The influence of the bottleneck protocol on the adaptation rate and predictability

Abstract

Bottlenecks are evolutionary events that reduce the population size. In experimental evolution, the cultivation of microorganisms in laboratory conditions follows a serial passaging protocol in which periodic bottlenecks are an inherent aspect. In this work, we study a computational model of microbial evolution to understand the influence of the bottleneck protocol on the rate of adaptation and predictability of the population. To address these questions, evolutionary simulations are run using a standard Wright-Fisher model integrated with a bottleneck regime, and the implementation of a fitness landscape with varying ruggedness. The rate of adaptation is analyzed as a function of bottleneck and population sizes at different time units. It is found that the rate of adaptation depends monotonically on bottleneck size when time is expressed per generation, but is maximal at intermediate bottleneck size for times expressed per bottleneck and per birth. Additionally, fitness landscapes allow to register the change in the population’s genetic composition as trajectories over the genotypic space. Under this path-dependent perspective, an ensemble of trajectories is generated in many independent runs, for which statistical and computational measurements of predictability are inferred. Irrespective of the timing of population bottlenecks, we find that predictability increases with population size. We also find that predictability of the adaptive pathways increases in increasingly rugged fitness landscapes.