# --------------------
# Set up Model
# --------------------
# Parameters found by stepwise fitting GAB mean data
# Note: can remove multiplicative factors on all K1, K2, K4 and still get very good fit to data (worst is 5 min beta)
initial_parameters = {'k_a1': 4.98E-14 * 2, 'k_a2': 8.30e-13 * 2, 'k_d4': 0.006 * 3.8,
'kpu': 0.00095,
'ka2': 4.98e-13 * 2.45, 'kd4': 0.3 * 2.867,
'kint_a': 0.000124, 'kint_b': 0.00086,
'krec_a1': 0.0028, 'krec_a2': 0.01, 'krec_b1': 0.005, 'krec_b2': 0.05}
dual_parameters = {'kint_a': 0.00052, 'kSOCSon': 6e-07, 'kint_b': 0.00052, 'krec_a1': 0.001, 'krec_a2': 0.1,
'krec_b1': 0.005, 'krec_b2': 0.05}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.1})
# ---------------------------------
# Make theory dose response curves
# ---------------------------------
# Make predictions
times = [2.5, 5.0, 7.5, 10.0, 20.0, 60.0]
test_doses = list(logspace(-1, 5.2))
dradf = Mixed_Model.mixed_dose_response(times, 'TotalpSTAT', 'Ia', test_doses,
'krec_a2': 0.01,
'krec_b1': 0.005,
'krec_b2': 0.05
}
dual_parameters = {
'kint_a': 0.00052,
'kSOCSon': 6e-07,
'kint_b': 0.00052,
'krec_a1': 0.001,
'krec_a2': 0.1,
'krec_b1': 0.005,
'krec_b2': 0.05
}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.2})
Mixed_Model.model_1.default_parameters = Mixed_Model.model_1.parameters
Mixed_Model.model_2.default_parameters = Mixed_Model.model_2.parameters
tspan = [2.5, 5.0, 7.5, 10.0, 20.0, 60.0]
alpha_doses = [10, 100, 300, 1000, 3000, 10000, 100000]
beta_doses = [0.2, 6, 20, 60, 200, 600, 2000]
# Load experimental data to which model will be fit.
def run_smooth_trajectories(cell_densities,
IFNAlpha_panel,
IFNBeta_panel,
times,
IFN_in_concentration=True):
# Experimental parameters
volume_panel = [1 / i for i in cell_densities]
# --------------------
# Set up Model
# --------------------
# Parameters found by stepwise fitting GAB mean data
# Note: can remove multiplicative factors on all K1, K2, K4 and still get very good fit to data (worst is 5 min beta)
initial_parameters = {
'k_a1': 4.98E-14 * 2,
'k_a2': 8.30e-13 * 2,
'k_d4': 0.006 * 3.8,
'kpu': 0.00095,
'ka2': 4.98e-13 * 2.45,
'kd4': 0.3 * 2.867,
'kint_a': 0.000124,
'kint_b': 0.00086,
'krec_a1': 0.0028,
'krec_a2': 0.01,
'krec_b1': 0.005,
'krec_b2': 0.05
}
dual_parameters = {
'kint_a': 0.00052,
'kSOCSon': 6e-07,
'kint_b': 0.00052,
'krec_a1': 0.001,
'krec_a2': 0.1,
'krec_b1': 0.005,
'krec_b2': 0.05
}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.1})
ka1 = Mixed_Model.model_1.parameters['ka1']
ka2 = Mixed_Model.model_1.parameters['ka2']
k_a1 = Mixed_Model.model_1.parameters['k_a1']
k_a2 = Mixed_Model.model_1.parameters['k_a2']
# ---------------------------------
# Make theory dose response curves
# ---------------------------------
# Make predictions
predictions = {}
