Mapping history dependence along evolutionary trajectories

Evolutionary trajectories often follow complex paths through genotype–phenotype landscapes. When initiated from identical starting conditions, replicate populations fix different mutations, but it remains unclear whether such trajectories are functionally interchangeable. In particular, we do not know whether points at similar fitness along different evolutionary paths correspond to equivalent underlying phenotypes and positions in genotype–phenotype space, or instead reflect distinct evolutionary solutions. Here, we address this question using high-throughput CRISPR-based methods to introduce large numbers of targeted genomic perturbations. We evolved many parallel Saccharomyces cerevisiae populations under identical conditions and systematically reintroduced mutations that naturally arose during these trajectories both into their original line and into other replicate lines. This design allows us to measure the effects of real, historically accumulated mutations in diverse genetic backgrounds and at multiple evolutionary stages. By quantifying how the consequences of the same mutation depend on where and when it occurs, we reveal the structure of functional divergence among trajectories and track how local genotype–phenotype landscapes change over the course of evolution.