def draw_clonesize_dist(name2obj, attr, outfile, outfmt='pdf', dpi=300):
# name2stat: key = sample name, val = CloneSizeStat
# attr = numclones/counts/numclones_cumul/counts_cumul/topfreqs/
# topfreqs_cumul
assert len(name2obj) > 0
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
obj0 = name2obj.values()[0]
xlabels = [str(x) for x in obj0.freqs]
if attr == 'topfreqs' or attr == 'topfreqs_cumul':
xlabels = [str(x + 1) for x in xrange(len(obj0.topfreqs))]
xdata = range(0, len(xlabels))
numtop = len(xdata)
if attr == 'numclones' or attr == 'numclones_attr':
axes.set_yscale('log')
linenames = []
lines = []
for name, obj in name2obj.iteritems():
ydata = obj[attr]
if len(xdata) != len(ydata): # HACK
continue
line, = axes.plot(xdata, ydata, color=obj.color, marker=obj.marker,
markeredgecolor=obj.color, linestyle='-')
lines.append(line)
linenames.append(obj.name)
drawcommon.set_grid(axes)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.edit_spine(axes)
drawcommon.set_xticks(axes, xdata, xlabels)
labels = cs_get_attr_plot_labels(attr, numtop)
drawcommon.set_labels(axes, labels[0], labels[1], labels[2])
drawcommon.write_image(fig, pdf, outfmt, outfile, dpi)
def draw_diversity_plot_hacktimeseries(group2names, name2obj, attr, outfile,
outfmt='pdf', dpi=300):
'''Instead of box plot for each group, keep each sample separately
'''
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
xlabels = ['BL', 'PC', 'W6', 'W14', 'W20']
xdata = range(1, len(xlabels) + 1)
cat2group2names = {'None': {}, 'IL7': {}}
for catgroup, names in group2names.iteritems():
items = catgroup.split('-')
cat = items[0]
group = items[1]
cat2group2names[cat][group] = names
#HACK
#sam2color = {'LiBr': "#6baed6", 'BrBu': "#3182bd", 'MeRi': "#08519c", # blue
# 'LaBo': "#74c476", 'FoCh': "#31a354", 'JaMa': "#006d2c"}
sam2color = {'LiBr': "#525252", 'BrBu': "#969696", 'MeRi': "#cccccc", # gray
'LaBo': "#2171b5", 'FoCh': "#6baed6", 'JaMa': "#bdd7e7"} # blue
sam2data = {}
sam2marker = {}
name2cat = {}
for cat, g2n in cat2group2names.iteritems():
for i, group in enumerate(xlabels): # each timepoint
names = g2n[group]
if i == 0:
name2cat[names[0].split('-')[0]] = cat
for name in names:
sam = name.split('-')[0]
y = name2obj[name][attr]
if sam not in sam2data:
sam2data[sam] = [y]
#sam2color[sam] = name2obj[name].color
sam2marker[sam] = name2obj[name].marker
else:
sam2data[sam].append(y)
lines = []
linenames = []
for sam, ydata in sam2data.iteritems():
l, = axes.plot(xdata, ydata, color=sam2color[sam],
marker=sam2marker[sam], linestyle='-',
markeredgecolor=sam2color[sam])
if sam in name2cat: #legend
lines.append(l)
linenames.append(name2cat[sam])
drawcommon.set_grid(axes)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.edit_spine(axes)
drawcommon.set_xticks(axes, xdata, xlabels)
drawcommon.adjust_ticklabels(axes, xrotation=30)
axes.set_xlim(0.5, len(xdata) + 0.5)
drawcommon.set_labels(axes, xlabel="Group", ylabel=attr.title())
axes.ticklabel_format(style='sci', axis='y', scilimits=(0,0))
drawcommon.write_image(fig, pdf, outfmt, outfile, dpi)
def draw_rarefaction(name2size2sampling, groups, index, outfile,
outformat='pdf', dpi=300):
# xaxis: sampling size; yaxis: index value +- std
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outformat,
dpi=dpi)
lines = []
linenames = []
for name, size2sampling in name2size2sampling.iteritems():
if not size2sampling:
continue
xdata = sorted(size2sampling.keys())
ydata = []
stddata = []
for x in xdata:
sampling = size2sampling[x]
y = sampling[index]
ydata.append(y)
stdindex = "%s_std" % index
if stdindex in sampling.getitems():
std = sampling[stdindex]
stddata.append(std)
sampling0 = size2sampling[xdata[0]]
color = sampling0.color
marker = sampling0.marker
if stddata:
axes.errorbar(xdata, ydata, yerr=stddata, color=color,
markeredgecolor=color, fmt='.')