for vidx, volume in enumerate(volume_panel):
Mixed_Model.set_global_parameters({
'volEC': volume,
'ka1': ka1 * 1E-5 / volume,
'ka2': ka2 * 1E-5 / volume,
'k_a1': k_a1 * 1E-5 / volume,
'k_a2': k_a2 * 1E-5 / volume
})
dradf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ia',
IFNAlpha_panel,
parameters={'Ib': 0},
sf=scale_factor)
drbdf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ib',
IFNBeta_panel,
parameters={'Ia': 0},
sf=scale_factor)
drIadf = Mixed_Model.mixed_dose_response(times,
'Free_Ia',
'Ia',
IFNAlpha_panel,
parameters={'Ib': 0},
sf=scale_factor)
drIbdf = Mixed_Model.mixed_dose_response(times,
'Free_Ib',
'Ib',
IFNBeta_panel,
parameters={'Ia': 0},
sf=scale_factor)
if IFN_in_concentration:
for d in drIadf.loc['Alpha'].index:
for t in drIadf.loc['Alpha'].columns:
drIadf.loc['Alpha', t][d] = (drIadf.loc['Alpha', t][d][0] /
(6.022E23 * volume * 1E-9),
drIadf.loc['Alpha', t][d][1])
for d in drIbdf.loc['Beta'].index:
for t in drIbdf.loc['Beta'].columns:
drIbdf.loc['Beta', t][d] = (drIbdf.loc['Beta', t][d][0] /
(6.022E23 * volume * 1E-9),
drIbdf.loc['Beta', t][d][1])
draIfnData = IfnData('custom',
df=dradf,
conditions={'Alpha': {
'Ib': 0
}})
drbIfnData = IfnData('custom',
df=drbdf,
conditions={'Beta': {
'Ia': 0
}})
drIaIfnData = IfnData('custom',
df=drIadf,
conditions={'Alpha': {
'Ib': 0
}})
drIbIfnData = IfnData('custom',
df=drIbdf,
conditions={'Beta': {
'Ia': 0
}})
predictions[cell_densities[vidx]] = [
draIfnData, drbIfnData, drIaIfnData, drIbIfnData
]
return predictions
def testing_specific_values():
volEC1 = 1 / (25E9)
volEC2 = 1E-5
# --------------------
# Set up Model
# --------------------
# Parameters found by stepwise fitting GAB mean data
# Note: can remove multiplicative factors on all K1, K2, K4 and still get very good fit to data (worst is 5 min beta)
initial_parameters = {
'k_a1': 4.98E-14 * 2,
'k_a2': 8.30e-13 * 2,
'k_d4': 0.006 * 3.8,
'kpu': 0.00095,
'ka2': 4.98e-13 * 2.45,
'kd4': 0.3 * 2.867,
'kint_a': 0.000124,
'kint_b': 0.00086,
'krec_a1': 0.0028,
'krec_a2': 0.01,
'krec_b1': 0.005,
'krec_b2': 0.05
}
dual_parameters = {
'kint_a': 0.00052,
'kSOCSon': 6e-07,
'kint_b': 0.00052,
'krec_a1': 0.001,
'krec_a2': 0.1,
'krec_b1': 0.005,
'krec_b2': 0.05
}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.1})
# ---------------------------------
# Make theory dose response curves
# ---------------------------------
# Make predictions
times = np.arange(0, 120, 0.5) # [2.5, 5.0, 7.5, 10.0, 20.0, 60.0]
alpha_doses_20190108 = [0, 10, 100, 300, 1000, 3000, 10000, 100000]
beta_doses_20190108 = [0, 0.2, 6, 20, 60, 200, 600, 2000]
# 1 uL
ka1 = Mixed_Model.model_1.parameters['ka1']
ka2 = Mixed_Model.model_1.parameters['ka2']
k_a1 = Mixed_Model.model_1.parameters['k_a1']
k_a2 = Mixed_Model.model_1.parameters['k_a2']
Mixed_Model.set_global_parameters({
