Open Science, Governed Data: Legal Interoperability Across the Precision-Biomedicine Lifecycle
A response to the Davos conversations at Science House
Marietjie Botes
University of Texas at Austin, Department of Psychology, Pestilli Neuroscience Lab, Austin, United States
Published on September 11th, 2026
The Frontiers Policy Labs conversations on advancing precision biomedicine and omic mapping identify an important inflection point: scientific discovery is no longer the principal constraint. The harder task is building health systems capable of translating genomic, multi-omic, AI-enabled and highly personalised interventions into routine care through effective diagnostics, manufacturing, reimbursement, regulatory capacity and long-term monitoring. Legal infrastructure is integral to that task. Rather than treating data governance as end-stage compliance added after the scientific system has been designed, policymakers should recognise it as part of the system’s operating architecture.
1. From innovation to implementation
Precision biomedicine depends on large, diverse, and longitudinal datasets moving between laboratories, hospitals, regulators, repositories, manufacturers and, increasingly, AI developers. Scientific interoperability allows these systems to exchange and interpret data. But implementation also depends on whether institutions are legally permitted to collect, combine, transfer, and reuse those data. A technically interoperable system that cannot establish lawful access, define responsibility or manage cross-border transfers is not implementation-ready.
2. Open science does not mean unrestricted data
The call in The promise of omic mapping for open, standardised sharing of publicly funded research is compelling but “open” requires legal differentiation. Publications, research protocols, interoperability standards, metadata, and genuinely non-identifying aggregate outputs should ordinarily be open. Pseudonymised individual-level, genomic, and linked clinical data should normally be accessible through proportionate controls. Exceptionally sensitive or readily re-identifiable datasets may require trusted research environments, federated analysis, or other arrangements in which authorised analysis travels to the data.
This is consistent with the FAIR principles where accessibility may include authentication and authorisation. Open science is therefore not defeated by access control. Rather, responsible openness makes data discoverable and potentially reusable while ensuring that access conditions reflect privacy, research ethics, cybersecurity, intellectual-property, community and public-interest considerations.
3. Legal interoperability is scientific infrastructure
Legal interoperability concerns the compatibility of the rules under which data move and are reused. It requires clarity about which institution is the controller, custodian, processor or recipient; the lawful grounds for primary and secondary use; the scope of consent; cross-border transfer conditions; data-security obligations; intellectual-property and publication rights; and mechanisms for audit, enforcement and redress.
These requirements must be translated into usable infrastructure, including data-access committees, data- and material-transfer (and access) agreements, model contractual clauses, data-protection and transfer-impact assessments, audit trails and regulatory oversight. Work on the Model Law on Health Data Governance and Implementation Guide, South African Data Transfer Agreement 1.1, the BRIDGE legal toolkit for international neurodata sharing (in press), and South Africa’s regulatory whitepaper on advanced therapeutic products demonstrates a practical lesson: legal certainty can enable responsible sharing rather than obstruct it.
4. Governance must follow data through transformation
Precision-biomedicine governance must extend beyond a narrow focus on “raw data.” Genomic, imaging, or clinical data are linked and curated, transformed into inferences, risk scores and digital biomarkers, incorporated into AI models and embeddings, and ultimately used to guide therapies and coverage decisions. Each transformation can create new value and new risks, yet legal protection often becomes uncertain once personal data become inferential data or model outputs. Technical transformation should not automatically extinguish purpose restrictions, provenance requirements, accountability or benefit-sharing obligations. At each stage, a data-lifecycle approach should determine what the object has become, what it can reveal, who controls it, and which downstream uses remain permissible.
5. From advanced therapies to long-term outcome governance
The Frontiers Policy Lab’s discussion of regulatory T-cell therapies shows why lifecycle governance matters. Living and potentially durable therapies challenge conventional regulatory and reimbursement systems and require long-term outcome registries, real-world evidence and international data sharing. Similar issues arise across cell and gene therapies.
However, a registry designed for post-market surveillance should not automatically become an unrestricted research, payer, or commercial database. Law should distinguish clinical care, regulatory surveillance, health-technology assessment, reimbursement, and secondary research. It should define minimum necessary datasets, retention periods, access rules, withdrawal limits, permitted linkages, and safeguards against discriminatory use. Outcome-based reimbursement may be innovative, but it must not turn patients receiving advanced therapies into permanently surveyed data subjects.
6. From representation to reciprocity
The Frontiers Policy Lab rightly stresses globally representative research and equitable access. But representation in a dataset is not the same as equity. Populations in Africa and other Global South regions may become visible as data sources while remaining marginal in decisions about access, analysis, intellectual property, and commercialisation.
Equitable precision biomedicine therefore requires reciprocity, including local governance authority, meaningful community participation, research and analytical capacity, fair licensing, technology transfer, benefit-sharing arrangements, and investment in local regulatory and manufacturing infrastructure. Representation without governance risks reproducing extraction in a more technologically sophisticated form. Trust arises when individuals and communities can see not only how their data are protected, but how scientific value and health benefits return to them.
7. An operational legal rule
Policymakers should therefore link open-science mandates, regulatory approval, long-term outcome collection and coverage decisions to legally interoperable, independently governed, and privacy-protective data environments. These environments should combine open standards and non-identifying outputs with proportionate controls for person-level data, continuity of protection across the data-to-model lifecycle, and enforceable accountability for downstream use.
The practical rule is simple: open where safe, controlled where necessary, and accountable throughout.
Copyright: © 2026 [author(s)]. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in Frontiers Policy Labs is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
