grid.newpage()
grid.viewport(layout=grid.layout(1, 2)).push()
gp = ggplot2.ggplot(dataf_rnorm)
vp = grid.viewport(**{'layout.pos.col': 1, 'layout.pos.row': 1})
pp = gp + \
ggplot2.aes_string(x='value', col='factor(mean)') + \
ggplot2.geom_freqpoly()
pp.plot(vp=vp)
vp = grid.viewport(**{'layout.pos.col': 2, 'layout.pos.row': 1})
#-- ggplot2geomfreqpolyfillcyl-begin
pp = gp + \
ggplot2.aes_string(x='value', fill='factor(mean)') + \
ggplot2.geom_density(alpha = 0.5)
#-- ggplot2geomfreqpolyfillcyl-end
pp.plot(vp=vp)
grdevices.dev_off()
grdevices.png('../../_static/graphics_ggplot2geompointandrug.png',
width=612,
height=612,
antialias="subpixel",
type="cairo")
#-- ggplot2geompointandrug-begin
gp = ggplot2.ggplot(mtcars)
pp = gp + \
ggplot2.aes_string(x='wt', y='mpg') + \
grid.newpage()
grid.viewport(layout=grid.layout(1, 2)).push()
gp = ggplot2.ggplot(dataf_rnorm)
vp = grid.viewport(**{'layout.pos.col':1, 'layout.pos.row': 1})
pp = gp + \
ggplot2.aes_string(x='value', col='factor(mean)') + \
ggplot2.geom_freqpoly()
pp.plot(vp = vp)
vp = grid.viewport(**{'layout.pos.col':2, 'layout.pos.row': 1})
#-- ggplot2geomfreqpolyfillcyl-begin
pp = gp + \
ggplot2.aes_string(x='value', fill='factor(mean)') + \
ggplot2.geom_density(alpha = 0.5)
#-- ggplot2geomfreqpolyfillcyl-end
pp.plot(vp = vp)
grdevices.dev_off()
grdevices.png('../../_static/graphics_ggplot2geompointandrug.png',
width = 612, height = 612, antialias="subpixel", type="cairo")
#-- ggplot2geompointandrug-begin
gp = ggplot2.ggplot(mtcars)
pp = gp + \
ggplot2.aes_string(x='wt', y='mpg') + \
ggplot2.geom_point() + \
def _plt_percountr(dat, independentpdf=False, fname='xpercount.pdf'):
def _filt_dat(dat, item, getlabel=True):
df = pd.DataFrame(dat[item].value_counts())
df.columns = ['count']
if getlabel:
df['label'] = [
list(dat[dat[item] == i]['label'])[0] for i in df.index
]
n = len(df)
mx = max(df['count'])
return df, n, mx
dat = dat[dat['label'] != 'NA']
## NUMBER OF MIRNA PER TSS
df, n, mx = _filt_dat(dat, 'tss', False)
df = {'count': robjects.IntVector(df['count'])}
df = robjects.DataFrame(df)
pt = ggplot2.ggplot(df) + \
ggplot2.geom_histogram(binwidth=1, origin=-.5, alpha=.5, position="identity") + \
ggplot2.xlim(-.5, mx+1) + \
ggplot2.aes_string(x='count') + \
ggplot2.ggtitle('TSS [Total = %s]' % n) + \
ggplot2.labs(x='Number of miRNA per TSS (max = %s)' % mx)
pt_den = ggplot2.ggplot(df) + \
ggplot2.aes_string(x='count', y='..density..') + \
ggplot2.geom_density(binwidth=1, alpha=.5, origin=-.5) + \
ggplot2.geom_histogram(binwidth=1, alpha=.33, position='identity', origin=-.5) + \
ggplot2.ggtitle('TSS [Total = %s]' % n) + \
ggplot2.labs(x='Number of miRNA per TSS (max = %s)' % mx)
## NUMBER OF TSS PER MIRNA
df, n, mx = _filt_dat(dat, 'mirna')
df = {
'count': robjects.IntVector(df['count']),
'label': robjects.StrVector(df['label'])
}
df = robjects.DataFrame(df)
_pm = ggplot2.ggplot(df) + \
ggplot2.geom_histogram(binwidth=1, origin=-.5, alpha=.5, position="identity") + \
ggplot2.xlim(-.5, mx+1) + \
ggplot2.ggtitle('miRNA [Total = %s]' % n)
_pm_den = ggplot2.ggplot(df) + \
ggplot2.geom_density(binwidth=1, alpha=.5, origin=-.5) + \
ggplot2.geom_histogram(binwidth=1, alpha=.33, position='identity', origin=-.5) + \
ggplot2.ggtitle('miRNA [Total = %s]' % n)
## not split by label
pm = _pm + ggplot2.aes_string(x='count')
pm_den = _pm_den + ggplot2.aes_string(x='count', y='..density..')
## split by label
pms = _pm + ggplot2.aes_string(x='count', fill='label')
pm_dens = _pm_den + ggplot2.aes_string(
x='count', fill='label', y='..density..')
