Fig. 4

Evolution of mutational landscape towards beneficial mutations. We put individuals in the LOD that lived just before a change of environment (see Methods ancestor tracing) in the next environment and let them undergo a single round of mutation. In this alternative environment we measure their fitness before and after the random mutation and define the fitness effect of the mutation as the gain relative to the maximum possible gain (absolute gain/(1−fitness before mutation)). Different colour graphs are for different periods of the evolutionary simulations. For environment 1 (a) and environment 2 (b) measures for all populations and all replicate runs are averaged