Regensburg 2022 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
BP: Fachverband Biologische Physik
BP 23: Evolution
BP 23.1: Talk
Thursday, September 8, 2022, 10:00–10:15, H13
New phenotypes appear in an evolving population in non-Poissonian bursts — •Nora S. Martin1, Steffen Schaper1, Chico Q. Camargo1, 2, and Ard A. Louis1 — 1Department of Physics, University of Oxford, Oxford, UK — 2Department of Computer Science, University of Exeter, Exeter, UK
For adaptive evolution, a central question is when and how frequently random mutations produce the specific rare adaptive phenotypes that have a selective advantage. A widely studied scenario is the following: a population starts with a given initial phenotype and accumulates neutral mutations, and new phenotypes are also introduced at a certain rate. Many theories implicitly assume that new phenotypes appear through simple stochastic processes which lead to Poissonian statistics. In this contribution, we use simulations on the biophysically motivated computational genotype-phenotype map from RNA sequences to secondary structures and show that new structures appear in highly non-Poissonian ``bursts''. In other words, if a new structure appears once, it is highly likely to appear multiple times in a relatively small number of generations. We show that there are several sources for this non-Poissonian behaviour, for example correlations in the mappings from genotypes to phenotypes, which may be a generic property of realistic genotype to phenotype maps. We find that these bursts can affect probabilities of fixation, especially when there are multiple competing adaptive phenotypes.