def series_scatter(s1, s2, ax=None, ann='p', filename=None, **plot_args):
fig, ax = init_ax(ax, figsize=(6, 4))
if 's' not in plot_args:
plot_args['s'] = 75
if 'alpha' not in plot_args:
plot_args['alpha'] = .5
ax.scatter(*match_series(s1, s2), **plot_args)
ax.set_xlabel(s1.name)
ax.set_ylabel(s2.name)
if ann == 'p':
ax.annotate('p = {0:.1e}'.format(Tests.spearman_pandas(s1, s2)['p']), (.95, -.02),
xycoords='axes fraction', ha='right', va='bottom', size=14)
if ann == 'fancy_p':
ax.annotate('$p = {}$'.format(latex_float(Tests.spearman_pandas(s1, s2)['p'])), (.95, -.02),
xycoords='axes fraction', ha='right', va='bottom', size=14)
if filename is not None:
fig.savefig(filename)
def series_scatter(s1, s2, ax=None, ann='p', filename=None, **plot_args):
fig, ax = init_ax(ax, figsize=(6,4))
if 's' not in plot_args:
plot_args['s'] = 75
if 'alpha' not in plot_args:
plot_args['alpha'] = .5
ax.scatter(*match_series(s1, s2), **plot_args)
ax.set_xlabel(s1.name)
ax.set_ylabel(s2.name)
if ann == 'p':
ax.annotate('p = {0:.1e}'.format(Tests.spearman_pandas(s1, s2)['p']), (.95, -.02),
xycoords='axes fraction', ha='right',va='bottom', size=14)
if ann == 'fancy_p':
ax.annotate('$p = {}$'.format(latex_float(Tests.spearman_pandas(s1, s2)['p'])), (.95, -.02),
xycoords='axes fraction', ha='right',va='bottom', size=14)
if filename is not None:
fig.savefig(filename)
def violin_plot_pandas(bin_vec,
real_vec,
ann='p',
order=None,
ax=None,
filename=None):
"""
http://pyinsci.blogspot.com/2009/09/violin-plot-with-matplotlib.html
Wrapper around matplotlib's boxplot function to add violin profile.
Inputs
bin_vec: Series of labels
real_vec: Series of measurements to be grouped according to bin_vec
"""
fig, ax = init_ax(ax)
ax.set_ylabel(real_vec.name)
ax.set_xlabel(bin_vec.name)
bin_vec, real_vec = match_series(bin_vec, real_vec)
try:
if order is None:
categories = bin_vec.value_counts().index
else:
categories = order
_violin_plot(ax, [real_vec[bin_vec == num] for num in categories],
pos=categories,
bp=True)
ax.set_xticklabels(
[str(c) + '\n(n=%i)' % sum(bin_vec == c) for c in categories])
except:
box_plot_pandas(bin_vec, real_vec, ax=ax)
#if type(bin_vec.name) == str:
# ax.set_title(str(bin_vec.name) + ' x ' + str(real_vec.name))
p_value = Stats.kruskal_pandas(bin_vec, real_vec)['p']
if ann == 'p_fancy':
ax.annotate('$p = {}$'.format(latex_float(p_value)), (.95, -.02),
xycoords='axes fraction',
ha='right',
va='bottom',
size=14)
if ann == 'p':
ax.annotate('p = {0:.1e}'.format(p_value), (.95, .02),
xycoords='axes fraction',
ha='right',
va='bottom',
size=12)
elif ann is not None:
ax.annotate(ann, (.95, .02),
xycoords='axes fraction',
ha='right',
va='bottom',
size=12)
if filename is not None:
fig.savefig(filename)
return
def paired_boxplot_tumor_normal(df,
sig=True,
cutoffs=[.01, .00001],
order=None,
ax=None):
"""
Draws a paired boxplot given a DataFrame with both tumor and normal
samples on the index. '01' and '11' are hard-coded as the ids for
tumor/normal.
