The PFMG2025 has produced tens of thousands of clinical results; their use is already operational for defined indications, while their expansion remains underway. Two national sequencing laboratories, AURAGEN and SeqOIA, are operational; €239 million has been invested in the France Médecine Génomique 2025 program as of December 31, 2023, and the French genomics market is already worth $0.54 billion in 2025. Interoperability, financing, and tool deployment remain challenges, but dedicated national infrastructures and tools exist. Governance constitutes one of the challenges in deployment.
The Essentials
- France has invested heavily in genomic sequencing capacity; the absence of digital interoperability prevents its routine clinical use.
- Twelve SeqOIA and AURAGEN centers operate within the PFMG2025 program, funded to the tune of €239 million (source: Ministry of Health / SEQOGEN project, Annals of Oncology).
- The blockage is institutional: genomic platforms, patient records, and billing systems operate in separate circuits.
- The United Kingdom solved this problem by planning digital architecture upstream, via Genomics England; France and Germany illustrate the cost of failing to make this decision.
- For 2030-2035, it remains to be determined whether an infrastructure designed for research can be repurposed as a mass health service or whether a different architecture is required.
France Built Twelve Laboratories Without Building the Network That Connects Them
The France Médecine Génomique 2025 program developed national sequencing capacity. The two operational platforms are SeqOIA in Île-de-France and AURAGEN in Auvergne-Rhône-Alpes; AURAGEN notably covers the Grand-Est region. For patients with rare cancers, undiagnosed genetic diseases, or tumors resistant to standard treatments, this infrastructure represents a genuine medical resource.
The problem emerges at the next step. Recording the data produced by the platforms in medical records may not be automatic. Permanent pricing for comprehensive genomic testing is undergoing evaluation; the complete absence of hospital coding cannot be ruled out. The data is intended to feed the national CAD; the infrastructure is being deployed and may not necessarily yet cover all therapeutic monitoring uses. The modalities for data transmission and integration vary depending on care pathways and deployed tools.
SEQOGEN is an evaluation project for the platforms; available sources do not permit citing it as precise proof of an interoperability ceiling. The technical infrastructure exists. Systemic integration was designed into the program, but its operational implementation is progressive and incomplete.
Comparison with Genomics England
The United Kingdom had crossed the threshold of 100,000 genomes, not one million, in 2018-2019. This figure is often cited as a measure of British success. It is a consequence, not an explanation.
The United Kingdom created Genomics England in 2013 and developed a national genomic research infrastructure; this does not correspond to a unique centralized architecture for NHS clinical records.
Data infrastructures were a factor in British deployment, among other institutional, financial, regulatory, and clinical factors. The PFMG2025 planned a national data infrastructure from its design; the CAD exists and is being progressively deployed, even if its functional coverage may still be incomplete.
Germany finds itself in a similar situation. Its national medical genomics program has also produced genuine capacities and encounters the same silos between Länder, university hospitals, and insurance systems. German health federalism even amplifies the problem: each Land has its own interoperability standards, or lacks standards entirely.
Interoperability Is a Political Choice, Not a Technical Question
It would be convenient to explain this blockage as a technical difficulty. Genomic data are voluminous, heterogeneous, subject to particularly strict protection constraints, GDPR applies in full, with additional requirements for health data. All of this is real.
The United Kingdom has developed significant national capacities but continues to address interoperability problems. The NHS continues to deploy tools intended to improve connections between care and patient data. Technical challenges add to governance and deployment challenges.
Governance constitutes one of the challenges in deployment. Building an interoperability system between genomic platforms, medical records, and billing databases requires resolving questions that concerned actors often leave open: who controls the data, which organization bears legal responsibility in case of error, and under what terms do public hospitals, private groups, and university platforms share a common infrastructure.
Billing for genomic procedures must also be integrated into a nomenclature designed for standardized medical procedures, which has not been resolved.
These questions are uncomfortable because they redistribute power and responsibility among actors who do not necessarily have an incentive to coordinate spontaneously. The public program finances technical capacity and organizes coordination; in France, this coordination was piloted at the national level, even if its execution and operational maturity can be debated.
The Concrete Cost of Silos for Patients
One figure helps measure the gap between ambition and results. The French genomics market reaches $0.54 billion in 2025. A fraction of this volume corresponds to procedures genuinely integrated into standardized care pathways. The rest belongs to research protocols, exceptional situations handled manually, patients for whom access to sequencing depends on the willingness and resources of their local care team.
For a patient with a rare cancer, the delay between the request for sequencing and receipt of a result usable by their oncologist can vary depending on care pathways and deployed tools. In some cases, this delay is clinically neutral. In others, it is not.
