Skip to content

DeeDeeExperiment

An S4 class for managing and exploring omics analysis results — extends SingleCellExperiment with dedicated slots for storing differential-expression and functional-enrichment outputs alongside the data they came from, so a downstream user can ship one object instead of a folder of CSVs.

Talk at EuroBioC2025

  • Speaker: Najla Abassi (IMBEI Mainz)
  • Day / session: Day 1 (Wed Sep 17, 2025) — afternoon short-talk session, after the Susan Holmes keynote
  • Talk title: "DeeDeeExperiment class for omics analysis results"
  • Slides: Zenodo 10.5281/zenodo.17184376
  • Video: not recorded — short talks were not part of the plenary YouTube playlist
  • Abstract / schedule: eurobioc2025 schedule

What it does

DeeDeeExperiment is a thin extension on top of SingleCellExperiment that adds dedicated, schema-validated slots for differential-expression results (e.g. DESeq2, edgeR, limma output tables) and functional-enrichment results (e.g. GSEA, ORA). Instead of a downstream user keeping the SCE alongside loose result data.frames, everything travels together in one object — which makes apps like iSEE, gating pipelines, and reproducibility wrappers trivially aware of "what tests were run on this data."

How it works

Core idea. DeeDeeExperiment is an S4 class that inherits from SingleCellExperiment, adding named slots for DEA (differential expression) and FEA (functional enrichment) result tables — so the result objects ride inside the same container as the counts, colData, rowData, and reduced dimensions they were derived from.

Inputs / outputs. Construction takes an existing SingleCellExperiment plus a named list of result tables; the returned DeeDeeExperiment preserves every SCE component and exposes the results through dedicated accessors. Downstream tools that already speak SCE (iSEE, scran, scater) keep working unchanged.

Key innovation. A schema-validated, single-object contract for DEA/FEA results. Prior practice scattered result data.frames as loose files alongside the SCE; DeeDeeExperiment standardizes the contract so consumers can introspect what tests were run without bespoke parsing.

Parameters / API surface worth knowing. - DeeDeeExperiment(sce = ..., de_results = ...) — constructor; accepts an SCE plus a named list of DEA tables. - getDEA(), getDEANames() — accessors that retrieve named DEA result tables (analogous accessors exist for FEA). - Expected single-cell colData columns when working with pseudobulk: sample_id, cluster_id, group_id. - Naming convention for results: "contrast_celltype" (e.g. "stim-ctrl_NK cells"), so per-cluster contrasts are addressable by name.

Canonical example. The single-cell vignette walks SCE → QC with scater → prepSCE() → pseudobulk via muscat's aggregateData() and pbDS() → conversion with muscat_list_for_dde() → dde <- DeeDeeExperiment(sce = sce, de_results = muscat_list). The resulting object is then queried with getDEA() / getDEANames() to inspect contrasts by cell type.

Where it fits in the corpus

  • AACR 2026: no current dossier — this is plumbing rather than science; relevant if any AACR talk uses iSEE / SCE infrastructure to ship results
  • iSEE (entry) — natural consumer; DDE's DEA/FEA slots are exactly the result shape that iSEE custom panels could surface
  • ISMB 2026: scaffold
  • GBCC 2025: queued — IMBEI Mainz is an active workshop contributor at both events
  • Nextflow Summit 2026: conceptually adjacent (single-object shipping is also a Nextflow design pattern); no direct mention in the Nextflow vault

Notes

The kind of package that's invisible until you don't have it. Bioconductor has long had SingleCellExperiment as the canonical container, but result-level schema (DEA / FEA outputs) has been ad-hoc — every downstream tool reinvents its own contract. DeeDeeExperiment is a bet that standardizing that schema unlocks better composability — Federico Marini's iSEE workshop on Day 2 is a natural consumer.