def tab(f, sample2row):
#Print total:
for sample in sorted(sample2row.keys()):
f.write("%s" % iseqlib.properName(sample))
row = sample2row[sample]
for i, cell in enumerate(row):
if cell in ['+', '-']:
cell = "$%s$" %cell
f.write( " & %s " %(cell) )
f.write("\\\\\n")
f.write("\\hline\n")
def tab0(f, clones, clone2hits, group2keywords, options):
for clone in clones:
vs, js, ds, sample2size = getCloneInfo(clone)
if clone in clone2hits:
hits = clone2hits[clone]
hitsWithKeyword = [] # list of hits that have at least 1 keyword
for hit in hits:
for g, kw in group2keywords.iteritems():
if g == "b27" or g == "pathogen":
continue
if checkKeywords(hit[0], kw):
hitsWithKeyword.append(hit)
break
if len(hitsWithKeyword) == 0: # no hit with keyword
continue
seq = hits[0][2]
samples = sorted([iseqlib.properName(s) for s in sample2size.keys()])
numrow = max([len(samples), len(hitsWithKeyword)])
f.write(
"\\multirow{%d}{*}{%s} & \\multirow{%d}{*}{%s} & \\multirow{%d}{*}{%s} & "
% (numrow, vs, numrow, seq, numrow, js)
) # Write V, CDR3, J
# f.write("%s & %d & %s & %s & %s \\\\\n " %( samples[0], sample2size[iseqlib.properName2name(samples[0])], hitsWithKeyword[0][4], hitsWithKeyword[0][3], hitsWithKeyword[0][0] )) #first row
f.write(
"%s & %d & %s & %s & %s \\\\\n "
% (
samples[0],
sample2size[iseqlib.properName2name(samples[0])],
hitsWithKeyword[0][4],
hitsWithKeyword[0][3],
parsePaperInfo(hitsWithKeyword[0][0]),
)
)
for i in xrange(1, numrow):
f.write("\\cline{4-8}\n")
f.write(" & & & ")
if i < len(samples):
s = samples[i]
f.write(" %s & %d &" % (s, sample2size[iseqlib.properName2name(s)]))
else:
f.write(" & & ")
if i < len(hitsWithKeyword):
h = hitsWithKeyword[i]
# f.write( "%s & %s & %s \\\\\n" %(h[4], h[3], h[0]) )
f.write("%s & %s & %s \\\\\n" % (h[4], h[3], parsePaperInfo(h[0])))
else:
f.write(" & & \\\\\n")
f.write("\\hline\n")
def tab(f, colnames, sample2row):
#Print total:
totalRow = sample2row['Total']
f.write("%s & %s \\\\\n" % ("Total", " & ".join( ["%s & 100.00" %cell for cell in totalRow] )) )
f.write("\\hline\n")
for sample in sorted(sample2row.keys()):
if sample == 'Total' or sample == 'controls' or sample == 'patients' :
continue
f.write("%s" % iseqlib.properName(sample))
row = sample2row[sample]
for i, cell in enumerate(row):
total = totalRow[i]
pc = iseqlib.getPc( int(cell), int(total) )
f.write( " & %s & %.2f " %(cell, pc) )
f.write("\\\\\n")
f.write("\\hline\n")
def drawVJdata( fig, axes, sample, vgenes, jgenes, options, minvj, maxvj ):
#Draw heatmap:
#data = sample.normalizeVJusage()
if options.abs:
data = sample.intersectVJusage
else:
data = sample.normIntersectVJusage
#Normalize data to the range minvj-maxvj:
if options.heatmapNoScale:
hmaxes = axes.imshow( data, interpolation='nearest' )
else:
norm = mpl.colors.Normalize(vmin=minvj, vmax=maxvj)
cmap = mpl.cm.get_cmap('rainbow', 20)
