Sequence composition and conservation predict the phenotypic relevance of transposable elements

Abstract

Transposable elements (TEs) are powerful drivers of genome evolution, in part through their ability to rapidly rewire the host regulatory network by creating transcription factor binding sites that can potentially be turned to the hosts advantage in a process called exaptation. However, the effects on the host phenotype vary widely among different TEs. Here, we classify TEs based on their contribution to the human host phenotypes at both molecular and macroscopic scales. TE contributions to chromatin accessibility, gene expression, and the heritability of complex traits are strongly correlated to each other, confirming that the main mechanism through which TEs affect the host phenotype is through the rewiring of the regulatory network. TE sequence and evolutionary features are able to explain a large fraction of the variance in phenotypic relevance, and in particular more than 50% of the variance of their contribution to the heritability of complex traits. A conspicuous exception to this pattern is represented by TEs of the ERV1 family, whose phenotypic impact cannot be explained by our model: In particular, this family includes a set of relatively young TEs whose phenotypic relevance is much larger than would be expected based on their sequence and evolutionary parameters. These TEs are involved in fast-evolving biological processes related to the interaction of the organism with its environment. In conclusion, our results confirm quantitatively that TE insertions affect the host phenotype mostly through the rewiring of its regulatory network; identify a signature of phenotypic relevance based on sequence and conservation properties; and highlight several TEs as promising candidates for functional studies.