def drawData( axes, stats, options ):
#halfsize = len(stats)/2 + len(stats)%2
#colors = libplot.getColors2( halfsize )
#colors = libplot.getColors2( len(stats) )
#styles = { 0:'-', 1:'--' }
colors = libplot.getColors1()
if len(stats) < 1:
return
if stats[0].reference == "reference":
colors.pop(0)
elif stats[0].reference == 'hg19':
colors.pop(1)
#===========
#dash = 0
colorindex = -1
lines = []
sampleNames = []
ymin = float('inf')
ref = ''
for sample in stats:
sampleNames.append(sample.name)
if ref == '':
ref = sample.reference
xdata = []
ydata = []
for bucket in sample:
xdata.append( bucket.mid )
if options.includeCov:
ydata.append( bucket.correctPerSample )
else:
ydata.append( bucket.correctPerAligned )
#if not dash:
# colorindex += 1
#if colorindex == 1:
# colorindex += 1
colorindex +=1
ymin = min([ymin, min(ydata)])
l = axes.plot( xdata, ydata, color=colors[colorindex], linewidth=1 )
#l = axes.plot( xdata, ydata, color=colors[colorindex], linestyle=styles[dash], linewidth=0.5 )
lines.append(l)
#dash = not dash
libplot.editSpine( axes )
title = options.title
if ref != '':
title += ', %s' % libplot.properName(ref)
axes.set_title(title)
pyplot.xlabel("Distance")
pyplot.ylabel("Correct proportion")
return lines, sampleNames, ymin
def drawData( axesList, samples, samplesPerPlot, options, proportion, culm ):
largeIns = [] #List of proportion of total indel bases that indels >= 1000bp take up, each element is for each sample
largeDels = []
if len(axesList) %2 != 0:
sys.stderr.write( 'Number of axes must be even. Got %d\n' %len(axesList) )
sys.exit( 1 )
colors = libplot.getColors1()
if len(samples) < 1:
return
if samples[0].attrib["referenceName"] == "reference":
colors.pop(0)
elif samples[0].attrib["referenceName"] == 'hg19':
colors.pop(1)
#styles = []
c = -1
textsize = 'x-small'
linesDict = {}
labelsDict = {}
xmax = float('-inf')
ymax = float('-inf')
xmin = float('inf')
ymin = float('inf')
for i in range( len(axesList)/2 ):
inslines = []
dellines = []
sampleNames = []
insAxes = axesList[ i ]
delAxes = axesList[ i + len(axesList)/2 ]
startIndex = i * samplesPerPlot
endIndex = min( [startIndex + samplesPerPlot, len(samples)] )
for j in range( startIndex, endIndex ):
sample = samples[j]
sampleNames.append( sample.attrib[ 'sampleName' ] )
insDist = [int(val) for val in sample.attrib[ 'insertionSizeDistribution' ].split()]
#insXdata, insYdata = getFreq( insDist, options.xlogscale, options.ylogscale )
insXdata, insYdata = getFreq( insDist, proportion, culm )
delDist = [int(val) for val in sample.attrib[ 'deletionSizeDistribution' ].split()]
#delXdata, delYdata = getFreq( delDist, options.xlogscale, options.ylogscale )
delXdata, delYdata = getFreq( delDist, proportion, culm )
#LARGE INDELS, FOR paper STATS, not related to the plot:
if proportion and culm:
largeIns.append( getLargeIndelProp(insXdata, insYdata) )
largeDels.append( getLargeIndelProp(delXdata, delYdata) )
c += 1
il = insAxes.plot( insXdata, insYdata, color=colors[c] )
dl = delAxes.plot( delXdata, delYdata, color=colors[c] )
inslines.append( il )
dellines.append( dl )
insXmax = xmax
delXmax = xmax
if len(insXdata) >0:
insXmax = max(insXdata)
if len(delXdata) > 0:
delXmax = max(delXdata)
xmax = max( [xmax, insXmax, delXmax] )
insYmax = ymax
delYmax = ymax
if len(insYdata) >0:
insYmax = max(insYdata)
if len(delYdata) > 0:
delYmax = max(delYdata)
ymax = max( [ymax, insYmax, delYmax] )
insXmin = xmin
delXmin = xmin
if len(insXdata) >0:
insXmin = min(insXdata)
if len(delXdata) > 0:
