def tabHeader(f, ref1, ref2):
f.write("\\begin{table}\n")
f.write("\\begin{center}\n")
f.write("\\scalebox{1}{%\n")
f.write("\\begin{tabular}{c|r|r|r|r}\n")
#f.write("\\multicolumn{6}{c}{%s} \\\\\n" %title)
#f.write("\\hline\n")
f.write("\\hline\n")
f.write("Sample & Repeat & %s & %s & Total\\\\\n" %(libplot.properName(ref1), libplot.properName(ref2)))
#f.write("Sample & Reads & Total Bases & \\%%Repeats & SNP Rate & Overal Snp Rate\\\\\n")
f.write("\\hline\n")
def drawPlot(samplesList, sampleNames, options):
options.out = os.path.join(options.outdir, "nonLinearBp")
fig, pdf = libplot.initImage(12.0, 8.0, options)
axes = fig.add_axes([0.09, 0.2, 0.9, 0.6])
list1 = samplesList[0]
list2 = samplesList[1]
if len(list1) < 1 or len(list2) < 1:
return
refname1 = list1[0].attrib['referenceName']
refname2 = list2[0].attrib['referenceName']
lines = []
barwidth = 0.3
y1data = []
y2data = []
for sample in sampleNames:
for s in list1:
if sample == s.attrib['sampleName']:
y1data.append( int(s.attrib['totalIntraJoin']) )
for s in list2:
if sample == s.attrib['sampleName']:
y2data.append( int(s.attrib['totalIntraJoin']) )
x1data = range( len(y1data) )
x2data = [ x+ barwidth for x in x1data]
colors =["#1F78B4", "#E31A1C"]
l1 = axes.bar( x1data, y1data, barwidth, color = colors[0], ec='w')
lines.append(l1[0])
l2 = axes.bar( x2data, y2data, barwidth, color = colors[1], ec='w')
lines.append(l2[0])
libplot.editSpine(axes)
axes.set_title("Non-linear Breakpoints")
#set ticks
xlabels = [ libplot.properName(name) for name in sampleNames ]
fontP = FontProperties()
fontP.set_size('small')
pyplot.xticks(x2data, xlabels, rotation=45, fontproperties=fontP)
pyplot.yticks( fontproperties = fontP )
pyplot.xlabel("Samples")
pyplot.ylabel("Number of breakpoints")
axes.xaxis.set_ticks_position('bottom')
axes.yaxis.set_ticks_position('left')
axes.yaxis.grid(b=True, color="#A8A8A8", linestyle='-', linewidth=0.25)
legend = axes.legend( lines, [libplot.properName(refname1), libplot.properName(refname2)], prop=fontP, loc="best" )
legend._drawFrame = False
libplot.writeImage(fig, pdf, options)
def tab( f, samplesList, sampleNames ):
refname1 = samplesList[0][0].attrib['referenceName']
refname2 = samplesList[1][0].attrib['referenceName']
getNonLinearOps(samplesList[0])
getNonLinearOps(samplesList[1])
for s in sampleNames:
#altColor = 1
for altColor in [1,0]:
#Get #Deletions, #Non-linearOps
numDels = -1
numDelsPerAlignedBase = -1
numNonLinearOps = -1
numNonLinearOpsPerAlignedBase = -1
for sample in samplesList[altColor]:
if sample.attrib['sampleName'] == s:
numDels = int( sample.attrib['totalInsertion'] )
numDelsPerAlignedBase = float( sample.attrib['totalInsertionPerAlignedBase'] )
numDelsPerAlignedBase = prettyFloat(numDelsPerAlignedBase)
#numNonLinearOps = int(sample.attrib['totalIntraJoin']) + int(sample.attrib['totalInterJoin'])
#numNonLinearOpsPerAlignedBase = float(sample.attrib['totalInterJoinPerAlignedBase']) + float(sample.attrib['totalIntraJoinPerAlignedBase'])
numNonLinearOps = int(sample.attrib['totalIntraJoin'])
numNonLinearOpsPerAlignedBase = float(sample.attrib['totalIntraJoinPerAlignedBase'])
numNonLinearOpsPerAlignedBase = prettyFloat(numNonLinearOpsPerAlignedBase)
break
#Get the Snps#
numSnps = -1
numSnpsPerAlignedBase = -1
for sample in samplesList[altColor + 2]:
if sample.attrib['sampleName'] == s:
numSnps = int(sample.attrib['totalErrors'])
numSnpsPerAlignedBase = '0'
if float( sample.attrib['totalCalls'] ) != 0:
numSnpsPerAlignedBase = numSnps/float( sample.attrib['totalCalls'])
numSnpsPerAlignedBase = prettyFloat( numSnpsPerAlignedBase )
break
if altColor == 1:
f.write("\\multirow{2}{*}{%s} &\\cellcolor[gray]{0.9} %s & \\cellcolor[gray]{0.9} %d (%s) & \\cellcolor[gray]{0.9} %d & \\cellcolor[gray]{0.9} %d (%s) \\\\\n" % \
( libplot.properName(s), libplot.properName(refname2), numDels, numDelsPerAlignedBase, numNonLinearOps, numSnps, numSnpsPerAlignedBase))
else:
f.write("& %s & %d (%s) & %d & %d (%s) \\\\\n" %\
(libplot.properName(refname1), numDels, numDelsPerAlignedBase, numNonLinearOps, numSnps, numSnpsPerAlignedBase))
f.write("\\hline\n\n")
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 tab( f, exps, rexps, samples ):
for sample in samples:
expList = copy.copy(exps[sample])
expList.sort()
expList.append( rexps[sample] )
#sys.stderr.write('expList for sample %s: %s\n' %(sample, '\t'.join([ '%s%d' %(e.ref, e.weight)for e in expList])))
f.write( "\\multirow{%d}{*}{%s} " %( len(expList), sample ) )
#f.write( "\\multirow{%d}{*}{%s} " %( len(expList) -1, sample ) ) #HACK
for e in expList:
#if e.ref == 'cactusRef' and e.weight == 1: #HACK
# continue
ref = libplot.properName(e.ref)
if re.search('cactusRef', e.ref):
r = e.ref.lstrip('cactusRef')
ref = "%s %s" % (libplot.properName('cactusRef'), r)
if e.ref != 'cactusRef':
f.write("& %s & %s & %s & %s & %s & %s \\\\\n" %(ref, libplot.prettyInt(e.mapped), libplot.prettyInt(e.uniquelyMapped), libplot.prettyInt(e.properlyPaired), libplot.prettyInt(e.uniquelyMappedAndProperlyPaired), libplot.prettyInt(e.snps)))
elif e.ref == 'cactusRef' and e.weight == 2:
f.write("& \\cellcolor{cyan!30} %s%d & \\cellcolor{cyan!30} %s & \\cellcolor{cyan!30} %s & \\cellcolor{cyan!30} %s & \\cellcolor{cyan!30} %s & \\cellcolor{cyan!30} %s \\\\\n" %(ref, e.weight, libplot.prettyInt(e.mapped), libplot.prettyInt(e.uniquelyMapped), libplot.prettyInt(e.properlyPaired), libplot.prettyInt(e.uniquelyMappedAndProperlyPaired), libplot.prettyInt(e.snps)))
else:
f.write("& %s%d & %s & %s & %s & %s & %s \\\\\n" %(ref, e.weight, libplot.prettyInt(e.mapped), libplot.prettyInt(e.uniquelyMapped), libplot.prettyInt(e.properlyPaired), libplot.prettyInt(e.uniquelyMappedAndProperlyPaired), libplot.prettyInt(e.snps)))
f.write("\\hline\n")
def drawCnvPlot( sample, options ):
