In [1]:
# PyABC imports
from pyabc import (ABCSMC, Distribution, RV,
History, MedianEpsilon)
from pyabc.populationstrategy import ConstantPopulationSize, AdaptivePopulationSize
from pyabc.epsilon import MedianEpsilon
from pyabc.sampler import MulticoreEvalParallelSampler
In [2]:
# Custom imports
from ionchannelABC import (ion_channel_sum_stats_calculator,
IonChannelAcceptor,
IonChannelDistance,
EfficientMultivariateNormalTransition,
plot_parameter_sensitivity,
plot_distance_weights,
plot_parameters_kde)
In [3]:
# Other necessary imports
import numpy as np
import subprocess
import pandas as pd
import io
import os
import tempfile
In [4]:
# Plotting imports
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
%config Inline.Backend.figure_format = 'retina'
In [6]:
from channels.ina import ina as model
#model.sample({})
In [7]:
measurements = model.get_experiment_data()
obs = measurements.to_dict()['y']
exp = measurements.to_dict()['exp']
errs = measurements.to_dict()['errs']
In [7]:
limits = dict(g_Na=(0, 100),
p1=(0, 100),
p2=(-100, 0),
p3=(-10, 10),
p4=(0, 1000),
p5=(-1, 0),
p6=(0, 1),
p7=(0, 100),
q1=(0, 100),
q2=(0, 10),
q4=(0, 100),
q5=(-100, 0),
q6=(0, 10),
q7=(0, 10),
q8=(0, 100),
q9=(-100, 0),
q10=(0, 10),
q11=(0, 1.0),
q12=(0, 10),
q13=(0, 1),
r2=(-1, 0),
r3=(0, 100),
r4=(-10, 0),
r5=(0, 10),
r6=(0, 100),
r7=(0, 1),
r8=(0, 100),
r9=(-10, 0),
r10=(0, 1),
r11=(0, 10),
r12=(-1, 1),
r13=(0, 1),
r14=(-1, 1),
r15=(-1, 1),
r16=(0, 100))
prior = Distribution(**{key: RV("uniform", a, b - a)
for key, (a,b) in limits.items()})
In [8]:
parameters = ['ina.'+k for k in limits.keys()]
In [9]:
distance_fn=IonChannelDistance(
obs=obs,
exp_map=exp,
err_bars=errs,
err_th=0.1)
In [11]:
sns.set_context('talk')
g = plot_distance_weights(model, distance_fn)
g.savefig('results/ina/dist_weights.pdf')
In [12]:
sns.set_context('talk')
grid1, grid2 = plot_parameter_sensitivity(
model,
parameters,
distance_fn,
sigma=0.1,
n_samples=1000,
plot_cutoff=0.1)
In [13]:
grid1.savefig('results/ina/sensitivity.pdf')
grid2.savefig('results/ina/sensitivity_fit.pdf')
In [8]:
limits = dict(g_Na=(0, 100),
p1=(0, 100),
p2=(-100, 0),
p3=(-10, 10),
p4=(0, 1000),
q1=(0, 100),
q5=(-100, 0),
q11=(0, 1.0),
q13=(0, 1),
r8=(0, 100),
r11=(0, 10),
r12=(-1, 1),
r13=(0, 1),
r16=(0, 100))
prior = Distribution(**{key: RV("uniform", a, b - a)
for key, (a,b) in limits.items()})
In [9]:
db_path = ('sqlite:///' +
os.path.join(tempfile.gettempdir(), "hl-1_ina.db"))
print(db_path)
In [10]:
# Let's log all the sh!t
import logging
logging.basicConfig()
abc_logger = logging.getLogger('ABC')
abc_logger.setLevel(logging.DEBUG)
eps_logger = logging.getLogger('Epsilon')
eps_logger.setLevel(logging.DEBUG)
In [11]:
abc = ABCSMC(models=model,
parameter_priors=prior,
distance_function=IonChannelDistance(
obs=obs,
exp_map=exp,
err_bars=errs,
err_th=0.1),
population_size=AdaptivePopulationSize(
start_nr_particles=2500,
mean_cv=0.4,
max_population_size=5000,
min_population_size=500),
summary_statistics=ion_channel_sum_stats_calculator,
transitions=EfficientMultivariateNormalTransition(),
eps=MedianEpsilon(),
sampler=MulticoreEvalParallelSampler(n_procs=8),
acceptor=IonChannelAcceptor())
In [12]:
abc_id = abc.new(db_path, obs)
In [20]:
history = abc.run(minimum_epsilon=0.05, max_nr_populations=10, min_acceptance_rate=0.01)
In [8]:
db_path = 'sqlite:///results/ina/hl-1_ina.db'
history = History(db_path)
history.all_runs()
Out[8]:
In [9]:
history.id = 1
In [21]:
sns.set_context('talk')
evolution = history.get_all_populations()
grid = sns.relplot(x='t', y='epsilon', size='samples', data=evolution[evolution.t>=0])