"""
n = df.groupby(level=0).size() == 2
df = df.ix[n[n].index]
if order is None:
o = df.xs('11', level=1).median().order().index
df = df[o[::-1]]
else:
df = df[order]
l1 = list(df.xs('01', level=1).as_matrix().T)
l2 = list(df.xs('11', level=1).as_matrix().T)
boxes = [x for t in zip(l1, l2) for x in t]
ax1, bp = paired_boxplot(boxes, ax)
test = lambda v: Stats.ttest_rel(v.unstack()['01'], v.unstack()['11'])
res = df.apply(test).T
p = res.p
if sig:
pts = [(i * 3.5 + .5, 18) for i, n in enumerate(p) if n < cutoffs[1]]
if len(pts) > 0:
s1 = ax1.scatter(*zip(*pts),
marker='$**$',
label='$p<10^{-5}$',
s=200)
else:
s1 = None
pts = [(i * 3.5 + .5, 18) for i, n in enumerate(p)
if (n < cutoffs[0]) and (n > cutoffs[1])]
if len(pts) > 0:
s2 = ax1.scatter(*zip(*pts),
marker='$*$',
label='$p<10^{-2}$',
s=30)
else:
s2 = None
ax1.legend(bp['boxes'][:2] + [s2, s1],
('Tumor', 'Normal', '$p<10^{-2}$', '$p<10^{-5}$'),
loc='best',
scatterpoints=1)
else:
ax1.legend(bp['boxes'][:2], ('Tumor', 'Normal'), loc='best')
ax1.set_xticklabels(df.columns)
def violin_plot_pandas(bin_vec, real_vec, ann='p', order=None, ax=None,
filename=None):
"""
http://pyinsci.blogspot.com/2009/09/violin-plot-with-matplotlib.html
Wrapper around matplotlib's boxplot function to add violin profile.
Inputs
bin_vec: Series of labels
real_vec: Series of measurements to be grouped according to bin_vec
"""
fig, ax = init_ax(ax)
ax.set_ylabel(real_vec.name)
ax.set_xlabel(bin_vec.name)
bin_vec, real_vec = match_series(bin_vec, real_vec)
try:
if order is None:
categories = bin_vec.value_counts().index
else:
categories = order
_violin_plot(ax, [real_vec[bin_vec == num] for num in categories],
pos=categories, bp=True)
ax.set_xticklabels([str(c) + '\n(n=%i)' % sum(bin_vec == c)
for c in categories])
except:
box_plot_pandas(bin_vec, real_vec, ax=ax)
#if type(bin_vec.name) == str:
# ax.set_title(str(bin_vec.name) + ' x ' + str(real_vec.name))
p_value = Stats.kruskal_pandas(bin_vec, real_vec)['p']
if ann == 'p_fancy':
ax.annotate('$p = {}$'.format(latex_float(p_value)), (.95, -.02),
xycoords='axes fraction', ha='right', va='bottom', size=14)
if ann == 'p':
ax.annotate('p = {0:.1e}'.format(p_value), (.95, .02),
xycoords='axes fraction', ha='right', va='bottom', size=12)
elif ann is not None:
ax.annotate(ann, (.95, .02), xycoords='axes fraction', ha='right',
va='bottom', size=12)
if filename is not None:
fig.savefig(filename)
return
def paired_boxplot_tumor_normal(df, sig=True, cutoffs=[.01, .00001],
order=None, ax=None):
"""
Draws a paired boxplot given a DataFrame with both tumor and normal
samples on the index. '01' and '11' are hard-coded as the ids for
tumor/normal.
"""
n = df.groupby(level=0).size() == 2
df = df.ix[n[n].index]
if order is None:
o = df.xs('11', level=1).median().order().index
df = df[o[::-1]]
else:
df = df[order]
l1 = list(df.xs('01', level=1).as_matrix().T)
l2 = list(df.xs('11', level=1).as_matrix().T)
boxes = [x for t in zip(l1, l2) for x in t]
ax1, bp = paired_boxplot(boxes, ax)
test = lambda v: Stats.ttest_rel(v.unstack()['01'], v.unstack()['11'])
res = df.apply(test).T
p = res.p
if sig:
pts = [(i * 3.5 + .5, 18) for i, n in enumerate(p) if n < cutoffs[1]]
if len(pts) > 0:
s1 = ax1.scatter(*zip(*pts), marker='$**$', label='$p<10^{-5}$', s=200)
else:
s1 = None
pts = [(i * 3.5 + .5, 18) for i, n in enumerate(p)
if (n < cutoffs[0]) and (n > cutoffs[1])]
if len(pts) > 0:
s2 = ax1.scatter(*zip(*pts), marker='$*$', label='$p<10^{-2}$', s=30)
else:
s2 = None
ax1.legend(bp['boxes'][:2] + [s2, s1],
('Tumor', 'Normal', '$p<10^{-2}$', '$p<10^{-5}$'),
loc='best', scatterpoints=1)
else:
ax1.legend(bp['boxes'][:2], ('Tumor', 'Normal'), loc='best')
ax1.set_xticklabels(df.columns)
def boxplot_panel(hit_vec, response_df):
"""
Draws a series of paired boxplots with the rows of the response_df
split according to hit_vec.
