def plot_metric_fit_cre_depth(data_input, metric_fit, label):
'''creates and saves a violinplot of metric for Cre/layers in VISp
Parameters
----------
data_input: pandas dataframe. If dataframe does not have a column "cre_depth" it will create it, which is slow
metric: string of the name of the metric
stimulus_suffix: string of the stimulus abbreviation (eg. 'dg','sg','ns')
label: string for the y-axis label
'''
cre_depth_list = [('Emx1-IRES-Cre',100), ('Slc17a7-IRES2-Cre', 100),('Cux2-CreERT2',100),('Vip-IRES-Cre',100),
('Emx1-IRES-Cre',200),('Slc17a7-IRES2-Cre', 200),('Cux2-CreERT2',200),
('Rorb-IRES2-Cre',200),('Scnn1a-Tg3-Cre',200),('Nr5a1-Cre',200),('Sst-IRES-Cre',200),('Vip-IRES-Cre',200),
('Emx1-IRES-Cre',300),('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100',300),('Fezf2-CreER',300),('Tlx3-Cre_PL56',300),('Sst-IRES-Cre',300),
('Ntsr1-Cre_GN220',500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
visp = data_input[(data_input.area=='VISp')&(data_input[metric_fit]>=0)&(data_input[metric_fit]<=4)]
if not np.issubdtype(visp[metric_fit].dtype, np.number):
visp[metric_fit] = visp[metric_fit].astype(float)
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6,16))
ax = fig.add_subplot(111)
ax = sns.violinplot(y='cre_depth', x=metric_fit, data=visp, inner='quartile', cut=0,scale='area',
linewidth=1, saturation=0.7, order=cre_depth_list, palette=cre_depth_palette)
for i,c in enumerate(cre_depth_list):
ax.plot(visp[(visp.tld1_name==c[0])&(visp.depth_range==c[1])][metric_fit].median(), [i], 'bo')
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.yticks(range(19), cre_list)
plt.tick_params(labelsize=18)
ax.yaxis.tick_right()
plt.ylabel("")
plt.xlabel(label, fontsize=20)
# plt.text(0.96,1.5, "2/3", fontsize=14)
# plt.text(0.99,7.5, "4", fontsize=14)
# plt.text(0.99,14.5, "5", fontsize=14)
# plt.text(0.99,18.1, "6", fontsize=14)
# plt.title("VISp", fontsize=16)
plt.axhspan(-0.5,3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5,11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5,17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5,19, color='#BDBDBD', alpha=0.4)
fig.tight_layout()
figname = r'/Users/saskiad/Documents/CAM/paper figures/'+metric_fit+'_visp_cre'
save_figure(fig, figname)
def plot_strip_plot(data_input,
area,
plot_key,
x_label,
fig_base_dir='/allen/aibs/mat/gkocker/bob_platform_plots',
figname='dg_decoding_performance',
Nticks=10,
xlim=None,
box=False,
bar=False,
point=False):
cre_depth_list = [('Emx1-IRES-Cre', 100), ('Slc17a7-IRES2-Cre', 100),
('Cux2-CreERT2', 100), ('Vip-IRES-Cre', 100),
('Emx1-IRES-Cre', 200), ('Slc17a7-IRES2-Cre', 200),
('Cux2-CreERT2', 200), ('Rorb-IRES2-Cre', 200),
('Scnn1a-Tg3-Cre', 200), ('Nr5a1-Cre', 200),
('Sst-IRES-Cre', 200), ('Vip-IRES-Cre', 200),
('Emx1-IRES-Cre', 300), ('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100', 300), ('Fezf2-CreER', 300),
('Tlx3-Cre_PL56', 300), ('Sst-IRES-Cre', 300),
('Ntsr1-Cre_GN220', 500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
exp_area = data_input[data_input.area == area]
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6, 16))
ax = fig.add_subplot(111)
ax = sns.stripplot(y='cre_depth',
x=plot_key,
data=exp_area,
palette=cre_depth_palette,
order=cre_depth_list,
size=10)
if box:
ax = sns.boxplot(y='cre_depth',
x=plot_key,
data=exp_area,
palette=cre_depth_palette,
order=cre_depth_list,
whis=np.inf)
if bar:
ax = sns.barplot(y='cre_depth',
x=plot_key,
data=exp_area,
palette=cre_depth_palette,
order=cre_depth_list,
ci=None,
alpha=0.5)
if point:
ax = sns.pointplot(y='cre_depth',
x=plot_key,
data=exp_area,
palette=cre_depth_palette,
order=cre_depth_list,
ci=None,
markers='|',
markersize=30,
join=False)
# if np.amax(exp_area[plot_key].values[np.isfinite(exp_area[plot_key])]) <= 1:
# plt.xticks([0,0.2,0.4, 0.6,0.8,1],range(0,101,20))
# plt.xlim(-0.05, 1.05)
# else:
if xlim is None:
