exploreDE

Install and run

install.packages("BiocManager")
BiocManager::install("fgcz/exploreDE")   # pulls in the Bioconductor dependencies

library(exploreDE)
exploreDE()                   # landing page: upload a file or load the demo
exploreDE("my_data.h5ad")     # or open straight into your own file

With no file, the app opens on its landing page, where you can upload an exploreDE .h5ad (see below) or load the demo dataset that ships with the package. At FGCZ the app is also reachable from SUSHI with ?data=<result directory>.

Bring your own data

exploreDE reads one file format: an .h5ad (AnnData) file built by build_explore_h5ad(), for "rnaseq", "proteomics" or "crispr" data from any pipeline. Only counts, sample metadata and the omics type are required; a DE table and pathway results are optional, and the app shows the tabs your data supports.

build_explore_h5ad(
  counts,                        # features x samples matrix
  out     = "my_data.h5ad",
  omics   = "rnaseq",
  coldata = coldata,             # one row per sample, rownames = colnames(counts)
  rowdata = annotation,          # optional: gene IDs, symbols, GO terms, ...
  de      = list(Treat_vs_Ctrl = de_table)   # optional: one table per contrast
)

You don't need to rename columns: a built-in dictionary recognises the usual DESeq2, edgeR, limma, prolfqua and SAINTexpress column names, and warns when it can't place one.

• Full guide: vignette("bring-your-own-data", package = "exploreDE") covers counts-only, proteomics and RNA-seq + pathway examples (all runnable on the data shipped with the package), checking a file, and what to do when a column isn't recognised.
• Using an AI assistant: give it system.file("skills/exploreDE-byod/SKILL.md", package = "exploreDE"). It's written so an assistant can write the import script for your data.
• Argument reference: ?build_explore_h5ad.

Citing exploreDE

See CITATION.cff.