## add xlabelling (need to be added after aes_string)
_xlab = ggplot2.labs(x='Number of TSS per miRNA (max = %s)' % mx)
pm += _xlab
pm_den += _xlab
pms += _xlab
pm_dens += _xlab
if independentpdf:
grdevices = importr('grDevices')
grdevices.pdf(fname)
pt.plot()
pt_den.plot()
pm.plot()
pm_den.plot()
pms.plot()
pm_dens.plot()
grdevices.dev_off()
else:
pt.plot()
pt_den.plot()
pm.plot()
pm_den.plot()
pms.plot()
pm_dens.plot()
return
def plot(self,
fn,
x='x',
y='y',
col=None,
group=None,
w=1100,
h=800,
size=2,
smooth=True,
point=True,
jitter=False,
boxplot=False,
boxplot2=False,
title=False,
flip=False,
se=False,
density=False,
line=False):
df = self.df
#import math, datetime
grdevices = importr('grDevices')
if not title:
title = fn.split("/")[-1]
grdevices.png(file=fn, width=w, height=h)
gp = ggplot2.ggplot(df)
pp = gp
if col and group:
pp += ggplot2.aes_string(x=x, y=y, col=col, group=group)
elif col:
pp += ggplot2.aes_string(x=x, y=y, col=col)
elif group:
pp += ggplot2.aes_string(x=x, y=y, group=group)
else:
pp += ggplot2.aes_string(x=x, y=y)
if boxplot:
if col:
pp += ggplot2.geom_boxplot(ggplot2.aes_string(fill=col),
color='blue')
else:
pp += ggplot2.geom_boxplot(color='blue')
if point:
if jitter:
if col:
pp += ggplot2.geom_point(ggplot2.aes_string(fill=col,
col=col),
size=size,
position='jitter')
else:
pp += ggplot2.geom_point(size=size, position='jitter')
else:
if col:
pp += ggplot2.geom_point(ggplot2.aes_string(fill=col,
col=col),
size=size)
else:
pp += ggplot2.geom_point(size=size)
if boxplot2:
if col:
pp += ggplot2.geom_boxplot(ggplot2.aes_string(fill=col),
color='blue',
outlier_colour="NA")
else:
pp += ggplot2.geom_boxplot(color='blue')
if smooth:
if smooth == 'lm':
if col:
pp += ggplot2.stat_smooth(ggplot2.aes_string(col=col),
size=1,
method='lm',
se=se)
else:
pp += ggplot2.stat_smooth(col='blue',
size=1,
method='lm',
se=se)
else:
if col:
pp += ggplot2.stat_smooth(ggplot2.aes_string(col=col),
size=1,
se=se)
else:
pp += ggplot2.stat_smooth(col='blue', size=1, se=se)
if density:
pp += ggplot2.geom_density(ggplot2.aes_string(x=x, y='..count..'))
if line:
pp += ggplot2.geom_line(position='jitter')
pp += ggplot2.opts(
**{
'title': title,
'axis.text.x': ggplot2.theme_text(size=24),
'axis.text.y': ggplot2.theme_text(size=24, hjust=1)
})
#pp+=ggplot2.scale_colour_brewer(palette="Set1")
pp += ggplot2.scale_colour_hue()
if flip:
pp += ggplot2.coord_flip()
pp.plot()
grdevices.dev_off()
print ">> saved: " + fn
def plot(self, fn, x='x', y='y', col=None, group=None, w=1100, h=800, size=2, smooth=True, point=True, jitter=False, boxplot=False, boxplot2=False, title=False, flip=False, se=False, density=False, line=False):
df=self.df
#import math, datetime
grdevices = importr('grDevices')
if not title:
title=fn.split("/")[-1]
grdevices.png(file=fn, width=w, height=h)
gp = ggplot2.ggplot(df)
pp = gp
if col and group:
pp+=ggplot2.aes_string(x=x, y=y,col=col,group=group)
elif col:
pp+=ggplot2.aes_string(x=x, y=y,col=col)
elif group:
pp+=ggplot2.aes_string(x=x, y=y,group=group)
else:
pp+=ggplot2.aes_string(x=x, y=y)
if boxplot:
if col:
pp+=ggplot2.geom_boxplot(ggplot2.aes_string(fill=col),color='blue')
else:
pp+=ggplot2.geom_boxplot(color='blue')
if point:
if jitter:
if col:
pp+=ggplot2.geom_point(ggplot2.aes_string(fill=col,col=col),size=size,position='jitter')
else:
pp+=ggplot2.geom_point(size=size,position='jitter')
else:
if col:
pp+=ggplot2.geom_point(ggplot2.aes_string(fill=col,col=col),size=size)
else:
pp+=ggplot2.geom_point(size=size)
if boxplot2:
if col:
pp+=ggplot2.geom_boxplot(ggplot2.aes_string(fill=col),color='blue',outlier_colour="NA")
else:
pp+=ggplot2.geom_boxplot(color='blue')
if smooth:
if smooth=='lm':
if col:
pp+=ggplot2.stat_smooth(ggplot2.aes_string(col=col),size=1,method='lm',se=se)
else:
pp+=ggplot2.stat_smooth(col='blue',size=1,method='lm',se=se)
else:
if col:
pp+=ggplot2.stat_smooth(ggplot2.aes_string(col=col),size=1,se=se)
else:
pp+=ggplot2.stat_smooth(col='blue',size=1,se=se)
if density:
pp+=ggplot2.geom_density(ggplot2.aes_string(x=x,y='..count..'))
if line:
pp+=ggplot2.geom_line(position='jitter')
pp+=ggplot2.opts(**{'title' : title, 'axis.text.x': ggplot2.theme_text(size=24), 'axis.text.y': ggplot2.theme_text(size=24,hjust=1)} )
#pp+=ggplot2.scale_colour_brewer(palette="Set1")
pp+=ggplot2.scale_colour_hue()
if flip:
pp+=ggplot2.coord_flip()
pp.plot()
grdevices.dev_off()
print ">> saved: "+fn
# 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)
for _ in xrange(numestimates)]
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()
pp.plot()