xmax = int(
np.ceil(
np.amax(exp_area[plot_key].values[np.isfinite(
exp_area[plot_key])])))
xmin = int(
np.floor(
np.amin(exp_area[plot_key].values[np.isfinite(
exp_area[plot_key])])))
x_spacing = max(int(np.round((xmax - xmin) / Nticks)), 1)
xmin -= np.remainder(xmin, x_spacing)
else:
xmin, xmax = xlim
xscale = max(abs(xmin), abs(xmax))
xmin, xmax = xmin / xscale, xmax / xscale
x_spacing = max(int(np.round((xmax - xmin) / Nticks)), 1) * xscale
xmin, xmax = xmin * xscale, xmax * xscale
plt.xticks(np.arange(xmin, xmax + x_spacing, x_spacing))
plt.xlim(xmin - 0.1, xmax + .1)
if xmax > 0 and xmin <= 0:
plt.axvline(x=0, lw=1, color='k')
plt.yticks(range(19), cre_list)
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.axhspan(-0.5, 3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5, 11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5, 17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5, 19, color='#BDBDBD', alpha=0.4)
plt.ylabel("")
ax.yaxis.tick_right()
plt.tick_params(labelsize=20)
plt.xlabel(x_label, fontsize=24)
plt.title(area, fontsize=24)
fig.tight_layout()
figname = os.path.join(fig_base_dir, figname + '_' + area)
save_figure(fig, figname)
def plot_metric_box_cre_depth(data_input,
area,
metric,
stimulus_suffix,
label,
xlim=None):
'''creates and saves a boxplot of metric for Cre/layers in specified area
Parameters
----------
data_input: pandas dataframe.
area: string of the visual area to plot
metric: string of the name of the metric
stimulus_suffix: string of the stimulus abbreviation (eg. 'dg','sg','ns')
label: string for the y-axis label
'''
cre_depth_list = [('Emx1-IRES-Cre', 100), ('Slc17a7-IRES2-Cre', 100),
('Cux2-CreERT2', 100), ('Vip-IRES-Cre', 100),
('Emx1-IRES-Cre', 200), ('Slc17a7-IRES2-Cre', 200),
('Cux2-CreERT2', 200), ('Rorb-IRES2-Cre', 200),
('Scnn1a-Tg3-Cre', 200), ('Nr5a1-Cre', 200),
('Sst-IRES-Cre', 200), ('Vip-IRES-Cre', 200),
('Emx1-IRES-Cre', 300), ('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100', 300), ('Fezf2-CreER', 300),
('Tlx3-Cre_PL56', 300), ('Sst-IRES-Cre', 300),
('Ntsr1-Cre_GN220', 500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
responsive = 'responsive_' + stimulus_suffix
visp = data_input[(data_input['area'] == area)
& (data_input[responsive] == True)]
if not np.issubdtype(visp[metric].dtype, np.number):
visp[metric] = visp[metric].astype(float)
visp = visp[['cre_depth', metric, 'tld1_name', 'depth_range']]
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6, 16))
ax = fig.add_subplot(111)
sns.boxplot(y='cre_depth',
x=metric,
data=visp,
palette=cre_depth_palette,
order=cre_depth_list,
saturation=0.7,
width=0.7)
for i, c in enumerate(cre_depth_list):
ax.plot(
visp[(visp.tld1_name == c[0])
& (visp.depth_range == c[1])][metric].median(), [i], 'bo')
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.yticks(range(19), cre_list)
plt.tick_params(labelsize=22) #26#18
ax.yaxis.tick_right()
plt.ylabel("")
plt.xlabel(label, fontsize=20)
plt.axhspan(-0.5, 3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5, 11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5, 17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5, 19, color='#BDBDBD', alpha=0.4)
plt.title(area_dict[area], fontsize=26) #30#20
# plt.xlim(0,1250)#rf area
plt.xticks([0, 0.2, 0.4, 0.6, 0.8, 1])
# plt.xticks([0,25,50,75,100])
# plt.xticks([0,250,500,750,1000,1250]) #rf area
if xlim != None:
plt.xlim(xlim[0], xlim[1])
fig.tight_layout()
figname = r'/Users/saskiad/Documents/CAM/paper figures/revision/' + metric + '_' + area + '_cre_box'
save_figure(fig, figname)
def plot_metric_box_cre_depth_noresp(data_input,
area,
metric,
label,
xlim=None):
'''creates and saves a boxplot of metric for Cre/layers in specified area with not responsiveness criteria
Parameters
----------
data_input: pandas dataframe.
