Long-read cDNA Sequencing Enables a ‘Gene-Like’ Transcript Annotation of Arabidopsis Transposable Elements

Abstract

High-quality transcript-based annotations of genes facilitates both genome-wide analyses and detailed single locus research. In contrast, transposable element (TE) annotations are rudimentary, consisting of only information on location and type of TE.When analyzing TEs, their repetitiveness and limited annotation prevents the ability to distinguish between potentially functional expressed elements and degraded copies. To improve genome-wide TE bioinformatics, we performed long-read Oxford Nanopore sequencing of cDNAs from Arabidopsis lines deficient in multiple layers of TE repression. We used these uniquely-mapping transcripts to identify the set of TEs able to generate mRNAs, and created a new transcript-based annotation of TEs that we have layered upon the existing high-quality community standard TAIR10 annotation.The improved annotation enables us to test specific standing hypotheses in the TE field.We demonstrate that inefficient TE splicing does not trigger small RNA production, and the cell more strongly targets DNA methylation to TEs that have the potential to make mRNAs. This work provides a transcript-based TE annotation for Arabidopsis, and serves as a blueprint to reduce the genomic complexity associated with repetitive TEs in any organism.

Comprehensive analysis of both long and short read transcriptomes of a clonal and a seed-propagated model plant species reveal the prerequisites for transcriptional activation of autonomous and non-autonomous transposons in plants

“…This nding indicates that the so-called 'TE activation' commonly seen from short-read RNAseq may not necessarily represent transcriptional activation leading to autonomous TE mobilisation and suggests a rare possibility of a TE burst occurred in wild-type populations in a natural environment [15]. A recent study in A. thaliana pooled six ONT cDNA sequencing libraries, which included A. thaliana wild-type and four epigenetically compromised genotypes before applying an expression threshold at ve ONT reads to annotate an active TE locus [55]; this generated more than 5 million reads for the model organism with a genome size of ~ 135 Mb. With a similar N50, it would require alignment depth at least 2.3 times deeper than the current Vv_mock12h or Vv_Yeast12h libraries that we have interrogated.…”

“…With a similar N50, it would require alignment depth at least 2.3 times deeper than the current Vv_mock12h or Vv_Yeast12h libraries that we have interrogated. The A. thaliana genotypes used in Panda and Slotkin [55] include those which were compromised in epigenetic silencing that allows the constant de-repression of TE loci, whereas the grapevine embryogenic callus in our study was initiated from wild-type P. noir clone UCD5 and was subjected to stress treatment. In such a fully active epigenetic silencing environment, the likelihood of high levels of transcription of autonomous elements is low.…”

“…Since our V. vinifera system has shown transcriptional activation of TEs under stressed conditions and demonstrated the characteristics of individual TE loci in the pool of expression candidates, we then aimed at identifying TE loci that potentially contribute full-length TE transcripts necessary for autonomous mobilisation. Due to the ability of ONT sequencing technology to sequence full-length transcripts, ONT cDNA libraries have the potential to reveal, if any, competent transcription of autonomous TE and decipher the origins of these transcripts [55]. The sequence integrity and structure of TE loci were rstly screened for structurally autonomous TE loci annotated in the genome.…”