InitialAssignment issue

[aidinbii]

Hello, there is an issue with InitialAssignment

• The sbml model modelDef_PK_1C_constr_2.txt

When I run:

#With petab
from amici.petab.petab_import import import_petab_problem
amici_model = import_petab_problem(petab_problem,)

simulation_time = 10
number_timepoints = 10
amici_model.setTimepoints(np.linspace(0, simulation_time, number_timepoints)) 

problem_parameters_ex = {
 'Ka_3': 1.14,
 'Ka_30': 9.63,
 'Ka_100': 9.63,
 'Ke_30': 0.64,
 'Ke_100': 0.07,
 'VdF_3': 2800,
 'VdF_30': 25800,
 'VdF_100': 32800,
 }

simulat_petab_ex = amici.petab.simulations.simulate_petab(petab_problem, amici_model,
problem_parameters=problem_parameters_ex)

It returns:

KeyError                                  Traceback (most recent call last)
File [~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:596](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:596), in _apply_parameter_table(par_mapping, scale_mapping, parameter_df, scaled_parameters, fill_fixed_parameters)
    [594](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:594) try:
    [595](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:595)     # the overridee is a model parameter
--> [596](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:596)     par_mapping[problem_par] = par_mapping[sim_par]
    [597](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/petab/parameter_mapping.py:597)     scale_mapping[problem_par] = scale_mapping[sim_par]

KeyError: 'Ke_3'

During handling of the above exception, another exception occurred:

KeyError                                  Traceback (most recent call last)
File [~/miniconda3/envs/modeling/lib/python3.9/site-packages/pandas/core/indexes/base.py:3790](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/pandas/core/indexes/base.py:3790), in Index.get_loc(self, key)
   [3789](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/pandas/core/indexes/base.py:3789) try:
-> [3790](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/pandas/core/indexes/base.py:3790)     return self._engine.get_loc(casted_key)
   [3791](https://file+.vscode-resource.vscode-cdn.net/home/aidin/Documents/tum_courses/hiwi_Atefeh/alivexbiotech/projects/igg4_python/notebooks/case_studies_1/~/miniconda3/envs/modeling/lib/python3.9/site-packages/pandas/core/indexes/base.py:3791) except KeyError as err:

File index.pyx:152, in pandas._libs.index.IndexEngine.get_loc()

File index.pyx:181, in pandas._libs.index.IndexEngine.get_loc()

File pandas/_libs/hashtable_class_helper.pxi:7080, in pandas._libs.hashtable.PyObjectHashTable.get_item()

File pandas/_libs/hashtable_class_helper.pxi:7088, in pandas._libs.hashtable.PyObjectHashTable.get_item()

KeyError: 'Ke_3'

The compiled model:

Model parameters: ['Ka', 'Ke', 'VdF', 'Ka_3', 'Ke_100', 'Ke_30', 'Ka_30', 'Ka_100', 'VdF_3', 'VdF_30', 'VdF_100', 'noiseParameter1_C_central'] 

Model const parameters: ['preequilibration_indicator', 'initial_mu_1_preeq', 'initial_mu_1_sim'] 

Model outputs:    ['C_central'] 

Model states:     ['mu_1', 'mu_2'] 

But, if I change the InitialAssignment so that there are no other symbols, for example:

<initialAssignment symbol="Ke_3">
        <math xmlns="http://www.w3.org/1998/Math/MathML">
          <apply>
            <plus/>
           <cn type="real"> 1.79 </cn> 
            <cn type="real"> 2.21 </cn>
          </apply>
        </math>
      </initialAssignment>

It works without errors and the compiled model adds Ke_3 to Model const parameters

I'm using up-to-date versions of packages

Thank you

[FFroehlich]

Is it possible this is a typo of Ke_3 vs Ka_3. Hard to say anything else without the full PEtab problem

[aidinbii]

@FFroehlich , unfortunately not

Here are the PEtab tables:

• parameters_modelDef_PK_1C_constr_2.csv
• experimental_conditions_modelDef_PK_1C_constr_2.csv
• measurement_table_PK_1C.csv
• observables_modelDef_PK_1C_constr_2.csv
• problem_PK_1C_constr_2.txt

This is related to ICB-DCM/pyPESTO#1334

[dweindl]

I can confirm that overriding some parameter A via the condition table with some parameter B, where B has a parameter-dependent initial assignment is currently not supported in amici. We should either implement that, or produce a more informative error.

[aidinbii]

@dweindl , thank you, I see.
Then there is no way to specify parameter constraints for parameter estimation.
For example, if I have a parameter vector: (p1, p2, p3, p4). And I want a constraint: p2 <= p3.

See: ICB-DCM/pyPESTO#1334

[FFroehlich]

If both parameters are estimated on a log scale, you can reparametrize the problem as p3 =p2 + r or p3=p2 * r with r>0 or r>1 respectively.

[aidinbii]

@FFroehlich, that's what I did before. Actually, this brought me to this InitialAssignment issue

Please, see ICB-DCM/pyPESTO#1334