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_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 diff_plot(x2yvec, outbase, xlabels=[], label="", xmin=None, xmax=None, ymin=None, ymax=None, ylabel=""):
if not x2yvec:
return
axes, fig, pdf = drawcommon.get_axes(outfile=outbase)
if not xlabels:
xlabels = sorted(x2yvec.keys())
xdata = range(len(xlabels))
ydata = [x2yvec[x] for x in xlabels]
axes.boxplot(ydata)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
# Set limits
if xmin:
if isinstance(xlabels[0], numbers.Number):
xmin += xdata[0] - xlabels[0] + 1
else:
xmin = -0.5
if xmax:
if isinstance(xlabels[0], numbers.Number):
xmax += xdata[0] - xlabels[0] + 1
else:
xmax = len(xlabels) + 0.5
if isinstance(xlabels[0], numbers.Number) or isinstance(xlabels[0], tuple):
xlabels = [str(x) for x in xlabels]
xlabels = [x.replace("TRB", "") for x in xlabels]
drawcommon.set_xticks(axes, [x + 1 for x in xdata], xlabels)
axes.plot([xmin, xmax], [0, 0], ls="-", color="#252525")
axes.set_xlim(xmin, xmax)
if ymin and ymax:
axes.set_ylim(ymin, ymax)
elif ymin:
# axes.set_ylim(bottom=-0.005)
axes.set_ylim(bottom=ymin)
drawcommon.adjust_ticklabels(axes, xrotation=75)
if not ylabel:
ylabel = "Difference in %s" % label
drawcommon.set_labels(axes, xlabel=label, ylabel=ylabel)
drawcommon.write_image(fig, pdf, "pdf", outbase, 300)
def diff_plot(n2diff, genes, outbase):
axes, fig, pdf = drawcommon.get_axes(outfile=outbase)
xdata = range(len(genes))
data = []
#for name, ydata in n2diff.iteritems():
#axes.plot(xdata, ydata, linestyle='None', marker='.')
for x in xdata:
ydata = [n2diff[name][x] for name in n2diff]
data.append(ydata)
axes.boxplot(data)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
xlabels = [g.lstrip("TRB") for g in genes]
drawcommon.set_xticks(axes, [x + 1 for x in 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="Difference in gene usage")
drawcommon.write_image(fig, pdf, 'pdf', outbase, 300)
def draw_diversity_plot(group2names, name2obj, attr, outfile, outfmt='pdf',
dpi=300):
axes, fig, pdf = drawcommon.get_axes(outfile=outfile, outfmt=outfmt,
dpi=dpi)
xdata = range(1, len(group2names) + 1)
xlabels = sorted(group2names.keys())
data = []
for group in xlabels:
names = group2names[group]
vec = [name2obj[name][attr] for name in names]
data.append(vec)
axes.boxplot(data)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
drawcommon.set_xticks(axes, xdata, xlabels)
drawcommon.adjust_ticklabels(axes, xrotation=30)
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 pair_group_cmp(medvec, vecs, genes, outfile):
axes, fig, pdf = drawcommon.get_axes(outfile=outfile)
xdata = range(len(genes))
data = []
#for vec in vecs:
#ydata = [v - medvec[i] for i, v in enumerate(vec)]
#axes.plot(xdata, ydata, linestyle='None', marker='.')
for x in xdata:
ydata = [vec[x] - medvec[x] for vec in vecs]
data.append(ydata)
axes.boxplot(data)
drawcommon.set_grid(axes)
drawcommon.edit_spine(axes)
xlabels = [g.lstrip("TRB") for g in genes]
#drawcommon.set_xticks(axes, xdata, xlabels)
drawcommon.set_xticks(axes, [x + 1 for x in 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="Difference in gene usage")
drawcommon.write_image(fig, pdf, 'pdf', outfile, 300)
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)
