Skip to content

atena

Transposable-element expression quantification from RNA-seq — a Bioconductor package implementing multiple TE-quantification methods (ERVmap, TEtranscripts, Telescope) under a single API so analysts can choose, compare, or benchmark approaches for measuring transposable-element activity from bulk RNA-seq.

  • Maintainer: Robert Castelo (Universitat Pompeu Fabra) — robert.castelo@upf.edu
  • Authors: Beatriz Calvo-Serra (author), Robert Castelo (author/creator)
  • Bioconductor: atena v1.18.0
  • Source: git.bioconductor.org/packages/atena
  • Vignettes: Introduction
  • License: Artistic-2.0
  • Status at GBCC2025: lightning talk teaser

Talk at GBCC2025

  • Speaker: Robert Castelo (Universitat Pompeu Fabra)
  • Day / session: Day 3 — lightning talks
  • Talk title: "atena (transposable element analysis)" (lightning teaser)
  • Slides: to verify after publish
  • Video: GBCC2025 playlist
  • Abstract / schedule: GBCC2025 program

What it does

atena ingests RNA-seq alignments and quantifies expression of transposable elements (LINEs, SINEs, ERVs, etc.) under three published method backends — ERVmap, TEtranscripts, Telescope — exposed through one Bioconductor API. Method choice matters because TEs share sequence with ancestral copies and the assignment strategy (best-hit, EM, family-collapsed) drives the answer; atena lets users pick deliberately or compare estimates from the same reads.

How it works

Core idea. atena re-implements three published multi-mapping read-assignment strategies in R. ERVmap filters multi-mappers by hard/soft-clipping ratio (≥0.02), edit-distance ratio (≥0.02), and alignment-score gap (<5). TEtranscripts reassigns multi-mappers to TEs proportionally to estimated relative abundance via expectation–maximization. Telescope extends the TEtranscripts EM with per-TE reassignment parameters tuned by Bayesian MAP estimation.

Inputs / outputs. Input is RNA-seq BAM files plus a TE annotation in RepeatMasker format or GTF; the annotation can be parsed with built-in functions such as OneCodeToFindThemAll. Output is a SummarizedExperiment of TE-level counts with coordinates as GRanges/GRangesList, which slots directly into a downstream DESeq2 / edgeR workflow.

Key innovation. A single R/Bioconductor API across three previously separate command-line tools — analysts can swap backends on the same BAM set and compare estimates without leaving R, which makes "did this TE signal survive a method change?" a one-line check.

Parameters worth knowing. - method — selects the annotation parser / quantification backend (e.g. rmskidentity, rmskbasicparser, OneCodeToFindThemAll, rmskatenaparser). - genome — genome assembly identifier (e.g. "dm6"). - parsefun — the annotation parsing function. - BPPARAM — BiocParallel backend for parallelization.

Canonical example. The vignette loads atena and builds a TE annotation with teann <- annotaTEs(genome="dm6", parsefun=OneCodeToFindThemAll, strict=FALSE, insert=500, BPPARAM=SerialParam(progress=FALSE)), then runs quantification against a BAM input under the chosen backend to produce a TE-count SummarizedExperiment.

Where it fits in the corpus

  • AACR 2026: axis = bioinfo-tools (TE biology — ERVs, LINE-1, retrotransposon de-repression — is increasingly mainstream in cancer transcriptomics, immunotherapy response, and aging-cancer interfaces)
  • GSVA / Bioc-to-Galaxy (entry) — Castelo also leads the GSVA Galaxy-wrapper CoFest deliverable; atena is a candidate next package for the same auto-integration pipeline
  • EuroBioC 2025: not presented
  • Nextflow Summit 2026: not mentioned

Notes

A lightning teaser at GBCC; the substantive value is the unified API across three published TE-quantification methods. For cancer transcriptomics groups exploring TE / ERV signatures (immunogenicity, replication stress, aging), atena removes the "which tool / will my numbers reproduce" hurdle by carrying all three backends.