hmaxes = axes.imshow( data, interpolation='nearest', cmap=cmap, norm=norm )
#Colorbar:
#from mpl_toolkits.axes_grid1.inset_locator import inset_axes
#axins = inset_axes(axes, width="5%", height="10%", loc=3, bbox_to_anchor=(1.05, 0, 1, 1), bbox_transform=axes.transAxes, borderpad=0)
#matplotlib.pyplot.colorbar(hmaxes, cax=axins, ticks=[0, 0.5, 1])
cbar = fig.colorbar(hmaxes, shrink=0.3)
#cbar.ax.set_yticklabels(['0', '0.5', '1'])
immunoseqLib.editSpine( axes )
axes.set_title( "VJ usage of sample %s" % immunoseqLib.properName(sample.name) )
xticks = [ x for x in range( len(jgenes) ) ]
axes.xaxis.set_ticks( xticks )
for i, xlabel in enumerate(jgenes):
items = xlabel.split('|')
items = [item.lstrip('TRB') for item in items]
jgenes[i] = '|'.join(items)
axes.xaxis.set_ticklabels( jgenes )
yticks = [ y for y in range( len(vgenes) ) ]
axes.yaxis.set_ticks( yticks )
for i, ylabel in enumerate(vgenes):
items = ylabel.split('|')
items = [item.lstrip('TRB') for item in items]
vgenes[i] = '|'.join(items)
axes.yaxis.set_ticklabels( vgenes )
textsize = 'x-small'
for label in axes.get_xticklabels():
label.set_fontsize( textsize )
label.set_rotation( 80 )
for label in axes.get_yticklabels():
label.set_fontsize( textsize )
def drawAll(options, outdir, rowname2cells, index2colname):
if options.infile == '-':
outname = 'all'
else:
outname = os.path.basename(options.infile).split('.')[0]
options.out = os.path.join(outdir, outname)
fig, pdf = iseqlib.initImage(10.0, 8.0, options)
axes = fig.add_axes( [0.12, 0.15, 0.85, 0.75] )
lines = []
#rownames = sorted( rowname2cells.keys() )
rownames = ['as11D', 'as16D', 'as1D', 'asBD', 'as20D', 'as15D', 'as8D']
name2color = sample2color(rownames)
#xdata = sorted( [ int(x) for x in colname2index.keys() ] )
xmax = 0.0
xmin = float('inf')
#ymax = 0
markersize = 12.0
xindices = []
#for rowname in rownames:
r = 0
while r < len(rownames):
rowname = rownames[r]
row = rowname2cells[rowname]
means = []
stds = []
xdata = []
for i, m in enumerate( row ):
if i % 2 == 0 and m != 'NA' and m != '' and m != '-':
colname = index2colname[i]
try:
colname = int(colname)
xdata.append( colname )
except:
xdata.append(i/2)
if i not in xindices:
xindices.append(i)
means.append( float(m) )
elif i%2 == 1 and m != 'NA' and m != '' and m != '-':
stds.append( float(m) )
xmax = max([xmax, max(xdata)])
xmin = min([xmin, min(xdata)])
#HACK
#if rowname == 'uniqClones':
#if rowname == 'mountford' or rowname == 'horn':
#if rowname != 'horn':
#if rowname == 'manhattan' or rowname == 'euclidean' or rowname == 'binomial':
exceptions = ['manhattan', 'euclidean', 'binomial', 'kulczynski', 'canberra', 'jaccard']
if rowname in exceptions:
rownames.remove(rowname)
continue
#means = [m/min(means) for m in means]
#stds = [0.0 for s in stds]
#END HACK
color = name2color[rowname]
axes.errorbar(xdata, means, yerr=stds, color=color, markeredgecolor=color, markersize=markersize, fmt='.')
