Example Data & Provenance

SeqDesk offers example data in three forms. They serve different purposes and are labelled separately so synthetic preview records cannot be mistaken for scientific source data.

Choose the right example

Dataset	Access	Contents	Intended use
Hosted demo workspace	Open demo.seqdesk.org . Use `/demo` for the researcher view or `/demo/admin` for the facility-admin view.	Disposable, pre-populated orders, studies, samples, configurable metadata, read records, and inspectable example pipeline results.	Browse researcher workflows and inspect facility administration without installing SeqDesk. Facility-admin pages are read-only; downloads, external ENA submission, and pipeline execution are disabled for safety.
Deterministic local fixture	Admin → Settings → Demo data, or `seqdesk demo-data install` on the server after storage setup.	Two studies, four orders, samples, metadata, read rows, and deterministic synthetic gzipped FASTQ files covering short- and long-read layouts in the configured storage path.	Exercise local metadata, file discovery, pipeline selection, screenshots, and demos. The files are synthetic, not scientific data.
Public ENA examples	Named example-data and CI seed paths only; they are not installed by `seqdesk demo-data`.	Selected real public FASTQ files with source accessions and normalized provenance.	Extended/opt-in pipeline integration where external databases, private packages, compute, or credentials are available.

Install the local fixture

First choose a real, writable sequencing-data directory:


seqdesk storage configure "$HOME/seqdesk/data"
seqdesk storage status

Then check and install the fixture:


seqdesk demo-data status
seqdesk demo-data install

The same action is available in the facility-admin panel at Admin → Settings → Demo data. Installation is idempotent: rerunning it does not create duplicates. If more than one facility administrator exists, select the owner:


seqdesk demo-data install --user-email admin@example.org --yes --json

Remove only that administrator’s fixture with:


seqdesk demo-data remove --user-email admin@example.org --yes --json

Removal refuses to bypass linked pipeline runs or ENA submission history and remembers the original storage path. It does not wipe unrelated SeqDesk data. Delete linked runs through Pipeline Runs first so their cancellation and output cleanup can complete.

Real public-read sample manifest

The integration fixtures select a small, stable subset rather than downloading whole projects.

SeqDesk order	Source	Assay stored in SeqDesk	Instrument	Samples and runs
`DEV-MOUSE-PRJDB6165-001`	PRJDB6165	PCR amplicon sequencing, 16S V3–V4	Illumina MiSeq	`MGB-01` `DRR099973`; `MGB-02` `DRR099974`; `MGB-03` `DRR099975`; `MGB-04` `DRR099976`; `MGB-05` `DRR099977`; `MGB-06` `DRR099978`; `MGB-07` `DRR099979`; `MGB-08` `DRR099980`
`DEV-HUMAN-PRJEB54724-001`	PRJEB54724	Paired-end WGS metagenomics	Illumina MiSeq	`HGM-01` `ERR10009592`; `HGM-02` `ERR10009593`; `HGM-03` `ERR10009594`; `HGM-05` `ERR10009595`; `HGM-08` `ERR10009590`; `HGM-09` `ERR10009591`; `HGM-10` `ERR10009596`
`DEV-HUMAN-PRJEB54724-002`	PRJEB54724	Paired-end WGS metagenomics	NextSeq 550	`HGM-04` `ERR10009610`; `HGM-06` `ERR10009623`; `HGM-07` `ERR10009639`; `HGM-11` `ERR10009608`; `HGM-12` `ERR10009632`

ENA declares library_strategy=WGS for the PRJDB6165 run records, while the associated publication describes 16S V3–V4 PCR amplicons sequenced on MiSeq. SeqDesk stores the scientific assay as amplicon sequencing and retains ENA’s conflicting source declaration in explicit provenance fields instead of discarding it.

How CI uses these inputs

Input	Test path	What a green result means
Small deterministic synthetic reads	Required public Simulate Reads, checksum, FastQC, Study Demo Report, and reduced nf-core/mag wiring smoke	Packaging and application wiring work on the tested commit. The reduced MAG run is not a full biological MAG analysis.
Small deterministic synthetic reads	Private core matrix for `fastq-checksum`, `fastqc`, `multiqc`, `nanoplot`, `reads-qc`, `simulate-reads`, and `study-demo-report`, locally and in real outer Slurm plus an installed app	Only the matching green private run, propagated to the public mirror check, proves the Slurm acceptance gate.
Mouse PRJDB6165 selection above	`fastq-checksum`, FastQC, `reads-qc`, and Study Demo Report	Manual opt-in in the private Slurm workflow. These continue-on-error diagnostics are not part of the required `main` gate.
Human PRJEB54724 selections above	Kraken2/Bracken on both instrument-specific orders; explicitly requested SubMG/MEGAHIT submission for the shared study	Manual opt-in. The taxonomy path needs the runner database; SubMG also needs ENA test credentials and fails closed without a valid MEGAHIT assembly.
nf-core/mag `test_minigut` public test pair (`DEV-MAG-ILMN-001`)	Reduced MAG/MEGAHIT assembly smoke; optionally reused for read-cleaning diagnostics	Manual extended Alma workflow and reduced public wiring checks. It is a tiny integration fixture, not a full biological MAG analysis.
Five-sample Gemma ONT profile fixture (`DEV-GEMMA-ONT-001`)	FastQC, checksum, Study Demo Report, and MetaxPath	Manual private hosted-profile workflow. Its gated source is supplied by the CI install profile; it is not one of the public ENA selections above.
Deterministic host/microbial spike (`DEV-RC-SPIKE-001`)	Read-cleaning contamination-removal count	Manual private extension, conditional on a staged Kraken2 database and a sufficiently large node.

Optional MAG and MetaxPath wrapper legs say SKIPPED when disabled and say OK only after validating a real Slurm PipelineRun. The SubMG real-assembly path requires successful MEGAHIT output and never silently substitutes a synthetic FASTA; explicitly synthetic contract fixtures remain labelled as such.

For repository-level detail, see the example-dataset and CI provenance reference .