- author = "John R. Koza and William Mydlowec and Guido Lanza and Jessen Yu and Martin A. Keane",
- title = "Reverse engineering of metabolic pathways from observed data by means of genetic programming",
- booktitle = "First International Conference on Systems Biology (ICSB)",
- year = "2000",
- address = "Tokyo",
- month = "14-16 " # nov,
- organisation = "Japan Society for Bioinformatics",
- keywords = "genetic algorithms, genetic programming, Biology, metabolic pathways, reverse engineering",
-
URL = "
http://www.genetic-programming.com/jkpdf/icsb2000metabolic.pdf",
- abstract = "Recent work has demonstrated that genetic programming is capable of automatically creating complex networks and structures (e.g., analog electrical circuits, controllers, and antennas) whose behavior is governed by linear and non-linear differential equations and whose behavior matches prespecified data values. The concentrations of substances (substrates, products, and catalysts) participating in networks of chemical reactions are described by non-linear continuous-time differential equations (e.g., Michaelis-Menten equations). This paper demonstrates that it is possible to automatically create (reverse engineer) a network of chemical reactions from observed time-domain data. Genetic programming starts with observed time-domain concentrations of substances and automatically creates both the topology and sizing (i.e., the rates of each reaction) of a network whose behavior matches observed time-domain data. Specifically, genetic programming automatically created a metabolic pathway involving four chemical reactions that consume glycerol and fatty acids as input, used ATP as a cofactor, and produced diacyl-glycerol as the final product. The metabolic pathway was created from 270 data points. The automatically created metabolic pathway contains three key topological features, including an internal feedback loop, a bifurcation point where one substance is distributed to two different reactions, and an accumulation point where one substance is accumulated from two sources. The topology and sizing of the entire metabolic pathway was automatically created using only the time-domain concentration values of diacyl-glycerol (the final product).",
-
notes = "ICSB-2000 E-CELL. Population size 100000.
This paper was subsequently published in 2001 as chapter in book (edited by Kitano) \cite{koza:2001:FSB}",
