Dynamic modeling and parameter estimation with dMod

Play Dynamic modeling and parameter estimation with dMod
Sign in to queue


useR!2017: Dynamic modeling and parameter estimatio...

Keywords: Parameter Estimation, ODEs, Systems Biology, Maximum-Likelihood, Profile-Likelihood
Webpages: https://github.com/dkaschek/dMod, https://github.com/dkaschek/cOde
ODE models to describe and understand interactions in complex dynamical systems are widely used in the physical sciences and beyond. In many situations, the model equations depend on parameters. When parameters are not known from first principle, they need to be estimated from experimental data.
The dMod package for R provides a framework for formulating complex reaction networks and estimating the inherent reaction parameters from experimental data. By design, different experimental conditions as well as explicit or implicit equality constraints, e.g., steady-state constraints, are formulated by parameter transformations which thereby take a central role in dMod. Since, in general, the observed reaction dynamics is a non-linear function of the reaction parameters, profile-likelihood methods are implemented to assess non-linear parameter dependencies and estimate parameter- and prediction confidence intervals.
Here, we present the abilities and particularities of our modeling framework. The methods are illustrated based on a minimal systems biology example.





Right click to download this episode

The Discussion

Add Your 2 Cents