nf-core pipeline · dna-methylation, fiber-seq, long-read
nf-core/methylong
Extract DNA methylation calls, phase variants, and identify differentially methylated regions from long reads.
What nf-core/methylong is
nf-core/methylong is a Nextflow bioinformatics pipeline developed by Jin Yan Khoo at Ludwig-Maximilians University Munich and YiJin Xiong at Central South University. It is designed to analyse DNA methylation patterns directly from long-read sequencing technologies, specifically Oxford Nanopore Technologies (ONT) and Pacific Biosciences (PacBio) HiFi platforms.
Native long-read sequencing can detect modified bases directly from raw electrical or optical signals without requiring bisulfite conversion, which damages DNA and produces short fragments. nf-core/methylong provides an end-to-end framework to process these datasets, spanning from basecalling raw instrument files to generating methylation pileups, phasing variants, and identifying differentially methylated regions across haplotypes and cohorts.
Beyond standard CpG methylation calling, the pipeline accommodates Fiber-seq workflows to examine chromatin accessibility by calling m6A modifications and inferring nucleosome positioning alongside methylation data.
What the pipeline does
For ONT data, the pipeline accepts raw pod5 files or modification-basecalled BAM (modBAM) files. pod5 inputs are basecalled to modBAM using Dorado with default 5mC and 5hmC models. For modBAM inputs, the pipeline optionally trims adapters and barcodes using Porechop, restores MM and ML tags with modkit repair, and aligns reads against a reference genome using either Dorado aligner or Minimap2. Methylation pileups are then generated at a minimum of 5x coverage using modkit pileup to produce bedMethyl and BEDGRAPH formats.
For PacBio data, unaligned BAM or modBAM inputs undergo optional modification calling with Jasmine or ccsmeth before alignment with pbmm2 or Minimap2. Methylation pileups are created using pb-CpG-tools or modkit pileup. Downstream of alignment for both platforms, the pipeline calls single-nucleotide variants (SNVs) using Clair3, phases reads using WhatsHap, and performs differential methylation analysis across haplotypes or populations using DSS or modkit dmr. For Fiber-seq data, it handles m6A calling, nucleosome inference, and bedMethyl extraction via modkit and fibertools.
Key facts
- Supported platforms
- Oxford Nanopore Technologies (ONT) and Pacific Biosciences (PacBio) HiFi
- Input formats
- Raw ONT pod5, raw BAM, unaligned modBAM, and FASTA genome reference
- Output formats
- BAM, BAI, VCF, bedMethyl (BED), BEDGRAPH, BigWig, and text reports
- Variant caller
- Clair3
- Phasing tool
- WhatsHap
- Differential methylation tools
- DSS and modkit dmr
- Base coverage threshold
- Minimum 5x for pileup and bedgraph outputs
- Pipeline engine
- Nextflow (nf-core)
What you can use it for
- Genome-wide CpG methylation profiling
- Process raw ONT pod5 or PacBio BAM files to yield 5mC and 5hmC bedMethyl and BEDGRAPH tracks aligned to a reference assembly.
- Haplotype-resolved methylation and DMR calling
- Call single-nucleotide variants with Clair3, phase reads with WhatsHap, and calculate allele-specific differentially methylated regions using DSS.
- Population-scale differential methylation analysis
- Compare methylation profiles between experimental cohorts or sample groups using statistical testing in DSS or modkit dmr.
- Chromatin accessibility and Fiber-seq analysis
- Infer nucleosome boundaries and single-molecule footprinting by predicting or filtering m6A modification calls.
Run it here
The research agent runs nf-core/methylong with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
Strengths and limits
Good at
- Supports end-to-end processing starting from raw instrument files (pod5 or raw BAM) through to downstream statistical DMR testing.
- Incorporates a repair step for ONT modBAM files that preserves MM and ML modification tags during adapter and barcode trimming.
- Provides automated file-type detection to choose whether basecalling needs to be triggered on ONT data.
- Generates unified MultiQC summary reports across read quality metrics and alignment flag statistics.
Watch out for
- Dorado and pb-CpG-tools cannot be installed or run through Conda environments within this workflow.
- Sites with base coverage lower than 5x are omitted from the default pileup and bedgraph outputs.
- Requires reference genomes supplied in FASTA format, preventing reference-free analysis.
Questions
Can I run nf-core/methylong without my own cluster?
The research agent runs nf-core/methylong with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
What licence is nf-core/methylong under?
MIT, like every nf-core pipeline.
Can I use already aligned modBAM files as input?
Yes. If your input modBAM has previous alignment records, you can supply the --reset parameter to strip old alignment information before realigning to your target reference.
Which differential methylation methods are available?
The pipeline provides DSS by default to identify statistically significant CpG sites and DMRs, with modkit dmr available as an alternative.
How are forward and reverse strand methylation signals handled for PacBio data?
While pb-CpG-tools merges signals on CpG dyads to the forward strand by default, the workflow tags forward and reverse reads using HP tags so that strand-specific signals can be output separately.
Can the pipeline call variants alongside methylation?
Yes. The downstream workflow includes single-nucleotide variant calling with Clair3, followed by read haplotagging and phasing using WhatsHap.
How to cite it
10.5281/zenodo.15366448
Written from nf-core/methylong's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.