line = axes.plot(xdata, means, color=color, linestyle='-', linewidth=4.0)
lines.append(line)
r += 1
#axes.set_xscale('log')
#axes.set_yscale('log')
iseqlib.editSpine(axes)
#axes.set_title("%s index across different sampling sizes" %outname, size='xx-large')
#axes.set_xlabel("Sampling size (number of reads)", size='large' )
#axes.set_ylabel("%s index" %outname, size='large')
#HACK
axes.set_title("Sequencing Saturation", size='xx-large', weight='bold')
axes.set_xlabel("Sampling size (number of sequences, in millions) ", size='x-large', weight='bold' )
#axes.set_xlabel("Sampling size (number of sequences, in thousands) ", size='x-large', weight='bold' )
axes.set_ylabel("Number of clones (in thousands)", size='x-large', weight='bold')
fontP = FontProperties()
fontP.set_size('medium')
#rownames = ["A", "B"]
rownames = [iseqlib.properName(n) for n in rownames]
legend = axes.legend( lines, rownames, numpoints = 1, loc='best', ncol = 1, prop=fontP)
legend._drawFrame = False
if len(xindices) > 0:
xticklabels = []
for i in sorted(xindices):
xticklabels.append( index2colname[i] )
axes.xaxis.set_ticks( xrange(len(xticklabels)) )
axes.xaxis.set_ticklabels( xticklabels )
#HACK:
#xticks = [ 10000, 50000, 100000, 200000, 300000, 400000, 500000]
#xticklabels = [ str(x/1000) for x in xticks]
xticks = [0, 1000000,2000000,3000000,4000000,5000000,6000000,7000000,8000000,9000000]
xticklabels = [ str(x) for x in xrange(0, 10) ]
axes.xaxis.set_ticks(xticks)
axes.xaxis.set_ticklabels( xticklabels )
#yticks = xrange(0, 121000, 20000)
#yticklabels = [ str(y) for y in xrange(0, 121, 20) ]
yticks = xrange(0, 250000, 50000)
yticklabels = [ str(y) for y in xrange(0, 250, 50) ]
axes.yaxis.set_ticks(yticks)
axes.yaxis.set_ticklabels(yticklabels)
for label in axes.get_xticklabels():
label.set_fontsize('large')
label.set_fontweight ('bold')
#label.set_rotation(45)
for label in axes.get_yticklabels():
label.set_fontsize('large')
label.set_fontweight ('bold')
axes.xaxis.grid(b=True, color='#3F3F3F', linestyle='-', linewidth=0.05)
axes.yaxis.grid(b=True, color='#3F3F3F', linestyle='-', linewidth=0.05)
#HACK
#axes.set_ylim(0.996, 1)
xspan = xmax - xmin
#axes.set_xlim(xmin - xspan*0.01, xmax + xspan*0.01)
#axes.set_ylim(10000, 121000)
#axes.set_xlim(-10, 110000)
#axes.set_ylim(-10, 60000)
axes.set_xlim(-10, 5100000)
axes.set_ylim(-10, 205000)
iseqlib.writeImage(fig, pdf, options)
def drawCloneSizeData( axesList, samples, samplesPerPlot, options, isAbs, yaxisPcReads, yaxisPcClones, cumulative ):
if len( samples ) <= 0:
return
colors = getColors6()
s2c, s2m, s2cLight = sample2color( colors )
#markers = ['o', '^']
#markers=['o', 'o', 'o', 'o', 'o', 'o', 'o', 'o', '^', '^']
#c = -1
#textsize = 'x-large'
textsize = 'medium'
fontP = FontProperties()
fontP.set_size(textsize)
linesDict = {}
labelsDict = {}
if isAbs:
xtickLabels = samples[0].countArr
else:
xtickLabels = samples[0].percentArr
#get x location
xdata = range( 0, len(xtickLabels), 1 )