sampleName = sample.attrib[ 'sampleName' ]
#print sampleName
options.out = os.path.join( options.outdir, 'cnv_%s' %sampleName )
fig, pdf = libplot.initImage( 11.0, 3.25, options )
title = "Copy Number Variation between %s and %s" % ( libplot.properName(sampleName), libplot.properName(sample.attrib['referenceName']) )
cnvDict, minCn, maxCn = getSampleData( sample )
axDict = setAxes( fig, cnvDict.keys(), options )
for r in axDict:
if r != 'bg':
drawOneCnvPlot( r, axDict[ r ], cnvDict[ r ], options, minCn, maxCn )
drawAxisLabels( axDict, cnvDict, options, title, maxCn )
setAxisLimits( axDict, minCn, maxCn )
libplot.writeImage( fig, pdf, options )
def drawLegend( axes, lines, sampleNames, options ):
fontP = FontProperties()
fontP.set_size('small')
box= axes.get_position()
axes.set_position([box.x0, box.y0, box.width * 0.8, box.height])
#legend = pyplot.legend( lines, sampleNames, numpoints = 1, prop= fontP, loc="best", bbox_to_anchor=(1, 0.5))
if not options.legendElements:
legend = pyplot.legend( lines, [ libplot.properName(n) for n in sampleNames ], prop= fontP, loc="best", bbox_to_anchor=(1,0.5))
legend._drawFrame=False
elif len(lines) == len(options.legendElements):
legend = pyplot.legend( lines, options.legendElements, prop= fontP, loc="best", bbox_to_anchor=(1,0.5) )
legend._drawFrame=False
else:
sys.stderr.write('Number of items in --legendElements is different '
'from the number of lines plotted\n' )
def tab(f, stats, sample2repeat):
altColor = 1
for sample in stats:
repeat = 'NA'
repeatPc = ''
if sample.name in sample2repeat:
repeat = libplot.prettyInt( sample2repeat[sample.name][1] )
repeatPc = "(%.2f \\%%)" %sample2repeat[sample.name][2]
otherRef = libplot.prettyInt(sample.otherReferenceBasesMapped)
otherRefPc = "%.2f" % (100.0*sample.otherReferenceBasesMapped/sample.totalBases)
ref = libplot.prettyInt(sample.referenceBasesMapped)
refPc = "%.2f" % (100.0*sample.referenceBasesMapped/sample.totalBases)
total = libplot.prettyInt(sample.totalBases)
sampleName = libplot.properName(sample.name)
if altColor == 1:
f.write("%s & %s %s & %s (%s \\%%) & %s (%s \\%%) & %s \\\\\n" %(sampleName, repeat, repeatPc, otherRef, otherRefPc, ref, refPc, total ))
else:
f.write("\\cellcolor[gray]{0.9} %s & \\cellcolor[gray]{0.9} %s %s & \\cellcolor[gray]{0.9} %s (%s \\%%) & \\cellcolor[gray]{0.9} %s (%s \\%%) & \\cellcolor[gray]{0.9} %s \\\\\n" %(sampleName, repeat, repeatPc, otherRef, otherRefPc, ref, refPc, total ))
altColor = 1 - altColor
f.write("\\hline\n")
def drawN50Plot( options, samples ):
#sort samples:
samples = sorted( samples, key=lambda s: int(s.attrib[ options.sortkey ]), reverse=True )
sampleNames = getSampleNames( samples )
if len(samples) < 1:
return
refname = samples[0].attrib[ 'referenceName' ]
options.out = os.path.join( options.outdir, options.prefix + '_' + refname )
fig, pdf = libplot.initImage( 8.0, 10.0, options )
axes = fig.add_axes( [0.12, 0.1, 0.85, 0.85] )
title = "N50"
lines = drawN50data( axes, samples, options )
axes.set_title(title)
#Legend
fontP = FontProperties()
fontP.set_size( 'small' )
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( [ libplot.properName(n) for n in 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_xlim( -0.5, len(samples) - 0.5 )
#axes.set_ylim( -20, 6000 )
libplot.writeImage( fig, pdf, options )
def drawPlot(rexps, exps, options, outfile, type):
options.out = outfile
fig, pdf = libplot.initImage( 8.0, 10.0, options )
axes = fig.add_axes( [0.12, 0.14, 0.85, 0.8] )
#Set title:
titleDict = {'mapped':'Mapped reads', 'uniquelyMapped':'Uniquely Mapped Reads', 'properlyPaired':'Properly Paired Reads', 'uniquelyMappedAndProperlyPaired':'Uniquely Mapped And Properly Paired Reads', 'snps':'SNPs'}
axes.set_title( titleDict[type] )
if len(rexps) < 1:
return
sampleNotherRefmapped = []
ref = ''
for sample in rexps:
if sample == 'average':
continue
exp = rexps[sample]
ref = exp.ref
sampleNotherRefmapped.append( (sample, exp.total) )
otherRefName = libplot.properName( ref )
sampleNotherRefmapped = sorted( sampleNotherRefmapped, key=lambda item: item[1], reverse=True )
samples = [ item[0] for item in sampleNotherRefmapped]
samples.append( 'average' )
xdata = range( 0, len(samples) )
colors = libplot.getColors4()
#c = -1
c = 0
lines = []
ydataList, miny, maxy = getData(samples, exps, rexps, type)
#print ydataList
refs = sorted( ydataList.keys() )
#miny = float('inf')
#maxy = 0
#offset = 0.075
offset = 0.12
#if type != 'snps':
# offset = 0
#axes.set_yscale('log')
scale = -1
if miny > 1000:
scale = len( str(int(miny)) ) - 1
#Draw line connecting the data for each sample (each bin):
binXdataList = [ [] for x in xdata ]
binYdataList = [ [] for x in xdata ]
for i, ref in enumerate(refs):
xdatai = [ x + offset*i for x in xdata ]
ydata = ydataList[ref]
if scale > 0:
ydata = [ float(y)/10**scale for y in ydata ]
for j, x in enumerate(xdatai):
binXdataList[j].append(x)
binYdataList[j].append( ydata[j] )
for i in xrange( len(binXdataList) ):
axes.plot( binXdataList[i], binYdataList[i], color="#CCCCCC", linestyle='-', linewidth=0.005 )
#Draw main plots:
for i, ref in enumerate(refs):
xdatai = [ x + offset*i for x in xdata ]
ydata = ydataList[ref]
if scale > 0:
ydata = [ float(y)/10**scale for y in ydata ]
c += 1
l = axes.plot( xdatai, ydata, color=colors[c], marker='.', markersize=16.0, linestyle='none')
lines.append(l)
if scale > 0:
miny = float(miny)/10**scale
maxy = float(maxy)/10**scale
#Draw horizontal line at y = 0:
xmin = -0.4
xmax = len(samples) - 1 + offset*len(refs) + offset
axes.plot( [xmin, xmax], [0,0], color="#6B6B6B", linewidth=0.005)
fontP = FontProperties()
fontP.set_size('x-small')
yrange = maxy - miny
miny = miny - yrange*0.05
maxy = maxy + yrange*0.2
#Draw vertical lines to separate each sample:
