MetaGEAR Pipeline

MetaGEAR Pipeline is the Nextflow / nf-core pipeline at the core of the platform. It takes shotgun metagenomic data through a set of workflows that run independently or chain together, with outputs in a consistent schema designed to be easy to post-process.

→ schirmer-lab/metagear-pipeline ↗ — source, releases, and the canonical pipeline reference.

The workflows

You pick one with --workflow. The first four run from reads; the rest read an earlier run’s outputs from the same directory.

Workflow	What it produces	Builds on
`download_databases`	Installs every reference database the others need	—
`qc_dna` / `qc_rna`	Trimmed, host-decontaminated reads and a QC report	raw reads
`microbial_profiles`	MetaPhlAn taxonomic and HUMAnN functional profiles	clean reads
`genes`	Assembly, gene and protein catalogs, abundance matrices	clean reads
`virus`	Viral and plasmid catalogs, annotation, host prediction, lifestyle calls	clean reads
`classification`	Per-contig classification and per-sample bacterial bins	clean reads
`mag`	Cohort MAG catalog with GTDB-Tk taxonomy and MAG×sample abundance	`classification`
`msp`	MetaSpecies Pangenomes from co-abundance, with taxonomy	`genes`
`structures`	Structural-homology annotation for proteins sequence search leaves unassigned	`genes`/`virus`

Each has its own reference page under Pipeline reference, mirrored from the pipeline repository so it always matches the released code.

Presets

Most cohorts do not need one workflow at a time. MetaGEAR Tools ships presets that run several in order in one workspace, each reusing what the ones before it produced:

Preset	Runs	Use it when
`profiles`	`microbial_profiles`	Reference-based profiles only, without assembling the cohort.
`genomes`	`genes` → `classification` → `mag` → `msp`	You want the assembled bacterial picture.
`microbiome`	`genomes` + `virus`	You want everything, viruses and plasmids included.

metagear microbiome --input samplesheet.csv --outdir results/

A preset is not a separate analysis — it is the same workflows in dependency order, with --reuse-outputs switched on so nothing is recomputed. Running them by hand in the same order gives the same result.

→ Presets — what each one produces, what it costs, and how to preview a run.

How you run it

Directly with Nextflow, if you already use it:

nextflow run schirmer-lab/metagear-pipeline \
  -profile docker \
  --workflow genes \
  --input samplesheet.csv \
  --outdir results/

--workflow is required — without it the run validates the samplesheet and exits without analysing anything.

The recommended path, especially if you are not a regular Nextflow user, is MetaGEAR Tools: it installs the pipeline, manages the reference databases, applies sensible resource defaults, and adds the presets above.

Inputs and outputs

A CSV samplesheet pointing at paired-end FASTQ files:

sample,fastq_1,fastq_2
SAMPLE-01,/path/to/sample1_R1.fastq.gz,/path/to/sample1_R2.fastq.gz

classification also reads an optional biome column between sample and fastq_1, used to pick the SemiBin2 model.

Outputs go to one directory shared by every workflow, organised by artifact rather than by workflow — catalogs/, abundance/, annotations/, assemblies/ — which is what lets a later workflow discover what an earlier one produced.

Built on nf-core

MetaGEAR Pipeline uses the nf-core framework, inheriting community conventions for reproducibility, containerization and testing. It runs with Docker, Singularity or any nf-core-supported execution profile, on a laptop or an HPC cluster.

Where to go next

Pipeline reference — a page per workflow, mirrored from the pipeline repository
MetaGEAR Tools (CLI) — install and drive the pipeline from one command
Pipeline repository — source, releases, full parameter reference
Open an issue — bug reports and feature requests