delXmin = min(delXdata)
xmin = min( [xmin, insXmin, delXmin] )
insYmin = ymin
delYmin = ymin
if len(insYdata) >0:
insYmin = min(insYdata)
if len(delYdata) > 0:
delYmin = min(delYdata)
ymin = min( [ymin, insYmin, delYmin] )
#xmax = max([xmax, max(insXdata), max(delXdata)])
#ymax = max([ymax, max(insYdata), max(delYdata)])
linesDict[ i ] = inslines
labelsDict[ i ] = sampleNames
linesDict[ i + len(axesList)/2 ] = dellines
labelsDict[ i + len(axesList)/2 ] = sampleNames
#fontp = FontProperties()
#fontp.set_size( 'x-small' )
if i == 0:
insAxes.set_title( 'Insertions' )
delAxes.set_title( 'Deletions' )
for i in range( len(axesList) ):
axes = axesList[ i ]
if options.xlogscale == "true":
axes.set_xscale('log')
#if options.ylogscale == "true" and not proportion:
if options.ylogscale == "true":
axes.set_yscale('log')
libplot.editSpine( axes )
axes.set_xlabel('Length (bp)', size = textsize)
if not proportion:
axes.set_ylabel('Event number', size = textsize)
else:
axes.set_ylabel('Number of positions', size = textsize)
axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
#if options.xlogscale == "true":
# axes.set_xlabel('Log 2 of length (bp)', size = textsize)
#else:
# axes.set_xlabel('Length (bp)', size = textsize)
#if options.ylogscale == "true":
# axes.set_ylabel('Log 2 of count', size = textsize)
#else:
# axes.set_ylabel('Count', size = textsize)
#Legend
legend = axes.legend( linesDict[ i ], [ libplot.properName(n) for n in labelsDict[ i ]], 'upper right', ncol=3 )
for t in legend.get_texts():
t.set_fontsize('x-small')
legend._drawFrame = False
if options.xlogscale == "true":
scale = len(str(xmax)) -1
xticks = [ 10**x for x in range(scale + 1) ]
axes.set_xticks( xticks )
#if options.ylogscale == "true" and not proportion:
if options.ylogscale == "true":
scale = len(str(ymax)) -1
yticks = [ 10**y for y in range(scale + 1) ]
axes.set_yticks( yticks )
for label in axes.get_xticklabels():
#label.set_rotation(75)
label.set_fontsize( textsize )
for label in axes.get_yticklabels():
label.set_fontsize( textsize )
#box = axes.get_position()
#axes.set_position( [box.x0, box.y0, box.width*0.8, box.height] )
#legend = pyplot.legend( lines, options.keys, numpoints=1, prop=fontP, loc="best", bbox_to_anchor=(1, 0.9) )
#legend._drawFrame = False
#libplot.setTicks( axes )
#axes.set_xticks( range( 0, len(samples) ) )
#axes.set_xticklabels( sampleNames )
#for label in axes.xaxis.get_ticklabels():
# label.set_rotation( 90 )
#axes.xaxis.set_ticks_position( 'bottom' )
#axes.yaxis.set_ticks_position( 'left' )
axes.set_ylim( ymin, ymax )
if proportion and not culm:
axes.set_xlim( xmin, 100 )
else:
axes.set_xlim( xmin, xmax )
#PRINT THE LARGE INDEL STATS:
if proportion and culm:
sys.stderr.write("largeIndelStats\n")
sys.stderr.write("Large insertions: %f\n" %( sum(largeIns)/len(largeIns) ))
sys.stderr.write("Large deletions: %f\n" %( sum(largeDels)/len(largeDels) ))
largeIndels = [ (largeIns[i] + largeDels[i])/2.0 for i in range(len(largeIns)) ]
sys.stderr.write("IndelsAverage: %f\n" %( sum(largeIndels)/len(largeIndels) ))
return
def drawCompareData( axesList, xstats, ystats, options ):
#Only draw the overlapped samples:
#colors = libplot.getColors2( len(xstats) )
colors = libplot.getColors1()
#colorindex = -1
#colorindex = 0
colorindex = 1
lines = []
sampleNames = []
p0axes = axesList[0] #plot 0 axes (see def 'setCompareAxes')
aggData = [] #data points (buckets) of all samples
minval = float('inf')
for xsample in xstats:
ysample = getSample( ystats, xsample.name )