#offset = 0.01
offset = 0.0
maxy = 0
for i in range( len(axesList) ):
lines = []
sampleNames = []
axes = axesList[i]
#HACK
#if not yaxisPcReads and not yaxisPcClones:
if not yaxisPcReads:
axes.set_yscale('log')
#axes.set_yscale('log')
startIndex = i*samplesPerPlot
endIndex = min( [startIndex + samplesPerPlot, len(samples)] )
for j in range( startIndex, endIndex ):
sample = samples[j]
#sampleNames.append( "%s" % (sample.name))
sampleNames.append( "%s" % (iseqlib.properName(sample.name)))
ydata = sample.clonesPerCount
if not yaxisPcReads and not isAbs:
ydata = sample.clonesPerPercent
elif yaxisPcReads and isAbs:
ydata = sample.readsPerCount
elif yaxisPcReads and not isAbs:
ydata = sample.readsPerPercent
if yaxisPcClones:
ydata = [ 100*y/sample.totalClones for y in ydata ]
maxy = max( [maxy, max(ydata)] )
#if (i*len(axesList) + j) %2 == 0:
# c += 1
#l = axes.plot( xdata, ydata, color=colors[c], marker=markers[(i*len(axesList) + j)%2], markersize=6.0, linestyle='none' )
#axes.plot( xdata, ydata, color=colors[c], linestyle='-', linewidth=0.01 )
currxdata = [x + offset*(j - startIndex) for x in xdata]
if isAbs:
markersize = 8.0
#markersize = 10.0
else:
markersize = 8.0
#markersize = 10.0
if s2m[sample.name] == '*':
markersize += 2.0
elif s2m[sample.name] == 's':
markersize -= 2.0
print sample.name
print ydata
#l = axes.plot( xdata, ydata, color=s2c[sample.name], marker=s2m[sample.name], markeredgecolor=s2c[sample.name], markersize=6.0, linestyle='none' )
l = axes.plot( currxdata, ydata, color=s2c[sample.name], marker=s2m[sample.name], markeredgecolor=s2c[sample.name], markersize=markersize, linestyle='none' )
#l = axes.plot( currxdata, ydata, color="#377EB8", marker='o', markeredgecolor=s2c[sample.name], markersize=markersize, linestyle='none' )
#axes.plot( xdata, ydata, color=s2cLight[sample.name], linestyle='-', linewidth=0.3 )
axes.plot( currxdata, ydata, color=s2cLight[sample.name], linestyle='-', linewidth=0.5 )
#axes.plot( currxdata, ydata, color="#377EB8", linestyle='-', linewidth=0.7 )
lines.append( l )
#if yaxisPcReads or yaxisPcClones:
# axes.plot( [-0.5, len(xtickLabels)], [90, 90], linestyle='-', linewidth=0.2, color="#CCCCCC" )
linesDict[i] = lines
labelsDict[i] = sampleNames
if cumulative:
#axes.set_title( 'Cumulative Distribution of Clones', size="xx-large", weight='bold' )
#axes.set_title( 'Cumulative Distribution of Clones', size="large", weight='bold' )
axes.set_title( 'A. Cumulative Distribution of Clones', size="large", weight='bold' )
if yaxisPcReads:
#axes.set_title( 'Cumulative Distribution of Sequences', size="xx-large", weight='bold')
axes.set_title( 'B. Cumulative Distribution of Sequences', size="large", weight='bold')
else:
axes.set_title( 'Clone Size Distribution', size="xx-large" )
for i in range( len(axesList) ):
axes = axesList[ i ]
#libplot.editSpine( axes )
iseqlib.editSpine( axes )
axes.set_xlabel('Clone size (number of sequences)', size = textsize)
if not isAbs:
