Skip to content

Tidyomics

Tidy-syntax ecosystem for Bioconductor data classes — a meta-package umbrella (~10 member packages) that lets analysts use dplyr / tidyverse verbs directly on SummarizedExperiment, SingleCellExperiment, SpatialExperiment, and Seurat objects without converting away from the canonical S4 representation. The GBCC2025 talk pitches Tidyomics as the substrate for democratizing Human Cell Atlas–scale single-cell analysis.

  • Lead author: Stefano Mangiola (Adelaide University) — mangiolastefano@gmail.com
  • Co-authors: Michael Love (UNC), William Hutchison, plus broader tidyomics community
  • Ecosystem hub: tidyomics on Bioconductor (umbrella, v1.8.0) plus member packages: tidySingleCellExperiment, tidybulk, tidySummarizedExperiment, tidyseurat, tidySpatialExperiment, plyranges (each on Bioconductor)
  • Source: github.com/tidyomics/tidyomics
  • License: MIT
  • Status at GBCC2025: ecosystem talk spanning multiple constituent packages

Talk at GBCC2025

  • Speaker: Stefano Mangiola (Adelaide University)
  • Day / session: Day 3 — Oral Session 4 (chair: Mike Love), talk #6; plus a Day 3 lightning-talk teaser
  • Talk title: "Democratising the Analysis of the Human Cell Atlas with Bioconductor" (Oral 4 #6)
  • Slides: to verify after publish
  • Video: GBCC2025 playlist
  • Abstract / schedule: GBCC2025 program

What it does

Tidyomics overloads dplyr/tidyverse verbs (filter, mutate, group_by, summarise, nest, pivot_*, etc.) so they operate directly on Bioconductor S4 containers without first coercing to a data frame. tidySingleCellExperiment makes a SingleCellExperiment look like a tibble (one row per cell, columns from colData + reduced dims), tidybulk wraps bulk RNA-seq, tidySummarizedExperiment handles the general SE case, tidyseurat covers Seurat objects, and plyranges provides tidy verbs over GRanges. All members share API conventions so a user who learns one transfers fluently to the others. The Mangiola GBCC talk frames this as the ergonomics layer that lets cell-atlas-scale analyses be written and read as readable pipelines.

How it works

Core idea. tidyomics is an umbrella meta-package: loading it attaches the member packages in one call and guarantees they expose consistent dplyr/tidyverse method dispatch on top of Bioconductor's S4 containers. The S4 object stays the ground truth — tidy verbs are implemented as methods that intercept dispatch and route to the appropriate slot, so no coercion / flatten step is needed.

Inputs / outputs. Inputs are unchanged Bioconductor objects (SummarizedExperiment, SingleCellExperiment, SpatialExperiment, Seurat, GRanges). Outputs are the same objects after verb application — the tibble-like printing is a view, not a representation change.

Key innovation. Ecosystem-level coordination. Each member package independently provides a tibble abstraction for its respective S4 class, but they share API conventions and a single install/load entry point, so a user who learns tidySingleCellExperiment transfers fluently to tidySpatialExperiment or tidybulk. The 2024 Nature Methods paper benchmarks this by running a 7.5M-PBMC Human Cell Atlas analysis across six data frameworks and ten tools using shared tidy syntax (Hutchison, Keyes, Crowell et al., Nat Methods 21:1166–1170, 2024).

Parameters / API surface worth knowing. - library(tidyomics) — attaches tidySummarizedExperiment, tidySingleCellExperiment, tidySpatialExperiment, tidyseurat, plyranges (and others) in one call. - Verb coverage: filter, mutate, select, group_by, summarise, nest, pivot_*, plus the package-specific extensions (e.g. tidybulk's differential-expression wrappers). - Philosophy — "tidy data principles on top of S4 objects": the container is canonical, the syntax is tidyverse.

Canonical example. Not specified in vignette (the umbrella package's own page mainly orchestrates loading); member-package vignettes carry the worked examples — e.g. tidySingleCellExperiment lets a user write sce |> filter(cell_type == "T cell") |> ggplot(aes(UMAP_1, UMAP_2)) + geom_point() directly on an SCE.

Where it fits in the corpus

  • AACR 2026: axes = single-cell-spatial-omics + bioinfo-tools (any AACR talk that ships an SCE / SE result is a candidate downstream user)
  • plyxp (entry) — both bridge Bioc S4 containers to dplyr-style syntax; plyxp is the newer, lazier SummarizedExperiment companion. Tidyomics is the umbrella; plyxp is a specific evaluation strategy in the same lineage
  • EuroBioC 2025: not a dedicated talk, but tidyomics packages are referenced throughout the schedule
  • Nextflow Summit 2026: not mentioned (R-side ecosystem)

Notes

Tidyomics is a meta-ecosystem covered by a single vault page; constituent packages get dedicated pages only if they carried their own talks. The GBCC framing — "Democratising the Analysis of the Human Cell Atlas" — places this work alongside the cross-platform / cross-language integration storylines that anchor the joint Galaxy-Bioconductor event.