Best Practices

Practical guidelines for getting the most out of SeqDesk — whether you are a researcher submitting orders or an admin running a sequencing facility.

For Researchers

Write descriptive order names

Order numbers are auto-generated (ORD-YYYYMMDD-XXXX), but a clear name like “Soil metagenome — Field Trial B 2025” makes it much easier to find orders later. Your facility admin sees these names too.

Fill in taxonomy early

Every sample needs a valid NCBI taxonomy ID for ENA submission. Use the organism autocomplete when adding samples — searching by common name or scientific name both work. For environmental samples, use metagenome terms like “soil metagenome” (taxid 410658).

Use the Excel import for large sample sets

If you have more than a handful of samples, prepare them in a spreadsheet first. SeqDesk supports Excel import with column mapping, which is faster and less error-prone than manual entry.

Keep sample aliases unique

Sample aliases must be unique across your orders. If you leave the alias blank, SeqDesk generates one automatically. If you supply your own, use a consistent naming scheme (e.g. PROJECT-SITE-001).

For Facility Admins

Set up departments before inviting users

Researchers select a department during registration. Having the department list ready avoids having to reassign users later. You can also enable department sharing in settings to let researchers within the same department see each other’s orders.

Configure the form builder early

The default order form works out of the box, but customizing it for your facility saves time later. Remove fields you don’t need, add custom ones (billing codes, grant numbers), and arrange them in a logical order.

Use test mode for ENA

Always start with ENA test mode enabled. The test server accepts the same XML but does not create real accessions. Once you have confirmed that submissions go through cleanly, switch to production.

Organize data storage paths

Set a clear, consistent base path for sequencing data (e.g. /data/sequencing). SeqDesk scans this directory for FASTQ files — keeping a predictable directory structure (by run date or instrument) makes file discovery faster and auto-detection more accurate.

Monitor pipeline runs, don’t just launch them

SeqDesk provides real-time pipeline monitoring with a DAG visualization and event feed. Check the progress view after launching a pipeline — early failures (missing inputs, configuration issues) show up quickly and are easier to fix than waiting for the full run to fail.

Metadata Quality

Follow MIxS standards

When creating studies, choose the right MIxS checklist type (e.g. Soil, Wastewater, Host-Associated). This determines which metadata fields are required and ensures your submission meets ENA standards.

Be consistent with units

MIxS fields often expect specific units (e.g. depth in meters, temperature in Celsius). SeqDesk stores units alongside values — fill them in consistently to avoid ENA validation errors.

Provide complete sample descriptions

ENA requires study descriptions of at least 50 characters and titles of at least 10 characters. Be descriptive — these become the public-facing metadata for your submitted data.

Pipeline Runs

Check read assignment before launching

Pipelines require reads assigned to samples. Before launching, verify that all samples in your study have their FASTQ files linked. Unassigned samples will be skipped or cause the pipeline to fail.

One study per analysis context

Create separate studies for different analysis contexts rather than putting everything in one study. For example, if you are running MAG on soil samples and a separate analysis on water samples, use two studies. This keeps results organized and makes ENA submissions cleaner.

Review assembly quality thresholds

After a MAG pipeline run, review the assembly and bin results. SeqDesk flags high-quality bins (≥90% completeness, less than 5% contamination) and lets you select a preferred assembly per sample. Taking a moment to review these before submission ensures only quality data reaches the archive.