petab.core¶

PEtab core functions

Functions

`create_condition_df`(parameter_ids, condition_ids)	Create empty condition dataframe
`create_measurement_df`()	Create empty measurement dataframe
`create_parameter_df`(sbml_model, …)	Create a new PEtab parameter table
`flatten_timepoint_specific_output_overrides`(…)	If the PEtab problem definition has timepoint-specific observableParameters or noiseParameters for the same observable, replace those by replicating the respective observable.
`get_condition_df`(condition_file_name)	Read the provided condition file into a pandas.Dataframe
`get_default_condition_file_name`(model_name, …)	Get file name according to proposed convention
`get_default_measurement_file_name`(…)	Get file name according to proposed convention
`get_default_parameter_file_name`(model_name, …)	Get file name according to proposed convention
`get_default_sbml_file_name`(model_name, folder)	Get file name according to proposed convention
`get_measurement_df`(measurement_file_name)	Read the provided measurement file into a pandas.Dataframe.
`get_measurement_parameter_ids`(measurement_df)	Return list of ID of parameters which occur in measurement table as observable or noise parameter overrides.
`get_model_parameters`(sbml_model)	Return list of SBML model parameter IDs which are not AssignmentRule targets for observables or sigmas
`get_noise_distributions`(measurement_df)	Returns dictionary of cost definitions per observable, if specified.
`get_notnull_columns`(df, candidates)	Return list of df-columns in candidates which are not all null/nan.
`get_observable_id`(parameter_id)	Get observable id from sigma or observable parameter_id
`get_observables`(sbml_model, remove)	Returns dictionary of observable definitions.
`get_optimization_parameters`(parameter_df)	Get list of optimization parameter ids from parameter dataframe.
`get_parameter_df`(parameter_file_name)	Read the provided parameter file into a pandas.Dataframe.
`get_placeholders`(formula_string, …)	Get placeholder variables in noise or observable definition for the given observable ID.
`get_priors_from_df`(parameter_df)	Create list with information about the parameter priors
`get_required_parameters_for_parameter_table`(…)	Get set of parameters which need to go into the parameter table
`get_rows_for_condition`(measurement_df, …)	Extract rows in measurement_df for condition according to ‘preequilibrationConditionId’ and ‘simulationConditionId’ in condition.
`get_sigmas`(sbml_model, remove)	Returns dictionary of sigma definitions.
`get_simulation_conditions`(measurement_df)	Create a table of separate simulation conditions.
`measurements_have_replicates`(measurement_df)	Tests whether the measurements come with replicates
`parameter_is_offset_parameter`(parameter, formula)	Returns true if parameter parameter is an offset parameter with positive sign in formula formula.
`parameter_is_scaling_parameter`(parameter, …)	Returns true if parameter parameter is a scaling parameter in formula formula.
`sample_from_prior`(prior, n_starts)	Creates samples based on prior
`sample_parameter_startpoints`(parameter_df, …)	Create numpy.array with starting points for an optimization
`sbml_parameter_is_observable`(sbml_parameter)	Returns whether the libsbml.Parameter sbml_parameter matches the defined observable format.
`sbml_parameter_is_sigma`(sbml_parameter)	Returns whether the libsbml.Parameter sbml_parameter matches the defined sigma format.
`split_parameter_replacement_list`(…)	Split values in observableParameters and noiseParameters in measurement table.

Classes

Problem(sbml_model, sbml_reader, …) PEtab parameter estimation problem as defined by - SBML model - condition table - measurement table - parameter table [optional]

class petab.core.Problem(sbml_model: libsbml.Model = None, sbml_reader: libsbml.SBMLReader = None, sbml_document: libsbml.SBMLDocument = None, condition_df: pandas.core.frame.DataFrame = None, measurement_df: pandas.core.frame.DataFrame = None, parameter_df: pandas.core.frame.DataFrame = None)¶

Bases: object

PEtab parameter estimation problem as defined by - SBML model - condition table - measurement table - parameter table [optional]

condition_df¶: @type pandas.DataFrame

measurement_df¶: @type pandas.DataFrame

parameter_df¶: @type pandas.DataFrame

sbml_reader¶: @type libsbml.SBMLReader Stored to keep object alive.

sbml_document¶: @type libsbml.Document Stored to keep object alive.

sbml_model¶: @type libsbml.Model

create_parameter_df(*args, **kwargs)¶

Create a new PEtab parameter table

See create_parameter_df

static from_files(sbml_file: str = None, condition_file: str = None, measurement_file: str = None, parameter_file: str = None) → petab.core.Problem¶

Factory method to load model and tables from files.

