Identification of transposable elements using multiple alignments of related genomes

Abstract

Accurate genome-wide cataloging of transposable elements (TEs) will facilitate our understanding of mobile DNA evolution, expose the genomic effects of TEs on the host genome, and improve the quality of assembled genomes. Using the availability of several nearly complete Drosophila genomes and developments in whole genome alignment methods, we introduce a large-scale comparative method for identifying repetitive mobile DNA regions. These regions are highly enriched for transposable elements. Our method has two main features distinguishing it from other repeat-finding methods. First, rather than relying on sequence similarity to determine the location of repeats, the genomic artifacts of the transposition mechanism itself are systematically tracked in the context of multiple alignments. Second, we can derive bounds on the age of each repeat instance based on the phylogenetic species tree. We report results obtained using both complete and draft sequences of four closely related Drosophila genomes and validate our results with manually curated TE annotations in the Drosophila melanogaster euchromatin. We show the utility of our findings in exploring both transposable elements and their host genomes: In the study of TEs, we offer predictions for novel families, annotate new insertions of known families, and show data that support the hypothesis that all known TE families in D. melanogaster were recently active; in the study of the host, we show how our findings can be used to determine shifts in the eu-heterochromatin junction in the pericentric chromosome regions.