312 lines
18 KiB
Python
312 lines
18 KiB
Python
"""
|
|
Script to analyse and collect data from sensitivity analysis simulations
|
|
|
|
|
|
"""
|
|
__author__ = "Nils A. Koch"
|
|
__copyright__ = 'Copyright 2022, Nils A. Koch'
|
|
__license__ = "MIT"
|
|
|
|
import numpy as np
|
|
import h5py
|
|
import pandas as pd
|
|
import os
|
|
from ast import literal_eval
|
|
import json
|
|
import scipy.stats as stats
|
|
|
|
# folder
|
|
top_dir = '../Sensitivity_Analysis/Data'
|
|
|
|
#%% create JSON files for each alt type for each model in '../Sensitivity_Analysis/Data/SA_summary_df/file.json'
|
|
for dir_name in next(os.walk(top_dir))[1]: # for each folder in top_file_dir (each model)
|
|
folder = os.path.join(top_dir, dir_name)
|
|
print(folder)
|
|
# for each alt_type create pandas dataframe
|
|
shift_AUC = pd.DataFrame()
|
|
shift_rheo = pd.DataFrame()
|
|
shift_fI = pd.DataFrame(dtype=object)
|
|
shift_I_mag = pd.DataFrame(dtype=object)
|
|
|
|
slope_AUC = pd.DataFrame()
|
|
slope_rheo = pd.DataFrame()
|
|
slope_fI = pd.DataFrame(dtype=object)
|
|
slope_I_mag = pd.DataFrame(dtype=object)
|
|
|
|
g_AUC = pd.DataFrame()
|
|
g_rheo = pd.DataFrame()
|
|
g_fI = pd.DataFrame(dtype=object)
|
|
g_I_mag = pd.DataFrame(dtype=object)
|
|
|
|
|
|
for root, dirs, files in os.walk(folder):
|
|
for file in files:
|
|
if file.endswith('.hdf5'):
|
|
with h5py.File(os.path.join(folder, file), "r+") as f:
|
|
alt = f['data'].attrs['alteration']
|
|
|
|
test = f['data'].attrs['alteration_info'].replace(' ', ',')
|
|
alt_info = literal_eval(test)
|
|
var = alt_info[0]
|
|
alt_type = alt_info[1]
|
|
if alt_type == 'shift':
|
|
shift_AUC.loc[alt, var] = f['analysis']['AUC'][()]
|
|
try:
|
|
shift_rheo.loc[alt, var] = f['analysis']['rheobase'][()]
|
|
except:
|
|
print('shift', var, alt)
|
|
shift_fI.loc[alt, var] = 0
|
|
shift_fI = shift_fI.astype(object)
|
|
shift_fI.at[alt, var] = f['analysis']['F_inf'][:].tolist()
|
|
shift_I_mag.loc[alt, var] = 0
|
|
shift_I_mag = shift_I_mag.astype(object)
|
|
shift_I_mag.at[alt, var] = ((np.arange(f['data'].attrs['I_low'], f['data'].attrs['I_high'],
|
|
(f['data'].attrs['I_high'] - f['data'].attrs['I_low']) /
|
|
f['data'].attrs['stim_num'])) * 1000).tolist() #nA
|
|
elif alt_type == 'slope':
|
|
slope_AUC.loc[alt, var] = f['analysis']['AUC'][()]
|
|
try:
|
|
slope_rheo.loc[alt, var] = f['analysis']['rheobase'][()]
|
|
except:
|
|
print('slope', var, alt)
|
|
slope_fI.loc[alt, var] = 0
|
|
slope_fI = slope_fI.astype(object)
|
|
slope_fI.at[alt, var] = f['analysis']['F_inf'][:].tolist()
|
|
slope_I_mag.loc[alt, var] = 0
|
|
slope_I_mag = slope_I_mag.astype(object)
|
|
slope_I_mag.at[alt, var] = ((np.arange(f['data'].attrs['I_low'], f['data'].attrs['I_high'],
|
|
(f['data'].attrs['I_high'] - f['data'].attrs['I_low']) /
|
|
f['data'].attrs['stim_num'])) * 1000).tolist()
|
|
elif alt_type == 'g':
|
|
g_AUC.loc[alt, var] = f['analysis']['AUC'][()]
|
|
try:
|
|
g_rheo.loc[alt, var] = f['analysis']['rheobase'][()]
|
|
except:
|
|
print('g', var, alt)
|
|
g_fI.loc[alt, var] = 0
|
|
g_fI = g_fI.astype(object)
|
|
g_fI.at[alt, var] = f['analysis']['F_inf'][:].tolist()
|
|
g_I_mag.loc[alt, var] = 0
|
|
g_I_mag = g_I_mag.astype(object)
|
|
g_I_mag.at[alt, var] = ((np.arange(f['data'].attrs['I_low'], f['data'].attrs['I_high'],
|
|
(f['data'].attrs['I_high'] - f['data'].attrs['I_low']) /
|
|
f['data'].attrs['stim_num'])) * 1000).tolist()
|
|
else:
|
|
print(file, 'Unknown alteration type')
|
|
|
|
#save df with folder+alt_type
|
|
save_folder = os.path.join(top_dir, 'SA_summary_df')
|
|
if not os.path.isdir(save_folder):
|
|
os.makedirs(save_folder)
|
|
shift_AUC.to_json(os.path.join(save_folder, '{}_shift_AUC.json'.format(dir_name)))
|
|
shift_rheo.to_json(os.path.join(save_folder, '{}_shift_rheo.json'.format(dir_name)))
|
|
shift_fI.to_json(os.path.join(save_folder, '{}_shift_fI.json'.format(dir_name)))
|
|
shift_I_mag.to_json(os.path.join(save_folder, '{}_shift_I_mag.json'.format(dir_name)))
|
|
|
|
slope_AUC.to_json(os.path.join(save_folder, '{}_slope_AUC.json'.format(dir_name)))
