Efficient detection of transposable element insertion polymorphisms between genomes using short-read sequencing data

Transposable elements (TEs) are powerful generators of major-effect mutations, most of which are deleterious and maintained at very low frequencies within natural populations. As a result, the reference genome for a given species only captures a minor fraction of such variants. With the advent of whole genome sequencing using short reads, it became possible to detect TE insertion polymorphisms (TIPs) between reference and non-reference genomes. Here, we present an improved version of the SPLITREADER pipeline we developed to detect TE insertions that are only present in non-reference genomes. We benchmark our method on twenty non-reference Arabidopsis thaliana genomes and demonstrate its high specificity and sensitivity. As short-read data still vastly outnumber long-read data, SPLITREADER can readily provide a first and accurate assessment of transposition activity in natural or experimental populations.