Parameters:	sbml_file – PEtab SBML model condition_file – PEtab condition table measurement_file – PEtab measurement table parameter_file – PEtab parameter table

static from_folder(folder: str, model_name: str = None) → petab.core.Problem¶

Factory method to use the standard folder structure and file names, i.e.

${model_name}/

+– experimentalCondition_${model_name}.tsv +– measurementData_${model_name}.tsv +– model_${model_name}.xml +– parameters_${model_name}.tsv

Parameters:	folder – Path to the directory in which the files are located. model_name – If specified, overrides the model component in the file names. Defaults to the last component of folder.

get_constant_parameters()¶: Provide list of IDs of parameters which are fixed (i.e. not subject to optimization, no sensitivities w.r.t. these parameters are required).

get_dynamic_simulation_parameters()¶: See get_model_parameters

get_noise_distributions()¶: See get_noise_distributions.

get_observables(remove: bool = False)¶: Returns dictionary of observables definitions See assignment_rules_to_dict for details.

get_optimization_parameters()¶

Return list of optimization parameter IDs.

See get_optimization_parameters.

get_optimization_to_simulation_parameter_mapping(warn_unmapped: bool = True)¶: See get_simulation_to_optimization_parameter_mapping.

get_sigmas(remove: bool = False)¶: Return dictionary of observableId => sigma as defined in the SBML model. This does not include parameter mappings defined in the measurement table.

get_simulation_conditions_from_measurement_df()¶: See petab.get_simulation_conditions

lb¶: Parameter table lower bounds

sample_parameter_startpoints(n_starts: int = 100)¶

Create starting points for optimization

See sample_parameter_startpoints

ub¶: Parameter table upper bounds

x_fixed_indices¶: Parameter table non-estimated parameter indices

x_fixed_vals¶: Nominal values for parameter table non-estimated parameters

x_ids¶: Parameter table parameter IDs

x_nominal¶: Parameter table nominal values

petab.core.create_condition_df(parameter_ids: Iterable[str], condition_ids: Iterable[str] = None) → pandas.core.frame.DataFrame¶

Create empty condition dataframe

Parameters:	parameter_ids – the columns condition_ids – the rows
Returns:	An pandas.DataFrame with empty given rows and columns and all nan values

petab.core.create_measurement_df() → pandas.core.frame.DataFrame¶: Create empty measurement dataframe

petab.core.create_parameter_df(sbml_model: libsbml.Model, condition_df: pandas.core.frame.DataFrame, measurement_df: pandas.core.frame.DataFrame, parameter_scale: str = 'log10', lower_bound: Iterable[T_co] = None, upper_bound: Iterable[T_co] = None) → pandas.core.frame.DataFrame¶

Create a new PEtab parameter table

All table entries can be provided as string or list-like with length matching the number of parameters

Parameters:	sbml_model – @type libsbml.Model condition_df – @type pandas.DataFrame measurement_df – @type pandas.DataFrame parameter_scale – parameter scaling lower_bound – lower bound for parameter value upper_bound – upper bound for parameter value

petab.core.flatten_timepoint_specific_output_overrides(petab_problem: petab.core.Problem) → None¶

If the PEtab problem definition has timepoint-specific observableParameters or noiseParameters for the same observable, replace those by replicating the respective observable.

This is a helper function for some tools which may not support such timepoint-specific mappings.

Parameters:	petab_problem – PEtab problem to work on

petab.core.get_condition_df(condition_file_name: str) → pandas.core.frame.DataFrame¶

Read the provided condition file into a pandas.Dataframe

Conditions are rows, parameters are columns, conditionId is index.

petab.core.get_default_condition_file_name(model_name: str, folder: str = '')¶: Get file name according to proposed convention

petab.core.get_default_measurement_file_name(model_name: str, folder: str = '')¶: Get file name according to proposed convention

petab.core.get_default_parameter_file_name(model_name: str, folder: str = '')¶: Get file name according to proposed convention

petab.core.get_default_sbml_file_name(model_name: str, folder: str = '')¶: Get file name according to proposed convention

petab.core.get_measurement_df(measurement_file_name: str) → pandas.core.frame.DataFrame¶: Read the provided measurement file into a pandas.Dataframe.

petab.core.get_measurement_parameter_ids(measurement_df: pandas.core.frame.DataFrame) → list¶: Return list of ID of parameters which occur in measurement table as observable or noise parameter overrides.

petab.core.get_model_parameters(sbml_model: libsbml.Model) → List[str]¶: Return list of SBML model parameter IDs which are not AssignmentRule targets for observables or sigmas

petab.core.get_noise_distributions(measurement_df: pandas.core.frame.DataFrame) → dict¶

Returns dictionary of cost definitions per observable, if specified.

