Parameter Identification in Synthetic Biological Circuits Using Multi-Objective Optimization

[EN] Synthetic biology exploits the of mathematical modeling of synthetic circuits both to predict the behavior of the designed synthetic devices, and to help on the selection of their biological coin portents. The increasing complexity of the circuits being designed requires performing approximatio...

ver descrição completa

Detalhes bibliográficos
Autores: Boada-Acosta, Yadira Fernanda|||0000-0001-5677-2702, Vignoni, Alejandro|||0000-0001-9977-7132, Picó, Jesús|||0000-0003-4144-3521, Reynoso Meza, Gilberto
Tipo de documento: artigo
Data de publicação:2016
País:España
Recursos:Universitat Politècnica de València (UPV)
Repositório:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglês
OAI Identifier:oai:riunet.upv.es:10251/150443
Acesso em linha:https://riunet.upv.es/handle/10251/150443
Access Level:Acceso aberto
Palavra-chave:Biological circuits
Kinetic parameters
Parameter identification
Multi-objective
Optimization
INGENIERIA DE SISTEMAS Y AUTOMATICA
Descrição
Resumo:[EN] Synthetic biology exploits the of mathematical modeling of synthetic circuits both to predict the behavior of the designed synthetic devices, and to help on the selection of their biological coin portents. The increasing complexity of the circuits being designed requires performing approximations and model reductions to get handy models. Parameter estimation in these models remains a challenging problem that has usually been addressed by optimizing the weighted combination of different prediction errors to obtain a single solution. The single-objective approach is inadequate to incorporate different kinds of experiments, and to identify parameters for an ensemble of biological circuit models. We present a methodology based on multi-objective optimization to perform parameter estimation that can fully harness to ensembles of local models for biological circuits. The methodology uses a global multi-objective evolutionary algorithm and a multi-criteria decision making strategy to select the most suitable solutions. Our approach finds an approximation to the Pareto optimal set of model parameters that correspond to each experimental scenario. Then, the Pareto set was clustered according to the experimental scenarios. This, in turn, allows to analyze the sensitivity of model parameters for different scenarios. Finally, we show the methodology applicability through the case study of a genetic incoherent feed-forward circuit, under different concentrations of the inducer input signal. (C) 2016 IFAC (International Federation Of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.