ChimeraTE: A pipeline to detect chimeric transcripts derived from genes and transposable elements

Abstract

ABSTRACTTransposable elements (TEs) are structural variants considered an important source of genetic diversity, which may arise in the transcriptome when TEs are transcribed in the same RNA molecule as genes, producing what we hereafter call chimeric transcripts. The presence of chimeric transcripts has been associated with adaptive traits in several species, but their identification remains hindered due to the lack of tools to detect them on a transcriptome-wide scale. Previous bioinformatics tools were developed to identify chimeric transcripts derived from TEs present in a reference genome. Nevertheless, different individuals/cells/strains might harbor different TE insertions generating such chimeric transcripts. Therefore, we have developed ChimeraTE, a pipeline that uses paired-end RNA-seq reads to identify chimeric transcripts with or without a reference genome, in a transcriptome-wide manner. ChimeraTE has two Modes: Mode 1 is a genome-guided approach that employs the canonical method of genome alignment, whereas Mode 2 identifies chimeric transcripts without a reference genome, being able to predict chimeras derived from fixed or polymorphic TEs. We have used both Modes with Illumina RNA-seq reads from ovarian tissues of Drosophila melanogaster wild-type strains, and found that ∼3% of all genes generate chimeric transcripts. Approximately ∼9% of all detected chimeras were absent from the D. melanogaster’s reference genome, corresponding to polymorphic insertions in the wild-type strains. ChimeraTE is the first pipeline with the ability to automatically uncover chimeric transcripts without a reference genome.