def dmr_cluster_count_by_class(blocks,
show_labels=True,
set_labels=None,
outdir=None,
figname="dmr_cluster_count_by_class",
ax=None):
"""
Venn diagram showing the distribution of clusters amongst the classes for a single result.
:param blocks: dictionary, keyed by class, values are iterables of IDs
:param outdir: If supplied, write plot to a file.
:param figname:
:return:
"""
created = False
if ax is None:
fig = plt.figure()
ax = fig.add_subplot(111)
created = True
# get classes
# all_classes = dmr_results.classes
sl, blocks = zip(*blocks.items())
# blocks = []
# sl = []
# for cls in all_classes:
# d = dmr_results[cls]
# blocks.append(d.keys())
# sl.append(cls)
if set_labels is None:
set_labels = sl
if not show_labels:
set_labels = None
venn.venn_diagram(*blocks, ax=ax, set_labels=set_labels)
if created:
fig.tight_layout()
if outdir is not None:
ax.figure.savefig(os.path.join(outdir, "%s.png" % figname), dpi=200)
ax.figure.savefig(os.path.join(outdir, "%s.pdf" % figname))
return ax
def dmr_overlap(dmr_results,
pids=None,
comparisons=('matched', 'gibco'),
comparison_titles=('Isogenic', 'Reference'),
outdir=None,
figname='dmr_individual_overlap'):
"""
Plot a Venn diagram showing the overlap of DMRs in the the individuals. One plot per comparison (default)
behaviour is to include two comparisons: isogenic and reference.
This can be limited to a single DMR probe class, or (default) include all.
:param dmr_results:
:param pids:
:param comparisons:
:param comparison_titles:
:param outdir:
:param figname:
:return:
"""
if len(comparisons) != len(comparison_titles):
raise AttributeError(
"Length of comparisons and comparison_titles must be equal.")
if pids is None:
pids = dmr_results.keys()
fig, axs = plt.subplots(1, len(comparisons), num=figname)
for i, cmp in enumerate(comparisons):
inputs = [dmr_results[pid][cmp] for pid in pids]
# n_level = 3
# if probe_class is not None:
# inputs = [dmr.dict_by_sublevel(t, 2, probe_class) for t in inputs]
# n_level = 2
# hasher = lambda x: tuple(x)
# else:
# hasher = lambda x: (x[0], x[2])
#
# inputs = [
# set([hasher(t) for t, _ in dmr.dict_iterator(x, n_level=n_level)]) for x in inputs
# ]
venn.venn_diagram(set_labels=pids, ax=axs[i], *inputs)
axs[i].set_title(comparison_titles[i])
fig.tight_layout()
if outdir is not None:
fig.savefig(os.path.join(outdir, "%s.png" % figname), dpi=200)
fig.savefig(os.path.join(outdir, "%s.pdf" % figname))
def venn_dmr_counts(dmr_results,
pids=None,
outdir=None,
figname='dmr_venn',
comparisons=('matched', 'gibco'),
set_labels=('Isogenic', 'Reference')):
"""
For each supplied patient (top level of keys in the supplied results), generate 3 Venn diagrams
- all DMR count
- hypermethylated DMR counts
- hypomethylated DMR counts
In each case, splitting according to `comparisons`
:param dmr_results: Dictionary of results, e.g. from the results (results_significant) attribute of DmrResults.
:param outdir:
:param comparisons: Iterable giving the comparisons to use.
:param set_labels: Iterable giving the set labels to use for the plot. Coresponds to comparisons.
