Structsearch#

The structsearch tool is a tool to find the best structural model given a base model and a specific model scope to test. This tool will always return a model of the specified type. Currently supported model types are PKPD, TMDD and drug metabolite.

Running#

The structsearch tool is available both in Pharmpy/pharmr.

The code to initiate structsearch for a PKPD model in Python/R is stated below:

from pharmpy.modeling import read_model
from pharmpy.tools read_modelfit_results, run_structsearch

start_model = read_model('path/to/model')
start_model_results = read_modelfit_results('path/to/model')
res = run_structsearch(type='pkpd',
                        search_space="DIRECTEFFECT(*)",
                        model=start_model,
                        results=start_model_results)

from pharmpy$tools read_modelfit_results, run_structsearch

start_model <- read_model('path/to/model')
start_model_results <- read_modelfit_results('path/to/model')
res <- run_structsearch(type='pkpd',
                        search_space="DIRECTEFFECT(*)",
                        model=start_model,
                        results=start_model_results)

This will take an input model model with a search_space that includes all direct effect PKPD models.

Arguments#

The arguments of the structsearch tool are listed below. Which arguments are mandatory or optional depends on the type of model. More information can be found on the respective page.

Argument	Description
`type`	Type of model. Can be either pkpd, drug_metabolite or tmdd (The model types).
`search_space`	Search space of models to test.
`b_init`	Initial estimate for baseline effect (only for PKPD models). Default is 0.1.
`emax_init`	Initial estimate for E _max parameter (only for PKPD models). Default is 0.1.
`ec50_init`	Initial estimate for EC ₅₀ parameter (only for PKPD models). Default is 0.1.
`met_init`	Initial estimate for mean equilibration time (only for PKPD models). Default is 0.1.
`model`	PK start model.
`results`	ModelfitResults of the start model.
`strictness`	Strictness criteria for model selection. Optional. Default is “minimization_successful or (rounding_errors and sigdigs>= 0.1)”.
`extra_model`	Extra model for TMDD structsearch (only for TMDD). Optional.
`extra_model_results`	ModelfitResults object for the extra model for TMDD structsearch (TMDD only). Optional.

The model types#

Structsearch is currently available for PKPD, drug metabolite and TMDD models. For more detailed information about the model types see the references.

type	Description
`pkpd`	PKPD models
`drug_metabolite`	Drug metabolite models
`TMDD`	TMDD models

The search space#

The search space is a set of possible model types that are allowed for the final model (e.g. only direct effect models for PKPD models). The search space is given as a string with a specific grammar, according to the Model Feature Language (MFL) (see detailed description).

For detailed information about the search space for the different model types in structsearch please see the respective page.

Results#

The results object contains various summary tables which can be accessed in the results object, as well as files in .csv/.json format. The name of the selected best model (based on the input selection criteria) is also included.

Example results for the different model types can be found on the respective pages.