def plot_chipseq_distribution(boundreads, unboundreads, title=None):
colors = ['#FF6A00','#1049A9']
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
hist_bnd = subplot.hist(np.sqrt(boundreads), bins=100, histtype='step', normed=True, \
linestyle='solid', color=colors[0], linewidth=1, label='bound')
hist_ubnd = subplot.hist(np.sqrt(unboundreads), bins=100, histtype='step', normed=True, \
linestyle='dashed', color=colors[1], linewidth=1, label='unbound')
ymax = max([0, hist_bnd[0].max(), hist_ubnd[0].max()])
xmax = max([0, np.sqrt(boundreads).max(), np.sqrt(unboundreads).max()])
subplot.axis([0, xmax, 0, ymax])
leghandle = subplot.legend(loc=1)
for text in leghandle.texts:
text.set_fontsize(8)
leghandle.set_frame_on(False)
subplot.set_xlabel('sqrt(reads) (chipseq)', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_chipseq_distribution(boundreads, unboundreads, title=None):
colors = ['#FF6A00', '#1049A9']
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
hist_bnd = subplot.hist(np.sqrt(boundreads), bins=100, histtype='step', normed=True, \
linestyle='solid', color=colors[0], linewidth=1, label='bound')
hist_ubnd = subplot.hist(np.sqrt(unboundreads), bins=100, histtype='step', normed=True, \
linestyle='dashed', color=colors[1], linewidth=1, label='unbound')
ymax = max([0, hist_bnd[0].max(), hist_ubnd[0].max()])
xmax = max([0, np.sqrt(boundreads).max(), np.sqrt(unboundreads).max()])
subplot.axis([0, xmax, 0, ymax])
leghandle = subplot.legend(loc=1)
for text in leghandle.texts:
text.set_fontsize(8)
leghandle.set_frame_on(False)
subplot.set_xlabel('sqrt(reads) (chipseq)', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_mnaseprofile(profiles, bounds, motiflen=None, title=None):
num = len(profiles)
if type(bounds[0]) is tuple:
labels = ['%d - %d' % (bound[0], bound[1]) for bound in bounds[:-1]]
labels.append('> %d' % bounds[-1][0])
else:
labels = bounds
colors = [
colorsys.hsv_to_rgb(h, 0.9, 0.7)
for h in np.linspace(0, 1, num + 2)[:-1]
]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
ymax = 0
for num, (profile, color, label) in enumerate(zip(profiles, colors,
labels)):
width = profile.size
xvals = np.arange(-width / 2, width / 2)
subplot.plot(xvals,
profile,
color=color,
linestyle='-',
linewidth=1,
label=label)
ymax = max([ymax, profile.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), 0, ymax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
if motiflen:
subplot.axvline(motiflen - 1, linestyle='--', c='g', linewidth=0.2)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_dnaseprofile(footprints, bounds, motiflen=None, title=None):
num = len(footprints)
if type(bounds[0]) is tuple:
labels = ['%d - %d'%(bound[0],bound[1]) for bound in bounds[:-1]]
labels.append('> %d'%bounds[-1][0])
else:
labels = bounds
colors = [colorsys.hsv_to_rgb(h,0.9,0.7) for h in np.linspace(0,1,num+2)[:-1]]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
fwdmax = 0
revmax = 0
for num, (footprint,color,label) in enumerate(zip(footprints,colors,labels)):
width = footprint.size/2
fwd = footprint[:width]
rev = footprint[width:]
xvals = np.arange(-width/2,width/2)
alpha = 0.7
subplot.plot(xvals, fwd, color=color, linestyle='-', linewidth=1, label=label, alpha=alpha)
subplot.plot(xvals, -1*rev, color=color, linestyle='-', linewidth=1, label="_nolabel_", alpha=alpha)
fwdmax = max([fwdmax, fwd.max()])
revmax = max([revmax, rev.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), -1*revmax, fwdmax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
if motiflen:
subplot.axvline(motiflen-1, linestyle='--', c='g', linewidth=0.2)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_readhist(Reads, title=None):
import scipy.stats as stats
num = len(Reads)
labels = ('Py', 'R')
colors = ['#FF6A00','#1049A9']
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
ymax = 0
xmin = 0
xmax = 0
for num, (reads,color,label) in enumerate(zip(Reads,colors,labels)):
alpha = 0.7
# read histogram for unbound sites
if reads[0].shape[0]>0:
hist = subplot.hist(reads[0]**0.5, bins=100, color=color, linestyle='dashed', histtype='step', \
linewidth=1, normed=True, alpha=alpha, label=label+'/ubnd')
xmax = max([xmax, reads[0].max()**0.5])
ymax = max([ymax, hist[0].max()])