:return:
"""
if len(set_labels) != len(comparisons):
raise AttributeError(
"set_labels must be the same length as comparisons")
if pids is None:
pids = dmr_results.keys()
# isogenic vs reference in each patient, all, hyper and hypomethylated
fig, axs = plt.subplots(3, len(pids), figsize=(11, 7), num=figname)
for i, pid in enumerate(pids):
inputs = [dmr_results[pid][k] for k in comparisons]
inputs_all = [set(x.keys()) for x in inputs]
inputs_up = [
set([k for k, v in x.items() if v['median_change'] > 0])
for x in inputs
]
inputs_down = [
set([k for k, v in x.items() if v['median_change'] < 0])
for x in inputs
]
sl = set_labels if i == 0 else None
venn.venn_diagram(set_labels=sl, ax=axs[0, i], *inputs_all)
axs[0, i].set_title("GBM%s" % pid)
venn.venn_diagram(set_labels=sl, ax=axs[1, i], *inputs_up)
venn.venn_diagram(set_labels=sl, ax=axs[2, i], *inputs_down)
fig.tight_layout()
if outdir is not None:
fig.savefig(os.path.join(outdir, "%s.png" % figname), dpi=200)
fig.savefig(os.path.join(outdir, "%s_venn.pdf" % figname))
def plot_venn_de_directions(
logfc,
set_colours_dict,
ax=None,
set_labels=('Hypo', 'Hyper'),
fontsize=16
):
if ax is None:
fig = plt.figure(figsize=(5., 3.3))
ax = fig.add_subplot(111)
vv, vs, vc = venn.venn_diagram(
*[logfc[k].index[logfc[k]['consistent'].astype(bool)] for k in set_labels],
set_labels=set_labels,
set_colors=[set_colours_dict[t] for t in set_labels],
ax=ax
)
ax.figure.tight_layout()
# modify labels based on direction
this_members = collections.OrderedDict()
for k in set_labels:
this_res = logfc[k]
this_ix = this_res['consistent'].astype(bool)
this_res = np.sign(this_res.loc[this_ix].astype(float).mean(axis=1))
this_members["%s up" % k] = this_res.index[this_res > 0].difference(vs['11'])
this_members["%s down" % k] = this_res.index[this_res < 0].difference(vs['11'])
# get the corresponding label
lbl = vv.get_label_by_id(setops.specific_sets(set_labels)[k])
## FIXME: font family seems to change from old text to new - related to LaTeX rendering?
lbl.set_text(
lbl.get_text()
+ '\n'
+ r'$%d\uparrow$' % len(this_members["%s up" % k])
+ '\n'
+ r'$%d\downarrow$' % len(this_members["%s down" % k]),
)
plt.setp(common.get_children_recursive(ax, type_filt=plt.Text), fontsize=fontsize)
return ax
all_gs_dict[('mTOR', )] = mtor_geneset
for_export = pd.DataFrame(index=range(
max((len(t) for t in all_gs_dict.values()))),
columns=[])
for k, v in all_gs_dict.items():
the_key = '_'.join(k)
for_export.loc[range(len(v)), the_key] = sorted(v)
for_export.fillna('', inplace=True)
for_export = for_export.sort_index(axis=1)
for_export.to_excel(os.path.join(outdir, "gene_sets.xlsx"), index=False)
# Venn diagram showing various mTOR signature options and overlap between them
fig, ax = plt.subplots()
venn.venn_diagram(*mtor_gs_dict.values(),
set_labels=mtor_gs_dict.keys(),
ax=ax)
fig.tight_layout()
ax.set_facecolor('w')
fig.savefig(os.path.join(outdir, "venn_mtor_genesets.png"), dpi=200)
basedir = os.path.join(HGIC_LOCAL_DIR, 'current/input_data/tcga')
brennan_s7_fn = os.path.join(basedir, "brennan_s7.csv")
brennan_s7 = pd.read_csv(brennan_s7_fn, header=0, index_col=0)
if rnaseq_type == 'counts':
rnaseq_dat_fn = os.path.join(basedir, 'rnaseq.xlsx')
