nonmem.utils

The goal of {nonmem.utils} is to provide a toolbox that includes helper functions, shiny apps and quarto template reports to help viewing, investigating and reporting Nonmem.

[!CAUTION] This repository is currently under development

Installation

You can install the development version of nonmem.utils like so:

remotes::install_github("pchelle/nonmem.utils")

Example

Run the following code to open the toolbox (also available in the RStudio Addins menu):

nonmem.utils::shiny_toolbox()

Select the desired tool from the toolbox (for instance, covariate-analysis) and click on the button .

The appropriate code will be added to the R console to start the corresponding tool.

Usage

Most of the functions use the following arguments data and meta_data.

data

The data argument is a data.frame that contains the data to be analyzed.

In Nonmem, the input datasets are usually provided as .csv files. In such cases, the tools will read them using readr::read_csv().

data <- readr::read_csv("path/to/your/data.csv")

Nonmem output files are usually provided as .tab files with 2 lines header. In such cases, the tools will read them using readr::read_table().

data <- readr::read_table("path/to/your/data.tab", skip = 1)

meta_data

The meta_data argument is a dictionary data.frame that contains the information about the dataset.

The meta_data is expected to be provided as .csv files. The tools will read them using readr::read_csv().

meta_data <- readr::read_csv("path/to/your/dictionary.csv")

It should contain the following columns:

  • Name: name of the dataset variable
  • Type: type of the dataset variables in lower cases.
    • The type can be one of the following: id, occ, time, tad, dv, amt, evid, mdv.
    • If the variable is a continuous covariate, use cov.
    • If the variable is a categorical covariate, use cat.
    • If the variable is a model BSV/BOV parameters, use eta.
  • Label: displayed label of the dataset variable
  • Unit: displayed unit of the dataset variable
    • For categorical covariates (Type is cat), see explanations and example below.
  • Min: the minimum expected value of the dataset variable
  • Max: the maximum expected value of the dataset variable

Mapping between the categorical covariate values and labels can be provided in the Label column. The jsonlite package is leveraged to parse the mapping. As a consequence, the mapping is expected to be provided as "value":"label" split by pipes |.

For instance, to map a variable named sex with 0 coding for male and 1 coding for female, the Label cell should be have the following value:

"0":"male"|"1":"female"