nanoseq completeness audit (2026-08-21)
Upstream: nf-core/nanoseq @ 3.1.0 · Port: oxo-flow-nanoseq (live-verified).
Headline: NO basecalling at this tag (guppy removed as EULA breach, dorado never added) — pre-basecalled inputs only. No methylation/polyA.
Mode matrix (upstream — param-driven, one entry)
| mode | selecting params |
|---|---|
| demux | --input_path + --barcode_kit → QCAT (off via --skip_demultiplexing) |
| alignment | --aligner minimap2(default)/graphmap2; --protocol DNA/cDNA/directRNA; BAM input via --skip_alignment |
| short VC (DNA only) | --call_variants + --variant_caller medaka(default)/deepvariant/pepper_margin_deepvariant |
| SV (DNA only) | --structural_variant_caller sniffles(default)/cutesv |
| quantification + DE | cDNA/directRNA; --quantification_method bambu(default)/stringtie2; DESeq2/DEXSeq |
| RNA modification | directRNA + fast5/fastq dirs; nanopolish eventalign + xpore + m6anet |
| fusion | cDNA/directRNA; JAFFAL (+ figshare ref download) |
| QC | NanoLyse / NanoPlot / FastQC / MultiQC |
Gap tiers vs the port
P0: graphmap2 aligner; BAM-input mode; deepvariant + pepper_margin callers (container-only); cutesv; stringtie2 + featureCounts quant path; DEXSeq; RNA modification chain (nanopolish/xpore/m6anet — container-only, fast5 inputs); JAFFAL fusion; NanoLyse; demux via QCAT + barcode kits.
P1: ONT basecalling EULA (user-side prerequisite — document, not portable); deepvariant/pepper/m6anet/dexseq are container-only (conda profile hard-fails — porting requires container path or documented constraint).
P2: --skip_* gates, --stranded, qcat score options, jaffal ref dir, nanolyse fasta, iGenomes refs. Dead params at this tag: --gpu_device, --split_mnps, --phase_vcf, --save_align_intermeds, --trim_barcodes (declared, unused).
External deps: figshare JAFFAL ref, GitHub lambda.fasta, nf-core test datasets, S3 iGenomes.
Verdict
Moderate surface: the ported default (minimap2 + medaka/sniffles + bambu) leaves ~5 feature modes as P0, with the container-only VC callers as the main porting constraint.
Re-verification (2026-08-23, light group)
Engine: latest main (post-v0.14.1) · Box: tx-ubuntu · Mode: real CLI run, not dry-run.
29/29 rules succeeded, exit 0, one round, zero failures at the latest commit. No fixes needed and no new failure classes — the default path runs clean against the reference data with no intervention.