Looks through all parameters satisfying sbml_parameter_is_cost and return as dictionary.

Parameters:	measurement_df – PEtab measurement table
Returns:	cost definition}
Return type:	{observableId

petab.core.get_notnull_columns(df: pandas.core.frame.DataFrame, candidates: Iterable[T_co])¶: Return list of df-columns in candidates which are not all null/nan. The output can e.g. be used as input for pandas.DataFrame.groupby.

petab.core.get_observable_id(parameter_id: str) → str¶

Get observable id from sigma or observable parameter_id

e.g. for observable_obs1 -> obs1: sigma_obs1 -> obs1

petab.core.get_observables(sbml_model: libsbml.Model, remove: bool = False) → dict¶: Returns dictionary of observable definitions. See assignment_rules_to_dict for details.

petab.core.get_optimization_parameters(parameter_df: pandas.core.frame.DataFrame) → List[str]¶: Get list of optimization parameter ids from parameter dataframe.

petab.core.get_parameter_df(parameter_file_name: str) → pandas.core.frame.DataFrame¶: Read the provided parameter file into a pandas.Dataframe.

petab.core.get_placeholders(formula_string: str, observable_id: str, override_type: str) → set¶

Get placeholder variables in noise or observable definition for the given observable ID.

Parameters:	formula_string – observable formula (typically from SBML model) observable_id – ID of current observable override_type – ‘observable’ or ‘noise’, depending on whether formula is for observable or for noise model
Returns:	(Un-ordered) set of placeholder parameter IDs

petab.core.get_priors_from_df(parameter_df: pandas.core.frame.DataFrame)¶

Create list with information about the parameter priors

Parameters:	parameter_df – @type pandas.DataFrame

petab.core.get_required_parameters_for_parameter_table(sbml_model: libsbml.Model, condition_df: pandas.core.frame.DataFrame, measurement_df: pandas.core.frame.DataFrame) → Set[str]¶: Get set of parameters which need to go into the parameter table

petab.core.get_rows_for_condition(measurement_df: pandas.core.frame.DataFrame, condition: Union[pandas.core.frame.DataFrame, dict]) → pandas.core.frame.DataFrame¶

Extract rows in measurement_df for condition according to ‘preequilibrationConditionId’ and ‘simulationConditionId’ in condition.

Returns:	cur_measurement_df – The subselection of rows in measurement_df for the condition condition.
Return type:	pd.DataFrame

petab.core.get_sigmas(sbml_model: libsbml.Model, remove: bool = False) → dict¶

Returns dictionary of sigma definitions.

Keys are observable IDs, for values see assignment_rules_to_dict for details.

petab.core.get_simulation_conditions(measurement_df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame¶

Create a table of separate simulation conditions. A simulation condition is a specific combination of simulationConditionId and preequilibrationConditionId.

Parameters:	measurement_df – PEtab measurement table
Returns:	Dataframe with columns ‘simulationConditionId’ and ‘preequilibrationConditionId’. All-NULL columns will be omitted.

petab.core.measurements_have_replicates(measurement_df: pandas.core.frame.DataFrame) → bool¶

Tests whether the measurements come with replicates

Parameters:	measurement_df – Measurement table
Returns:	True if there are replicates, False otherwise

petab.core.parameter_is_offset_parameter(parameter: str, formula: str) → bool¶: Returns true if parameter parameter is an offset parameter with positive sign in formula formula.

petab.core.parameter_is_scaling_parameter(parameter: str, formula: str) → bool¶: Returns true if parameter parameter is a scaling parameter in formula formula.

petab.core.sample_from_prior(prior: tuple, n_starts: int)¶

Creates samples based on prior

Parameters:	prior – @type tuple n_starts – @type int

petab.core.sample_parameter_startpoints(parameter_df: pandas.core.frame.DataFrame, n_starts: int = 100)¶

Create numpy.array with starting points for an optimization

Dimension of output: n_optimization_parameters x n_startpoints

Parameters:	parameter_df – @type pandas.DataFrame n_starts – @type int

petab.core.sbml_parameter_is_observable(sbml_parameter: libsbml.Parameter) → bool¶: Returns whether the libsbml.Parameter sbml_parameter matches the defined observable format.

petab.core.sbml_parameter_is_sigma(sbml_parameter: libsbml.Parameter) → bool¶: Returns whether the libsbml.Parameter sbml_parameter matches the defined sigma format.

petab.core.split_parameter_replacement_list(list_string: Union[str, numbers.Number], delim: str = ';') → List[T]¶

Split values in observableParameters and noiseParameters in measurement table. Convert numeric values to float.

Parameters:	delim – delimiter list_string – delim-separated stringified list