# read histogram for bound sites
if reads[1].shape[0]>0:
hist = subplot.hist(reads[1]**0.5, bins=100, color=color, linestyle='solid', histtype='step', \
linewidth=1, normed=True, alpha=alpha, label=label+'/bnd')
xmax = max([xmax, reads[1].max()**0.5])
ymax = max([ymax, hist[0].max()])
subplot.axis([xmin, xmax, 0, ymax])
legend = subplot.legend(loc=1)
for text in legend.texts:
text.set_fontsize('8')
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_readhist(Reads, title=None):
import scipy.stats as stats
num = len(Reads)
labels = ('Py', 'R')
colors = ['#FF6A00', '#1049A9']
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
ymax = 0
xmin = 0
xmax = 0
for num, (reads, color, label) in enumerate(zip(Reads, colors, labels)):
alpha = 0.7
# read histogram for unbound sites
if reads[0].shape[0] > 0:
hist = subplot.hist(reads[0]**0.5, bins=100, color=color, linestyle='dashed', histtype='step', \
linewidth=1, normed=True, alpha=alpha, label=label+'/ubnd')
xmax = max([xmax, reads[0].max()**0.5])
ymax = max([ymax, hist[0].max()])
# read histogram for bound sites
if reads[1].shape[0] > 0:
hist = subplot.hist(reads[1]**0.5, bins=100, color=color, linestyle='solid', histtype='step', \
linewidth=1, normed=True, alpha=alpha, label=label+'/bnd')
xmax = max([xmax, reads[1].max()**0.5])
ymax = max([ymax, hist[0].max()])
subplot.axis([xmin, xmax, 0, ymax])
legend = subplot.legend(loc=1)
for text in legend.texts:
text.set_fontsize('8')
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_auc(Values, positive, negative, labels=None, title=None):
ilabels = np.zeros((positive.sum()+negative.sum(),),dtype=float)
ilabels[:positive.sum()] = 1
TPR = []
FPR = []
for allvalues in Values:
values = np.hstack((allvalues[positive],allvalues[negative]))
sorted = np.sort(values)[::-1]
tpr = np.zeros((values.size,),dtype=float)
fpr = np.zeros((values.size,),dtype=float)
for i,value in enumerate(sorted):
pos = ilabels[values>=value]
fpr[i] = (pos==0).sum()/float((ilabels==0).sum())
tpr[i] = (pos==1).sum()/float((ilabels==1).sum())
TPR.append(tpr)
FPR.append(fpr)
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
for tpr,fpr,color,label in zip(TPR,FPR,colors,labels):
subplot.plot(fpr, tpr, color=color, linewidth=1, label=label)
subplot.plot([0,1],[0,1], color='k', linewidth=1, alpha=0.5)
leghandle = subplot.legend(loc=4)
for text in leghandle.texts:
text.set_fontsize(6)
leghandle.set_frame_on(False)
subplot.axis([0,1,0,1])
subplot.set_xlabel('1-Specificity', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
subplot.set_ylabel('Sensitivity', fontsize=8)
for tick in subplot.get_yticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_auc(Values, positive, negative, labels=None, title=None):
ilabels = np.zeros((positive.sum() + negative.sum(), ), dtype=float)
ilabels[:positive.sum()] = 1
TPR = []
FPR = []
for allvalues in Values:
values = np.hstack((allvalues[positive], allvalues[negative]))
sorted = np.sort(values)[::-1]
tpr = np.zeros((values.size, ), dtype=float)
fpr = np.zeros((values.size, ), dtype=float)
for i, value in enumerate(sorted):
pos = ilabels[values >= value]
fpr[i] = (pos == 0).sum() / float((ilabels == 0).sum())
tpr[i] = (pos == 1).sum() / float((ilabels == 1).sum())
TPR.append(tpr)
FPR.append(fpr)
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
for tpr, fpr, color, label in zip(TPR, FPR, colors, labels):
subplot.plot(fpr, tpr, color=color, linewidth=1, label=label)
subplot.plot([0, 1], [0, 1], color='k', linewidth=1, alpha=0.5)
leghandle = subplot.legend(loc=4)
for text in leghandle.texts:
text.set_fontsize(6)
leghandle.set_frame_on(False)
subplot.axis([0, 1, 0, 1])
subplot.set_xlabel('1-Specificity', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
subplot.set_ylabel('Sensitivity', fontsize=8)
for tick in subplot.get_yticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_scoreprofile(boundscores, bounds, title=None):
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
num = len(bounds)
labels = ['%d - %d' % (bound[0], bound[1]) for bound in bounds[:-1]]
labels.append('> %d' % bounds[-1][0])
colors = [
colorsys.hsv_to_rgb(h, 0.9, 0.7)
for h in np.linspace(0, 1, num + 2)[:-1]
]
ymax = 0
ymin = 0
xvals = np.arange(len(boundscores[0][0]))