'volEC': volEC1,
'ka1': ka1 * 1E-5 / volEC1,
'ka2': ka2 * 1E-5 / volEC1,
'k_a1': k_a1 * 1E-5 / volEC1,
'k_a2': k_a2 * 1E-5 / volEC1
})
dradf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ia',
list(logspace(1, 5.2)),
parameters={'Ib': 0},
sf=scale_factor)
drbdf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ib',
list(logspace(-1, 4)),
parameters={'Ia': 0},
sf=scale_factor)
dra15uL = IfnData('custom', df=dradf, conditions={'Alpha': {'Ib': 0}})
drb15uL = IfnData('custom', df=drbdf, conditions={'Beta': {'Ia': 0}})
# 1 mL
Mixed_Model.set_global_parameters({
'volEC': volEC2,
'ka1': ka1 * 1E-5 / volEC2,
'ka2': ka2 * 1E-5 / volEC2,
'k_a1': k_a1 * 1E-5 / volEC2,
'k_a2': k_a2 * 1E-5 / volEC2
})
dradf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ia',
list(logspace(1, 5.2)),
parameters={'Ib': 0},
sf=scale_factor)
drbdf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ib',
list(logspace(-1, 4)),
parameters={'Ia': 0},
sf=scale_factor)
dra5mL = IfnData('custom', df=dradf, conditions={'Alpha': {'Ib': 0}})
drb5mL = IfnData('custom', df=drbdf, conditions={'Beta': {'Ia': 0}})
# -------------------------------
# Plot model dose response curves
# -------------------------------
alpha_palette = sns.color_palette("deep", 6)
beta_palette = sns.color_palette("deep", 6)
new_fit = DoseresponsePlot((2, 2))
alpha_mask = [2.5, 5.0, 10.0, 20.0, 60.0]
beta_mask = [2.5, 5.0, 10.0, 20.0, 60.0]
# Add fits
color_counter = -1
for idx, t in enumerate(times):
if t in alpha_mask:
color_counter += 1
new_fit.add_trajectory(dra15uL,
t,
'plot',
alpha_palette[color_counter], (0, 0),
'Alpha',
label=str(t) + ' min',
linewidth=2)
new_fit.add_trajectory(dra5mL,
t,
'plot',
alpha_palette[color_counter], (1, 0),
'Alpha',
label=str(t) + ' min',
linewidth=2)
if t in beta_mask:
new_fit.add_trajectory(drb15uL,
t,
'plot',
beta_palette[color_counter], (0, 1),
'Beta',
label=str(t) + ' min',
linewidth=2)
new_fit.add_trajectory(drb5mL,
t,
'plot',
beta_palette[color_counter], (1, 1),
'Beta',
label=str(t) + ' min',
linewidth=2)
dr_fig, dr_axes = new_fit.show_figure()
dr_axes[0][0].set_title(r'IFN$\alpha$ at {} L'.format(volEC1))
dr_axes[0][1].set_title(r'IFN$\beta$ at {} L'.format(volEC1))
dr_axes[1][0].set_title(r'IFN$\alpha$ at {} L'.format(volEC2))
dr_axes[1][1].set_title(r'IFN$\beta$ at {} L'.format(volEC2))
dr_fig.tight_layout()
dr_fig.savefig('varying_reaction_volume_dr.pdf')
# -------------------------------
# Plot model dose response curves
# -------------------------------
tc_figure = TimecoursePlot((1, 2))
alpha_testdose = list(logspace(1, 5.2))[24]
beta_testdose = list(logspace(-1, 4))[17]
print("Using {} for IFNalpha".format(alpha_testdose))
print("Using {} for IFNbeta".format(beta_testdose))
tc_figure.add_trajectory(dra15uL,
'plot',
'-',