#for i in xrange(1, len(samples)):
# d = (1 - offset*len(refs))/2.0
# x = [i - d, i - d]
# y = [miny , maxy]
# axes.plot(x,y, color="#CCCCCC", linewidth=0.005)
axes.set_xlim(xmin, xmax )
axes.set_ylim( miny, maxy )
libplot.editSpine( axes )
axes.set_xticks( [ i + offset*(len(refs)/2.0) for i in range(0, len(samples))] )
axes.set_xticklabels( samples )
for label in axes.xaxis.get_ticklabels():
label.set_rotation(90)
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
properRefs = []
for r in refs:
if re.search('cactusRef', r):
r = r.lstrip('cactusRef')
properRefs.append( "%s %s" %(libplot.properName('cactusRef'), r))
else:
properRefs.append( libplot.properName(r) )
legend = pyplot.legend( lines,properRefs, numpoints=1, loc='best', prop=fontP)
legend._drawFrame = False
axes.set_xlabel( 'Samples' )
axes.set_ylabel( 'Percentage of mapping difference between C. Ref. and %s' % otherRefName)
if scale > 0:
axes.set_ylabel( 'Event counts (x%d)' %(10**scale) )
#axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
libplot.writeImage( fig, pdf, options )
def drawSamplePlot(rexps, exps, options, outfile, type):
options.out = outfile
fig, pdf = libplot.initImage( 11.2, 10.0, options )
axes = fig.add_axes( [0.14, 0.12, 0.8, 0.8] )
#Set title:
axes.set_title( "SNP Rate Using BWA Mapping" )
sampleNsize = []
if len(rexps) < 1:
return
ref = ''
for sample in rexps:
if sample == 'average':
continue
exp = rexps[sample]
ref = exp.ref
#sampleNsize.append( (sample, exp.snps) )
sampleNsize.append( (sample, exp.snprate) )
otherRefName = ref
sampleNsize = sorted( sampleNsize, key=lambda item: item[1], reverse=True )
samples = [ item[0] for item in sampleNsize]
samples.append( 'average' )
#Get ydata:
ydata1 = [] #otherRef (hg19, apd, ...)
ydata2 = [] #cactusRef2
for sample in samples:
explist = exps[sample]
otherRef = rexps[sample]
ydata1.append( otherRef.snprate )
for e in explist:
if e.ref == 'cactusRef' and e.weight == 2:
ydata2.append( e.snprate )
miny = min([min(ydata1), min(ydata2)])
maxy = max([max(ydata1), max(ydata2)])
xdata = range( 0, len(samples) )
#colors = ["#E31A1C", "#1F78B4"] #red, blue
colors = ["#1F78B4", "#E31A1C"] #red, blue
scale = -1
if miny > 1000:
scale = len( str(int(miny)) ) - 1
if scale > 0:
ydata1 = [ float(y)/10**scale for y in ydata1 ]
ydata2 = [ float(y)/10**scale for y in ydata2 ]
lines = []
lines.append( axes.plot(xdata, ydata1, color=colors[0], marker=".", markersize=16.0, linestyle='none') )
lines.append( axes.plot(xdata, ydata2, color=colors[1], marker=".", markersize=16.0, linestyle='none') )
if scale > 0:
miny = float(miny)/10**scale
maxy = float(maxy)/10**scale
fontP = FontProperties()
fontP.set_size('x-small')
axes.set_xlim(-0.4, len(samples) - 0.6 )
yrange = maxy - miny
miny = miny - yrange*0.05
maxy = maxy + yrange*0.1
axes.set_ylim( miny, maxy )
libplot.editSpine( axes )
axes.set_xticks( xdata )
axes.set_xticklabels( samples )
for label in axes.xaxis.get_ticklabels():
label.set_rotation(90)
axes.yaxis.set_ticks_position( 'left' )
axes.xaxis.set_ticks_position( 'bottom' )
legend = pyplot.legend( lines, [libplot.properName(otherRefName), libplot.properName("cactusRef")], numpoints=1, loc='best', prop=fontP)
legend._drawFrame = False
axes.set_xlabel( 'Samples' )
axes.set_ylabel( 'SNPs Per Site' )
if scale > 0:
axes.set_ylabel( 'Snp counts (x%d)' %(10**scale) )
axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
libplot.writeImage( fig, pdf, options )
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 drawSnpPlot(options, samples1, samples2):
#All the samples sorted indecreasing order of SNP rate, then average, then chimp
samples1 = sorted( samples1, key=lambda s:s.errPerSite, reverse=True )
if len( samples1 ) < 1:
return
chimpSample = None
#remove chimpSample
for i, s in enumerate(samples1):
if s.name == "panTro3":
chimpSample = samples1.pop(i)
break
refname1 = samples1[0].refname
refname2 = samples2[0].refname
y1data = [ s.errPerSite for s in samples1 ]
xticklabels = [s.name for s in samples1]
y1data.append(0) #snps of refname1 w.r.t itself (which is 0)
xticklabels.append(refname1)
y2data = []
for name in xticklabels:
if name == refname2:
y2data.append(0) #snps of refname2 w.r.t itself
for s2 in samples2:
if s2.name == name:
y2data.append(s2.errPerSite)
break
if len(y1data) != len(y2data):
sys.stderr.write("Input files have different number of samples: %d, %d\n" %(len(y1data), len(y2data)))
sys.exit(1)
#add the average column:
if len(y1data) >= 2:
y1avr = sum(y1data)/float(len(y1data) -1)
y1data.append(y1avr)
y2avr = sum(y2data)/float(len(y2data) -1)
y2data.append(y2avr)
#Print summary stats to stderr:
sys.stderr.write("%s\t%f\t%f\t%f\n" %( refname1, sorted(y1data)[1] , max(y1data), y1avr ))
sys.stderr.write("%s\t%f\t%f\t%f\n" %( refname2, sorted(y2data)[1] , max(y2data), y2avr ))
xticklabels.append('average')
#add chimp:
samples1.append(chimpSample)
y1data.append( chimpSample.errPerSite )
for s in samples2:
if s.name == 'panTro3':
y2data.append( s.errPerSite )
xticklabels.append( 'panTro3' )
#Min, max values:
num = options.numOutliners
numcols = len(y1data)
minOutlier = min( [ min(y1data[numcols -num:]), min(y2data[numcols - num:]) ] ) - 0.001
maxOutlier = max( [ max(y1data[numcols -num:]), max(y2data[numcols - num:]) ] ) + 0.001
minMajority = min( [min(y1data[:numcols - num]), min(y2data[: numcols - num])] ) - 0.001
maxMajority = max( [max(y1data[:numcols - num]), max(y2data[: numcols - num])] ) + 0.001
if minMajority < 0:
minMajority = -0.0001
#Set up
basename = os.path.basename(options.files[0])
options.out = os.path.join(options.outdir, '%s' %(basename.lstrip('snpStats').lstrip('_').rstrip('.xml')) )
fig, pdf = libplot.initImage( 11.2, 10.0, options )
ax, ax2 = setAxes(fig, maxOutlier - minOutlier, maxMajority - minMajority)
