nf-core pipeline · epigenome, multi-omics-integration, segmentation
nf-core/epigenomesegmentation
A chromatin segmentation pipeline annotating functional genomic states from histone, chromatin, and methylation data.
What nf-core/epigenomesegmentation is
nf-core/epigenomesegmentation is a bioinformatics workflow designed for chromatin segmentation across multi-omics datasets. The pipeline employs hidden Markov models (HMMs) to partition the genome and annotate regions into functional categories such as enhancers and promoters. It maps combinatorial patterns of epigenetic modifications while capturing spatial interactions through state transition probabilities.
The underlying framework relies on EpiSegMix and EpiSegMeth, developed by researchers at Saarland University, and is an approved de.NBI service. The workflow standardises the processing of sequencing files, including histone marks, open chromatin, and DNA methylation measurements, allowing researchers to evaluate complex epigenetic landscapes across diverse sample cohorts.
By integrating read count distribution fitting with options for topology or state duration modelling, the pipeline helps scientists discern regulatory architectures. It unifies data preparation, counting, model training, and state decoding into reproducible runs across genomes.
What the pipeline does
The pipeline starts by partitioning a reference genome into non-overlapping genomic windows (defaulting to 200 base pairs on the hg38 build). For histone modifications, aligned BAM files are indexed, reheadered, and converted into binned count matrices. For DNA methylation or coverage markers, BED files are processed and mapped across the genomic bins using Bedtools. The pipeline standardises and merges these matrices so that all modalities align to identical genomic coordinates.
Following count matrix generation and configuration preparation, the pipeline trains an EpiSegMix hidden Markov model. By default, it applies linear topology modelling (LDM), but users can enable duration modelling (DM), run DNA methylation-only segmentation, or evaluate best-fitting statistical count distributions (such as Negative Binomial or Binomial distributions) per mark. After model training, state decoding and reporting modules assign functional states across the genome and produce summary reports.
Key facts
- Pipeline framework
- Nextflow DSL2
- Underlying tool
- EpiSegMix / EpiSegMeth
- Modelling approaches
- Topology-modelling (LDM), duration-modelling (DM), standard modelling, and methylation modelling
- Input formats
- BAM (.bam, .bam.gz) for histones; BED (.bed, .bed.gz) for methylation/coverage; precomputed count matrices
- Default reference genome
- hg38
- Default bin size
- 200 bp
- Original developers
- Johanna Elena Schmitz, Nihit Aggarwal, AbdulRahman Salhab, Jörn Walter, Sven Rahmann, and Aaryan Jaitly (Saarland University)
What you can use it for
- Multi-assay chromatin state annotation
- Integrating histone ChIP-seq BAM files and DNA methylation BED files to annotate regulatory elements like promoters and enhancers across a reference genome.
- Methylation-only segmentation
- Annotating genomic regions exclusively using coverage markers or whole-genome bisulfite sequencing (WGBS) data using the dedicated methylation workflow.
- Epigenetic distribution fitting
- Evaluating candidate count distributions across assay modalities to identify the best statistical models (such as Negative Binomial or Binomial) before final segmentation.
- Genomic window count generation
- Extracting and standardising binned read count matrices across histone and methylation markers without running model training or state decoding.
Run it here
The research agent runs nf-core/epigenomesegmentation 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 multiple hidden Markov model architectures, including standard, topology-modelling, and duration-modelling approaches.
- Integrates both histone modification BAM files and DNA methylation BED files into unified genomic bin representations.
- Provides an automated evaluation workflow to determine the optimal count distribution for each epigenetic marker.
- Accepts precomputed count matrices to skip early-stage BAM and BED processing steps.
Watch out for
- Requires specific file types per assay modality, accepting only BAM or BAM.GZ for histones and BED or BED.GZ for methylation data.
- Joint training across shared models is disabled by default and requires explicit configuration.
- Distribution fitting requires running a dedicated workflow mode rather than evaluating distributions concurrently during standard segmentation.
Questions
Can I run nf-core/epigenomesegmentation without my own cluster?
The research agent runs nf-core/epigenomesegmentation 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/epigenomesegmentation under?
MIT, like every nf-core pipeline.
What input file formats does the pipeline accept?
Histone modifications must be provided as BAM files (.bam or .bam.gz). Methylation and coverage markers must be provided in BED format (.bed or .bed.gz). Precomputed count matrices can also be supplied directly.
What is the default bin size for genome segmentation?
The pipeline defaults to a genomic window size of 200 base pairs, typically on the hg38 reference genome, but this can be adjusted using parameter settings.
Can the pipeline be run on methylation data alone?
Yes. Passing the methylation flag enables a dedicated modelling workflow specifically tailored for coverage and methylation markers without histone BAM inputs.
Which statistical distributions can be used for read counts?
The pipeline supports discrete count distributions such as Negative Binomial (NBI) and Binomial (BI). A dedicated fitting mode can also be executed to detect the best-fitting distribution for each mark.
How to cite it
Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.
Written from nf-core/epigenomesegmentation's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.