subplot_idx=(0, 0),
label=r'IFN$\alpha$ at {} L'.format(volEC1),
doseslice=alpha_testdose,
dose_species='Alpha',
color=alpha_palette[1])
tc_figure.add_trajectory(dra5mL,
'plot',
'-',
subplot_idx=(0, 0),
label=r'IFN$\alpha$ at {} L'.format(volEC2),
doseslice=alpha_testdose,
dose_species='Alpha',
color=alpha_palette[4])
tc_figure.add_trajectory(drb15uL,
'plot',
'-',
subplot_idx=(0, 1),
label=r'IFN$\beta$ at {} L'.format(volEC1),
doseslice=beta_testdose,
dose_species='Beta',
color=beta_palette[0])
tc_figure.add_trajectory(drb5mL,
'plot',
'-',
subplot_idx=(0, 1),
label=r'IFN$\beta$ at {} L'.format(volEC2),
doseslice=beta_testdose,
dose_species='Beta',
color=beta_palette[2])
tc_figure.show_figure(save_flag=True,
save_dir='varying_reaction_volume_tc.pdf')
# ------------------------------------
# Now let's quickly look at IFN levels
# ------------------------------------
# 1 uL
Mixed_Model.set_global_parameters({
'volEC': volEC1,
'ka1': ka1 * 1E-5 / volEC1,
'ka2': ka2 * 1E-5 / volEC1,
'k_a1': k_a1 * 1E-5 / volEC1,
'k_a2': k_a2 * 1E-5 / volEC1
})
dradf = Mixed_Model.mixed_dose_response(times,
'Free_Ia',
'Ia',
list(logspace(1, 5.2)),
parameters={'Ib': 0},
sf=scale_factor)
drbdf = Mixed_Model.mixed_dose_response(times,
'Free_Ib',
'Ib',
list(logspace(-1, 4)),
parameters={'Ia': 0},
sf=scale_factor)
dra15uL = IfnData('custom', df=dradf, conditions={'Alpha': {'Ib': 0}})
drb15uL = IfnData('custom', df=drbdf, conditions={'Beta': {'Ia': 0}})
# 1 mL
Mixed_Model.set_global_parameters({
'volEC': volEC2,
'ka1': ka1 * 1E-5 / volEC2,
'ka2': ka2 * 1E-5 / volEC2,
'k_a1': k_a1 * 1E-5 / volEC2,
'k_a2': k_a2 * 1E-5 / volEC2
})
dradf = Mixed_Model.mixed_dose_response(times,
'Free_Ia',
'Ia',
list(logspace(1, 5.2)),
parameters={'Ib': 0},
sf=scale_factor)
drbdf = Mixed_Model.mixed_dose_response(times,
'Free_Ib',
'Ib',
list(logspace(-1, 4)),
parameters={'Ia': 0},
sf=scale_factor)
dra5mL = IfnData('custom', df=dradf, conditions={'Alpha': {'Ib': 0}})
drb5mL = IfnData('custom', df=drbdf, conditions={'Beta': {'Ia': 0}})
tc_figure = TimecoursePlot((1, 2))
tc_figure.add_trajectory(dra15uL,
'plot',
'-',
subplot_idx=(0, 0),
label=r'IFN$\alpha$ at {} L'.format(volEC1),
doseslice=alpha_testdose,
dose_species='Alpha',
color=alpha_palette[1])
tc_figure.add_trajectory(dra5mL,
'plot',
'-',
subplot_idx=(0, 0),
label=r'IFN$\alpha$ at {} L'.format(volEC2),
doseslice=alpha_testdose,
dose_species='Alpha',
color=alpha_palette[4])
tc_figure.add_trajectory(drb15uL,
'plot',
'-',
subplot_idx=(0, 1),
label=r'IFN$\beta$ at {} L'.format(volEC1),
doseslice=beta_testdose,
dose_species='Beta',
color=beta_palette[0])
tc_figure.add_trajectory(drb5mL,
'plot',
'-',
subplot_idx=(0, 1),
label=r'IFN$\beta$ at {} L'.format(volEC2),
doseslice=beta_testdose,
dose_species='Beta',
color=beta_palette[2])