#Plot the outliers:
l2 = ax.plot( y2data, marker='.', markersize=14.0, linestyle='none', color="#E31A1C" )#Red
l1 = ax.plot( y1data, marker='.', markersize=14.0, linestyle='none', color="#1F78B4" )#Blue
ax2.plot( y2data, marker='.', markersize=14.0, linestyle='none', color="#E31A1C" )
ax2.plot( y1data, marker='.', markersize=14.0, linestyle='none', color="#1F78B4" )
#Legend
fontP = FontProperties()
fontP.set_size("x-small")
legend = ax.legend([l1, l2], [libplot.properName(refname1), libplot.properName(refname2)], 'upper left', numpoints=1, prop=fontP)
legend._drawFrame = False
d = .0001 # how big to make the diagonal lines in axes coordinates
ax.plot( (-1, 0), (minOutlier +d, minOutlier - d), color = "k", clip_on=False )
ax2.plot( (-1, 0), (maxMajority +d, maxMajority - d), color = "k", clip_on=False )
ax.set_ylim( minOutlier, maxOutlier ) # outliers only
ax.set_xlim( -0.5, len(xticklabels) -0.5 )
dummyxticklabels = [ "" for l in xticklabels ]
ax.set_xticklabels( dummyxticklabels )
#Make sure the y ticks of the top plot (the outlier plot) is the same with the other plot:
step = 0.001
ytickpositions = []
ytickpos = 0
while ytickpos < maxOutlier:
if ytickpos >= minOutlier:
ytickpositions.append(ytickpos)
ytickpos += step
ax.set_yticks( ytickpositions )
ax2.set_ylim( minMajority, maxMajority )
ax2.set_xlim( -0.5, len(xticklabels) -0.5 )
# hide the spines between ax and ax2
ax.spines['bottom'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.yaxis.set_ticks_position( 'left' )
ax.xaxis.set_ticks_position( 'none' )
ax2.spines['top'].set_visible(False)
ax2.spines['right'].set_visible(False)
ax2.xaxis.tick_bottom()
ax2.yaxis.set_ticks_position( 'left' )
ax2.set_xticks( range( 0, len(xticklabels) ) )
properxticklabels = [ libplot.properName(l) for l in xticklabels ]
ax2.set_xticklabels( properxticklabels )
#Make sure the x ticks of the top plot is the same with the other plot:
ax.set_xticks( range(0, len(xticklabels)) )
for label in ax2.xaxis.get_ticklabels():
label.set_rotation( 90 )
ax.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.set_xlabel( 'Samples' )
ax2.set_ylabel( 'SNPs per site' )
title = 'SNPs'
ax.set_title( title )
libplot.writeImage( fig, pdf, options )
def drawPlot( options, samples1, samples2, type ):
#Sorted in decreasing order of errorPerSite in samples1
if type == 'insertion':
samples1 = sorted( samples1, key=lambda s:s.ins, reverse=True )
else:
samples1 = sorted( samples1, key=lambda s:s.dels, reverse=True )
if len( samples1 ) < 1:
return
#remove chimpSample:
chimpSample = None
for i, s in enumerate(samples1):
if s.name == 'panTro3':
chimpSample = samples1.pop(i)
break
refname1 = samples1[0].refname
refname2 = samples2[0].refname
y1data = [ s.ins for s in samples1 ]
if type == 'deletion':
y1data = [ s.dels for s in samples1 ]
xticklabels = [ s.name for s in samples1 ]
#indel of refname1 w.r.t itself (0)
y1data.append(0)
xticklabels.append(refname1)
y2data = []
for name in xticklabels:
if name == refname2:#indel of refname2 w.r.t itself (0)
y2data.append(0)
for s in samples2:
if s.name == name:
if type == 'insertion':
y2data.append(s.ins)
else:
y2data.append(s.dels)
break
if len(xticklabels) != len(y2data):
sys.stderr.write("Input file 1 and 2 do not have the same set of samples\n")
sys.exit( 1 )
#add the average column:
num = 1
y1avr = sum(y1data)/float(len(y1data) - 1)
y1data.append(y1avr)
xticklabels.append('average')
y2avr = sum(y2data)/float(len(y2data) - 1)
y2data.append(y2avr)
print "%s Average: %s %f, %s %f" %(type, refname1, y1avr, refname2, y2avr)
#Add chimp:
samples1.append(chimpSample)
if type == 'insertion':
y1data.append( chimpSample.ins )
else:
y1data.append( chimpSample.dels )
for s in samples2:
if s.name == 'panTro3':
if type == 'insertion':
y2data.append(s.ins)
else:
y2data.append(s.dels)
xticklabels.append("panTro3")
minMajority = min( [min(y2data), min(y1data)] ) - 0.0001
maxMajority = max( [max(y2data), max(y1data)] ) + 0.0001
basename = os.path.basename(options.files[0])
options.out = os.path.join( options.outdir, '%s_%s' %( type, basename.lstrip('pathStats').lstrip('_').rstrip('.xml') ) )
fig, pdf = libplot.initImage( 11.2, 10.0, options )
#ax, ax2 = setAxes(fig, maxOutlier - minOutlier, maxMajority - minMajority)
ax2 = fig.add_axes( [0.15, 0.15, 0.8, 0.8] )
l2 = ax2.plot( y2data, marker='.', markersize=14.0, linestyle='none', color="#E31A1C" )
l1 = ax2.plot( y1data, marker='.', markersize=14.0, linestyle='none', color="#1F78B4" )
#Legend
fontP = FontProperties()
fontP.set_size("x-small")
legend = ax2.legend([l1, l2], [libplot.properName(refname1), libplot.properName(refname2)], 'upper right', numpoints=1, prop=fontP)
legend._drawFrame = False
ax2.set_ylim( minMajority, maxMajority )
ax2.set_xlim( -0.5, len(xticklabels) -0.5 )
ax2.spines['top'].set_visible(False)
ax2.spines['right'].set_visible(False)
ax2.xaxis.tick_bottom()
ax2.yaxis.set_ticks_position( 'left' )
ax2.set_xticks( range( 0, len(xticklabels) ) )
properxticklabels = [ libplot.properName(l) for l in xticklabels ]
ax2.set_xticklabels( properxticklabels )
for label in ax2.xaxis.get_ticklabels():
label.set_rotation( 90 )
ax2.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.set_xlabel( 'Samples' )
title = 'Deletions'
#if type == 'insertion':
if type == 'insertion':
ax2.set_ylabel( 'Insertions per site' )
title = 'Insertions'
else:
ax2.set_ylabel( 'Deletions per site' )
ax2.set_title( title )
libplot.writeImage( fig, pdf, options )
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
def drawCompareData2( axes, options, stats, isAbs ):
if len(stats) == 0:
return
#if isAbs, draw absolute values. If not, draw proportion (relative values)
lines = []
linenames = [ stats[0].otherReferenceName, stats[0].referenceName, "total" ]
#X data:
