Chromatography / Data integration
From instrument evidence to connected records.
Bring chromatography curves, peak tables and run reports into your PD experiments. Seal connects the exported evidence to the experiments, samples and review workflows your team already uses.
- ExportThe vendor software saves the measurement or evaluated result.
- CollectSeal IoT on the site computer uploads the completed file.
- MapA Seal Script validates the data and writes your chosen fields.
Instrument acquisition, analysis and control stay in the vendor software. This connection uses the existing Seal file and entity APIs. Configure the collection and mapping for your installed software and export layout; selecting the instrument in the directory does not install a parser or start acquisition.
What evidence can ÄKTA UNICORN provide?
- Source
- Cytiva / ÄKTA UNICORN
- Evidence to collect
- ASCII curves and peak tables; keep the UNICORN result archive separately.
Export options depend on the installed software and licences. The ÄKTA UNICORN setup guide covers source preparation and collection in Seal.
Curves, peak tables and the complete UNICORN result serve different purposes. Link the expected files to one run and check their identities before declaring the import complete. Converting a volume boundary to time requires the relevant flow history, not an assumed constant flow.
Keep one representative export and its matching instrument report. Record the software version, licensed export options, chosen columns or sheets, and which results the file actually contains. Check the layout after a software or report-template change.
Vendor reference: Cytiva UNICORN 7.9 Evaluation Manual. Product pages describe the source software; verify detailed export options against the manual for your installed version.
Choose the records you need.
One chromatography run linked to the experiment, column and source files; add fraction or peak records only when you need to track them individually.
These are suggested fields, not a required schema. Reuse your existing sample, experiment, equipment and method references. Preserve complete exports as File content; store selected values as typed fields when people need to filter, compare, calculate or review them.
| Information | Seal fields | Mapping rule |
|---|---|---|
| Run and fractions | Reference and text fields | Resolve experiment, column, run and tracked fraction identities. |
| Detector curves | File and unit fields | Keep channel names and time or volume axes; preserve cropping and export reduction. |
| Peak summaries | Number and unit fields | Map selected evaluated peaks with integration boundaries and method context. |
| Source and revision | File and text fields | Retain original exports, source identity and the evaluation revision used for these results. |
Context to retain: Run identity, column and method revision, channel names and units, time or volume axis, fraction identity, and evaluation revision. A missing, failed or qualified result should stay distinguishable from a valid zero. Require explicit sample identities and reject unexpected layouts before the Script writes results.
Ready to configure the connection?
The platform docs contain the setup steps: prepare the source export, choose your System and template, configure collection, and test a mapping Script against your actual fields.
Follow the ÄKTA UNICORN setup guide ↗
Start with one representative export and its matching vendor report. The on-site equipment docs cover IoT installation, credentials and the first upload. The instrument mapping docs explain File content, typed records and ordinary Seal Scripts.
The site computer connects to Seal over outbound HTTPS. A cloud Script cannot read the instrument’s local drive. Collection frequency follows the export workflow: a folder watcher observes saved files and does not request new measurements.
Keep the source files until delivery and mapping are verified. For frequent or large datasets, choose useful summaries or intervals, retain coverage information and measure the intended volume before unattended operation.
Use the evidence in your existing workflow.
Imported records use normal Seal permissions and references. Compare typed results in views and formulas, link the analysis to a procedure, and use @ references to give neil the accessible samples, experiments and source evidence needed for a question.
Compare these ÄKTA runs and the selected peak and fraction results. Keep the method revision, axes and units visible.
neil can help write the mapping Script from a representative file and the actual fields in your System. Ask for a proposed mapping and a dry run first. Review its sample matching and scientific assumptions before allowing writes. An instrument name alone does not define a file layout or an analysis method.
Define which fields the import owns. Preserve user-entered notes, review decisions and formula-controlled values. Keep source acquisition time, analysis revision and import time separate so later reanalysis remains understandable.
Prove the workflow with representative data.
A useful first acceptance check
Use a completed run with two detector channels and a known peak. Compare first and last coordinates, exported point counts and the peak result. Repeat with a volume axis and a revised evaluation.
- Compare the result, not only the upload. Check the destination sample, selected values, units, qualifiers and linked File against the vendor report.
- Replay identical input. Verify the Script recognizes the source identity and revision before creating another result. File-upload calls alone do not prevent duplicate scientific records.
- Reprocess the analysis. Export a changed evaluation and check that it retains its prior evidence. Decide explicitly whether this creates a linked analysis or updates allowed fields under your normal review rules.
- Interrupt and recover. With disposable data, test a network interruption and restart. Also test failure after upload but before mapping completes. Confirm retained files, visible failures and correct retry behavior for the installed IoT app and Script.
Reconcile the source run or sample list with the imported analyses. A watcher cannot recover measurements the instrument never exported. Retain the checks and their evidence with the configured connection before relying on it for your intended use.
When the result is missing or looks wrong.
- No file arrives
- Confirm the instrument created a completed export in the watched folder. Check the IoT account’s file access, upload logs, destination System and template permissions.
- The File exists, but result fields are empty
- Check whether the mapping Script ran and recognized the export. Resolve missing sample identities or fields explicitly; uploading bytes does not populate an analysis automatically.
- The values differ from the instrument report
- A volume axis is not a time axis. Export reduction, cropping and normalization change the curve. Record those choices; never silently treat a reduced export as the original acquisition. Compare the exact export, report and evaluation revision using the field rules above.
- Results are repeated or an analysis is missing
- Check export selection, source identities and revision matching. Reconcile at the analysis level; a filename or successful network response alone cannot establish complete coverage.