area: string of the cortical area
metric: string of the name of the metric
label: string for the y-axis label
'''
cre_depth_list = [('Emx1-IRES-Cre', 100), ('Slc17a7-IRES2-Cre', 100),
('Cux2-CreERT2', 100), ('Vip-IRES-Cre', 100),
('Emx1-IRES-Cre', 200), ('Slc17a7-IRES2-Cre', 200),
('Cux2-CreERT2', 200), ('Rorb-IRES2-Cre', 200),
('Scnn1a-Tg3-Cre', 200), ('Nr5a1-Cre', 200),
('Sst-IRES-Cre', 200), ('Vip-IRES-Cre', 200),
('Emx1-IRES-Cre', 300), ('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100', 300), ('Fezf2-CreER', 300),
('Tlx3-Cre_PL56', 300), ('Sst-IRES-Cre', 300),
('Ntsr1-Cre_GN220', 500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
visp = data_input[(data_input.area == area)]
if not np.issubdtype(visp[metric].dtype, np.number):
visp[metric] = visp[metric].astype(float)
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6, 16))
ax = fig.add_subplot(111)
sns.boxplot(y='cre_depth',
x=metric,
data=visp,
palette=cre_depth_palette,
order=cre_depth_list,
saturation=0.7,
width=0.7)
for i, c in enumerate(cre_depth_list):
ax.plot(
visp[(visp.tld1_name == c[0])
& (visp.depth_range == c[1])][metric].median(), [i], 'bo')
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.yticks(range(19), cre_list)
plt.tick_params(labelsize=18)
ax.yaxis.tick_right()
plt.ylabel("")
plt.xlabel(label, fontsize=20)
# plt.text(0.96,1.5, "2/3", fontsize=14)
# plt.text(0.99,7.5, "4", fontsize=14)
# plt.text(0.99,14.5, "5", fontsize=14)
# plt.text(0.99,18.1, "6", fontsize=14)
# plt.title("VISp", fontsize=16)
plt.axhspan(-0.5, 3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5, 11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5, 17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5, 19, color='#BDBDBD', alpha=0.4)
plt.title(area_dict[area], fontsize=20)
# plt.xlim(0,1250)#rf area
if xlim != None:
plt.xlim(xlim[0], xlim[1])
fig.tight_layout()
figname = r'/Users/saskiad/Documents/CAM/paper figures/revision/' + metric + '_' + area + '_cre_box'
save_figure(fig, figname)
def plot_metric_cre_depth_run(data_input, metric, area, label):
'''creates and saves a violinplot of running modulation metric for Cre/layers in specified area
Parameters
----------
data_input: pandas dataframe. If dataframe does not have a column "cre_depth" it will create it, which is slow
metric: string of the name of the metric
area: string of the cortical area
label: string for the y-axis label
'''
cre_depth_list = [('Emx1-IRES-Cre', 100), ('Slc17a7-IRES2-Cre', 100),
('Cux2-CreERT2', 100), ('Vip-IRES-Cre', 100),
('Emx1-IRES-Cre', 200), ('Slc17a7-IRES2-Cre', 200),
('Cux2-CreERT2', 200), ('Rorb-IRES2-Cre', 200),
('Scnn1a-Tg3-Cre', 200), ('Nr5a1-Cre', 200),
('Sst-IRES-Cre', 200), ('Vip-IRES-Cre', 200),
('Emx1-IRES-Cre', 300), ('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100', 300), ('Fezf2-CreER', 300),
('Tlx3-Cre_PL56', 300), ('Sst-IRES-Cre', 300),
('Ntsr1-Cre_GN220', 500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
visp = data_input[(data_input.area == area)
& (data_input.run_pval_dg < 0.05)]
if not np.issubdtype(visp[metric].dtype, np.number):
visp[metric] = visp[metric].astype(float)
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6, 16))
ax = fig.add_subplot(111)
ax = sns.violinplot(y='cre_depth',
x=metric,
data=visp,
inner='quartile',
cut=0,
scale='area',
linewidth=1,
saturation=0.7,
order=cre_depth_list,
palette=cre_depth_palette)
for i, c in enumerate(cre_depth_list):
ax.plot(
visp[(visp.tld1_name == c[0])
& (visp.depth_range == c[1])][metric].median(), [i], 'bo')
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.yticks(range(19), cre_list)
plt.tick_params(labelsize=22)
ax.yaxis.tick_right()
plt.ylabel("")
plt.xlabel(label, fontsize=20)
plt.title(area_dict[area], fontsize=26)
plt.axhspan(-0.5, 3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5, 11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5, 17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5, 19, color='#BDBDBD', alpha=0.4)
fig.tight_layout()
figname = r'/Users/saskiad/Documents/CAM/paper figures/revision/' + metric + area + '_cre'
save_figure(fig, figname)
def plot_metric_fit_cre_depth_box(data_input, area, metric_fit, label):
'''creates and saves a boxplot of a fit metric for Cre/layers in specified area
Parameters
----------
data_input: pandas dataframe.