line, = axes.plot(xdata, ydata, color=color, mec=color,
marker=marker, linestyle='-')
if not groups:
lines.append(line)
linenames.append(sampling0.name)
else:
if sampling0.group not in linenames:
lines.append(line)
linenames.append(sampling0.group)
drawcommon.set_grid(axes)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.edit_spine(axes)
# Labeling:
title = "%s Rarefaction Curve" % index.title()
xlabel = "Sampling size (number of sequences)"
ylabel = index.title()
drawcommon.set_labels(axes, title, xlabel, ylabel)
axes.ticklabel_format(style='sci', axis='x', scilimits=(0,0))
axes.ticklabel_format(style='sci', axis='y', scilimits=(0,0))
drawcommon.write_image(fig, pdf, outformat, outfile, dpi)
def draw_gene_usage(name2obj, attr, type, outbase, genes, opts=None):
'''xaxis: genes
yaxis: relative usage
one line/ sample
'''
if opts:
axes, fig, pdf = drawcommon.get_axes(outfile=outbase,
outfmt=opts.plotformat, dpi=opts.dpi)
else:
axes, fig, pdf = drawcommon.get_axes(outfile=outbase)
#genes = sorted(genes)
genes = libcommon.sort_by_gene_number(genes)
xdata = range(len(genes))
group2line = {}
lines = []
linenames = []
for name, obj in name2obj.iteritems():
ydata = gu_get_sample_data(obj, attr, type, genes)
line, = axes.plot(xdata, ydata, color=obj.color, marker=obj.marker,
markeredgecolor=obj.color, linestyle='-')
lines.append(line)
linenames.append(obj.name)
if obj.group not in group2line:
group2line[obj.group] = line
if len(linenames) > 10:
linenames = sorted(group2line.keys())
lines = [group2line[group] for group in linenames]
drawcommon.set_legend(axes, lines, linenames)
drawcommon.adjust_ticklabels(axes, xrotation=75)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
xlabels = [g.lstrip("TRB") for g in genes]
drawcommon.set_xticks(axes, xdata, xlabels)
axes.set_xlim(-0.5, len(xlabels) + 0.5)
axes.set_ylim(bottom=-0.005)
drawcommon.adjust_ticklabels(axes, xrotation=75)
drawcommon.set_labels(axes, xlabel="Gene", ylabel="%% of total %s" % attr)
if opts:
drawcommon.write_image(fig, pdf, opts.plotformat, outbase, opts.dpi)
else:
drawcommon.write_image(fig, pdf, outname=outbase)
def draw_lendist(name2obj, attr, outfile, outfmt='pdf', dpi=300, bar=False):
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
group2line = {}
lines = []
linenames = []
barwidth = (1.0 - 0.35) / len(name2obj)
i = -1
minx = 30
maxx = 0
for name, obj in name2obj.iteritems():
i += 1
xdata, ydata = ld_get_sample_data(obj, attr)
minx = min(minx, min(xdata))
maxx = max(maxx, max(xdata))
if not bar:
line, = axes.plot(xdata, ydata, color=obj.color, marker=obj.marker,
markeredgecolor=obj.color, linestyle='-')
lines.append(line)
else:
group_xdata = [x + barwidth * i for x in xdata]
line = axes.bar(group_xdata, ydata, barwidth,
color=obj.color, ecolor="#424242") #,
#edgecolor=g2color[g], ecolor=g2color[g])
lines.append(line[0])
linenames.append(obj.name)
if obj.group not in group2line:
group2line[obj.group] = line
if bar:
xticks = [x + 0.325 for x in xrange(minx, maxx + 1)]
xlabels = [str(x) for x in xdata]
drawcommon.set_xticks(axes, xticks, xlabels)
axes.set_xlim(xmin=8, xmax=min(maxx, 30))
axes.set_ylim(bottom=-0.005)
if len(linenames) > 10:
linenames = sorted(group2line.keys())
lines = [group2line[group] for group in linenames]
drawcommon.set_grid(axes)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.edit_spine(axes)
drawcommon.set_labels(axes, xlabel='Length', ylabel='%% of total %s' % attr)
drawcommon.write_image(fig, pdf, outfmt, outfile, dpi)
