When the Grant Ends, the Obligation Doesn't: Closure Protocols for Genomic Research Data

A genomic research study rarely ends cleanly. The grant period expires, the final progress report is filed, and the laboratory doors close, but the freezer still holds aliquots, the sequencer still holds reads, and the consent forms still describe promises made to people who may not know the study has concluded. Federal policy has grown increasingly specific about how data should be managed and shared while an award is active. It has far less to say about who is responsible for those materials once the award is gone.

That asymmetry is the subject of this working paper. It is not a question of whether genomic data should be retained; long-term retention is often scientifically valuable and sometimes ethically required. The question is whether institutions have a documented, auditable answer to a simpler prompt: when this study ends, what happens next, and who decides?

The Gap Between Sharing and Stewardship

The NIH Data Management and Sharing Policy, effective since January 2023, pushed institutions to describe up front how data would be managed, preserved, and shared. That policy improved practice during the life of an award. It did not resolve the endgame. A data management and sharing plan typically addresses deposit timelines, repository selection, and access conditions. It less often addresses what happens when the principal investigator retires, the consortium's coordinating center closes, or the institution decides not to renew the underlying specimen repository contract.

The Common Rule and FDA human-subjects regulations govern informed consent and continuing review, but they do not prescribe a closure protocol. HIPAA governs covered entities and their business associates, not the downstream lifecycle of research data in every setting. Certificates of Confidentiality add protections against compelled disclosure, yet they do not designate a custodian after the study team disperses. The result is a patchwork in which the most consequential decisions—transfer, retention, destruction, and withdrawal—are often made informally, late, and by whoever happens to still be answering email.

Four Decisions Every Closure Plan Must Answer

A workable closure protocol need not be elaborate, but it should force four decisions into the open.

Retention. Will samples and sequences be kept, and for how long? If the answer is indefinite, the institution should be able to explain why, under what consent language, and at whose cost. Storage is not free, and unreviewed retention is not the same as stewardship.

Transfer. If data will move to a repository, another institution, or a commercial partner, the transfer should be governed by a data use agreement that carries forward the original consent terms, including any restrictions on commercial use or recontact. Transfer is where consent most often erodes, because the receiving party may never have seen the original form.

Destruction. If materials will be destroyed, the protocol should specify verification. A signed attestation that a freezer was emptied is not the same as a record showing which aliquots were destroyed, when, and by whom.

Withdrawal. Participants retain the ability to withdraw consent in most research contexts, subject to limits on withdrawing already-aggregated data. A closure plan should explain how a withdrawal request will be honored after the study team has disbanded, and which office will field it.

The Retirement Problem

Principal investigators retire, move, or change fields. When they do, the study's data often follow them physically or conceptually, even when the institution is the actual steward. Institutions should treat investigator departure as a trigger event, not an afterthought. That means a documented handoff: an inventory of samples and datasets, a consent audit, a named successor custodian, and a notification to the institutional review board and any repository partners.

Consortia face a harder version of the same problem. When a multi-site project dissolves, the coordinating center may hold the only consolidated record of who consented to what. If that center closes without transferring its governance records, the remaining sites may hold data they can no longer lawfully characterize. Dissolution agreements should therefore address governance records separately from scientific data.

Consent That Outlives the Grant

Consent language is frequently written for the study as imagined at enrollment, not as it will exist at closure. Broad consent helps, but breadth is not the same as durability. A participant who agreed to "future research" has not necessarily agreed to indefinite retention by an unnamed successor entity.

Institutions can narrow this gap by auditing consent forms at closure and mapping each promise to a post-closure practice. Where the original language is silent, the conservative course is to limit new uses rather than infer permission. Where recontact is feasible, a brief notice describing the study's closure and the fate of the data is good practice, even when not legally required.

A Model Closure Protocol

A defensible protocol can be built from seven steps: inventory all samples, sequences, and consent records; audit consent terms against proposed post-closure uses; designate a successor custodian with named authority; execute transfer agreements that incorporate original restrictions; document any destruction with verification; notify participants where consent or regulation supports it; and publish a closure record in the institution's research registry.

The final step matters more than it may appear. A public closure record gives participants, regulators, and future researchers a place to look. It also creates an institutional memory that survives staff turnover.

Where Clarification Would Help

Federal agencies could reduce uncertainty by issuing guidance on post-award retention periods, successor custodian requirements, and the treatment of withdrawal requests after study closure. Institutions need not wait for that guidance. The decisions are already being made; the only question is whether they are being made deliberately, documented, and made visible to the people whose genomes made the study possible.