nanopore-dna-modification-pipeline

From a directory of Oxford Nanopore DNA POD5 files to aligned modBAMs, per-site modification tables and QC, on the cluster, from one config file.

Documentation: https://rnabioco.github.io/nanopore-dna-modification-pipeline/ — start with the walkthrough if you have a run directory in hand.

POD5 run ──► dorado basecaller ──► dorado aligner ──► bam/final ──► modkit pileup · summary · DMR
             simplex + stacked        (minimap2)          │          samtools · mosdepth · summary table
             mod models (GPU)                             ├──► DNAscent: index → detect (GPU) → forkSense
MinKNOW bam_pass (already basecalled) ────────────────────┘└──► custom steps (your scripts)

Features

• dorado 2.1.2 with a config-resolved model stack: one simplex model plus ONT modified-base models (5mC_5hmC, 5mCG_5hmCG, 4mC_5mC, 6mA, by short code or full name) and your own dorado-format models, checked for conflicts before any GPU time is spent.
• Three kinds of input: plain POD5 runs, barcoded runs (one dorado pass with --kit-name, then dorado demux), and MinKNOW live-basecalled runs used as-is. POD5 is linked, never copied.
• Downstream: alignment with every tag preserved, modkit pileup / summary / per-read calls / bigWig / pairwise DMR, samtools and mosdepth QC, one summary table per project, a provenance manifest.
• DNAscent (optional): BrdU/EdU detection, a per-read table, forkSense origins and forks, from the maintainers' Singularity image.
• Custom steps: your commands as first-class rules with the pipeline's outputs as placeholders.
• Clusters: Slurm profiles for Bodhi and Alpine; dorado and DNAscent on the GPU queues. Everything installed and run through pixi.

Quick start

git clone https://github.com/rnabioco/nanopore-dna-modification-pipeline.git
cd nanopore-dna-modification-pipeline
pixi install                    # snakemake, samtools, minimap2, modkit, mosdepth ...
pixi run setup                  # dorado 2.1.2 + the configured models + escpod (once, from one node)
pixi run dl-test-data && pixi run dry-run     # 5 MB synthetic fixture; builds the DAG

Then a samples file and a project config:

# config/gdna-samples.tsv
wt_rep1    /beevol/data/nanopore/20260901_gDNA_wt
mut_rep1   /beevol/data/nanopore/20260901_gDNA_mut

# config/gdna.yml
samples: config/gdna-samples.tsv
output_directory: /beevol/home/me/results/gdna-2026-09
reference:
  fasta: /beevol/data/ref/chm13v2.0.fa
models:
  simplex: dna_r10.4.1_e8.2_400bps_sup@v5.2.0
  modified_bases: [5mC_5hmC, 6mA]

pixi run list-models --configfile config/gdna.yml         # what will be stacked
pixi run snakemake -n --configfile config/gdna.yml        # dry run
pixi run run-bodhi  --configfile config/gdna.yml          # or run-alpine, or run (local GPU)

Outputs land under output_directory: bam/final/<sample>.bam, summary/modkit/<sample>/<sample>.pileup.bed.gz, summary/qc/, summary/samples_summary.tsv, manifest.json.

Documentation map

Page	For
Installation	pixi, setup, DNAscent image, checking the install
Walkthrough	step by step from a run directory to results
Samples file, Configuration, Models	the inputs
Running and monitoring, Outputs	the run
DNAscent, Custom steps	the optional stages
Bodhi, Alpine	site specifics
Architecture, Contributing	for developers; CLAUDE.md for the conventions

License

MIT (see LICENSE). DNAscent is GPL-3.0 and is used as an external tool.