area: string of the cortical area
metric: string of the name of the metric
label: string for the y-axis label
'''
cre_depth_list = [('Emx1-IRES-Cre', 100), ('Slc17a7-IRES2-Cre', 100),
('Cux2-CreERT2', 100), ('Vip-IRES-Cre', 100),
('Emx1-IRES-Cre', 200), ('Slc17a7-IRES2-Cre', 200),
('Cux2-CreERT2', 200), ('Rorb-IRES2-Cre', 200),
('Scnn1a-Tg3-Cre', 200), ('Nr5a1-Cre', 200),
('Sst-IRES-Cre', 200), ('Vip-IRES-Cre', 200),
('Emx1-IRES-Cre', 300), ('Slc17a7-IRES2-Cre', 300),
('Rbp4-Cre_KL100', 300), ('Fezf2-CreER', 300),
('Tlx3-Cre_PL56', 300), ('Sst-IRES-Cre', 300),
('Ntsr1-Cre_GN220', 500)]
cre_color_dict = core.get_cre_colors()
cre_depth_colors = []
cre_list = []
for c in cre_depth_list:
cre_list.append(c[0].split('-')[0])
cre_depth_colors.append(cre_color_dict[c[0]])
cre_depth_palette = sns.color_palette(cre_depth_colors)
visp = data_input[(data_input.area == area) & (data_input[metric_fit] >= 0)
& (data_input[metric_fit] <= 4)]
visp = visp[['cre_depth', metric_fit, 'tld1_name', 'depth_range']]
if not np.issubdtype(visp[metric_fit].dtype, np.number):
visp[metric_fit] = visp[metric_fit].astype(float)
with sns.axes_style('whitegrid'):
fig = plt.figure(figsize=(6, 16))
ax = fig.add_subplot(111)
ax = sns.boxplot(y='cre_depth',
x=metric_fit,
data=visp,
palette=cre_depth_palette,
order=cre_depth_list,
saturation=0.7,
width=0.7)
for i, c in enumerate(cre_depth_list):
ax.plot(
visp[(visp.tld1_name == c[0])
& (visp.depth_range == c[1])][metric_fit].median(), [i],
'bo')
plt.axhline(y=3.5, lw=7, color='w')
plt.axhline(y=11.5, lw=7, color='w')
plt.axhline(y=17.5, lw=7, color='w')
plt.yticks(range(19), cre_list)
plt.yticks([]) #sup
if metric_fit == 'fit_tf_ind_dg':
plt.xticks(range(5), [1, 2, 4, 8, 15])
if metric_fit == 'fit_sf_ind_sg':
plt.xticks(range(5), [0.02, 0.04, 0.08, 0.16, 0.32])
plt.tick_params(labelsize=22) #26
ax.yaxis.tick_right()
plt.ylabel("")
plt.xlabel(label, fontsize=20)
plt.title(area_dict[area], fontsize=26) #30
plt.axhspan(-0.5, 3.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(3.5, 11.5, color='#BDBDBD', alpha=0.4)
plt.axhspan(11.5, 17.5, color='#DBDBDB', alpha=0.4)
plt.axhspan(17.5, 19, color='#BDBDBD', alpha=0.4)
fig.tight_layout()
figname = r'/Users/saskiad/Documents/CAM/paper figures/revision/' + metric_fit + '_' + area + '_cre_box_sup'
save_figure(fig, figname)
