nf-core pipeline · genomics, genotype, imputation
nf-core/phaseimpute
Phases and imputes genetic data across low-pass sequencing and reference panel datasets.
What nf-core/phaseimpute is
nf-core/phaseimpute is a bioinformatics pipeline designed to phase haplotypes and impute unobserved genotypes from genetic data. Phasing determines which genetic variants sit together along the same chromosome, whilst imputation statistically infers missing alleles using a reference panel. The workflow was originally developed by Louis Le Nézet and Anabella Trigila.
The software addresses the challenge of handling low-pass genomic sequencing and reference panels by streamlining data preparation, downsampling, variant estimation, and quality evaluation. It allows geneticists to prepare high-quality reference panels, simulate lower coverage depths, run imputation algorithms, and benchmark the resulting genotypes against truth datasets in an integrated framework.
What the pipeline does
The pipeline operates across five modular stages that can be executed together or independently. First, a chromosome name check performs quality control to ensure contig naming consistency between variant files and alignments. The simulation step allows users to downsample BAM or CRAM alignment files to lower read depths using samtools view. In the panel preparation step, reference panels are normalised, phased using Shapeit5, partitioned into genomic chunks, and processed to extract target marker coordinates.
During the imputation stage, genotypes are inferred using one of several integrated engines, including Glimpse1, Glimpse2, Stitch, Quilt, Beagle5, or Minimac4. Imputed genomic chunks are subsequently ligated into unified, per-sample VCF files across all chromosomes. Finally, the validation step evaluates imputation accuracy by comparing the imputed VCF against a high-quality truth dataset using the Glimpse2 concordance procedure.
Key facts
- Latest release
- 1.1.0
- Original authors
- Louis Le Nézet and Anabella Trigila
- Core framework
- Nextflow (nf-core)
- Phasing engine
- Shapeit5
- Supported imputation tools
- Glimpse1, Glimpse2, Stitch, Quilt, Beagle5, Minimac4
- Input formats
- BAM, CRAM, and VCF files with index files
- Simulation tool
- samtools view -s
- Validation tool
- Glimpse2 concordance
What you can use it for
- Reference panel preparation
- Normalising, phasing with Shapeit5, chunking chromosomes, and extracting variant positions from reference datasets.
- Low-pass sequence imputation
- Imputing genotypes from low-coverage alignment data using tools such as Glimpse, Stitch, Quilt, Beagle5, or Minimac4.
- Sequencing depth simulation
- Downsampling high-coverage BAM or CRAM files to lower coverage levels with samtools to simulate low-pass datasets.
- Imputation accuracy benchmarking
- Validating imputed variant calls by assessing concordance against a known, high-quality truth set with Glimpse2.
Run it here
The research agent runs nf-core/phaseimpute 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
- The workflow consists of five independent steps that can be run individually according to project requirements.
- It incorporates a broad selection of imputation algorithms, accommodating both array and low-pass sequencing workflows.
- Built-in simulation and validation procedures enable direct concordance benchmarking against ground-truth data.
Watch out for
- All input alignment and index files must have consistent file extensions across the provided samplesheet.
- Pipeline parameters cannot be supplied through custom Nextflow configuration files loaded via the -c option.
Questions
Can I run nf-core/phaseimpute without my own cluster?
The research agent runs nf-core/phaseimpute 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/phaseimpute under?
MIT, like every nf-core pipeline.
Do all pipeline steps need to be run sequentially?
No. The pipeline comprises five distinct steps that can be run separately and independently depending on project requirements.
Which software tools can be selected for imputation?
Imputation can be performed using Glimpse1, Glimpse2, Stitch, Quilt, Beagle5, or Minimac4.
What input formats are required for target datasets?
The target dataset requires a comma-separated samplesheet specifying sample identifiers alongside paths to BAM or CRAM files and their corresponding indices.
How are chromosome chunks combined after imputation?
The imputation process includes a ligation step that merges the imputed chunks into a unified per-sample VCF file covering all chromosomes.
How to cite it
10.5281/zenodo.14329225
Written from nf-core/phaseimpute's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.