x1data = []
currx = -1
for i,s in enumerate( stats ):
if s.name == 'all':
continue
if s.name == 'average' or s.name == 'panTro3':
currx += 1.5
else:
currx += 1
x1data.append( currx )
y1data = []
for sample in stats:
if sample.name == 'all':
continue
if isAbs:
y1data.append( sample.referenceBasesMapped )
else:
y1data.append( 100.0*sample.referenceBasesMapped/sample.totalBases )
barwidth = 0.6
#barwidth = 0.25
l1 = axes.bar( x1data, y1data, barwidth, color = "#E31A1C", ec="w" )
lines.append( l1[0] )
libplot.editSpine( axes )
axes.set_title("Sample Coverage") #TO BE NAMED
#set ticks:
samples = []
for sample in stats:
if sample.name == 'all':
continue
samples.append( libplot.properName(sample.name) )
fontP = FontProperties()
fontP.set_size('small')
pyplot.xticks( [x + barwidth/2.0 for x in x1data], samples, rotation=45, fontproperties=fontP )
pyplot.yticks( fontproperties=fontP )
#HACK:
yticks = range(2000000, 6000000, 500000)
yticklabels = [ float(y)/1000000 for y in yticks ]
axes.set_yticks(yticks)
axes.set_yticklabels(yticklabels)
pyplot.xlabel("Samples")
pyplot.ylabel("Number of positions (in millions)")
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
miny = min( y1data )
miny = miny*0.9
axes.set_ylim( miny, max(y1data) )
axes.set_xlim(-0.5, max(x1data) + 0.5 )
axes.yaxis.grid(b=True, color="#A8A8A8", linestyle='-', linewidth=0.25)
#Legend:
box = axes.get_position()
axes.set_position( [box.x0, box.y0, box.width*0.95, box.height*0.9] )
#legend = axes.legend( lines, [libplot.properName(n) for n in linenames], prop=fontP, loc="best", bbox_to_anchor=(0.2, 1) )
#legend._drawFrame=False
return
def drawPlot(exps, options, outfile, type):
options.out = outfile
fig, pdf = libplot.initImage( 11.2, 10.0, options )
axes = fig.add_axes( [0.12, 0.18, 0.85, 0.75] )
#Set title:
titleDict = {'tpfn':'Indel Overlap with dbSNP', 'tp':'True Positives According to dbSNP', 'tp2':'Indel Overlap with dbSNP', 'fn':'False Negatives According to dbSNP', 'total':'Total Indels Called'}
axes.set_title( titleDict[type] )
if 'All' not in exps:
return
samples = getSamplesOrder( exps['All'], type )
if len( samples ) < 1:
return
samples.append('average')
if type != 'fn':
samples.append('reference')
samples.append('panTro3')
xdata = range( 0, len(samples) )
colors = libplot.getColors6()
c = -1
lines = []
pointsize = 10.0
offset = 0.15
#exporder = ['All', 'No repeats', 'Wobble', 'Wobble, No repeats']
exporder = ['Wobble', 'Wobble, No repeats', 'All', 'No repeats']
if type == 'fn':
exporder = ['Wobble', 'All']
elif type == 'total':
exporder = ['All', 'No repeats']
elif type == 'tp2' or type == 'tpfn':
exporder = ['Wobble', 'Wobble, No repeats']
#pointsize = 16.0
#offset = 0.3
#Get ydata
ydataList, ymin, ymax = getData(samples, exps, type, exporder)
scale = -1
if ymin > 1000:
scale = len( str(int(ymin)) ) -1
if scale > 0:
for exp in ydataList:
ydataList[exp] = [ float(y)/10**scale for y in ydataList[exp]]
if type == 'tpfn':
for j,t in enumerate(['tp', 'fn']):
for i, exp in enumerate(exporder):
if i > 0 and t == 'fn':
continue
if t == 'tp':
xdatai = [x + offset*(j*2+i) for x in xdata]
else:
xdatai = [x + offset*(j*2+i) for x in xdata[: len(xdata) -3]]
ydata = ydataList["%s.%s" %(exp,t)]
c +=1
lines.append(axes.plot(xdatai, ydata, color=colors[c], marker='.', markersize=pointsize, linestyle='none'))
else:
for i, exp in enumerate(exporder):
xdatai = [x + offset*i for x in xdata]
ydata = ydataList[exp]
c += 1
l = axes.plot(xdatai, ydata, color=colors[c], marker='.', markersize=pointsize, linestyle='none')
if type == 'fn':
c += 1
lines.append(l)
xmin = -0.4
xmax = len(samples) - 1 + offset*len(exps) + offset*3
fontP = FontProperties()
fontP.set_size('x-small')
if scale > 0:
ymin = float(ymin)/10**scale
ymax = float(ymax)/10**scale
datarange = ymax -ymin
ymin = ymin - datarange*0.01
ymax = ymax + datarange*0.01
#Draw vertical lines to separate each sample:
for i in xrange(1, len(samples)):
d = (1 - offset*len(exporder))/2.0
x = [i - d, i - d]
y = [ymin , ymax]
axes.plot(x,y, color="#CCCCCC", linewidth=0.005)
axes.set_xlim(xmin, xmax)
axes.set_ylim(ymin, ymax)
libplot.editSpine( axes )
axes.set_xticks( [ i + offset*(len(exps)/2-1) for i in range(0, len(samples))] )
axes.set_xticklabels( [ libplot.properName(s) for s in samples] )
for label in axes.xaxis.get_ticklabels():
label.set_rotation(90)
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
if type == 'tp':
legend = pyplot.legend(lines, ['All', 'No repeats', 'No wobble', 'No wobble, No repeats'], numpoints=1, loc='best', prop=fontP)
elif type == 'fn':
legend = pyplot.legend( lines, ['All', 'No wobble'], numpoints=1, loc='best', prop=fontP)
elif type == 'tpfn':
legend = pyplot.legend(lines, ['All, TP', 'No repeats, TP', 'All, FN'], numpoints=1, loc='best', prop=fontP)
elif type == 'tp2':
legend = pyplot.legend( lines, ['All', 'No repeats'], numpoints=1, loc='best', prop=fontP)
else:
legend = pyplot.legend( lines, exporder, numpoints=1, loc='best', prop=fontP)
legend._drawFrame = False
axes.set_xlabel( 'Samples' )
ylabel = "Percentage"
if type == 'total':
ylabel = 'Number of indels'
if scale > 0:
ylabel += '(x%d)' %10**scale
axes.set_ylabel(ylabel)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
libplot.writeImage( fig, pdf, options )
def drawCompareN50data( axes, xsamples, ysamples, options ):
keys = options.keys
lineNames = []
colors = libplot.getColors6()
c = -1
lines = []
xrefname = xsamples[0].attrib[ 'referenceName' ]
yrefname = ysamples[0].attrib[ 'referenceName' ]
minval = inf
maxval = 0
for key in keys:
xdata = []
ydata = []
for xsample in xsamples:
name = xsample.attrib[ 'sampleName' ]
if name == yrefname:
continue
ysample = getSample( ysamples, name )
if ysample == None:
sys.stderr.write( "%s has %s sample, but %s doesn't\n" % (xrefname, name, yrefname) )
