Gene expression as phenotype - Many small-step changes leading to little long-term phenotypic evolution

Abstract

Unlike in genotypic evolution, there are few general rules governing phenotypic evolution with one of them being the small-step evolution. More specifically, natural selection tends to favor mutations of smaller phenotypic effects than of larger ones. This postulate can be viewed as a logical extension of Fishers Geometric Model (FGM). Testing this FGM postulate, however, is challenging as the test would require a large number of phenotypes, each with a clear genetic basis. For such a test, we treat the expression level of each gene as a phenotype. Furthermore, a mechanism of small-step expression evolution exists, namely via the control by microRNAs (miRNAs). Each miRNA in metazoans is known to weakly repress the expression of tens or hundreds of target genes. In our analysis of mammalian and Drosophila expression data, small step evolution via miRNA regulation happens frequently in long-term evolution. However, such small-step evolution does not lead to long-term phenotypic changes which would take too many such steps to accomplish. Furthermore, target site changes often cancel themselves out by continual gains and losses. The results suggest that the FGM postulate may be most appropriate for phenotypic fine-tuning near the expression optimum. In contrast, longterm expression evolution may occasionally take large steps (e.g., mutations in transcription factors) when big environmental shift happens. In another study (Lu et al. 2021), we further show how the small-step evolution of expression phenotypes is a manifestation of miRNAs role in developmental canalization. In conclusion, the rules of phenotypic evolution may depend crucially on the genetics of the phenotype, rather than its metric properties.