rnaseq_meta_fn = os.path.join(basedir, 'rnaseq.meta.xlsx')
sheet_name = 'htseq'
wang_fn = os.path.join(basedir, 'wang_classification',
'methylation.450k.meta.csv'),
'meth_27k':
os.path.join(DATA_DIR, 'methylation', 'tcga_gbm', 'primary_tumour',
'methylation.27k.meta.csv'),
}
meta = {}
for k, fn in meta_fn.items():
meta[k] = pd.read_csv(fn, header=0, index_col=0)
# microarray only
fig = plt.figure()
ax = fig.add_subplot(111)
venn.venn_diagram(meta['marr_u133'].case_id,
meta['marr_agilent1'].case_id,
meta['marr_agilent2'].case_id,
set_labels=('U133', 'Agilent 1', 'Agilent 2'),
ax=ax)
plt.tight_layout(rect=(0, 0, 1., 1.05))
plt.savefig(os.path.join(outdir, "venn_by_caseid_microarray.png"), dpi=200)
# rnaseq / methylation
fig = plt.figure()
ax = fig.add_subplot(111)
venn.venn_diagram(meta['rnaseq'].index,
meta['meth_450k'].index,
meta['meth_27k'].index,
set_labels=('RNA-Seq', 'Methylation 450K',
'Methylation 27K'),
ax=ax)
plt.tight_layout(rect=(0, 0, 1.05, 1.05))
de_in_all = reference_genomes.ensembl_to_gene_symbol(
setops.reduce_intersection(
*[t.index for t in de_res_separate.values()]))
# sort this by the avg logFC
logfc_in_all = pd.DataFrame.from_dict(
dict([(p, v.loc[de_in_all.index, 'logFC'])
for p, v in de_res_separate.items()]))
logfc_in_all = logfc_in_all.loc[logfc_in_all.mean(
axis=1).abs().sort_values(ascending=False).index]
general.add_gene_symbols_to_ensembl_data(logfc_in_all)
fig = plt.figure(figsize=(5, 5))
ax = fig.add_subplot(111, facecolor='w')
venn.venn_diagram(set_labels=de_res_separate.keys(),
*[t.index for t in de_res_separate.values()],
ax=ax)
fig.tight_layout()
fig.savefig(os.path.join(outdir,
"de_nsc_ss2-polya_combined_dispersion.png"),
dpi=200)
fig, axs = plt.subplots(nrows=2,
ncols=2,
sharex=True,
sharey=True,
figsize=(10, 10))
fig.tight_layout()
for i, p in enumerate(pids):
ax = axs.flat[i]
sp = sample_pairs[p]
gl261_bmdm = pd.read_csv(os.path.join(
indir, 'gl261_bmdm_vs_healthy_monocyte.csv'),
header=0,
index_col=0)
gemm_mg = pd.read_csv(os.path.join(indir, 'gemm_mg_vs_healthy_mg.csv'),
header=0,
index_col=0)
gemm_bmdm = pd.read_csv(os.path.join(indir,
'gemm_bmdm_vs_healthy_monocyte.csv'),
header=0,
index_col=0)
fig = plt.figure()
ax = fig.add_subplot(111)
v, sets, counts = venn.venn_diagram(gl261_mg.index,
gemm_mg.index,
set_labels=("GL261 MG", "GEMM MG"),
ax=ax)
fig.tight_layout()
fig.savefig(os.path.join(outdir, "gl261_mg-gemm_mg.png"), dpi=200)
fig.savefig(os.path.join(outdir, "gl261_mg-gemm_mg.pdf"))
fig = plt.figure()
ax = fig.add_subplot(111)
v, sets, counts = venn.venn_diagram(gl261_bmdm.index,
gemm_bmdm.index,
set_labels=("GL261 BMDM", "GEMM BMDM"),
ax=ax)
fig.tight_layout()
fig.savefig(os.path.join(outdir, "gl261_bmdm-gemm_bmdm.png"), dpi=200)
fig.savefig(os.path.join(outdir, "gl261_bmdm-gemm_bmdm.pdf"))
vert=True,
patch_artist=True,
medianprops=medianprops,
widths=0.7)
ax.set_xticks([])
for p, c in zip(bplot['boxes'], colours):
p.set_facecolor(c)
# add third column with Venn diagrams
ax = fig.add_subplot(gs[ax_i + 1, 2],
facecolor='none',
frame_on=False,
xticks=[],
yticks=[])
v = venn.venn_diagram(*u_hypo.values(),
ax=ax,
set_labels=u_hypo.keys(),
set_colors=[colours[1]] * 2)
ax = fig.add_subplot(gs[ax_i, 2],
facecolor='none',
frame_on=False,
xticks=[],
yticks=[])
