Multidimensional epistasis and the advantage of sex

Kondrashov and Kondrashov (2001) point out that, although common in population genetic models, epistatic systems where the fitness of a genotype is a non-linear function of the number of mutations it carries, which they term “unidimensional epistasis,” are a limited subset of possible epistatic landscapes. They claim that more general “multidimensional epistasis” usually confers a disadvantage for sex. However, the evolutionary computation (EC) literature contains models that lie within the space of multidimensional epistasis yet demonstrate an advantage to sex. Here we provide modifications of the Kondrashovs’ model that connect with EC results and help to explore the space of epistatic models between the two disciplines.

• Type: Conference
• Year: 2005
• Where: CEC
• Authors: Richard A. Watson, John Wakeley
• Keywords: Artificial Intelligence, CEC 2005, Epistatic, Multidimensional Epistasis, Possible Epistatic Landscapes