Culture trends and offline assays

Compare process trends and sample results for the same vessel and experiment. Preserve the timing and coverage behind the comparison.

Connect a culture trend with an offline glucose result

The same vessel position can occur in many studies. The combined key and separate timestamps make the comparison reproducible.

  1. Culture

    Experiment PD-12 / vessel 3

    Resolve a unique vessel record within the experiment.

  2. Trend window

    Culture hour 24–48

    Store channel, unit, aggregation and coverage with the selected summary.

  3. Sample

    PD-12-V3-S2 / collected at hour 36

    Reference the vessel and preserve the actual collection time.

  4. Offline assay

    Glucose / measured at hour 37

    Keep method, dilution and qualifiers; align by collection time for culture interpretation.

Start with the comparison your scientists need.

An Ambr experiment can combine vessel conditions, feeds, process measurements and offline cell-culture analysis. Decide whether the first workflow supports daily culture review, a condition comparison, a harvest decision or a final study report. That decision determines which summaries need queryable fields and which detailed exports should remain attached evidence.

Do not combine similarly named channels without checking their meaning. A dissolved-oxygen measurement, its setpoint and a controller output are separate observations. Keep the channel identity, unit and relevant method context. Where a reported value is calculated or corrected, retain enough source context for a scientist to understand that calculation.

Use experiment and vessel identity together.

A position such as ‘vessel 3’ is reused between experiments. Resolve the experiment first, then the vessel within it. Give samples their own identities and link each one to the parent vessel. Record sample collection time separately from the time a BioProfile, Vi-CELL or other offline analyzer measures it.

Define the time basis for comparisons. Preserve source timestamps and timezone information, and record the experiment event used as elapsed-time zero. Inoculation, culture start and the beginning of acquisition are not necessarily the same event. A timezone assumption can shift a feeding event against an assay result even when both imports succeed.

Choose export intervals and data volume deliberately.

Use the installed system’s supported export workflow and collect completed files through Seal IoT where the source is on site. A file watcher uploads saved evidence; it does not change the controller’s sampling frequency. Configure the Seal Script against a representative export from the actual software version and the channels selected for the study.

For dense trends, keep the full export as File content and map the intervals or summaries needed for the task. Specify the time window, aggregation method, expected coverage and missing-data treatment. A daily average of available points should not look complete when acquisition was absent for half the day. Measure the proposed import volume before deciding to create an entity for every observation.

Combine offline results without inventing continuity.

An offline glucose or viability measurement represents a collected sample. Link it to the collection event and retain dilution, method and quality flags from the analyzer. If a plot interpolates between offline results, make that calculation explicit and retain the underlying observations; the interpolation is not another measured value.

Use linked records to ask neil for comparisons across conditions and to prepare a review with references to the relevant exports. Define the population before comparing cultures: experiment, vessel set, time window and excluded results. A useful answer should expose missing samples or incomplete trend coverage rather than treating absence as a normal result.

Test it with your data.

Use these cases to agree and test the connection’s behaviour. They are proposed acceptance checks, not completed tests or automatic connector features.

Two sites use different time conventions

One export uses local time; an offline result uses UTC. Both relate to the same collection event.

Expected behaviour
Preserve the original timestamps and resolve their time basis before alignment. Do not infer a one-hour biological delay from a timezone difference.
Evidence to keep
Original times, timezone assumptions and the resulting shared timeline.

A daily mean hides an acquisition gap

A day’s trend contains measurements before and after a long gap.

Expected behaviour
Keep coverage and the missing interval beside the mean. Apply the study’s rule for whether that summary belongs in the comparison.
Evidence to keep
Expected window, observed coverage, aggregation method and any exclusion from the comparison.
More checks for this connection
  • Import two experiments that both contain vessel 3 and confirm their records remain separate.
  • Compare a known feed event and offline sample against the vendor timeline, including timezone and elapsed-time zero.
  • Remove a trend interval from a test export and verify that the mapped coverage shows the gap.
  • Replay a completed export after a connection interruption and confirm there are no duplicate observations or lost source files.

Before you connect

Does every measurement become a Seal entity?

Only if you configure that record model. Dense data can remain in source Files, with selected observations and summaries in typed fields. Choose the grain from the queries and review process, then test performance at the expected volume.

Can Seal control the bioreactor through this connection?

These export workflows collect process evidence. Acquisition and real-time control stay in the bioreactor or control system. Any separately supported dispatch or control interface needs its own scope and verification.

Check your system’s connection options.

Confirm the installed version, available exports and permissions. These pages cover the source interface and link to setup instructions.

For configuration: Seal Scripts and local-file collection.

Start with one export.

Bring the source evidence and the records it needs to connect to. For this workflow, a useful starting pack is:

  • One culture’s exported trends with channel names, units and timestamps.
  • Its sample collection log, offline results and feeding events.
  • The study’s comparison window, time origin and missing-data rules.
Discuss this connection