Biobank Chain of Custody: Data Fields for Specimen Tracking
Biobank chain of custody is the unbroken, time-stamped record of who held a specimen, where it was, what condition it was in and what was done to it, from the moment of collection to its final use or disposal. In practice it is a data model: a fixed set of fields captured at every handoff, so that any sample, and any result derived from it, can be traced and trusted.
This guide sets out the fields that a working biobank chain of custody record needs, how physical custody of a tube connects to digital custody of the data it produces, and how to design the model so it holds up under audit.
Why Chain of Custody Matters Now
A specimen is only as valuable as the confidence researchers have in its history. If a tube spent six hours at room temperature during a courier delay, or was relabeled after an aliquot was split, every downstream measurement inherits that uncertainty. Without a complete record, a study team cannot tell a biological signal from a handling artifact.
Three pressures make this sharper today. First, international standards such as ISO 20387 for biobanking and the ISBER Best Practices expect documented, traceable processes for collection, storage and distribution. Second, ethics committees and data access committees increasingly ask how consent travels with the sample. Third, most specimens now generate digital data: genomic sequences, proteomic profiles, imaging. Once that happens, custody does not end at the freezer door. It continues into every system where the derived data is stored and analyzed.
The May 2026 UK Biobank incident showed what happens when the digital half of custody is weak: approved researchers could move data out through a normal workflow. A biobank that tracks its tubes carefully but loses sight of its data has only half a chain.
The Core Data Fields Every Record Needs
Field names vary between laboratory information systems, but the content is consistent. Group the fields by the question each one answers.
Identity Fields
- Specimen ID: a unique, never-reused identifier, ideally encoded in a barcode or 2D matrix code on the container.
- Parent ID: for aliquots and derivatives, the ID of the specimen it came from.
- Participant pseudonym: a coded link to the donor, never a name or national identifier.
- Specimen type: whole blood, plasma, serum, tissue, saliva, DNA extract and so on, using a controlled list.
Collection Fields
- Collection date and time, to the minute.
- Collection site and collector role.
- Container and additive type (for example EDTA or heparin tubes).
- Time to processing and temperature during that window.
Storage Fields
- Storage location down to freezer, rack, box and position.
- Storage temperature class and any excursions recorded by monitoring.
- Freeze and thaw count.
Transfer and Shipment Fields
- Sender, receiver and courier, each with a timestamp at handoff.
- Shipping condition (dry ice, liquid nitrogen dry shipper, ambient) and the data logger reference.
- Condition on receipt, including any discrepancy between manifest and contents.
- Material transfer agreement reference.
Processing and Derivative Fields
- Each processing step (centrifugation, extraction, aliquoting) with operator, protocol version and time.
- Quality measures such as DNA yield or RNA integrity.
- Identifiers of any data files generated, for example a sequencing run ID.
Consent and Access Fields
- Consent version and scope, plus any withdrawal date.
- Approved projects that have received the specimen or its data.
- Disposal or return date, with the reason.
The last group is often missed. A record that knows where a tube is but not what the donor agreed to cannot support a release decision.
Physical Custody Versus Digital Custody
The moment a sample is sequenced, the biobank holds two linked assets: the remaining physical material and a digital record that can be copied endlessly. The controls for each look different, but the principle is the same: every movement is recorded and every recipient is approved.
| Dimension | Physical specimen custody | Digital data custody in a TRE |
|---|---|---|
| What moves | Tubes, slides, tissue blocks | Analysis code and approved results |
| Who handles it | Collectors, lab staff, couriers | Approved researchers working inside the environment |
| Proof of handoff | Signed manifest, scan at receipt | Audit log of every query and every export request |
| Main risk | Temperature excursion, mislabeling, loss | Uncontrolled copying of record-level data |
| Control point | Material transfer agreement and receipt check | Airlock review of outputs before release |
| Can it be undone | Material can sometimes be returned | A copied dataset cannot be recalled |
The last row is why digital custody deserves at least as much design effort as physical custody. A shipped tube is finite. A downloaded dataset is not.
Designing a Chain of Custody Data Model That Holds Up
A good model is boring in the best sense: consistent, complete and hard to bypass. Use these principles.
- Record events, not states. Store each handoff as an immutable event with who, what, where and when. The current location is then derived from the event history rather than overwritten.
- Use controlled vocabularies. Free-text fields for specimen type or shipping condition make later harmonization painful. Map local terms to shared standards early.
- Capture at the point of action. Barcode scans at collection, storage and receipt beat transcription from paper hours later.
- Link physical and digital IDs. When a specimen produces a data file, store the file identifier against the specimen and the specimen ID in the file metadata.
- Make consent a live field. If a participant withdraws, the change should flag every specimen and every derived dataset at once.
- Plan for audit queries. Ask in advance which questions an auditor will ask, such as “show every movement of this sample” or “list every project that received data from this donor”, and make sure the model answers them in one query.
When biobanks later pool data across sites, these choices matter even more. Consistent fields are what make data harmonization across collections possible without months of manual cleanup.
Extending Custody From the Freezer to the Analysis
Traditional practice was to ship data the way samples are shipped: extract a dataset, transfer it to the researcher and rely on an agreement to govern what happens next. That breaks the chain, because the biobank can no longer see what happens to the copy.
A federated model keeps the chain intact. In a Trusted Research Environment (TRE), approved researchers bring their analysis to the data instead of taking the data away. Record-level data never leaves the source, every action is logged, and only aggregate results that pass an airlock review can leave. The digital custody record then looks much like the physical one: every access is an event with a named actor, a time and a purpose.
For a biobank, this has a practical benefit. The same questions that the chain of custody answers for a tube (who had it, when and under which approval) can now be answered for every derived dataset. A Federated Trusted Research Environment extends the same logic across several biobanks, so a study can draw on collections in different institutions or countries while each one keeps custody of its own data.
Practical Next Steps
- Audit one specimen end to end. Pick a sample collected last year and try to reconstruct its full history, including every data file it produced. Note each gap.
- Agree the minimum field set. Use the groups above as a checklist and decide which fields are mandatory at each handoff.
- Standardize vocabularies. Replace free text with controlled lists for specimen type, container, shipping condition and processing step.
- Connect the digital record. Make sure every sequencing run or assay output carries the specimen ID, and that access to those outputs is logged.
- Review release controls. Check whether derived data can still leave your environment as record-level files. If it can, the chain of custody ends there.
Chain of custody is not paperwork for its own sake. It is the evidence that makes a biobank’s samples, and the data they generate, worth using.
Frequently asked questions
What is chain of custody in a biobank?
It is the complete, time-stamped record of every person, location and process a specimen passes through from collection to use or disposal, so that its handling can be verified at any point.
Which data fields are mandatory for specimen tracking?
At minimum: a unique specimen ID, parent ID for aliquots, participant pseudonym, specimen type, collection time and site, storage location, every handoff with timestamps, and the consent scope that governs use.
How is chain of custody different from sample inventory?
Inventory tells you where a sample is now. Chain of custody tells you everywhere it has been, who handled it and what happened to it along the way.
Does chain of custody apply to data as well as samples?
Yes. Once a specimen generates genomic or other data, that data needs the same traceability: who accessed it, for which approved project and what left the environment.
Which standards cover biobank chain of custody?
ISO 20387 sets general requirements for biobanking, and the ISBER Best Practices give detailed guidance on collection, storage, transport and documentation.
How does a Trusted Research Environment help?
It keeps record-level data at the source, logs every action by approved researchers and checks outputs before release, so the digital chain of custody stays unbroken.