continue
xval = int(xsample.attrib[key])
yval = int(ysample.attrib[key])
if xval > 0 and yval > 0:
xdata.append(xval)
ydata.append(yval)
#xdata.append( int(xsample.attrib[ key ]) )
#ydata.append( int(ysample.attrib[ key ]) )
if len(xdata) == 0:
continue
if options.logscale:
xdata = log10( array(xdata) )
ydata = log10( array(ydata) )
c += 1
l = axes.plot( xdata, ydata, color=colors[c], marker=".", markersize=10.0, linestyle='none' )
lines.append(l)
lineNames.append( key )
currmax = max( xdata.max(), ydata.max() )
if maxval < currmax:
maxval = currmax
currmin = min( xdata.min(), ydata.min() )
if minval > currmin:
minval = currmin
if minval == -inf:
minval = 0
#Draw y=x line
span = maxval - minval
#print "MaxVal: %f, MinVal: %f. Span: %f" % (maxval, minval, span)
x = [ minval - span*0.1, maxval + span*0.1 ]
y = [ minval - span*0.1, maxval + span*0.1 ]
axes.plot( x, y, color="0.9" )
libplot.editSpine( axes )
if options.logscale:
#pyplot.ylabel( 'Log 10 of N50' )
pyplot.xlabel( "%s (Log 10)" % libplot.properName(xrefname) )
pyplot.ylabel( "%s (Log 10)" %libplot.properName(yrefname) )
return lines, lineNames, maxval, minval
def drawScatter( axes, options, stats, type, cumulative ):
if len(stats) < 4:
return
title = "Distribution of Positions Shared Among Samples"
if cumulative:
title = "Cumulative Distribution of Positions Shared Among Samples"
axes.set_title(title) #TO BE NAMED
#samples = ["panTro3", "minusOtherReference", "average", "reference", "hg19"]
samples = ["reference", "hg19", "panTro3", "average"]
if type == 'noHg19':
samples = ["minusOtherReference"]
xdata = range( 0, len(stats) -4 )
#print xdata
ydataList = []
miny = float('inf')
maxy = float('-inf')
for name in samples:
for s in stats:
if s.name == name:
ydata = s.baseCoverages[: len(stats) -4]
if cumulative:
ydata = [ sum(ydata[i:]) for i in xrange( len(ydata) ) ]
ydataList.append( ydata )
miny = min( [miny, min(ydata)] )
maxy = max( [maxy, max(ydata)] )
break
lines = []
#colors = libplot.getColors0()
colors =["#E31A1C", "#1F78B4", "#3D3D3D", "#4DAF4A"] #ConsensusRef, GRCh37, chimp, average
c = -1
offset = 0.12
axes.set_yscale('log')
#if type == 'noHg19':
# axes.set_yscale('log')
for i in xrange( len(samples) ):
xdatai = [x + offset*i for x in xdata]
ydata = ydataList[i]
c += 1
if i == 0:
axes.plot(xdatai[1:], ydata[1:], color="#CCCCCC", linestyle='-', linewidth=0.002)
else:
axes.plot(xdatai, ydata, color="#CCCCCC", linestyle='-', linewidth=0.002)
l = axes.plot(xdatai, ydata, color=colors[c], marker='.', markersize=12.0, linestyle='none')
lines.append(l)
fontP = FontProperties()
fontP.set_size('x-small')
yrange = maxy - miny
miny = miny - 10
maxy = maxy + yrange*0.1
xmin = -0.4
xmax = len(stats) - 4 -1 + offset*len(samples) + offset
libplot.editSpine(axes)
axes.set_xticks( [ i + offset*(len(samples)/2.0 ) for i in range(0, len(stats) -4)] )
axes.set_xticklabels( range(1, len(stats) -2) )
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
scale = len(str( int(maxy) )) - 1
ylabel = "Number of positions"
if type == "noHg19":
yticks = [ 10**y for y in range(scale + 1) ]
else:
#yticks = [ 10**y for y in range(scale + 2) ]
yticks = []
for y in range(scale + 1):
for y1 in range(1,10):
yticks.append(y1*(10**y))
axes.set_yticks( yticks )
minorLocator = LogLocator( base=10, subs = range(1, 10) )
axes.yaxis.set_minor_locator( minorLocator )
#else:
# yticks = range(0, int(maxy), 10**scale)
# yticklabels = [ y/(10**scale) for y in yticks ]
# axes.set_yticks( yticks )
# axes.set_yticklabels( yticklabels )
# ylabel += " (x%s)" %( libplot.prettyInt(10**scale) )
#ylabel += " (in millions)"
axes.set_xlim(xmin, xmax)
if type == "noHg19":
axes.set_ylim(miny, maxy)
else:
axes.set_ylim(10000, 1000000)#HACK
if type != 'noHg19':
legend = pyplot.legend( lines, [libplot.properName(s) for s in samples], numpoints=1, loc='lower right', prop=fontP )
legend._drawFrame = False
axes.set_xlabel( 'Number of samples' )
#if type == "noHg19":
# ylabel += " (x %d)" %(10**(scale -1))
axes.set_ylabel( ylabel )
#axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
return
def drawRef2(rexps, exps, options, outfile, numCats):
options.out = outfile
fig, pdf = libplot.initImage( 8.0, 10.0, options )
axes = fig.add_axes( [0.12, 0.14, 0.85, 0.8] )
if len(rexps) < 1:
return
sampleNotherRefmapped = []
ref = ''
for sample in rexps:
if sample == 'average':
continue
e = rexps[sample]
ref = e.ref
sampleNotherRefmapped.append( (sample, e.total) )
otherRefName = libplot.properName( ref )
#Set title:
#axes.set_title("Mapability of C. Ref. in Comparison to %s" % otherRefName)
#HACK
axes.set_title("Mapability of C. Ref. in Comparison to GRCh37 haplotypes")
sampleNotherRefmapped = sorted( sampleNotherRefmapped, key=lambda item: item[1], reverse=True )
samples = [ item[0] for item in sampleNotherRefmapped]
samples.append( 'average' )
xdata = range( 0, len(samples) )
colors = libplot.getColors4()
c = -1
#c = 0
lines = []
#titleDict = {'mapped':'Mapped', 'uniquelyMapped':'Uniquely Mapped', 'properlyPaired':'Properly Paired', 'uniquelyMappedAndProperlyPaired':'Uniquely Mapped And Properly Paired', 'snps':'Snp'}
titleDict = {'mapped':'Mapped', 'properlyPaired':'Properly Paired', 'uniquelyMapped':'Uniquely Mapped', 'uniquelyMappedAndProperlyPaired':'Uniquely Mapped And Properly Paired'}
ydataList, miny, maxy = getData2(samples, rexps, exps, titleDict.keys())
#ydataList, miny, maxy = getData2(samples, exps, titleDict.keys())
#refs = sorted( ydataList.keys() )
offset = 0.12
scale = -1
if miny > 1000:
scale = len( str(int(miny)) ) - 1
linenames = []
categories = ["mapped", "properlyPaired", "uniquelyMapped", "uniquelyMappedAndProperlyPaired"]
cats = categories[:numCats]
for i, key in enumerate( cats ):