"""
b = response_df.copy()
b.columns = pd.MultiIndex.from_arrays([b.columns, hit_vec.ix[b.columns]])
b = b.T
v1, v2 = hit_vec.unique()
test = lambda v: Stats.anova(
v.reset_index(level=1)[v.index.names[1]],
v.reset_index(level=1)[v.name])
res = b.apply(test).T
p = res.p.order()
b = b.ix[:, p.index]
l1 = list(b.xs(v1, level=1).as_matrix().T)
l2 = list(b.xs(v2, level=1).as_matrix().T)
boxes = [x for t in zip(l1, l2) for x in t]
ax1, bp = paired_boxplot(boxes)
y_lim = (response_df.T.quantile(.9).max()) * 1.2
pts = [(i * 3.5 + .5, y_lim) for i, n in enumerate(p) if n < .00001]
if len(pts) > 0:
s1 = ax1.scatter(*zip(*pts), marker='$**$', label='$p<10^{-5}$', s=200)
else:
s1 = None
pts = [(i * 3.5 + .5, y_lim) for i, n in enumerate(p)
if (n < .01) and (n > .00001)]
if len(pts) > 0:
s2 = ax1.scatter(*zip(*pts), marker='$*$', label='$p<10^{-2}$', s=30)
else:
s2 = None
ax1.set_xticklabels(b.columns)
ax1.legend(bp['boxes'][:2] + [s2, s1],
(v1, v2, '$p<10^{-2}$', '$p<10^{-5}$'),
loc='best',
scatterpoints=1)
def boxplot_panel(hit_vec, response_df):
"""
Draws a series of paired boxplots with the rows of the response_df
split according to hit_vec.
"""
b = response_df.copy()
b.columns = pd.MultiIndex.from_arrays([b.columns, hit_vec.ix[b.columns]])
b = b.T
v1, v2 = hit_vec.unique()
test = lambda v: Stats.anova(v.reset_index(level=1)[v.index.names[1]],
v.reset_index(level=1)[v.name])
res = b.apply(test).T
p = res.p.order()
b = b.ix[:, p.index]
l1 = list(b.xs(v1, level=1).as_matrix().T)
l2 = list(b.xs(v2, level=1).as_matrix().T)
boxes = [x for t in zip(l1, l2) for x in t]
ax1, bp = paired_boxplot(boxes)
y_lim = (response_df.T.quantile(.9).max()) * 1.2
pts = [(i * 3.5 + .5, y_lim) for i, n in enumerate(p) if n < .00001]
if len(pts) > 0:
s1 = ax1.scatter(*zip(*pts), marker='$**$', label='$p<10^{-5}$', s=200)
else:
s1 = None
pts = [(i * 3.5 + .5, y_lim) for i, n in enumerate(p) if (n < .01)
and (n > .00001)]
if len(pts) > 0:
s2 = ax1.scatter(*zip(*pts), marker='$*$', label='$p<10^{-2}$', s=30)
else:
s2 = None
ax1.set_xticklabels(b.columns)
ax1.legend(bp['boxes'][:2] + [s2, s1],
(v1, v2, '$p<10^{-2}$', '$p<10^{-5}$'),
loc='best', scatterpoints=1)
def exp_change(s):
'''
Calculates an anova for the change in expression across a variable
on the second level of a MultiIndex. (eg. tumor/normal).
'''
return Tests.anova(pd.Series(s.index.get_level_values(1), s.index), s)
def exp_change(s):
'''
Calculates an anova for the change in expression across a variable
on the second level of a MultiIndex. (eg. tumor/normal).
'''
return Tests.anova(pd.Series(s.index.get_level_values(1), s.index), s)