grid.savefig('results/ina/eps_evolution.pdf')
In [22]:
df, w = history.get_distribution(m=0)
In [23]:
g = plot_parameters_kde(df, w, limits, aspect=5, height=.8)
In [14]:
g.savefig('results/ina/parameters_kde.pdf')
In [24]:
# Generate parameter samples
n_samples = 100
df, w = history.get_distribution(m=0)
th_samples = df.sample(n=n_samples, weights=w, replace=True).to_dict(orient='records')
In [25]:
# Generate sim results samples
samples = pd.DataFrame({})
for i, th in enumerate(th_samples):
output = model.sample(pars=th, n_x=50)
output['sample'] = i
output['distribution'] = 'post'
samples = samples.append(output, ignore_index=True)
In [26]:
from ionchannelABC import plot_sim_results
sns.set_context('talk')
g = plot_sim_results(samples, obs=measurements)
# Set axis labels
xlabels = ["voltage, mV", "voltage, mV", "voltage, mV", "time, ms", "time, ms"]
ylabels = ["current density, pA/pF", "activation", "inactivation", "recovery", "normalised current trace"]
for ax, xl in zip(g.axes.flatten(), xlabels):
ax.set_xlabel(xl)
for ax, yl in zip(g.axes.flatten(), ylabels):
ax.set_ylabel(yl)
In [27]:
g.savefig('results/ina/ina_sim_results.pdf')
In [28]:
# Require discrete samples for exact measurements at -20mV
discrete_samples = pd.DataFrame({})
for i, th in enumerate(th_samples):
output = model.sample(pars=th)
output['sample'] = i
output['distribution'] = 'post'
discrete_samples = discrete_samples.append(output, ignore_index=True)
In [29]:
# Amplitude at -20 mV
grouped = discrete_samples[discrete_samples['exp']==0].groupby('sample')
def get_amplitude(group):
return group.loc[group.x==-20]['y']
print(grouped.apply(get_amplitude).mean())
print(grouped.apply(get_amplitude).std())
In [31]:
import scipy.stats as st
peak_current = discrete_samples[discrete_samples['exp']==0].groupby('sample').apply(get_amplitude).tolist()
rv = st.rv_discrete(values=(peak_current, [1/len(peak_current),]*len(peak_current)))
print("median: {}".format(rv.median()))
print("95% CI: {}".format(rv.interval(0.95)))
In [32]:
# Voltage and slope factor at half-activation
from scipy.optimize import curve_fit
grouped = samples[samples['exp']==1].groupby('sample')
def fit_boltzmann(group):
def boltzmann(V, Vhalf, Shalf):
return 1/(1+np.exp((Vhalf-V)/Shalf))
guess = (-50, 10)
popt, _ = curve_fit(boltzmann, group.x, group.y, guess)
return popt
output = grouped.apply(fit_boltzmann).apply(pd.Series)
In [33]:
print(output.mean())
print(output.std())
In [34]:
Vhalf = output[0].tolist()
rv = st.rv_discrete(values=(Vhalf, [1/len(Vhalf),]*len(Vhalf)))
print("median: {}".format(rv.median()))
print("95% CI: {}".format(rv.interval(0.95)))
In [35]:
slope = output[1].tolist()
rv = st.rv_discrete(values=(slope, [1/len(slope),]*len(slope)))
print("median: {}".format(rv.median()))
print("95% CI: {}".format(rv.interval(0.95)))
In [36]:
# Voltage and slope factor at half-inactivation
grouped = samples[samples['exp']==2].groupby('sample')
def fit_boltzmann(group):
def boltzmann(V, Vhalf, Shalf):
return 1/(1+np.exp((V-Vhalf)/Shalf))
guess = (-50, 10)
popt, _ = curve_fit(boltzmann, group.x, group.y, guess)
return popt
output = grouped.apply(fit_boltzmann).apply(pd.Series)
In [37]:
print(output.mean())
print(output.std())
In [38]:
Vhalf = output[0].tolist()
rv = st.rv_discrete(values=(Vhalf, [1/len(Vhalf),]*len(Vhalf)))
print("median: {}".format(rv.median()))
print("95% CI: {}".format(rv.interval(0.95)))
In [39]:
slope = output[1].tolist()
rv = st.rv_discrete(values=(slope, [1/len(slope),]*len(slope)))
print("median: {}".format(rv.median()))
print("95% CI: {}".format(rv.interval(0.95)))
In [ ]: