Dissertation/Thesis Abstract

Agent based modeling of tissue metabolism
by Powers, Edwin Carey, M.S., Tufts University, 2010, 115; 1482850
Abstract (Summary)

Motivation. Metabolic network models serve as powerful tools to quantify and predict how changes in metabolite concentrations and other variables influence cellular function. However, these models are limited in their ability to explain cellular heterogeneity and incorporate qualitative observations and knowledge. Agent based modeling provides a framework to study the emergent behavior of a heterogeneous population of cells. Moreover, this framework supports rule-based models that incorporate qualitative observations through a series of agent heuristics. A comprehensive review of the literature suggests that little work has been done to date to incorporate the reaction kinetics of metabolic network models into agent based models.

Results. A novel platform was developed which supports the integration of metabolic network models into an agent based modeling framework. This platform couples the reaction kinetics of cellular metabolism with rule-based and probabilistic determination of cellular phenotype. Through this coupling, each cellular network can be modeled as an agent whose metabolic state depends on the biochemical cues established by the local environment. In this coupled model, each agent (cell) interprets the environment based on its present state of metabolism, thereby generating an emergent behavior for the system (tissue), which consists of an ensemble of such agents. As proof of concept, an agent based model of adipocyte formation, or adipogenesis, was developed. Each agent was allowed to model the metabolism of the adipocyte or preadipocyte, two distinct cell types (i.e. phenotypes) whose interactions are critical in adipose tissue development. Based on rules derived from qualitative knowledge of adipogenesis, the agent based model demonstrated realistic emergent behavior describing the induction of preadipocyte differentiation through the metabolic activity of the neighboring adipocytes.

Indexing (document details)
Advisor: Lee, Kyongbum
Commitee: Georgakis, Christos, Hassoun, Soha
School: Tufts University
Department: Chemical and Biological Engineering
School Location: United States -- Massachusetts
Source: MAI 49/02M, Masters Abstracts International
Subjects: Systematic, Bioinformatics
Keywords: Agent based simulations, Computational biology, Metabolic network models, Preadipocyte differentiation
Publication Number: 1482850
ISBN: 978-1-124-31879-0
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy