EVOLUTIONARY DYNAMICS AND THE CODING STRUCTURE OF SEQUENCES - MULTIPLE CODING AS A CONSEQUENCE OF CROSSOVER AND HIGH MUTATION-RATES

In this paper we explore the influence of the dynamics of evolution on coding structures of sequences. We show that, in systems with crossover, high mutation rates cause the most conserved subsequences to be preferentially used as recognition sites for newly evolving sequences. In other words: "multiple coding" evolves in these systems. Multiple coding often does not increase the fitness of the population; nevertheless it is selected. By contrast, in systems without crossover, a low mutation rate causes multiple coding to be avoided, so that only single coding evolves. Again this "choice" is not reflected in the fitness of the population, but is dictated by the evolutionary dynamics. We conclude that the genetic operator crossover turns evolutionary processes in pattern detectors rather than optimizers.