Can EU Seize the Chance of Pooling Genomic Data?

Around the industrialized world, the consensus is stronger every day that the future of healthcare is digital, and that effective data-sharing is pivotal to making it a reality. And in Europe, a genuinely innovative attempt to share genomic data across national borders is getting underway. But in the patchwork-quilt diversity of the old continent, every day is revealing just how much has to be resolved to make the concept work in practice.

A declaration of cooperation on cross-border access to genomic information, signed by nearly all the 28 European Union countries since 2018, is making some headway against multiple technical, administrative, and political barriers in pursuit of its goal of making one million genomes accessible in the EU by 2022. But it is still running into new impediments with each step that it takes, raising questions about the feasibility of its further targets for 2027 of seeing 100 million citizens with personalized health systems, and making data from 10 million citizens available for new discoveries.

The ambitions are noble, in terms of “supporting data infrastructure to advance research, disease prevention, and personalized health and care in key areas, including rare, infectious, and complex diseases.” The declaration speaks glowingly of how “the digital transformation of health and care and, in particular, the use of genomic medicine, will help health systems,” and ensure that the EU “remains competitive in the global race to advance personalized medicine.”

When the declaration was signed, its plan of opening up national and regional banks of genetic and other health data was hailed by EU officials as a game-changer for European research and clinical practice. “This initiative can boost the development of public health for the benefit of EU citizens,” said EU Commissioner for Health Vytenis Andriukaitis, vowing support for defining a governance model for cooperation, developing technical specifications, and promoting interoperability of registries and databases.

The rationale is not in dispute. There is broad agreement that digital applications and data have the potential to improve health and ease some of the well-known challenges faced by health systems: the budgetary pressure of ageing populations and chronic diseases, unequal quality and access to healthcare services, and shortages of health professionals.

The practical factors

But once the glow of the signing ceremony had faded, the real work began of identifying and solving the problems. Conspicuous among the challenges in whole gene sequencing were the different sequencing techniques commonly used, the lack of standardized measurements for data quality, quality of samples, or for storage, and the absence of clear methodologies for measuring the benefits.

The decision was made to tackle the deficiencies in both research and clinical practice, with a focus on capacity building, training, harmonization, policies for sequencing, intellectual property rights, interoperability, data governance, data quality control, patient engagement, ethics, and regulations.

Some best practices for genome sequencing practices were highlighted in national or regional initiatives such as the UK 100,000 Genomes Project, RD-Connect, and SciLifeLab sequencing facilities, and it was agreed that

genomics data should be findable, accessible, interoperable, and reusable using a federated database model, in which every country keeps their data but provides controlled access.

But questions remain over the type of data to be accessed-minimum genomic data sets, variants, or full genome, and what type of clinical or wider healthcare data. And procedural questions came to the fore-not least, for the one million genomes to be used for research purposes, it will be critical to obtain consent, and therefore to move toward some harmonization of consent forms.

Across a series of meetings, officials from the participating countries have taken stock of their growing awareness of the opportunities, reviewed the divergences in their current practices, and chiselled away at the issues they confronted that impeded closer cooperation. They have been gratified to find some degree of commonality in sequencing techniques, data/sample quality checks, and data representation standards used, giving rise to optimism that interoperability should in principle be easier. Most countries, it has emerged, have standards for interpretation of sequences, and most link genomic data to other health data, such as demographic information, patient registries, clinical reports, or biobanks. Many countries also have genomic training of healthcare professionals in place, often with functioning accreditation systems.

Drilling deeper

To advance the work, the countries have set up a series of working groups exploring specific areas ranging from broad governance to ethics, and from privacy to the development of infrastructure. 

They have also been holding discussions with an alphabet soup of organizations with a stake in this wide field, such as the European biobank and biomolecular organization  BBMRI-ERIC, the European Bioinformatics Institute (EMBL-EBI), GA4GH Mission (the global alliance for genomics and health), or ELIXIR, the connection among national bioinformatics centers, which is currently linked to projects being conducted in the Czech Republic, Denmark, France, the Netherlands, England, Finland, Estonia, Scotland, Sweden,  Slovenia, and Hungary.

The discussions are now, in the second half of 2019, moving on to defining genomics data and clinical information standards geared toward specific disease communities, designing common application programming interfaces to enable remote data discovery and access, and federated cloud computing environments with secure access to raw data and interoperable results. 

And now the focus is shifting to ways of ensuring the maximum impact of data, and timely access to the data for the three principal groups envisioned: healthcare bodies, universities, and commercial entities. Effective data management and security are under study, and agreement is being pursued on standards that are persistent, stable, and fit for purpose. The working groups are at last drafting deliverables for the short, mid and long term, as the starting point for a road map for the entire project. They are also identifying experts, and aiming to map available genomic data sets on a national level, so as to understand how far there is to go toward the goal of  more than one million human genomes.

The practical questions include whether extra efforts should be made to generate all the one million genomes through a self-standing new project, or whether it is enough to build on ongoing initiatives-for instance BBMRI already has 100 million samples. Decisions will also have to be made over what amount of clinical and phenotypic data would have to be linked to genomics data as a minimum for the data to count toward the ultimate goal. 

The working hypotheses developed will now be tested in three use cases-current suggestions include one covering personalized prevention through population-based sequencing, one on rare diseases, and a third on cancer. The aim is to demonstrate the potential benefits, iron out wrinkles in the chosen approaches, and develop incentives for the involvement of researchers and industry, as well as of citizens and healthcare providers. And over the rest of 2019 and 2020, a series of workshops and conferences are now being planned.

Funding and beyond

The inevitable issue of funding also preoccupies the participants. Possible financial support is being explored via the EU’s research fund, Horizon 2020, and its successor program now being finalized, as well as through the EU’s regional development funds. Some of the European Commission’s own digital health strategies also offer hope of gaining backing. Member states will be approached for assistance. Guidelines will be drafted and published for Good Genomics Practice, covering sample management, data generation, and data analysis, and plans are afoot for a meta data structure that details the processing, sequencing, and analysis work carried out on samples. 

But Europe is not alone in pursuing this course. There are plenty of other efforts underway around the world, and the window of opportunity for Europe is not going to be open indefinitely. One million genomes is a fine slogan and a worthy target-but it is only a start. 

Reflector is Pharmaceutical Executive’s correspondent in Brussels