ADGenKinetics: An Algorithmically Differentiated Library for Biochemical Networks Modeling via Simplified Kinetics Formats

Abstract

This work demonstrates the compact but powerful freely available Modelica library ADGenKinetics for descriptive modeling of biochemical reaction networks using simplified kinetics formats. While existing powerful works and guidelines for modeling biochemical reaction networks based on classical mechanistic kinetics already exist, in this work a first attempt of utilizing the power of Modelica constructs for providing a compact implementation of simplified kinetic formats with generalized structured formulas is presented. This gives the opportunity of realizing biochemical reaction networks using few number of reaction components, in contrast to libraries based on classical mechanistic kinetics which require hundreds of reaction components. Moreover, ADGenKinetics is the first algorithmically differentiated Modelica library that is enhanced with differentiated components by which parameter sensitivities are additionally computed with minimal efforts from the user perspective.