|
|
slope_rheo.to_json(os.path.join(save_folder, '{}_slope_rheo.json'.format(dir_name)))
|
|
slope_fI.to_json(os.path.join(save_folder, '{}_slope_fI.json'.format(dir_name)))
|
|
slope_I_mag.to_json(os.path.join(save_folder, '{}_slope_I_mag.json'.format(dir_name)))
|
|
|
|
g_AUC.to_json(os.path.join(save_folder, '{}_g_AUC.json'.format(dir_name)))
|
|
g_rheo.to_json(os.path.join(save_folder, '{}_g_rheo.json'.format(dir_name)))
|
|
g_fI.to_json(os.path.join(save_folder, '{}_g_fI.json'.format(dir_name)))
|
|
g_I_mag.to_json(os.path.join(save_folder, '{}_g_I_mag.json'.format(dir_name)))
|
|
|
|
|
|
#%% AUC Correlation analysis
|
|
alt_dict = {}
|
|
alt_dict['m'] = 'Na activation'
|
|
alt_dict['h'] = 'Na inactivation'
|
|
alt_dict['n'] = 'K activation'
|
|
alt_dict['s'] = '$K_V1.1$ activation'
|
|
alt_dict['u'] = '$K_V1.1$inactivation'
|
|
alt_dict['a'] = 'A activation'
|
|
alt_dict['b'] = 'A inactivation'
|
|
alt_dict['n_A'] = 'A activation'
|
|
alt_dict['h_A'] = 'A inactivation'
|
|
alt_dict['Na'] = 'Na'
|
|
alt_dict['Kd'] = 'K'
|
|
alt_dict['Kv'] = '$K_V1.1$'
|
|
alt_dict['A'] = 'A'
|
|
alt_dict['Leak'] = 'Leak'
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
shift_df = pd.DataFrame(columns=['model', 'corr', 'p_value', 'local corr', 'local p_value', 'ratio', '$\Delta V_{1/2}$', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df/{}_shift_AUC.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[0, :]) / df.loc[0, :]
|
|
zero_ind = np.argwhere(df.index == 0)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l / tau
|
|
shift_df = shift_df.append(pd.Series([model_names[mod], tau, p,tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=shift_df.columns), ignore_index=True)
|
|
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
slope_df = pd.DataFrame(columns=['model', 'corr', 'p_value','local corr', 'local p_value', 'ratio', 'Slope (k)', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df//{}_slope_AUC.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file, convert_dates=False, convert_axes=False)
|
|
df.index = df.index.map(float)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[1.0, :]) / df.loc[1.0, :]
|
|
zero_ind = np.argwhere(df.index == 1)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l/tau
|
|
slope_df = slope_df.append(pd.Series([model_names[mod], tau, p, tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=slope_df.columns), ignore_index=True)
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
g_df = pd.DataFrame(columns=['model', 'corr', 'p_value', 'local corr', 'local p_value', 'ratio', 'g', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df//{}_g_AUC.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file, convert_dates=False, convert_axes=False)
|
|
df.index = df.index.map(float)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[1.0, :]) / df.loc[1.0, :]
|
|
zero_ind = np.argwhere(df.index == 1)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l / tau
|
|
g_df = g_df.append(pd.Series([model_names[mod], tau, p,tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=g_df.columns), ignore_index=True)
|
|
|
|
shift_df.to_json('./Sensitivity_analysis/shift_box_kendall_corr.json')
|
|
slope_df.to_json('./Sensitivity_analysis/slope_box_kendall_corr.json')
|
|
g_df.to_json('./Sensitivity_analysis/g_kendall_corr.json')
|
|
|
|
|
|
#%% rheobase correlation analysis
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
shift_df = pd.DataFrame(columns=['model', 'corr', 'p_value', 'local corr', 'local p_value', 'ratio', '$\Delta V_{1/2}$', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df//{}_shift_rheo.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[0, :]) / df.loc[0, :]
|
|
zero_ind = np.argwhere(df.index == 0)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l / tau
|
|
shift_df = shift_df.append(pd.Series([model_names[mod], tau, p,tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=shift_df.columns), ignore_index=True)
|
|
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