for score, label, color in zip(boundscores, labels, colors):
subplot.errorbar(xvals, score[0], yerr=0, color=color, marker='o', markersize=10, \
markerfacecolor=color, markeredgewidth=0, capsize=0, \
linestyle='None', label=label)
ymax = max([ymax, score[0].max()])
ymin = min([ymin, score[0].min()])
subplot.axis([0, len(boundscores[0][0]), ymin - 0.5, ymax + 0.5])
leghandle = subplot.legend(loc=1)
for text in leghandle.texts:
text.set_fontsize(8)
leghandle.set_frame_on(False)
subplot.set_xlabel('Position', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
subplot.set_xlabel('LogLikehood Ratio', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_mnaseprofile(profiles, bounds, motiflen=None, title=None):
num = len(profiles)
if type(bounds[0]) is tuple:
labels = ['%d - %d'%(bound[0],bound[1]) for bound in bounds[:-1]]
labels.append('> %d'%bounds[-1][0])
else:
labels = bounds
colors = [colorsys.hsv_to_rgb(h,0.9,0.7) for h in np.linspace(0,1,num+2)[:-1]]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
ymax = 0
for num, (profile,color,label) in enumerate(zip(profiles,colors,labels)):
width = profile.size
xvals = np.arange(-width/2,width/2)
subplot.plot(xvals, profile, color=color, linestyle='-', linewidth=1, label=label)
ymax = max([ymax, profile.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), 0, ymax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
if motiflen:
subplot.axvline(motiflen-1, linestyle='--', c='g', linewidth=0.2)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_scoreprofile(boundscores, bounds, title=None):
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
num = len(bounds)
labels = ['%d - %d'%(bound[0],bound[1]) for bound in bounds[:-1]]
labels.append('> %d'%bounds[-1][0])
colors = [colorsys.hsv_to_rgb(h,0.9,0.7) for h in np.linspace(0,1,num+2)[:-1]]
ymax = 0
ymin = 0
xvals = np.arange(len(boundscores[0][0]))
for score,label,color in zip(boundscores,labels,colors):
subplot.errorbar(xvals, score[0], yerr=0, color=color, marker='o', markersize=10, \
markerfacecolor=color, markeredgewidth=0, capsize=0, \
linestyle='None', label=label)
ymax = max([ymax, score[0].max()])
ymin = min([ymin, score[0].min()])
subplot.axis([0, len(boundscores[0][0]), ymin-0.5, ymax+0.5])
leghandle = subplot.legend(loc=1)
for text in leghandle.texts:
text.set_fontsize(8)
leghandle.set_frame_on(False)
subplot.set_xlabel('Position', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
subplot.set_xlabel('LogLikehood Ratio', fontsize=8)
for tick in subplot.get_xticklabels():
tick.set_fontsize(8)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_footprint(footprints, labels, stderr=None, motif=None, title=None):
import os
from matplotlib.offsetbox import AnnotationBbox, OffsetImage
from matplotlib._png import read_png
num = len(footprints)
# colors = [colorsys.hsv_to_rgb(h,0.9,0.7) for h in np.linspace(0,1,num+2)[:-1]]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
fwdmax = 0
revmax = 0
for num, (footprint,color,label) in enumerate(zip(footprints,colors,labels)):
width = footprint.size/2
fwd = footprint[:width]
rev = footprint[width:]
xvals = np.arange(-width/2,width/2)
alpha = 0.7
subplot.plot(xvals, fwd, color=color, linestyle='-', linewidth=1, label=label, alpha=alpha)
subplot.plot(xvals, -1*rev, color=color, linestyle='-', linewidth=1, label="_nolabel_", alpha=alpha)
if stderr is not None:
if stderr[num] is not None:
subplot.fill_between(xvals, footprint[:width]-stderr[num][:width]/2, footprint[:width]+stderr[num][:width]/2, \
alpha=0.3, edgecolor=color, facecolor=color)
subplot.fill_between(xvals, -(footprint[width:]+stderr[num][width:]/2), \
-(footprint[width:]-stderr[num][width:]/2), alpha=0.3, edgecolor=color, \
facecolor=color)
fwdmax = max([fwdmax, fwd.max()])
revmax = max([revmax, rev.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), -1*revmax, fwdmax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
legend.set_frame_on(False)
if motif:
subplot.axvline(len(motif)-1, linestyle='--', c='g', linewidth=0.2)
if motif:
# overlap motif diagram over footprint
motif.has_instances = False
tag = '_'.join(title.split(' / '))
tmpfile = "/mnt/lustre/home/anilraj/linspec/fig/footprints/pwmlogo_%s.png"%tag
motif.weblogo(tmpfile)
zoom = 0.15*len(motif)/15.
