Meta-Analysis Suggests That Intron Retention Can Affect Quantification of Transposable Elements from RNA-Seq Data

Abstract

Transposable elements (TEs), also known as “jumping genes”, are repetitive sequences with the capability of changing their location within the genome. They are key players in many different biological processes in health and disease. Therefore, a reliable quantification of their expression as transcriptional units is crucial to distinguish between their independent expression and the transcription of their sequences as part of canonical transcripts. TEs quantification faces difficulties of different types, the most important one being low reads mappability due to their repetitive nature preventing an unambiguous mapping of reads originating from their sequences. A large fraction of TEs fragments localizes within introns, which led to the hypothesis that intron retention (IR) can be an additional source of bias, potentially affecting accurate TEs quantification. IR occurs when introns, normally removed from the mature transcript by the splicing machinery, are maintained in mature transcripts. IR is a widespread mechanism affecting many different genes with cell type-specific patterns. We hypothesized that, in an RNA-seq experiment, reads derived from retained introns can introduce a bias in the detection of overlapping, independent TEs RNA expression. In this study we performed meta-analysis using public RNA-seq data from lymphoblastoid cell lines and show that IR can impact TEs quantification using established tools with default parameters. Reads mapped on intronic TEs were indeed associated to the expression of TEs and influence their correct quantification as independent transcriptional units. We confirmed these results using additional independent datasets, demonstrating that this bias does not appear in samples where IR is not present and that differential TEs expression does not impact on IR quantification. We concluded that IR causes the over-quantification of intronic TEs and differential IR might be confused with differential TEs expression. Our results should be taken into account for a correct quantification of TEs expression from RNA-seq data, especially in samples in which IR is abundant.

Transposons Acting as Competitive Endogenous RNAs: In-Silico Evidence from Datasets Characterised by L1 Overexpression

“…As Jonsson and colleagues demonstrated, the abrogated activity of the most important maintenance DNA methyltransferase [ 56 ] erases a fundamental epigenetic repressive layer from the L1 defense mechanisms. Aiming to confirm whether L1 elements were autonomously transcribed [ 57 , 58 ] in this model, we used SQuIRE [ 38 ] to quantify the locus-specific expression of TEs. Our differential expression analysis highlighted that the KO of DNMT1 leads to a strong upregulation of TEs: 3015 TEs result upregulated and 277 downregulated.…”

Generalized nuclear localization of retroelement transcripts

“…Recently, it was reported that incidental signal from L1 sequences nested within retained introns of protein coding genes could interfere with accurate quantification of L1 element expression patterns [56]. We therefore considered the possibility that signals from retroelements nested within retained introns could contribute to the pattern of nuclear enrichment that we observe.…”

A concerted increase in readthrough and intron retention drives transposon expression during aging and senescence

“…Intron retention is one prominent example of dysfunctional splicing during aging, but the association with age-related transposon expression has not been evaluated. A recent study by Gualandi et al, 2022 using RNA-seq data from the 1000 Genomes project found a strong per sample correlation between intron and transposon expression in the general population. In line with these results, we showed that intron retention may explain a large fraction of transposon reads in aging and senescence datasets ( Figure 1 ).…”