Skip to contents

BayesPharma: Tools for Bayesian Analysis of Non-Linear Pharmacology Models

BayesPharma-package.Rd

The BayesPharma package builds on the Stan and brms to provide support for Bayesian regression modeling for foundational pharmacology models. For each model type, the user provides

formula:

Describing how the model parameters, treatment, and optional predictors lead to the measured response using functions provided by BayesPharma for each model type.

observed data:

response, treatment, and optional predictors as a data.frame

prior:

Initial distributions over the model parameters

The models that BayesPharma support are

sigmoid_model:

4-parameter Hill equation

MuSyC_model:

Bivariate synergy model with Bliss and Loewe interaction models as special cases

michaelis_menten_model:

Michaelis Menten enzyme kinetics ordinary differential equation model

tQ_model:

Generalization of enzyme progress curve kinetics ordinary differential equation

growth_sigmoid_model:

Sigmoid model for growth kinetics

growth_richards_model:

Generalized Richards model for growth kinetics

The BayesPharma package also provides a range of case studies as templates and examples for getting started at applying Bayesian modeling to pharmacology data analysis.

Building on rstan::stan brings the performance and stability of No-U-Turn Sampling (NUTs) Hamiltonian Monte Carlo and a whole ecosystem of tools for model assessment, and visualization (see https://mc-stan.org/).

Building on brms allows for compact formula based model specification adding complexity to the model incrementally, including handling missing data, measurement error, and other response distributions (see https://paul-buerkner.github.io/brms/).

See also

Useful links:

Author

Maintainer: Matthew O'Meara maom@umich.edu (ORCID)

Authors: