Differential inclusions and piecewise models in biochemical and gene regulatory networks
Year: 2010
Type: dissertation
Abstract: The simplification of a complex system is a widely accepted way of its modeling, the inevitable drawback of which is the ”resolution reduction”, i.e., the loss of information inherent to the real-world processes. In this thesis I want to address the ”resolution reduction” issue for some well-elaborated simplifications in systems biology. Broadly speaking, the current work is about determining and justifying mathematically the criteria of eligibility of the simplified frameworks.\nThe work presented in this thesis is aimed at studying how the theory of differential inclusions can be applied to some open mathematical problems stemming from biology. These problems are motivated by two basic simplification paradigms in systems biology at the micro level: the Boolean-like formalism and the power-law formalism. The Boolean-like formalism is widely used for describing gene regulatory networks. Roughly speaking, it consists in replacing smooth, yet inconvenient, steep sigmoidal nonlinearities with step functions. The power-law formalism assumes that unknown or numerically obtained relationships within biochemical reaction networks can be described by sums of power monomials. Note that both of these formalisms, in the form we consider them in the present work, lead to piecewise models. These models based on discontinuous differential equations can be treated in a rigorous mathematical way using the theory of differential inclusions, which was originally developed within theoretical and applied mathematics. Its application to systems biology has not yet been completely developed.\nFor the Boolean-like formalism in gene regulatory networks (Paper I, Paper II) we present a rigorous analysis of the simplified model based on differential inclusions; a mathematical justification for the similarities between the ”realworld” model (actually also representing a simplification of the real world processes) and the simplified model.\nFor the power-law formalism in biochemical reaction networks we introduce an algorithm for automated piecewise power-law approximations (Paper III). The algorithm solves the important problem of automated partition in piecewise models. We also obtain analytical results on the convergence and well-posedness of the approximations generated by this algorithm (Paper IV).\nIn Paper V the study of differential inclusions is extended to its generalization - the theory of functional differential inclusions, which covers also the models with delay effects. Delay effects are well-know phenomena in biological processes, e.g., time delays in gene expression, but are often omitted in modeling for the sake of simplicity. Here we focus mainly on analytical techniques and prove some basic theoretical results.
Source: BIBSYS Brage (BIBSYS (Norway))
Author: Anna Machina
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2601124175
