Steiert Bernhard, Kreutz Clemens, Raue Andreas, Timmer Jens
Institute of Physics, University of Freiburg, Freiburg, Germany.
Signalling Research Centres BIOSS and CIBSS, University of Freiburg, Freiburg, Germany.
Methods Mol Biol. 2019;1945:341-362. doi: 10.1007/978-1-4939-9102-0_16.
Mechanistic models of biomolecular processes are established research tools that enable to quantitatively investigate dynamic features of biological processes such as signal transduction cascades. Often, these models aim at describing a large number of states, for instance concentrations of proteins and small molecules, as well as their interactions. Each modeled interaction increases the number of potentially unknown parameters like reaction rate constants or initial amount of proteins. In order to calibrate these mechanistic models, the unknown model parameters have to be estimated based on experimental data. The complexity of parameter estimation raises several computational challenges that can be tackled within the Data2Dynamics modeling environment. The environment is a well-tested, high-performance software package that is tailored to the modeling of biological processes with ordinary differential equation models and using experimental biomolecular data.In this chapter, we introduce and provide "recipes" for the most frequent analyses and modeling tasks in the Data2Dynamics modeling environment. The presented protocols comprise model building, data handling, parameter estimation, calculation of confidence intervals, model selection and reduction, deriving prediction uncertainties, and designing informative novel experiments.
生物分子过程的机理模型是成熟的研究工具，能够定量研究生物过程的动态特征，如信号转导级联反应。通常，这些模型旨在描述大量状态，例如蛋白质和小分子的浓度及其相互作用。每个建模的相互作用都会增加潜在未知参数的数量，如反应速率常数或蛋白质的初始量。为了校准这些机理模型，必须根据实验数据估计未知的模型参数。参数估计的复杂性带来了几个计算挑战，这些挑战可以在Data2Dynamics建模环境中解决。该环境是一个经过充分测试的高性能软件包，专为使用常微分方程模型并结合实验生物分子数据对生物过程进行建模而量身定制。在本章中，我们介绍并提供Data2Dynamics建模环境中最常见分析和建模任务的“方法”。所展示的协议包括模型构建、数据处理、参数估计、置信区间计算、模型选择与简化、推导预测不确定性以及设计信息丰富的新实验。