xdatai = [ x + offset*i for x in xdata ]
ydata = ydataList[key]
if scale > 0:
ydata = [ float(y)/10**scale for y in ydata ]
c += 1
l = axes.plot( xdatai, ydata, color=colors[c], marker='.', markersize=16.0, linestyle='none')
lines.append(l)
linenames.append( titleDict[key] )
if scale > 0:
miny = float(miny)/10**scale
maxy = float(maxy)/10**scale
#Draw horizontal line at y = 0:
xmin = -0.4
xmax = len(samples) - 1 + offset*len(linenames) + offset
axes.plot( [xmin, xmax], [0,0], color="#6B6B6B", linewidth=0.005)
fontP = FontProperties()
fontP.set_size('x-small')
yrange = maxy - miny
miny = miny - yrange*0.05
maxy = maxy + yrange*0.2
#Draw vertical lines to separate each sample:
for i in xrange(1, len(samples)):
d = (1 - offset*len(linenames))/2.0
x = [i - d, i - d]
y = [miny , maxy]
axes.plot(x,y, color="#CCCCCC", linewidth=0.005)
axes.set_xlim(xmin, xmax )
axes.set_ylim( miny, maxy )
#HACK:
#axes.set_ylim( -2, 0 )
libplot.editSpine( axes )
axes.set_xticks( [ i + offset*(len(linenames)/2.0) for i in range(0, len(samples))] )
axes.set_xticklabels( samples )
for label in axes.xaxis.get_ticklabels():
label.set_rotation(90)
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
legend = pyplot.legend( lines, linenames, numpoints=1, loc='upper right', prop=fontP)
legend._drawFrame = False
axes.set_xlabel( 'Samples' )
axes.set_ylabel( 'Percentage of mapping difference between C. Ref. and %s' % otherRefName) #NEED TO DO
#axes.set_ylabel( 'Percentage of mapping difference between C. Ref. and GRCh37 haplotypes')
if scale > 0:
axes.set_ylabel( 'Event counts (x%d)' %(10**scale) )
#axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
axes.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
libplot.writeImage( fig, pdf, options )
def drawCompareData( axes, options, stats, isAbs ):
if len(stats) == 0:
return
#if isAbs, draw absolute values. If not, draw proportion (relative values)
lines = []
linenames = [ stats[0].otherReferenceName, stats[0].referenceName, "total" ]
barwidth = 0.25
#X data:
x3data = []
#avgIndex = -1
currx = -1
#xVer = [] #location (x) of vertical lines to separate between human samples | avr, all | chimp
for i,s in enumerate( stats ):
#if s.name == 'average':
# avgIndex = i
if s.name == 'average' or s.name == 'panTro3':
currx += 1 + 1.5*barwidth
#xVer.append( currx - (1.0 + 1.5*barwidth - 3*barwidth)/2.0 )
else:
currx += 1
x3data.append( currx )
#print x1data
x2data = [ x + barwidth for x in x3data ]
x1data = [ x + barwidth for x in x2data ]
if isAbs:
y1data = [ sample.otherReferenceBasesMapped for sample in stats ]
y2data = [ sample.referenceBasesMapped for sample in stats ]
y3data = [ sample.totalBases for sample in stats ]
else:
y1data = [ 100.0*sample.otherReferenceBasesMapped/sample.totalBases for sample in stats ]
y2data = [ 100.0*sample.referenceBasesMapped/sample.totalBases for sample in stats ]
y3data = [ 100.0*sample.totalBases/sample.totalBases for sample in stats ]
#Average aggregate data:
#if avgIndex > 0:
# y1data[ avgIndex ] /= float(avgIndex)
# y2data[ avgIndex ] /= float(avgIndex)
# y3data[ avgIndex ] /= float(avgIndex)
colors =["#1F78B4", "#E31A1C", "#4DAF4A"]
#colors =["#1B9E77", "#D95F02", "#7570B3"]
#colors =["#EDF8B1", "#7FCDBB", "#2C7FB8"]
#colors =["#A1DAB4", "#41B6C4", "#225EA8"]
l1 = axes.bar( x1data, y1data, barwidth, color = colors[0], ec="w" )
lines.append( l1[0] )
l2 = axes.bar( x2data, y2data, barwidth, color = colors[1], ec="w" )
lines.append( l2[0] )
l3 = axes.bar( x3data, y3data, barwidth, color = colors[2], ec="w" )
lines.append( l3[0] )
libplot.editSpine( axes )
axes.set_title("Sample Coverage") #TO BE NAMED
#set ticks:
samples = []
for sample in stats:
samples.append( libplot.properName(sample.name) )
fontP = FontProperties()
fontP.set_size('small')
#pyplot.xticks( x + barwidth/2., samples, rotation=45, fontproperties=fontP )
pyplot.xticks( x2data, samples, rotation=45, fontproperties=fontP )
pyplot.yticks( fontproperties=fontP )
#HACK:
yticks = range(2000000, 6000000, 500000)
yticklabels = [ float(y)/1000000 for y in yticks ]
axes.set_yticks(yticks)
axes.set_yticklabels(yticklabels)
pyplot.xlabel("Samples")
pyplot.ylabel("Number of positions (in millions)")
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
miny = min( [min(y1data), min(y2data), min(y3data)] )
miny = miny*0.9
maxy = max([max(y1data), max(y2data), max(y3data)])
#Draw vertical lines:
#for x in xVer:
# axes.plot([x, x], [miny, maxy], color="#A8A8A8")
axes.set_ylim( miny, maxy )
axes.set_xlim(-0.5, max(x1data) + 0.5 )
axes.yaxis.grid(b=True, color="#A8A8A8", linestyle='-', linewidth=0.25)
#Legend:
box = axes.get_position()
axes.set_position( [box.x0, box.y0, box.width*0.95, box.height*0.9] )
legend = axes.legend( lines, [libplot.properName(n) for n in linenames], prop=fontP, loc="best", bbox_to_anchor=(0.2, 1) )
legend._drawFrame=False
return
def drawPlot2(exps, options, outfile, type):
options.out = outfile
fig, pdf = libplot.initImage( 11.2, 10.0, options )
#Set title:
titleDict = {'total':'Total Indels Called'}
if 'All' not in exps:
return
samples = getSamplesOrder( exps['All'], type )
if len( samples ) < 1:
return
samples.append('average')
samples.append('reference')
samples.append('panTro3')
xdata = range( 0, len(samples) )
colors = libplot.getColors6()
c = -1
lines = []
pointsize = 10.0
offset = 0.15
exporder = ['All', 'No repeats']
#Get ydata
ydataList, ymin, ymax = getData(samples, exps, type, exporder)
yrange = ymax - ymin
#Get normal range and outlier range:
normalvals, outliers = getOutliers(ydataList)
minNormal = min(normalvals) - 0.05*yrange
maxNormal = max(normalvals) + 0.05*yrange
minOutlier = min(outliers) - 0.05*yrange
maxOutlier = max(outliers) + 0.05*yrange
if minNormal< 0:
minNormal = -0.5
#Set up the axes