def draw_pca(rownames, rows, outbase, name2obj, var1=None, var2=None):
axes, fig, pdf = drawcommon.get_axes(outfile=outbase)
lines = []
linenames = []
g2xdata = {}
g2ydata = {}
g2color = {}
for i, (name, group) in enumerate(rownames):
row = rows[i]
x = row[0]
y = row[1]
if group not in g2xdata:
g2xdata[group] = [x]
g2ydata[group] = [y]
g2color[group] = name2obj[name].color
else:
g2xdata[group].append(x)
g2ydata[group].append(y)
lines = []
linenames = sorted(g2xdata.keys())
for g in linenames:
xdata = g2xdata[g]
ydata = g2ydata[g]
color = g2color[g]
l = axes.plot(xdata, ydata, color=color, markersize=10.0, marker='.',
markeredgecolor=color, linestyle='none')
lines.append(l)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
if var1 and var2:
drawcommon.set_labels(axes, xlabel="PC1, %f%%" % var1,
ylabel="PC2, %f%%" % var2)
else:
drawcommon.set_labels(axes, xlabel="PC1", ylabel="PC2")
drawcommon.write_image(fig, pdf, outname=outbase)
def draw_clonesize_dist_avr(name2obj, attr, outfile, outfmt='pdf', dpi=300):
# name2stat: key = sample name, val = CloneSizeStat
# attr = numclones/counts/numclones_cumul/counts_cumul/topfreqs/
# topfreqs_cumul
assert len(name2obj) > 0
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
obj0 = name2obj.values()[0]
xlabels = [str(x) for x in obj0.freqs]
if attr == 'topfreqs' or attr == 'topfreqs_cumul':
xlabels = [str(x + 1) for x in xrange(len(obj0.topfreqs))]
xdata = range(0, len(xlabels))
numtop = len(xdata)
#if attr == 'numclones': # or attr == 'numclones_cumul':
# axes.set_yscale('log')
g2ydata = {}
g2color = {}
for name, obj in name2obj.iteritems():
ydata = obj[attr]
g = obj.group
if g not in g2ydata:
g2ydata[g] = [[y] for y in ydata]
g2color[g] = obj.color
else:
assert len(g2ydata[g]) == len(ydata)
for i, y in enumerate(ydata):
g2ydata[g][i].append(y)
linenames = []
lines = []
#boxwidth = 0.05
#offset = boxwidth + 0.01
offset = 0.05
for i, g in enumerate(sorted(g2ydata.keys())):
g_xdata = [x + 0.5 + offset * (i - 1) for x in xdata]
g_xdata[0] += 0.35
ydata = g2ydata[g]
#axes.boxplot(ydata, positions=g_xdata, widths=boxwidth)
mean_ydata = [np.mean(ylist) for ylist in ydata]
std_ydata = [np.std(ylist) for ylist in ydata]
if attr == "numclones_cumul" or attr == "numclones":
mean_ydata = [log10(y) if y > 0 else 2.1 for y in mean_ydata]
std_ydata = get_logstd(ydata)
line, = axes.plot(g_xdata, mean_ydata, color=g2color[g], linestyle='-',
markeredgecolor=g2color[g], marker='o', lw=2)
lines.append(line)
linenames.append(g)
axes.errorbar(g_xdata, mean_ydata, yerr=std_ydata, color=g2color[g],
linestyle="None", marker="None")
drawcommon.set_grid(axes)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.edit_spine(axes)
drawcommon.set_xticks(axes, xdata, xlabels)
axes.set_xlim(-0.5, len(xlabels) + 0.5)
if attr != "numclones_cumul" and attr != 'numclones':
axes.set_ylim(bottom=-0.005)
labels = cs_get_attr_plot_labels(attr, numtop)
drawcommon.set_labels(axes, labels[0], labels[1], labels[2])
drawcommon.write_image(fig, pdf, outfmt, outfile, dpi)
def draw_gene_usage_avr(name2obj, attr, type, outbase, genes, opts, bar=False):
'''xaxis: genes
yaxis: relative usage
one line/ sample
'''
axes, fig, pdf = drawcommon.get_axes(outfile=outbase,
outfmt=opts.plotformat, dpi=opts.dpi)
#genes = sorted(genes)
#genes = sorted(genes, key=lambda g: libcommon.get_gene_number(g))
genes = libcommon.sort_by_gene_number(genes)
xdata = range(len(genes))
g2ydata = {}
g2color = {}
