nf-core pipeline · alignment, long-read, pacbio
nf-core/pacvar
Processes PacBio long-read sequencing data for whole-genome variant discovery and tandem repeat characterisation.
What nf-core/pacvar is
nf-core/pacvar is a bioinformatics pipeline designed to process long-read Pacific Biosciences (PacBio) sequencing data. Originally developed by Tanya Sarkin Jain, with subsequent contributions from Chao-Jen Wong and Stijn Van de Sompele, the workflow applies PacBio tools to standardise variant analysis. It provides two core pathways: one tailored for whole-genome sequencing (WGS) and another designed for targeted reads generated with the PacBio PureTarget repeat expansion panel.
The pipeline enables researchers to detect multiple classes of genomic variations from high-fidelity long reads. It coordinates demultiplexing, read alignment, single-nucleotide variant (SNV) calling, structural variant (SV) calling, haplotype phasing, copy number variation (CNV) discovery, and per-CpG methylation extraction. It also provides targeted tandem repeat genotyping and an optional Fiber-seq pathway for chromatin architecture profiling.
What the pipeline does
Data analysis begins with optional read demultiplexing using lima, followed by alignment to a reference genome with pbmm2 and alignment indexing with SAMtools. In the whole-genome sequencing workflow, SNVs and small indels are called using either DeepVariant by default or GATK HaplotypeCaller. Structural variants are detected using sawfish by default or pbsv. Identified variants and alignments are subsequently phased with HiPhase, CNVs are identified with HiFiCNV, per-CpG methylation scores are extracted using pb-CpG-tools, and variant functional consequences are annotated with Ensembl VEP.
For the tandem repeat workflow targeting panels such as PureTarget, TRGT genotypes repeat expansions, outputs spanning alignment files and VCF records, and generates repeat motif plots. If a fail BAM is supplied in the samplesheet, mapped HiFi and fail reads are merged before repeat profiling. Additionally, the WGS workflow can be configured to process Fiber-seq data using fibertools-rs to predict m6A modifications from kinetic tags, add nucleosome and methyltransferase-sensitive patch tags, and extract nucleosome positions.
Key facts
- Input formats
- Unmapped BAM, optional PacBio BAM index (.pbi), CSV samplesheet
- Supported data types
- PacBio HiFi WGS, PureTarget repeat expansion panel, Fiber-seq
- Read aligner
- pbmm2
- Default SNV caller
- DeepVariant
- Default SV caller
- sawfish
- Tandem repeat caller
- TRGT
- Haplotype phasing
- HiPhase
- Methylation extraction
- pb-CpG-tools
- Original author
- Tanya Sarkin Jain
- Latest release
- 1.1.0 (2026-06-24)
What you can use it for
- Whole-genome variant discovery and phasing
- Call SNVs, indels, and structural variants from PacBio HiFi reads, followed by chromosome-scale haplotype phasing of variants and alignments using HiPhase.
- Targeted tandem repeat characterisation
- Genotype expanded tandem repeats, produce spanning alignment files, and generate motif summary plots from PacBio PureTarget panels using TRGT.
- Genome-wide CpG methylation extraction
- Extract per-CpG 5mC methylation probabilities and summary scores directly from aligned long-read BAM files using pb-CpG-tools.
- Fiber-seq chromatin accessibility profiling
- Predict m6A modifications from PacBio kinetic tags, record nucleosome and patch auxiliary tags, and extract single-molecule nucleosome positions with fibertools-rs.
- Integrated structural and copy number analysis
- Identify large-scale structural rearrangements and genomic copy number changes using sawfish, pbsv, or HiFiCNV.
Run it here
The research agent runs nf-core/pacvar 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
- Automates official PacBio toolchains for alignment, variant discovery, phasing, and methylation extraction within a single pipeline.
- Provides a dedicated sub-workflow for resolving complex repeat expansions from PureTarget panel sequencing using TRGT.
- Supports single-molecule chromatin structure analysis by integrating Fiber-seq processing alongside standard whole-genome variant calling.
- Allows consolidation of structural variant and copy number detection by pairing sawfish with optional disabling of HiFiCNV.
Watch out for
- The Ensembl VEP implementation does not bundle plugins or custom reference files.
- The default Ensembl VEP cache version 115 does not support the CHM13 human reference assembly, requiring annotation to be disabled when using this genome.
- Custom Ensembl VEP arguments cannot include --merged or --refseq when using the default S3 cache due to missing cache files.
Questions
Can I run nf-core/pacvar without my own cluster?
The research agent runs nf-core/pacvar 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/pacvar under?
MIT, like every nf-core pipeline.
Which variant callers are available for the whole-genome sequencing path?
For SNVs and small indels, users can choose between DeepVariant (the default) and GATK HaplotypeCaller. For structural variants, the pipeline provides sawfish (the default) and pbsv.
Can HiFiCNV be omitted when running sawfish?
Yes. Because sawfish consolidates structural variant and copy number events, setting --skip_hificnv true is recommended to avoid redundant CNV calculations.
How are failed reads incorporated into tandem repeat analyses?
Users can supply an optional fail BAM column in the samplesheet for the repeat workflow, which instructs the pipeline to merge the mapped HiFi reads and fail reads prior to repeat genotyping.
What settings are required to process Fiber-seq data?
Users must set --skip_fiberseq false. If the input BAM contains raw kinetic tags, --skip_m6A_predict false must be used to call m6A, whereas --skip_m6A_predict true should be set if the file already contains m6A tags.
How to cite it
Tanya Jain, Claire Clelland, nf-core/pacvar: a pipeline for analyzing longread PacBio whole genome and repeat expansion sequencing data, Bioinformatics, 2025;, btaf116, https://doi.org/10.1093/bioinformatics/btaf116
Written from nf-core/pacvar's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.