Evolutionary dynamics of piRNA clusters in Drosophila

Small RNAs produced from transposable element (TE)-rich sections of the genome, termed piRNA clusters, are a crucial component in the genomic defence against selfish DNA. In animals, it is thought the invasion of a TE is stopped when a copy of the TE inserts into a piRNA cluster, triggering the production of cognate small RNAs that silence the TE. Despite this importance for TE control, little is known about the evolutionary dynamics of piRNA clusters, mostly because these repeat-rich regions are difficult to assemble and compare. Here, we establish a framework for studying the evolution of piRNA clusters quantitatively. Previously introduced quality metrics and a newly developed software for multiple alignments of repeat annotations (Manna) allow us to estimate the level of polymorphism segregating in piRNA clusters and the divergence among homologous piRNA clusters. By studying 20 conserved piRNA clusters in multiple assemblies of four Drosophila species, we show that piRNA clusters are evolving rapidly. While 70%-80% of the clusters are conserved within species, the clusters share almost no similarity between species as closely related as D. melanogaster and D. simulans. Furthermore, abundant insertions and deletions are segregating within the Drosophila species. We show that the evolution of clusters is mainly driven by large insertions of recently active TEs and smaller deletions mostly in older TEs. The effect of these forces is so rapid that homologous clusters often do not contain insertions from the same TE families.