Somrit: The Somatic Retrotransposon Insertion Toolkit

Mobile elements, such as retrotransposons, have the ability to express and re-insert themselves into the genome, with over half the human genome being made up of mobile element sequence. Somatic mobile element insertions (MEIs) have been shown to cause disease, including some cancers. Accurate identification of where novel retrotransposon insertion events occur in the genome is crucial to understand the functional consequence of an insertion event. In this paper we describe somrit, a modular toolkit for detecting somatic MEIs from long reads aligned to a reference genome. We identify the initial read-to-reference mapping step as a potential source of error when the insertion is similar to a nearby repeat in the reference genome and develop a consensus-realignment procedure to resolve this. We show how somrit improves the sensitivity of detection for rare somatic retrotransposon insertion events compared to existing tools, and how the local realignment procedure can reduce false positive translocation calls caused by mis-mapped reads bearing MEIs. Somrit is openly available at: https://github.com/adcosta17/somrit