axes.set_xlabel('Clone size (% of total sequences)', size = textsize, weight='bold')
axes.set_ylabel('Frequency (number of clones)', size=textsize)
if yaxisPcReads:
axes.set_ylabel('Frequency (% of total sequences)', size=textsize, weight='bold')
if yaxisPcClones:
axes.set_ylabel('Frequency (% of total clones)', size = textsize, weight='bold')
#Legend
#legend = pyplot.legend( linesDict[i], labelsDict[i], numpoints=1, prop=fontP, loc="best" )
#legend = axes.legend( linesDict[i], labelsDict[i], numpoints = 1, "upper right", ncol=3)
if yaxisPcClones and not yaxisPcReads:
legend = axes.legend( linesDict[i], labelsDict[i], numpoints = 1, loc="best", ncol=1, prop=fontP)
#for t in legend.get_texts():
# t.set_fontsize(textsize)
legend._drawFrame = False
#HACK
#if isAbs:
if False:
xtickdata = xrange(10, 101, 10)
#grid:
for x in xtickdata:
axes.plot([x,x], [0.0001, maxy], color='#3F3F3F', linestyle='-', linewidth=0.1)
axes.xaxis.set_ticks(xtickdata)
minx = 0
maxx = 101
axes.set_xlim(minx, maxx)
axes.xaxis.set_ticklabels([str(x) for x in xtickdata] , size='medium')
else:
axes.xaxis.set_ticklabels( xtickLabels, size='medium' )
axes.xaxis.set_ticks([x + 0.5*offset*(samplesPerPlot-1) - 0.5 for x in xdata] )
minx = -0.5
maxx = len(xtickLabels) - 0.5
xticks = [x + 0.5*offset*(samplesPerPlot-1) - 0.5 for x in xdata]
for xi, x in enumerate(xticks):
if xi == 0:
continue
axes.plot([x,x], [0.0001, maxy], color='#3F3F3F', linestyle='-', linewidth=0.1)
axes.set_xlim(minx, maxx)
#END HACK
#axes.xaxis.set_ticks( xdata )
#axes.xaxis.set_ticks([x + 0.5*offset*(samplesPerPlot-1) for x in xdata] )
#axes.set_xlim(-0.5, len(xtickLabels))
#numTicks = 20
#yticks = [ float(t)/numTicks for t in range(numTicks +1) ]
#ytickLabels = []
#for y in yticks:
# ytickLabels.append( "%d" %(y*100) )
if not yaxisPcReads and not yaxisPcClones:
#yticks = [100, 1000, 5000, 10000, 15000, 20000]
#ytickLabels = ["100", "1k", "5k", "10k", "15k", "20k"]
yticks = [1, 10, 100, 1000, 10000, 50000, 100000, 150000]
#ytickLabels = ["1", "10", "100", "1k", "10k", "50k", "100k", "150k"]
ytickLabels = ["1", "10", "100", "1k", "10k", "50k", "100k", "150k"]
axes.yaxis.set_ticklabels( ytickLabels )
axes.yaxis.set_ticks( yticks )
for y in yticks:
if y in [1, 10, 100, 1000, 10000, 100000]:
#if y in [100, 1000, 5000, 10000, 15000, 20000]:
axes.plot([minx, maxx], [y, y], color='#3F3F3F', linestyle='-', linewidth=0.1)
else:
axes.plot([minx, maxx], [y, y], color='#3F3F3F', linestyle='-.', linewidth=0.1)
axes.set_ylim(0.8, maxy + 15000)
elif yaxisPcClones:
#elif yaxisPcClones or yaxisPcReads:
#yticks = [0.001, 0.01, 0.1, 1, 10, 25, 50, 75, 100]
yticks = [0.001, 0.01, 0.1, 1, 10, 100]
#yticks = xrange(0, 101, 20)
ytickLabels = [str(y) for y in yticks]
axes.yaxis.set_ticks(yticks)
axes.yaxis.set_ticklabels( ytickLabels )
for y in yticks:
if y in [0.001, 0.01, 0.1, 1, 10, 100]:
#if y in xrange(0, 101, 20):
axes.plot([minx, maxx], [y, y], color='#3F3F3F', linestyle='-', linewidth=0.1)