v = venn.venn_diagram(*u_hyper.values(),
ax=ax,
set_labels=u_hyper.keys(),
set_colors=[colours[0]] * 2,
alpha=0.7)
for i, (k1, obj) in enumerate(to_plot.items()):
axs[i, 0].set_ylabel(k1)
cols = []
for k, nm in zip(['res_1', 'res_2', 'res_4'], ['original', 'original_filter', 'group_dispersion']):
for m in methods:
cols.append("%s_%s" % (nm, m))
n_pair_only_intersect = pd.DataFrame(index=pids, columns=cols)
for k, nm in zip(['res_1', 'res_2', 'res_4'], ['original', 'original_filter', 'group_dispersion']):
for m in methods:
this_res = to_save[k][m]
# number of DE genes in Venn diagrams
fig, axs = plt.subplots(nrows=3, ncols=4, figsize=(10, 8))
for i, pid in enumerate(pids):
# number DE in the refs
ax = axs.flat[i]
venn.venn_diagram(*[this_res[pid][t].index for t in ['iNSC'] + refs], set_labels=['iNSC'] + refs, ax=ax)
ax.set_title(pid, fontsize=16)
for i in range(len(pids), 12):
ax = axs.flat[i]
ax.set_visible(False)
fig.subplots_adjust(left=0.02, right=0.98, bottom=0.02, top=0.95)
fig.savefig(os.path.join(outdir, "number_de_genes_ref_comparison_%s_%s.png" % (nm, m)), dpi=200)
# number of PO genes in Venn diagrams
fig, axs = plt.subplots(nrows=3, ncols=4, figsize=(10, 6))
for i, pid in enumerate(pids):
a = this_res[pid]['iNSC'].index
po = []
for ref in refs:
b = this_res[pid][ref].index
vs, vc = setops.venn_from_arrays(a, b)
# Venn diagrams (two ESC studies)
s1 = ['_'.join(t) for t in zip(pids, ['ours'] * len(pids))]
s2 = ['_'.join(t) for t in zip(['N2', '50'], ['kogut'] * 2)]
fig, axs = plt.subplots(ncols=3, nrows=3)
i = 0
for s in s1 + s2:
this_arr = []
for r in ref_labels:
k = ('iPSC_%s' % s, 'ESC_%s' % r)
if k in de_res_sign:
this_arr.append(de_res_sign[k])
if len(this_arr):
print "Found comparison %s" % s
venn.venn_diagram(*[t.index for t in this_arr],
ax=axs.flat[i],
set_labels=ref_labels)
axs.flat[i].set_title(s)
i += 1
else:
print "No comparison %s" % s
for j in range(i, axs.size):
axs.flat[i].axis('off')
fig.subplots_adjust(left=0.05, right=0.98, bottom=0.05, top=0.95)
fig.savefig(os.path.join(outdir, "ipsc_esc_venn.png"), dpi=200)
# core DE genes
ipsc_vs_esc = dict([(k, v) for k, v in de_res_sign.items()
if re.search('|'.join(ref_labels), k[1])
legend=False,
ax=ax)
if i == 0:
ax.set_ylabel('% DMRs')
else:
ax.yaxis.set_visible(False)
fig.tight_layout()
fig.savefig(os.path.join(outdir, "dmr_direction_all_groups.png"), dpi=200)
fig.savefig(os.path.join(outdir, "dmr_direction_all_groups.tiff"), dpi=200)
# is there much overlap in the gene sets between the two groups?
fig = plt.figure(figsize=(5, 3))
ax = fig.add_subplot(111)
venn.venn_diagram(genes_from_dmr_groups['Hyper'],
genes_from_dmr_groups['Hypo'],
set_labels=['Hyper', 'Hypo'],
ax=ax)
fig.tight_layout()
fig.savefig(os.path.join(outdir, "genes_from_dmr_groups_venn.png"),
dpi=200)
# no, but if we look at the intersection genes, are they in different directions (DE) between the two groups?
groups_inv = dictionary.complement_dictionary_of_iterables(groups,
squeeze=True)
in_both = setops.reduce_intersection(*genes_from_dmr_groups.values())
in_both_ens = reference_genomes.gene_symbol_to_ensembl(in_both)
# some of these will have no DE results
tmp = {}
for pid in pids:
tmp[pid] = de_res_s1[pid].reindex(in_both_ens).logFC.dropna()
