Repeatability of gene expression evolution in experimental environmental adaptation

Assessing the relative roles of chance and necessity in evolution is of wide interest, but it requires evolving the same population under the same environment multiple times—a virtually impossible task in nature that has been repeatedly accomplished in the laboratory. Capitalizing on the transcriptome data collected in 10 laboratory evolution studies conducted in 22 distinct environments, we investigate the evolutionary repeatability of a total of 182,103 gene expression traits in a prokaryotic and five eukaryotic species. The number of gene expression traits exhibiting concordant changes in direction and magnitude between two replicate populations typically exceeds the chance expectation by 10 to 100 standard deviations. Contrasting replicate evolution in the same environment with that in different environments suggests that the concordance in expression evolution is largely attributable to environment-specific selections, a finding that is further supported by comparing with the outcome of mutation accumulation experiments where the efficacy of selection is minimized. Additionally, genes controlled by more transcription factors tend to show more repeatable expression evolution, likely due to a higher certainty in the occurrence of mutations altering their expressions. In conclusion, contrary to almost unrepeatable genotypic evolution, phenotypic evolution during environmental adaptation is quite repeatable and deterministic, at least for gene expressions. Experimental evolution enables evaluation of the relative roles of chance and necessity in evolution. This study compiles transcriptomic data from experimental evolution of a prokaryotic and five eukaryotic species in 22 environments to reveal that gene expression evolution is often repeatable and deterministic.