ax, ax2 = setAxes(fig, maxOutlier - minOutlier, maxNormal - minNormal)
scale = -1
if minNormal > 1000:
scale = len( str(int(minNormal)) ) -1
if scale > 0:
for exp in ydataList:
ydataList[exp] = [ float(y)/10**scale for y in ydataList[exp]]
#PLOT
for i, exp in enumerate(exporder):
xdatai = [x + offset*i for x in xdata]
ydata = ydataList[exp]
c += 1
#Outlier plot
l = ax.plot(xdatai, ydata, color=colors[c], marker='.', markersize=pointsize, linestyle='none')
lines.append(l)
#Normal range plot
ax2.plot(xdatai, ydata, color=colors[c], marker='.', markersize=pointsize, linestyle='none')
xmin = -0.4
xmax = len(samples) - 1 + offset*len(exps) + offset*3
fontP = FontProperties()
fontP.set_size('x-small')
if scale > 0:
minNormal = float(minNormal)/10**scale
maxNormal = float(maxNormal)/10**scale
minOutlier = float(minOutlier)/10**scale
maxOutlier = float(maxOutlier)/10**scale
#Draw the Discontinue sign:
d = 0.2 #how big to make the diagonal lines in axes coordinates
if scale == -1:
d = 50
ax.plot( (-1, 0), (minOutlier +d, minOutlier - d), color = "k", clip_on=False )
ax2.plot( (-1, 0), (maxNormal +d, maxNormal - d), color = "k", clip_on=False )
#Draw vertical lines to separate each sample:
for i in xrange(1, len(samples)):
d = (1 - offset*len(exporder))/2.0
x = [i - d, i - d]
y = [minNormal , maxOutlier]
ax.plot(x,y, color="#CCCCCC", linewidth=0.005)
ax2.plot(x,y, color="#CCCCCC", linewidth=0.005)
xticklabels = [libplot.properName(s) for s in samples]
#Set limit for the top plot (outlier)
ax.set_ylim(minOutlier, maxOutlier)
ax.set_xlim(xmin, xmax)
ax.set_xticks( [ i + offset*(len(exps)/2-1) for i in range(0, len(samples))] )
dummyxticklabels = [ "" for l in xticklabels ]
ax.set_xticklabels(dummyxticklabels)
#Make sure the y ticks of the top plot is the same with the bottom plot:
step = 2
if scale == -1:
step = 500
ytickpositions = []
ytickpos = 0
while ytickpos < maxOutlier:
if ytickpos >= minOutlier:
ytickpositions.append(ytickpos)
ytickpos += step
ax.set_yticks(ytickpositions)
#Set limit for the bottom plot:
ax2.set_ylim(minNormal, maxNormal)
ax2.set_xlim(xmin, xmax)
#Hide the spines between ax and ax2:
ax.spines['bottom'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.yaxis.set_ticks_position('left')
ax.xaxis.set_ticks_position('none')
ax2.spines['top'].set_visible(False)
ax2.spines['right'].set_visible(False)
ax2.xaxis.tick_bottom()
ax2.yaxis.set_ticks_position( 'left' )
ax2.set_xticks( [ i + offset*(len(exps)/2-1) for i in range(0, len(samples))] )
ax2.set_xticklabels( xticklabels )
for label in ax2.xaxis.get_ticklabels():
label.set_rotation(75)
legend = pyplot.legend( lines, exporder, numpoints=1, loc='upper left', prop=fontP)
legend._drawFrame = False
ax2.set_xlabel( 'Samples' )
ylabel = 'Number of indels'
if scale > 0:
ylabel += '(x%d)' %10**scale
ax2.set_ylabel(ylabel)
ax.set_title( titleDict[type] )
ax.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
ax2.yaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
libplot.writeImage( fig, pdf, options )
def drawData( axes, stats, isAbs, ycutoff ):
#if isAbs, draw absolute values. If not, draw proportion (relative values)
lines = []
linenames = []
ydataList = []
#initialize ydataList:
#for i in range( len(stats[0].baseCoverages) - len(stats), len(stats[0].baseCoverages) ):
#for i in range( len(stats) -1 ): #each coverage level
for i in range( len(stats) -1 - 2 ): #each coverage level (num samples - average, reference, minusOtherReference
ydata = []
for j in range( len(stats) ):#each sample
if isAbs:
#if stats[j].name == 'aggregate':
# ydata.append( stats[j].baseCoverages[i]/(len(stats) -1) )
#else:
ydata.append( stats[j].baseCoverages[i] )
else:
ydata.append( stats[j].relativeBaseCoverages[i] )
ydataList.append(ydata)
#colors = libplot.getColors2( len(stats) )
colors = libplot.getColors3()
colorindex = 0
x = arange( len(stats) ) #x axis represents the samples
barwidth = 0.6
#add bottom-most bar (number of bases that are in all samples)
l = axes.bar( x, ydataList[ len(ydataList) - 1 ], barwidth, color = colors[colorindex], ec="w" )
lines.append( l[0] )
linenames.append( "%d" % len(ydataList) )
culmulativeList = ydataList[ len(ydataList) - 1 ]
for i in range( len(ydataList) - 2, -1, -1 ):
colorindex += 1
l = axes.bar( x, ydataList[i], barwidth, color = colors[colorindex], bottom=culmulativeList, ec="w" )
lines.append( l[0] )
linenames.append( "%d" % (i + 1) )
#Update cumulative list:
for j in range( len(culmulativeList) ):
culmulativeList[j] += ydataList[i][j]
#l = axes.fill_between( x=range(len(ydataList[i])), y1=ydataList[i], y2=[0] * len(ydataList[i]) , facecolor=colors[colorindex], linewidth = 0.0)
libplot.editSpine( axes )
axes.set_title("Sample Coverage") #TO BE NAMED!!!
pyplot.xlabel("Samples")
if isAbs:
pyplot.ylabel("Number of positions")
else:
pyplot.ylabel("Proportion of total positions")
#set ticks:
samples = []
for sample in stats:
samples.append( libplot.properName( sample.name ) )
fontP = FontProperties()
fontP.set_size('small')
pyplot.xticks( x + barwidth/2., samples, rotation=90, fontproperties=fontP )
pyplot.yticks( fontproperties=fontP )
#for label in axes.yaxis.get_ticklabels():
# label.fontproperties = fontP
# label.set_rotation( 45 )
axes.xaxis.set_ticks_position( 'bottom' )
axes.yaxis.set_ticks_position( 'left' )
miny = ycutoff
if not isAbs:
axes.set_ylim(ycutoff, 1)
#axes.set_ylim(0, 1)
axes.set_xlim(-0.5, len(stats) )
axes.yaxis.grid(b=True, color="#A8A8A8", linestyle='-', linewidth=0.25)
#Legend:
box = axes.get_position()
axes.set_position( [box.x0, box.y0, box.width*0.8, box.height] )
lines.reverse()
linenames.reverse()
legend = axes.legend( lines, [libplot.properName(n) for n in linenames], prop=fontP, loc="best", bbox_to_anchor=(1,0.75) )
legend._drawFrame=False
return lines, linenames