The Time Scale of Evolution

Abstract

Fitness is a measure of how quickly alleles change in frequency under natural selection. Time is always implicit in evolutionary models but its units are rarely made explicit. When measuring phenotypes such as absolute growth rate, the units of measurement need to be made explicit. By contrasting measures of fitness and growth rate, we uncovered a curious effect, by which evolutionary time runs at different speeds depending on how restricted population growth is. In other words, when the generation time of a population is externally imposed, relative fitness per generation is no longer an accurate measure of differences between genotypes. We explore this effect and describe how it affects selective sweeps, probability of fixation of beneficial mutations and adaptation dynamics. Moreover, we show that different populations cannot be compared unless they share a common reference and that our inference of epistasis can be biased by this temporal effect. Finally, we suggest less biased ways to measure selection in experimental evolution.