for ax in tc_figure.axes.flatten():
ax.set_yscale('log')
ifn_fig, ifn_axes = tc_figure.show_figure()
ifn_axes[0].set_ylabel('IFN-alpha2')
ifn_axes[1].set_ylabel('IFN-beta')
ifn_fig.savefig('varying_reaction_volume_tc_IFN.pdf')
'krec_a2': 0.01,
'krec_b1': 0.005,
'krec_b2': 0.05
}
dual_parameters = {
'kint_a': 0.00052,
'kSOCSon': 6e-07,
'kint_b': 0.00052,
'krec_a1': 0.001,
'krec_a2': 0.1,
'krec_b1': 0.005,
'krec_b2': 0.05
}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.1})
times = [10.0, 60.0]
# Make predictions
dradf = Mixed_Model.mixed_dose_response(times,
'TotalpSTAT',
'Ia',
list(logspace(1, 5.2)),
parameters={'Ib': 0},
parameter_names = pd.read_csv(output_dir + os.sep + 'descriptive_statistics.csv').columns.values[1:]
parameters_to_check = [parameters[i] for i in list(np.random.randint(0, high=len(parameters), size=num_checks))]
# Set up model
initial_parameters = {'k_a1': 4.98E-14 * 2, 'k_a2': 8.30e-13 * 2, 'k_d4': 0.006 * 3.8,
'kpu': 0.00095,
'ka2': 4.98e-13 * 2.45, 'kd4': 0.3 * 2.867,
'kint_a': 0.000124, 'kint_b': 0.00086,
'krec_a1': 0.0028, 'krec_a2': 0.01, 'krec_b1': 0.005, 'krec_b2': 0.05}
dual_parameters = {'kint_a': 0.00052, 'kSOCSon': 6e-07, 'kint_b': 0.00052, 'krec_a1': 0.001, 'krec_a2': 0.1,
'krec_b1': 0.005, 'krec_b2': 0.05}
scale_factor = 1.227
Mixed_Model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
Mixed_Model.model_1.set_parameters(initial_parameters)
Mixed_Model.model_1.set_parameters(dual_parameters)
Mixed_Model.model_1.set_parameters({'R1': 12000.0, 'R2': 1511.1})
Mixed_Model.model_2.set_parameters(initial_parameters)
Mixed_Model.model_2.set_parameters(dual_parameters)
Mixed_Model.model_2.set_parameters({'R1': 6755.56, 'R2': 1511.2})
times = [2.5, 5.0, 7.5, 10.0, 20.0, 60.0]
# Compute posterior sample trajectories
def posterior_prediction(parameter_vector, parameter_names=parameter_names):
# Make predictions
Mixed_Model.update_parameters(parameter_vector)
dradf = Mixed_Model.mixed_dose_response(times, 'TotalpSTAT', 'Ia', list(np.logspace(1, 5.2)),
'kpu': 0.0011,
'ka1': 3.3e-15,
'ka2': 1.22e-12,
'kd4': 0.86,
'kd3': 1.74e-05,
'kint_a': 0.000124,
'kint_b': 0.00086,
'krec_a1': 0.0028,
'krec_a2': 0.01,
'krec_b1': 0.005,
'krec_b2': 0.05
}
mixed_parameters = [{'R2': 570, 'R1': 180}, {'R2': 5700, 'R1': 8000}]
het_model = DualMixedPopulation('Mixed_IFN_ppCompatible', 0.8, 0.2)
het_model.model_1.set_parameters(base_parameters)
het_model.model_1.set_parameters(mixed_parameters[0])
het_model.model_2.set_parameters(base_parameters)
het_model.model_2.set_parameters(mixed_parameters[1])
# ----------------------
# Perform stepwise fit
# ----------------------
parameters_to_fit = {
'R1__1': (200, 12000),
'R1__2': (200, 12000),
'R2__1': (200, 12000),
'R2__2': (200, 12000),
'kSOCSon': (1E-07, 1E-06),
'kint_a': (0.0001, 0.002),