nf-core/methylseq

The nf-core/methylseq pipeline is a Nextflow-based workflow designed for processing methylation sequencing data. The specific output files generated depend on the alignment workflow selected (Bismark, bwa-meth, or bwa-mem) and whether additional protocols like TAPS (processed via Rastair) are used.

Core processing steps include:

• FastQC: Raw read quality control.
• TrimGalore: Adapter trimming.
• Alignment: Mapping reads to a reference genome.
• Deduplication: Removing duplicate reads.
• Methylation Extraction: Calling cytosine methylation.
• MultiQC: Aggregating results into a single report.

Optional steps include Reference Genome Preparation, Qualimap (alignment QC), Preseq (sample complexity estimation), and HS Metrics (target-capture performance).

The documentation for the nf-core/methylseq pipeline is organized into specific sections for usage and output. For general information regarding the installation, configuration, and execution of any nf-core pipeline, refer to the central nf-core website.

• Usage: Detailed overview of pipeline operation, execution instructions, and command-line flag descriptions.
• Output: Description of produced results and guidance on how to interpret them.

You can append additional arguments to a specific tool/module by modifying its custom configuration in conf/modules or conf/subworkflows.

Example Use Case: To bypass OS file handle limits (e.g., ulimit -a showing 1024) when working with genomes containing many scaffolds, you can add the --scaffolds option as ext.args in conf/modules/bismark_methylationextractor.config. This prevents methylation calls from being pre-sorted into individual chromosome files.

The pipeline supports TAPS data analysis, which preserves the original DNA sequence while converting methylated cytosines to thymine.

Key Features of TAPS Workflow

• Automatic Configuration: When the --taps flag is specified, the pipeline automatically uses Rastair for methylation calling, validates aligner compatibility (preventing Bismark usage), and configures Picard MarkDuplicates.
• Optimization: Using --aligner bwamem automatically triggers the use of Rastair for methylation calling, even if the --taps flag is not explicitly provided, as bwamem is optimized for TAPS data.

The default workflow uses Bismark with Bowtie2 for alignment. This is optimized for most DNA methylation workflows.

Using HISAT2 with Bismark

Starting with Bismark v0.21.0, you can use HISAT2 as an alternative aligner (useful for RNA-based analyses like SLAMseq) by using the --aligner bismark_hisat flag. For splice-aware alignment, you can provide known splice sites using the --known_splices parameter.

nextflow run nf-core/methylseq --aligner bismark_hisat

The preseq step estimates library complexity by predicting how many additional unique reads are sequenced as the total read count increases.

• A shallow curve indicates complexity saturation: further sequencing is unlikely to yield more unique reads.
• A dashed line represents a perfectly complex library where total reads equal unique reads.

Note: Values are predictive, not absolute. If the MultiQC plot shows extreme sequencing depth on the X axis, click and drag from the left side of the plot to zoom in on realistic values.

Output directory: results/preseq

• sample_ccurve.txt: Contains the plot values used for the MultiQC complexity curve.

To avoid specifying long lists of flags in the command line, you can use a yaml or json file via the -params-file flag. This is recommended for reproducibility and ease of use.

Warning: Do not use -c <file> to specify pipeline parameters; this will result in errors. The -c flag is reserved for Nextflow configuration files used for tuning resources or module arguments. Use -params-file for pipeline parameters like input, outdir, or genome.

nextflow run nf-core/methylseq -profile docker -params-file params.yaml

Example params.yaml:

input: './samplesheet.csv'
outdir: './results/'
genome: 'GRCh37'
<...>

To run the nf-core/methylseq pipeline, ensure you meet the following requirements:

• Nextflow: version >= 24.10.5
• Container Runtime: Docker, Singularity, Podman, Charliecloud, or Apptainer.
• Conda/Mamba: Supported for CPU workflows. Note that the Parabricks GPU pathway does not support Conda/Mamba.

Nextflow caches the pipeline code locally. To ensure you are running the latest version of the pipeline, you must manually pull the updates using the pull command.

nextflow pull nf-core/methylseq