A genetic approach for synthesizing metabolic models from time series

In this paper we introduce a new approach, based on genetic algorithms and multiple linear regression, for the synthesis of flux regulation functions in metabolic models from observed time series. Genetic algorithms are used as a variable selection technique to identify the best primitive functions for flux regulation, and multiple linear regression is employed to compute primitive function coefficients. Our methodology is here successfully applied to synthesize a set of regulation functions able to regenerate an observed dynamics for the mitotic oscillator in early amphibian embryos.