for name, obj in name2obj.iteritems():
ydata = gu_get_sample_data(obj, attr, type, genes)
g = obj.group
if g not in g2ydata:
g2ydata[g] = [[y] for y in ydata]
g2color[g] = obj.color
else:
assert len(g2ydata[g]) == len(ydata)
for i, y in enumerate(ydata):
g2ydata[g][i].append(y)
barwidth = (1.0 - 0.35) / len(g2ydata.keys())
lines = []
linenames = []
#for g, ydata in g2ydata.iteritems():
for i, g in enumerate(sorted(g2ydata.keys())):
ydata = g2ydata[g]
ydata = [ylist if ylist else [0.0] for ylist in ydata]
mean_ydata = [np.mean(ylist) for ylist in ydata]
std_ydata = [np.std(ylist) for ylist in ydata]
if not bar:
line, = axes.plot(xdata, mean_ydata, color=g2color[g],
linestyle='-', markeredgecolor=g2color[g],
marker='o')
lines.append(line)
axes.errorbar(xdata, mean_ydata, yerr=std_ydata, color=g2color[g],
linestyle="None", marker="None")
else:
group_xdata = [x + barwidth * i for x in xdata]
line = axes.bar(group_xdata, mean_ydata, barwidth, yerr=std_ydata,
color=g2color[g], ecolor="#424242",
edgecolor=g2color[g])
lines.append(line[0])
linenames.append(g)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
xlabels = [g.lstrip("TRB") for g in genes]
if not bar:
drawcommon.set_xticks(axes, xdata, xlabels)
else:
xticks = [x + 0.325 for x in xdata]
drawcommon.set_xticks(axes, xticks, xlabels)
axes.set_xlim(-0.5, len(xlabels) + 0.5)
axes.set_ylim(bottom=-0.005)
drawcommon.adjust_ticklabels(axes, xrotation=75)
drawcommon.set_labels(axes, xlabel="Gene", ylabel="%% of total %s" % attr)
drawcommon.write_image(fig, pdf, opts.plotformat, outbase, opts.dpi)
def draw_lendist_avr(name2obj, attr, outfile, outfmt='pdf', dpi=300, bar=False):
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
g2x2y = {}
g2numsam = {}
g2color = {}
minx = 30
maxx = 0
for name, obj in name2obj.iteritems():
xdata, ydata = ld_get_sample_data(obj, attr)
g = obj.group
if g not in g2numsam:
g2numsam[g] = 1
g2color[g] = obj.color
else:
g2numsam[g] += 1
minx = min(minx, min(xdata))
maxx = max(maxx, max(xdata))
for i, x in enumerate(xdata):
y = ydata[i]
if g not in g2x2y:
g2x2y[g] = {x: [y]}
elif x not in g2x2y[g]:
g2x2y[g][x] = [y]
else:
g2x2y[g][x].append(y)
lines = []
linenames = []
xdata = range(minx, maxx + 1)
barwidth = (1.0 - 0.35) / len(g2x2y.keys())
for i, g in enumerate(sorted(g2x2y.keys())):
numsam = g2numsam[g]
mean_ydata = []
std_ydata = []
for x in xdata:
if x not in g2x2y[g]:
mean_ydata.append(0)
std_ydata.append(0)
else:
ylist = g2x2y[g][x] + [0] * (numsam - len(g2x2y[g][x]))
mean_ydata.append(np.mean(ylist))
std_ydata.append(np.std(ylist))
if not bar:
line, = axes.plot(xdata, mean_ydata, color=g2color[g], marker='o',
markeredgecolor=g2color[g], linestyle='-')
axes.errorbar(xdata, mean_ydata, yerr=std_ydata, color=g2color[g],
linestyle='None', marker='None')
lines.append(line)
else:
group_xdata = [x + barwidth * i for x in xdata]
line = axes.bar(group_xdata, mean_ydata, barwidth, yerr=std_ydata,
color=g2color[g], ecolor="#424242") #,
#edgecolor=g2color[g], ecolor=g2color[g])
lines.append(line[0])
linenames.append(g)
if bar:
xticks = [x + 0.325 for x in xdata]
xlabels = [str(x) for x in xdata]
drawcommon.set_xticks(axes, xticks, xlabels)
axes.set_xlim(xmin=8, xmax=min(maxx, 30))
axes.set_ylim(bottom=-0.005)
drawcommon.set_legend(axes, lines, linenames)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
drawcommon.set_labels(axes, xlabel='Length', ylabel='%% of total %s' % attr)
drawcommon.write_image(fig, pdf, outfmt, outfile, dpi)