FAIR-bioHeaders Nextflow workflow

A minimal Nextflow DSL2 workflow around the two nf-core-style modules of this repository:

1. FAIRBIOHEADERS_COMBINE adds FAIR-bioHeaders metadata (FHR for FASTA and GFA, FHGFF3 for GFF3) to each sequence file and calculates its checksum;
2. FAIRBIOHEADERS_VERIFY checks the headed file against the schema and the checksum. The run fails if a file does not verify.

It is registered on Dockstore through .dockstore.yml at the repository root.

For the full pipeline (metadata creation, conversion, stripping, GFF3 validation and FHP protein FASTA), see FAIR-bioHeaders/FHR-Nextflow, currently a development prototype. This repository provides the reusable combine/verify modules and this minimal example workflow.

Input

Either a samplesheet, a CSV file with the columns id,metadata,sequence:

id,metadata,sequence
genome,genome.fhr.yaml,genome.fa.gz
graph,genome.fhr.json,graph.gfa
annotation,annotation.fhgff3.yaml,annotation.gff3

nextflow run FAIR-bioHeaders/FAIR-bioHeaders-Tools \
    -main-script nextflow/workflows/fairbioheaders/main.nf \
    -profile docker --input samplesheet.csv --outdir results

or one file:

nextflow run FAIR-bioHeaders/FAIR-bioHeaders-Tools \
    -main-script nextflow/workflows/fairbioheaders/main.nf \
    -profile docker --metadata genome.fhr.yaml --sequence genome.fa.gz --outdir results

| Parameter | Description | | --- | --- | | --input | Samplesheet (CSV: id,metadata,sequence). Relative paths are relative to the samplesheet, which may be a URL. | | --metadata, --sequence | Metadata (JSON, YAML, JSON-LD or HTML microdata) and one FASTA, GFA or GFF3 file, plain or gzip/BGZF, in place of --input. | | --id | Output name for --sequence (default: its file name without the extension). | | --outdir | Output directory (default results). |

The file type comes from the sequence file's extension: .gfa is GFA, .gff3 or .gff is GFF3, anything else is FASTA (.gz/.bgz ignored). In a clone of the repository, run nextflow run nextflow/workflows/fairbioheaders/main.nf ... in the same way.

Output

• <outdir>/combine/<id>.fhr.<ext> for FASTA and GFA, or <outdir>/combine/<id>.fhgff3.gff3 for GFF3: the sequence file with its metadata header (.gz when the input was compressed);
• <outdir>/verify/<id>.verify.txt: the verification report.

Profiles

| Profile | Runs bioheaders from | | --- | --- | | docker | the modules' BioContainers image, quay.io/biocontainers/fair-bioheaders:0.8.2--pyhdfd78af_0 | | singularity, apptainer | the modules' Galaxy depot Singularity image of the same build | | conda | environment.yml: bioconda::fair-bioheaders=0.8.2 | | test | the samplesheet in assets/test: a FASTA, a GFA and a GFF3 example, and the FASTA with expanded JSON-LD metadata | | none | bioheaders on the PATH (pip install "fair-bioheaders[jsonld]") |

Combine test with one of the others, for example -profile test,docker. assets/test/params.json is the same test as a parameter file (-params-file), reading the samplesheet from GitHub; Dockstore lists it as the test parameter file.

Container and conda environment. Both processes run the modules' container, the BioContainers image of fair-bioheaders 0.8.2 (or the Galaxy depot's Singularity image of it). The conda profile uses environment.yml, which pins the same Bioconda package as the modules; keep the two pins equal. From 0.8.2 the Bioconda package, and so the image, includes pyld, the optional dependency (the jsonld extra) that reading JSON-LD metadata in other than the canonical form needs, for example expanded JSON-LD, so the workflow no longer replaces the modules' image with the project's GHCR image.

The Nextflow workflow CI job checks that the workflow sets no container and that its Bioconda pin equals the modules', then runs -profile test,docker, including the expanded JSON-LD row of the test samplesheet.