For undiagnosed rare diseases, data can be preserved and reanalyzed; their transfer to the CAD for research depends on consent, while clinical results can feed into BNDMR unless opposition is registered. The interpretive and clinical value of a genome generally increases when genomic data can be linked to reliable and relevant clinical data. Comparison and reanalysis are planned in the system, even if they are not necessarily available uniformly for all cases.
This finding could be connected to the broader movement around public data as infrastructure for progress, a question that the Sino-American competition in quantum computing also raises, at another scale: investments in raw capacity produce their effects only if the usage ecosystem is built in parallel.
Toward 2030, Two Trajectories for French Genomic Infrastructure
For the coming decade, it remains to be established whether an infrastructure designed for research can be repurposed as a mass health service, or whether the architectural choices made between 2016 and 2025 have locked in a trajectory that cannot be exited without a clear break.
Two scenarios are emerging, and they are not symmetrical.
The first assumes that France decides, in the coming years, to treat interoperability as a public investment priority on the same level as sequencing itself. Several conditions make this scenario plausible. The digital health space, launched by the government, creates a shared medical record infrastructure that could serve as the backbone for genomic integration, provided that interoperability standards are defined and made mandatory for labeled platforms. Current work on certification of hospital software points in this direction. The Ministerial Delegation for Digital Health has identified interoperability as a priority workstream.
If these efforts converge and if an authority with a clear mandate is designated to pilot integration, on the model of Genomics England, France could achieve by 2030 a level of clinical integration comparable to what the United Kingdom has built over fifteen years.
The second scenario is that of persistent fragmentation. Platforms continue operating in semi-closed circuits, each developing its own interface tools with partner hospitals. The result would be a series of islands of excellence, a few well-connected university hospitals, a few well-organized cancer networks, and a large majority of patients for whom sequencing remains inaccessible in routine care. Public investment would have produced a prestigious infrastructure, useful for research and a limited number of patients, but without transforming the healthcare system.
The difference between the two trajectories hinges on a governance decision that neither technology nor financing can replace. It hinges on whether an actor with a mandate, sufficient authority, and resources is tasked with building the bridge between genomic platforms and the rest of the health system, and whether this actor can impose common standards on institutions that have until now functioned independently.
This type of decision resembles other industrial arbitrations where public power must choose between financing capacity and directing use. Export controls as a tool of industrial policy illustrates, in another domain, how a governance architecture thought out in advance radically changes the effect of an investment. In genomics, the architectural choice must be made now: the platforms are there, data is accumulating, and the window for deciding on coherent integration rather than multiplying disparate local solutions is closing as each actor consolidates its own systems.
Accumulated Data and Its Potential If It Circulates
The most immediate case is oncology. Rare solid tumors, cancers resistant to standard immunotherapies, atypical genomic profiles: all situations where cross-referencing genomic data with large-scale clinical data would allow identification of responder patient subgroups, refinement of protocols, and reduction of delays in access to targeted therapies. The United Kingdom has developed integrated data cohorts capable of supporting pharmacogenomics research.
The next case concerns rare diseases. France has approximately 3 million patients with rare genetic diseases, of which a significant portion remains without precise diagnosis. An integrated genomic information system would allow real-time comparison of sequencing results with clinical data from similar patients, an approach that European initiatives such as the ERDERA project actively develop, but which requires a national infrastructure capable of contributing to it.
The public health challenge by horizon 2030-2035 goes beyond rare diseases alone. Genomic medicine is becoming progressively relevant for common pathologies: certain breast, colon, and lung cancers, cardiovascular diseases with a genetic component. If sequencing costs continue to fall at the current rate—they have been divided by several thousands since the first human genome was sequenced in the early 2000s—the question will no longer be whether France can afford to sequence, but whether its health information system is capable of absorbing and using these data.
The question posed by the SEQOGEN project, implicitly, is ultimately broader than genomics. It concerns the capacity of public health systems to extract medical value from the investments they make in biomedical technologies. The integration of AI in complex organizations poses a structurally similar problem: the value of a sophisticated tool depends on its integration into existing workflows, not on its intrinsic performance. In genomics as elsewhere, technology is ready before the organization is.
Sources
- SEQOGEN project, Annals of Oncology: https://www.annalsofoncology.org/article/S0923-7534(25)03824-4/fulltext
- Ministry of Health, France Médecine Génomique 2025 Program (PFMG2025)
- Genomics England, annual reports, NHS England
- Ministerial Delegation for Digital Health, digital health space (Mon Espace Santé)
- ERDERA, European rare disease network, roadmap 2021-2027