slope_df = pd.DataFrame(columns=['model', 'corr', 'p_value','local corr', 'local p_value', 'ratio', 'Slope (k)', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df//{}_slope_rheo.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file, convert_dates=False, convert_axes=False)
|
|
df.index = df.index.map(float)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[1.0, :]) / df.loc[1.0, :]
|
|
zero_ind = np.argwhere(df.index == 1)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l/tau
|
|
slope_df = slope_df.append(pd.Series([model_names[mod], tau, p, tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=slope_df.columns), ignore_index=True)
|
|
|
|
# models = directory names in top_file_dir and what json files are named
|
|
models = ['RS_pyramidal', 'RS_inhib', 'FS', 'RS_pyramidal_Kv', 'RS_inhib_Kv', 'FS_Kv', 'Cb_stellate', 'Cb_stellate_Kv',
|
|
'Cb_stellate_Kv_only', 'STN', 'STN_Kv', 'STN_Kv_only']
|
|
# model_names = names of models for json files and ultimately csv files for plotting
|
|
model_names = ['RS Pyramidal','RS Inhibitory','FS', 'RS Pyramidal +$K_V1.1$','RS Inhibitory +$K_V1.1$', 'FS +$K_V1.1$',
|
|
'Cb stellate','Cb stellate +$K_V1.1$', 'Cb stellate $\Delta$$K_V1.1$','STN','STN +$K_V1.1$', 'STN $\Delta$$K_V1.1$']
|
|
|
|
g_df = pd.DataFrame(columns=['model', 'corr', 'p_value', 'local corr', 'local p_value', 'ratio', 'g', 'color']) # for boxplots
|
|
for mod in range(len(models)):
|
|
with open('../Sensitivity_Analysis/Data/SA_summary_df//{}_g_rheo.json'.format(models[mod])) as json_file:
|
|
df = pd.read_json(json_file, convert_dates=False, convert_axes=False)
|
|
df.index = df.index.map(float)
|
|
df.sort_index(inplace=True)
|
|
df.replace(0., np.NaN, inplace=True)
|
|
df = (df - df.loc[1.0, :]) / df.loc[1.0, :]
|
|
zero_ind = np.argwhere(df.index == 1)[0][0]
|
|
ind = [df.index[zero_ind - 1], df.index[zero_ind], df.index[zero_ind + 1]]
|
|
df2 = df.loc[ind, :]
|
|
for c in df.keys():
|
|
tau, p = stats.kendalltau(df.index, df[c], nan_policy='omit')
|
|
tau_l, p_l = stats.kendalltau(df2.index, df2[c], nan_policy='omit')
|
|
ratio_tau = tau_l / tau
|
|
g_df = g_df.append(pd.Series([model_names[mod], tau, p,tau_l, p_l, ratio_tau, alt_dict[c],clr_dict[models[mod]]], index=g_df.columns), ignore_index=True)
|
|
|
|
shift_df.to_json('./Sensitivity_analysis/rheo_shift_box_kendall_corr.json')
|
|
slope_df.to_json('./Sensitivity_analysis/rheo_slope_box_kendall_corr.json')
|
|
g_df.to_json('./Sensitivity_analysis/rheo_g_kendall_corr.json')
|
|
|
|
#%% create csv files for plotting
|
|
# | (index) | model | corr | p_value | g | color
|
|
#% AUC
|
|
AUC_shift_json = pd.read_json('./Sensitivity_analysis/shift_kendall_corr_rel.json', orient='records')
|
|
AUC_slope_json = pd.read_json('./Sensitivity_analysis/slope_kendall_corr_rel.json', orient='records') #, lines=True)
|
|
AUC_g_json = pd.read_json('./Sensitivity_analysis/g_kendall_corr_rel.json', orient='records')
|
|
AUC_shift_df = AUC_shift_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
AUC_slope_df = AUC_slope_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
AUC_g_df = AUC_g_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
AUC_shift_df.to_csv('AUC_shift_corr.csv')
|
|
AUC_slope_df.to_csv('AUC_scale_corr.csv')
|
|
AUC_g_df.to_csv('AUC_g_corr.csv')
|
|
|
|
#% rheo
|
|
rheo_shift_json = pd.read_json('./Sensitivity_analysis/rheo_shift_kendall_corr.json', orient='records')
|
|
rheo_slope_json = pd.read_json('./Sensitivity_analysis/rheo_slope_kendall_corr.json', orient='records') #, lines=True)
|
|
rheo_g_json = pd.read_json('./Sensitivity_analysis/rheo_g_kendall_corr.json', orient='records')
|
|
rheo_shift_df = rheo_shift_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
rheo_slope_df = rheo_slope_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
rheo_g_df = rheo_g_json[['model', 'corr', 'p_value', 'g', 'color']]
|
|
rheo_shift_df.to_csv('rheo_shift_corr.csv')
|
|
rheo_slope_df.to_csv('rheo_scale_corr.csv')
|
|
rheo_g_df.to_csv('rheo_g_corr.csv')
|