Skip to content

Selection Report — the founding twenty

Date: 2026-08-15 · Method: GitHub API data + snakemake catalog, scored by a selection committee agent against pre-registered criteria, then reviewed by the maintainer.

This report documents how the first twenty entries in the catalog were chosen. The catalog has since opened to original workflows and community submissions — see Curation & ratings.

Method

Data was collected from the live GitHub API on 2026-08-15:

  • nf-core: org:nf-core fork:false archived:false, sorted by stars (30 analysis pipelines after excluding infrastructure repos: modules 424★, tools 319★, test-datasets 164★, configs 110★, website 98★).
  • Snakemake: GitHub search snakemake workflow in:name,description stars:>30 merged with the official workflows.snakemake.org catalog listing.

Criteria (pre-registered weights)

Criterion Weight Definition
Usage & adoption 40% GitHub stars; snakemake-catalog presence
Maintenance health 20% pushed_at within ~18 months, not archived
Portability 25% default-parameters main path ≈ 5–45 rules, expressible in oxo-flow
Domain diversity 15% library-wide coverage across omics domains

Fidelity scope for every port: the default-parameters main execution path is replicated 100% (tools, versions, command logic); optional/alternate paths are documented as exclusions in each repo's README fidelity table.

Selection (20)

Port repo Source (engine) @ tag Domain Rules† Verify
oxo-flow-rnaseq nf-core/rnaseq @ 3.26.0 1349 bulk RNA-seq 14
oxo-flow-sarek nf-core/sarek @ 3.10.0 592 WGS/WES germline 12
oxo-flow-scrnaseq nf-core/scrnaseq @ 4.2.0 349 single-cell RNA-seq 8
oxo-flow-mag nf-core/mag @ 5.5.0 314 metagenomics 10
oxo-flow-ampliseq nf-core/ampliseq @ 2.18.0 256 amplicon (16S/ITS) 8
oxo-flow-chipseq nf-core/chipseq @ 2.1.0 242 ChIP-seq 10
oxo-flow-nanoseq nf-core/nanoseq @ 3.1.0 228 nanopore long-read 6
oxo-flow-atacseq nf-core/atacseq @ 2.1.2 226 ATAC-seq 9
oxo-flow-eager nf-core/eager @ 2.5.3 212 ancient DNA 12
oxo-flow-fetchngs nf-core/fetchngs @ 1.12.0 200 fetch/QC utility 6
oxo-flow-methylseq nf-core/methylseq @ 4.2.0 197 WGBS methylation 8
oxo-flow-viralrecon nf-core/viralrecon @ 3.0.0 167 viral genomics 12
oxo-flow-rnaseq-star-deseq2 snakemake-workflows/rna-seq-star-deseq2 @ v3.1.1 370 bulk RNA-seq 7
oxo-flow-snparcher harvardinformatics/snparcher @ v2.2 95 non-model variant calling 9
oxo-flow-varlociraptor snakemake-workflows/dna-seq-varlociraptor @ v6.10.0 90 small & structural variants 8
oxo-flow-enrichment epigen/enrichment_analysis @ v3.0.1 77 functional enrichment 6
oxo-flow-unsupervised epigen/unsupervised_analysis @ v4.0.2 67 post-analysis statistics 7
oxo-flow-genome-tracks epigen/genome_tracks @ v2.0.5 57 genome-browser tracks 6
oxo-flow-mixscape epigen/mixscape_seurat @ v2.0.3 56 CRISPR perturbation 7
oxo-flow-bgcflow NBChub/bgcflow @ v1.1.2 55 biosynthetic gene clusters 6

† Estimated rule count of the default main path. ✓ = fidelity verification agent assigned.

Notable non-selections (and why)

  • snakemake-workflows/dna-seq-gatk-variant-calling (267★) — the canonical GATK best-practices tutorial, but unmaintained since 2023-06; maintenance criterion (20%) disqualified it, and variant-calling is covered by sarek/snparcher/varlociraptor.
  • Hoohm/dropSeqPipe (149★) and snakemake-workflows/single-cell-rna-seq (107★) — stale (2023 / 2022); single-cell covered by scrnaseq.
  • maxplanck-ie/snakepipes (407★) — actively maintained but a collection of ~20 pipelines, not one workflow; unsuitable as a single template repo.
  • snakemake/snakemake (2847★) — the engine itself, not a workflow.
  • csoneson/ARMOR (168★) — RNA-seq preprocessing is covered by both rnaseq ports.

Domain coverage

bulk RNA-seq (rnaseq, rnaseq-star-deseq2) · single-cell (scrnaseq) · WGS/WES (sarek) · metagenomics + amplicon (mag, ampliseq) · epigenetics ChIP/ATAC/methyl (chipseq, atacseq, methylseq) · viral (viralrecon) · long-read (nanoseq) · ancient DNA (eager) · variant calling non-model/SV (snparcher, varlociraptor) · microbial natural products (bgcflow) · CRISPR perturbation (mixscape) · post-analysis (enrichment, unsupervised, genome-tracks) · fetch/QC (fetchngs).

Known gap: proteomics — the only strong candidate (nf-core/proteinfold) has a GPU-heavy few-rule default path with poor portability value; left for a future round.

Chair review notes

  1. All 20 sources are MIT-licensed, none archived, every one pushed within ~13 months of the selection date.
  2. All tags/commit SHAs in pipeline-specs.json were fetched live from the GitHub API at selection time — no invented versions.
  3. The snakemake half draws 4/8 entries from the maintained MrBiomics (epigen) suite: provenance is concentrated, but the four cover four distinct domains and each is catalog-listed and actively maintained. Accepted with this note.
  4. The 6 highest-complexity ports (rnaseq, sarek, viralrecon, eager, mag, chipseq) carry a mandatory fidelity-verification stage against upstream source.