if ysample is None:
continue
xsample, ysample = intersect(xsample, ysample)
#if len(xsample) != len(ysample):
# xsample, ysample = intersect(xsample, ysample)
# sys.stderr.write( "Error: Two xml files do not have the same number of buckets for sample %s\n" % xsample.name )
#sys.exit( 1 )
data = [] #list of (x,y) tuples
colorindex += 1
for i in range( len( xsample ) ): #each bucket
if xsample[i].mid != ysample[i].mid:
sys.stderr.write( "Two xml files have different buckets\n " )
sys.exit( 1 )
if options.includeCov:
data.append( (xsample[i].correctPerSample, ysample[i].correctPerSample) )
else:
data.append( (xsample[i].correctPerAligned, ysample[i].correctPerAligned) )
x2data = [ point[0] for point in data ]
y2data = [ point[1] for point in data ]
l = p0axes.plot( x2data, y2data, color=colors[colorindex], marker='.', markersize=4.0, linestyle='none' )
lines.append( l )
sampleNames.append( xsample.name )
aggData.extend( data )
minval = min( [min(x2data), min(y2data)] )
#Draw the y=x line
x = [0, 1]
y = [0, 1]
p0axes.plot(x, y, color="#919191")
fontP = FontProperties()
fontP.set_size('small')
libplot.editSpine( p0axes )
p0axes.set_title(options.title)
p0axes.set_xlabel( libplot.properName(xstats.refname) )
p0axes.set_ylabel( libplot.properName(ystats.refname) )
libplot.setTicks( p0axes )
for l in p0axes.xaxis.get_ticklabels():
l.set_fontsize('small')
for l in p0axes.yaxis.get_ticklabels():
l.set_fontsize('small')
#legend:
legend = p0axes.legend( lines, [ libplot.properName(n) for n in sampleNames], 'lower right', numpoints = 1, prop=fontP, ncol = 2)
legend._drawFrame = False
#p0axes.set_xlim( -0.005, 1.005 )
#p0axes.set_ylim( -0.005, 1.005 )
ycutoff = minval
if options.ycutoff:
ycutoff = options.ycutoff
p0axes.set_xlim( ycutoff - (1-ycutoff)*0.02, 1 + (1 - ycutoff)*0.01 )
p0axes.set_ylim( ycutoff - (1-ycutoff)*0.02, 1 + (1 - ycutoff)*0.01 )
#box = p0axes.get_position()
#p0axes.set_position([box.x0, box.y0, box.width * 0.8, box.height * 0.8])
#legend = pyplot.legend( lines, sampleNames, numpoints = 1, prop= fontP, loc="best", bbox_to_anchor=(1, 0.6))
#legend._drawFrame=False
#DRAW AGGREGATE DATA (plot 1 and plot 2):
nbins = 20
p1axes = axesList[1]
y1min, y1max = drawAggData( p1axes, aggData, 0, 0, 1, ycutoff, nbins )
y1lim = max( abs(y1min), abs(y1max) )
p1axes.set_ylim( -y1lim*1.1, y1lim*1.1 )
p1axes.set_xlim( ycutoff - (1-ycutoff)*0.02, 1 + (1-ycutoff)*0.01 )
#p1axes.set_ylim( y1min*1.1, y1max*1.1 )
for loc, spine in p1axes.spines.iteritems():
if loc == 'left':
spine.set_position( ( 'outward', 10 ) )
spine.set_color( 'none' )
p1axes.axhline( 0, color = '#000000' )
p1axes.xaxis.set_major_locator( NullLocator() )
p1axes.xaxis.set_major_formatter( NullFormatter() )
p1axes.yaxis.set_ticks([-y1lim, 0, y1lim])
for l in p1axes.yaxis.get_ticklabels():
l.set_fontsize('small')
p2axes = axesList[2]
x2min, x2max = drawAggData( p2axes, aggData, 1, 0, 1, ycutoff, nbins )
x2lim = max( abs(x2min), abs(x2max) )
p2axes.set_xlim( -x2lim*1.1, x2lim*1.1 )
p2axes.set_ylim( ycutoff - (1-ycutoff)*0.02, 1 + (1- ycutoff)*0.01 )
#p2axes.set_xlim( x2min*1.1, x2max*1.1 )
for loc, spine in p2axes.spines.iteritems():
if loc == 'bottom':
spine.set_position( ( 'outward', 10 ) )
spine.set_color( 'none' )
p2axes.axvline( 0, color = '#000000' )
p2axes.yaxis.set_major_locator( NullLocator() )
p2axes.yaxis.set_major_formatter( NullFormatter() )
p2axes.xaxis.set_ticks([-x2lim, 0, x2lim])
for l in p2axes.xaxis.get_ticklabels():
l.set_fontsize('small')
l.set_rotation( 45 )
return
