IRescue: uncertainty-aware quantification of transposable elements expression at single cell level

Abstract

Transposable elements (TEs) are interspersed and repetitive DNA sequences that contribute to the evolution of species by introducing novel regulatory sequences in the eukaryotic genome, and their transcription have been associated to both physiological functions and detrimental effects in human cells. Given their ubiquity and redundancy, the transcriptome analysis of TEs is notoriously challenging and requires specialized tools. While a plethora of software is available for bulk RNA-Seq, few attempts have been made for measuring the expression of TEs in scRNA-seq data so far. Here, we present IRescue (Interspersed Repeats single-cell quantifier), a software to quantify TE expression in scRNA-seq using a UMI deduplication algorithm to solve the allocation of reads ambiguously mapped on interspersed TEs. Using simulated data, we demonstrate that IRescue is precise at estimating the expression of TE subfamilies. Furthermore, by analysing normal and cancer cells from colorectal cancer (CRC) patients, we identify the TE expression signature of CRC cells, where we found young LINE1 elements to be specifically enriched in subpopulations of cancer cells, compared to normal cells. Finally, we show the heterogeneity and expression dynamics of CRC TE markers previously characterized at bulk-level only, and that the expression of such markers is the result of the transcription of tumor-specific alternative isoforms of human oncogenes. With IRescue, we long to enable adding the layer of TE expression in the multimodal analysis of single cell sequencing and to facilitate the discovery of novel markers from the high and increasing amount of scRNA-seq experiments data.