eager completeness audit (2026-08-21)
Upstream: nf-core/eager @ 2.5.3 · Port: oxo-flow-eager (live-verified).
Monolithic DSL1 pipeline (single 3617-line main.nf), 59 processes. Input types: TSV design file (11 columns: SeqType/Colour_Chemistry/Strandedness/ UDG_Treatment), FASTQ glob, BAM input.
Mode matrix (upstream)
| axis | options | default |
|---|---|---|
| mapper | bwaaln / bwamem / circularmapper / bowtie2 | bwaaln (aDNA) |
| dedup | dedup / markduplicates | markduplicates |
| damage calc | damageprofiler / mapdamage | damageprofiler |
| damage extras | mapdamage_rescaling, pmdtools (UDG-aware), bam_trim | off |
| genotyping | ug (GATK3) / hc (GATK4) / freebayes / pileupcaller / angsd; sources raw/trimmed/pmd/rescaled | off |
| genotyping extras | vcf2genome, multivcfanalyzer (ug-only) | off |
| human modes | mtnucratio, sexdeterrmine, nuclear_contamination (HapMapChrX) | off |
| metagenomic screening | malt (+maltextract) / kraken2; requires BAM filtering + fastq unmapped | off |
| input conversion | convertBam (BAM→FASTQ remap) vs pass-through BAM | — |
| host removal | hostremoval fastq export (2 modes) | off |
| filtering/QC | samtools_filter, fastp poly-G (Colour_Chemistry 2), bedtools coverage, preseq modes | off/on |
Gap tiers vs the port
P0: the 3 non-default mappers (bwamem, circularmapper — circulargenerator+circularmapper for mtDNA, bowtie2 with its alignmode/ sensitivity matrix); genotyping suite (5 tools + 3 sources + vcf2genome + multivcfanalyzer); metagenomic screening (malt + kraken2 + maltextract + complexity filter); pmdtools + bam_trim + rescaling chain; human modes (mtnucratio, sexdeterrmine, nuclear_contamination); convertBam BAM-input mode; hostremoval exports; fastp poly-G path.
P1: GATK3 (ug + MultiVCFAnalyzer) — academic-free/commercial-paid license model (document; hc/freebayes/angsd are the license-free alternatives).
P2: dedup choice, preseq modes, qualimap snpcapture bed, --large_ref CSI indexing, adapter_removal 9-branch PE/SE matrix, clip params, strandedness/UDG per-sample columns, test profile set. NOTE: 5 test profiles in config reference conf/test_*.config files that don't exist at this tag (broken refs — do not port).
External deps: MALT/Kraken2 DBs user-provisioned (free), nf-core configs fetch, iGenomes S3 (off by default).
Verdict
Wide toggle surface on a shared backbone — the ported aDNA default path (bwaaln + damageprofiler) leaves the genotyping suite and metagenomic screening as the two biggest P0 blocks, plus 3 mappers.
Re-verification (2026-08-23, 9-mini queue 2/9)
Engine: latest main (post-v0.14.1) · Box: tx-ubuntu (docker,
nfcore/eager:2.5.3) · Mode: real CLI run, not dry-run · -j 2
--keep-going.
22 succeeded, 6 skipped, 0 failed, exit 0, one round, zero repo fixes. Coverage chain: bwa index → fasta index → seq dict → fastqc → adapterremoval → dedup → preseq → damageprofiler → qualimap → multiqc. All 6 skips are condition gates + checkpoint reuse.
Box-side note: docker root disk at 96% broke the image pull's layer
commit ("failed commit on ref") — a leftover of the earlier ENOSPC;
docker system prune -af reclaimed 23GB (see failure catalog).