else:
axes.plot([minx, maxx], [y, y], color='#3F3F3F', linestyle='-.', linewidth=0.1)
axes.set_ylim(0.0005, maxy + 25)
#axes.set_ylim(0.001, maxy + 25)
else:
yticks = xrange(0, 101, 20)
ytickLabels = [str(y) for y in yticks]
axes.yaxis.set_ticks(yticks)
axes.yaxis.set_ticklabels( ytickLabels )
for y in yticks:
axes.plot([minx, maxx], [y, y], color='#3F3F3F', linestyle='-', linewidth=0.1)
axes.set_ylim(-1, maxy + 1)
#axes.set_ylim(-1, maxy + 1)
#if isAbs: #HACK
# axes.set_ylim(-0.1, 15)
#axes.set_ylim(-0.1, 50) #HACK
for label in axes.get_xticklabels():
#label.set_fontsize( 'large' )
label.set_fontsize( 'small' )
#label.set_rotation( 45 )
for label in axes.get_yticklabels():
#label.set_fontsize( 'large' )
label.set_fontsize( 'small' )
#axes.yaxis.grid(b=True, color="#BCBCBC", linestyle='-', linewidth=0.005)
#axes.xaxis.grid(b=True, color="#CCCCCC", linestyle='-', linewidth=0.005)
return
def drawPcaData(axes, rownames, transformedM, options):
group2xdata = {}
group2ydata = {}
for i,r in enumerate(transformedM):
sample = rownames[i][0]
group = rownames[i][1]
x = r[0]
y = r[1]
if group not in group2xdata:
group2xdata[group] = [x]
group2ydata[group] = [y]
else:
group2xdata[group].append(x)
group2ydata[group].append(y)
groups = sorted(group2xdata.keys())
if options.groupDrawInfo:
g2c = options.g2c
g2m = options.g2m
colors = [g2c[g] for g in groups]
markers = [g2m[g] for g in groups]
else:
colors = immunoseqLib.getColors6()
markers = ['^', 'o', 'd', 'p', 'v', '*', 's']
lines = []
minx = float('inf')
maxx = float('-inf')
miny = float('inf')
maxy = float('-inf')
for i, group in enumerate(groups):
if i >= len(colors):
raise ValueError("drawPcaData: Need more color!")
color = colors[i]
marker = markers[i]
xdata = group2xdata[group]
ydata = group2ydata[group]
minx = min(minx, min(xdata))
miny = min(miny, min(ydata))
maxx = max(maxx, max(xdata))
maxy = max(maxy, max(ydata))
l = axes.plot(xdata, ydata, color=color, marker=marker, markersize=15.0, markeredgecolor=color, linestyle='none' )
lines.append(l)
for label in axes.get_xticklabels():
label.set_fontsize( 'large' )
label.set_fontweight( 'bold' )
for label in axes.get_yticklabels():
label.set_fontsize( 'large' )
label.set_fontweight( 'bold' )
rangex = maxx - minx
axes.set_xlim(minx - rangex*0.1, maxx + rangex*0.1)
rangey = maxy - miny
axes.set_ylim(miny - rangey*0.1, maxy + rangey*0.1 )
linenames = []
for g in groups:
linenames.append( immunoseqLib.properName(g) )
#legend = pyplot.legend( lines, sorted(group2xdata.keys()), numpoints=1, loc="lower left" )
legend = pyplot.legend( lines, linenames, numpoints=1, loc="best" )
#legend._drawFrame = False
#axes.set_title( 'PCA', size='xx-large', weight='bold' )
axes.set_xlabel('PC1', size='xx-large', weight='bold')
axes.set_ylabel('PC2', size='xx-large', weight='bold')
axes.yaxis.grid(b=True, color="#3F3F3F", linestyle='-', linewidth=0.5)
axes.xaxis.grid(b=True, color="#3F3F3F", linestyle='-', linewidth=0.5)