try:
handle = read_png(tmpfile)
imagebox = OffsetImage(handle, zoom=zoom)
xy = [len(motif)/2-1,0]
ab = AnnotationBbox(imagebox, xy, xycoords='data', frameon=False)
subplot.add_artist(ab)
except RuntimeError:
print "Could not retrieve weblogo"
pass
os.remove(tmpfile)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_footprint(footprints, labels, stderr=None, motif=None, title=None):
import os
from matplotlib.offsetbox import AnnotationBbox, OffsetImage
from matplotlib._png import read_png
num = len(footprints)
# colors = [colorsys.hsv_to_rgb(h,0.9,0.7) for h in np.linspace(0,1,num+2)[:-1]]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
fwdmax = 0
revmax = 0
for num, (footprint, color,
label) in enumerate(zip(footprints, colors, labels)):
width = footprint.size / 2
fwd = footprint[:width]
rev = footprint[width:]
xvals = np.arange(-width / 2, width / 2)
alpha = 0.7
subplot.plot(xvals,
fwd,
color=color,
linestyle='-',
linewidth=1,
label=label,
alpha=alpha)
subplot.plot(xvals,
-1 * rev,
color=color,
linestyle='-',
linewidth=1,
label="_nolabel_",
alpha=alpha)
if stderr is not None:
if stderr[num] is not None:
subplot.fill_between(xvals, footprint[:width]-stderr[num][:width]/2, footprint[:width]+stderr[num][:width]/2, \
alpha=0.3, edgecolor=color, facecolor=color)
subplot.fill_between(xvals, -(footprint[width:]+stderr[num][width:]/2), \
-(footprint[width:]-stderr[num][width:]/2), alpha=0.3, edgecolor=color, \
facecolor=color)
fwdmax = max([fwdmax, fwd.max()])
revmax = max([revmax, rev.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), -1 * revmax, fwdmax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
legend.set_frame_on(False)
if motif:
subplot.axvline(len(motif) - 1, linestyle='--', c='g', linewidth=0.2)
if motif:
# overlap motif diagram over footprint
motif.has_instances = False
tag = '_'.join(title.split(' / '))
tmpfile = "/mnt/lustre/home/anilraj/linspec/fig/footprints/pwmlogo_%s.png" % tag
motif.weblogo(tmpfile)
zoom = 0.15 * len(motif) / 15.
try:
handle = read_png(tmpfile)
imagebox = OffsetImage(handle, zoom=zoom)
xy = [len(motif) / 2 - 1, 0]
ab = AnnotationBbox(imagebox, xy, xycoords='data', frameon=False)
subplot.add_artist(ab)
except RuntimeError:
print "Could not retrieve weblogo"
pass
os.remove(tmpfile)
if title:
plot.suptitle(title, fontsize=10)
return figure
def plot_dnaseprofile(footprints, bounds, motiflen=None, title=None):
num = len(footprints)
if type(bounds[0]) is tuple:
labels = ['%d - %d' % (bound[0], bound[1]) for bound in bounds[:-1]]
labels.append('> %d' % bounds[-1][0])
else:
labels = bounds
colors = [
colorsys.hsv_to_rgb(h, 0.9, 0.7)
for h in np.linspace(0, 1, num + 2)[:-1]
]
figure = plot.figure()
subplot = figure.add_subplot(111)
subplot = vizutils.remove_spines(subplot)
fwdmax = 0
revmax = 0
for num, (footprint, color,
label) in enumerate(zip(footprints, colors, labels)):
width = footprint.size / 2
fwd = footprint[:width]
rev = footprint[width:]
xvals = np.arange(-width / 2, width / 2)
alpha = 0.7
subplot.plot(xvals,
fwd,
color=color,
linestyle='-',
linewidth=1,
label=label,
alpha=alpha)
subplot.plot(xvals,
-1 * rev,
color=color,
linestyle='-',
linewidth=1,
label="_nolabel_",
alpha=alpha)
fwdmax = max([fwdmax, fwd.max()])
revmax = max([revmax, rev.max()])
subplot.axhline(0, linestyle='--', linewidth=0.2)
subplot.axvline(0, linestyle='--', linewidth=0.2)
subplot.axis([xvals.min(), xvals.max(), -1 * revmax, fwdmax])
legend = subplot.legend(loc=4)
for text in legend.texts:
text.set_fontsize('8')
if motiflen:
subplot.axvline(motiflen - 1, linestyle='--', c='g', linewidth=0.2)
if title:
plot.suptitle(title, fontsize=10)
return figure
