[WP1] Here, we studied the relationship between plasticity and evolvability. On the conceptual side, we presented our views in three ‘opinion’ papers that provide arguments in favour of a mechanistic view of phenotypic plasticity, epigenetic inheritance and evolvability. The ideas developed were later worked out in various modelling studies, which all arrived at the conclusion that the (epi)genetic system can evolve in such a way that plasticity and evolvability are enhanced simultaneously. Here, we only mention our studies on the evolution of the mutation process. It is a kind of dogma in molecular genetics that mutations are random and not related to the ‘evolutionary needs’ of an organism. By means of a surprisingly simple gene-regulatory network model, we could show that this view needs to be modified. Even if the mutations are random at the genetic level, the regulatory network systematically evolves in such a way that the phenotypic effects of these mutations are not random but biased towards higher-fitness outcomes, enhancing the evolvability of the population. Other models on the evolution of the mutation rate also arrived at the conclusion that evolvability is not constant but evolvable. These theoretical predictions were partly tested in the lab. We showed, for example, that the mutation rate towards antibiotic resistance is strongly temperature dependent, a finding of potential medical importance.
[WP2] Learning is one of the most important mechanisms to respond adaptively to novel conditions. . We made a major methodological advance in developing a new framework for the evolution of individual and social learning, which considers the evolution of neural networks that are capable of learning. Although learning is based on a simple (but biologically realistic) mechanism, effective learning rapidly evolved. The extension to social learning is of special importance for the field of cultural evolution, where progress is hampered by the lack of convincing approaches to the spread of cultural information. Our framework can be the basis for a new generation of models for cultural evolution.
[WP3] All our studies show that in mechanistic models evolution is much faster than previously thought, making the timescale of evolution similar to the timescale of ecological processes. This leads to an intriguing interplay of ecology and evolution with very different properties than previously thought. In various studies, we investigated the evolution of animal movement in response to resources, competitors, and predators. Our neural network model has the advantage that its outcome can be compared with movement patterns in the field. Surprisingly, even the simplest networks allow the rapid evolution of highly efficient (but previously unanticipated) movement patterns. Accordingly, environmental change can trigger a rapid change in movement strategies. For example, social distancing evolved within a few generations after the introduction of a novel pathogen. In addition, virtually all our mechanistic models lead to the emergence of systematic variation between individuals. In various studies (e.g. on the evolution of parental care strategies), we worked out in detail how and why such polymorphism changes the eco-evolutionary dynamics.
