L'évolution expérimentale in silico

Carole Knibbe 1, 2
2 BEAGLE - Artificial Evolution and Computational Biology
LBBE - Laboratoire de Biométrie et Biologie Evolutive - UMR 5558, Inria Grenoble - Rhône-Alpes, LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : In silico experimental evolution is a method aimed at studying the mechanisms of evolution and the constraints these mechanisms may impose on cellular organization. Researchers create virtual organisms and let them compete for reproduction and mutate inside the computer. To do so, one must choose a formalism to represent the genome of an organism, define an artificial biochemistry to compute the organism's form or behavior from its genome, set a task that the organisms must fulfill to survive and reproduce in their abstract environment, and finally define the mutation mechanisms that will happen when an organism will replicate its genome. Although this technique is algorithmically close to the genetic algorithms used in evolutionary computation, it is actually closer to the wet evolutionary experiments carried out by biologists on microorganisms, both in terms of scientific objectives and in terms of experimental design. In silico experimental evolution is used to stimulate the intuition, to generate plausible hypotheses on the mechanisms and effects of evolution. The strengths of this simulation approach are the easy way the parameters can be controlled, the possibility to run many repetitions, and the perfect fossil record. However, because simplifications are inevitably made when designing the virtual organisms and their biochemistry, the mechanisms discovered in the simulations are merely hypotheses as far as real organisms are concerned, and these hypotheses must be tested by wet evolutionary experiments. After having positioned in silico experimental evolution with respect to related approaches, this chapter describes the main families of formalisms used in the field. Finally, to illustrate more concretely the methods currently used, an example is given based on the aevol model, the way it is used and the scientific results it allowed for.
Complete list of metadatas

https://hal.inria.fr/hal-00850175
Contributor : Carole Knibbe <>
Submitted on : Monday, August 5, 2013 - 11:51:21 AM
Last modification on : Monday, February 10, 2020 - 4:36:52 PM

Identifiers

  • HAL Id : hal-00850175, version 1

Citation

Carole Knibbe. L'évolution expérimentale in silico. Franck Varenne and Marc Silberstein. Modéliser et simuler - Epistémologies et pratiques de la modélisation et de la simulation - Tome 1, Editions Matériologiques, pp.581-610, 2013, 978-2-919694-19-8. ⟨hal-00850175⟩

Share

Metrics

Record views

651