Population structure amplifies mutation load

The interplay of mutation and natural selection is key to evolution: Mutations introduce genetic variation, whereas selection favours higher-fitness individuals. The balance between the two is quantified by the mutation load—the reduction in mean population fitness due to deleterious mutations. In large well-mixed populations, selection dominates, and the mutation load is proportional to the mutation probability per reproduction event. However, the role of spatial structure in shaping mutation–selection dynamics is poorly understood. Here, we use evolutionary graph theory to show that heterogeneous spatial structures generically amplify the effect of mutation relative to selection. In contrast to well-mixed populations, the mutation load in heterogeneous structures can remain substantial even when the mutation rate is very small, signaling a breakdown of selection. For example, in star-shaped populations the mutation rate is effectively amplified by a factor proportional to the population size. When mutation is coupled to reproduction, this leads to the proliferation of less-fit types over fitter ones, a scenario of “survival of the weakest”. The reduction of natural selection may help explain, for example, tumour heterogeneity in spatially structured tissues. It also suggests that heterogeneous social networks can promote the persistence of unpopular opinions, thereby fostering diversity. By showing how spatial structure modulates the effect of selection, our work points to potentially new ways of steering evolution toward desired outcomes.