Fig. 3

Expected and observed correlation between the TE abundance in piRNA clusters and the rest of the genome. A Expected correlation based on simulations under the trap model (neutral insertions and \(u=0.1\)). B Observed correlation based on short reads aligned to the reference genome and the reference annotations of piRNA clusters. C Observed correlation based on strain-specific assemblies and a lift-over of the reference annotations of piRNA clusters using unique sequences flanking the clusters. D Observed correlation based on strain-specific assemblies and de novo annotations of piRNA clusters. All counts refer to copy numbers per haploid genome. For the observed data we averaged the counts over the five strains. Solely TE families active in the germline were considered. Kendall rank correlation coefficients are reported