Summary

• Horizon Europe is directing €65 million towards critical life sciences data and knowledge infrastructure.
• GENCODE, UniProt and Europe PMC account for €57 million, with another €8 million available through an open call.
• The intervention responds to concern that widely used scientific resources remain dependent on fragmented, short term funding.

The European Commission is directing €65 million of Horizon Europe funding towards scientific data resources that underpin modern life sciences, biotechnology and a growing share of biological AI.

Three established resources account for €57 million. GENCODE is due to receive €19 million for genomic annotation, UniProt €30 million for protein sequence and function data, and Europe PMC €8 million for access to life sciences literature and links between publications and underlying research. A further €8 million will be available through an open call for other resources judged critical to European research.

Although the sums are modest beside spending on laboratories or computing capacity, the intervention targets infrastructure that is difficult to replace once researchers and companies have built it into their workflows. Biological databases require continuous curation as new papers appear, annotations change and scientific understanding develops. Hosting the records is only one part of keeping them useful.

The funding also reflects how AI changes the value of that maintenance work. More capable models can process larger volumes of biological information, but they remain dependent on the quality, provenance and structure of the material they receive. Increased computing power does not correct an incomplete gene annotation or restore a scientific resource that has disappeared because its grant expired.

Maintenance becomes an infrastructure problem

GENCODE, UniProt and Europe PMC perform different functions, yet modern research increasingly connects information across all three. GENCODE helps describe how genomes are organised and expressed. UniProt provides structured information on proteins and their functions. Europe PMC gives researchers access to scientific literature while linking papers with datasets, grants and other research records.

Keeping those services dependable requires specialist scientific and technical teams. Records have to be curated, classifications revised, software maintained and interfaces kept compatible with the systems that query them. As the quantity of biological information expands, so does the work required to distinguish established knowledge from incomplete or conflicting evidence.

That work has often been financed through grants that are much shorter than the useful life of the infrastructure itself. A database can become embedded across academic research, pharmaceutical development and software pipelines while the organisation maintaining it continues to compete periodically for the money needed to keep it running.

The Commission says fragmented and short term financing has placed critical resources under pressure. The new package therefore treats continuity, curation and technical maintenance as research infrastructure expenditure rather than assuming that once a database has been created its operating costs become marginal.

Open access remains part of the model. The three named resources serve international scientific communities rather than closed European user groups, and the Commission intends the funding to preserve broad access. That places the programme between two policy goals that can sometimes pull in different directions: strengthening European scientific capability while keeping research infrastructure available globally.

AI raises the cost of weak foundations

Biotechnology companies are increasingly using AI to search biological data for relationships that would be difficult to examine manually. Drug discovery, protein engineering and genomic medicine all depend to some degree on the ability to connect experimental evidence with consistent representations of genes, proteins and published research.

As more commercial systems rely on those resources, their economic importance expands beyond academia. A widely used database can sit several layers below a product without the end user knowing it exists, much like an open source software component buried inside an enterprise application.

The €65 million package does not solve the long term financing question permanently. The three named resources are still being supported through Horizon Europe rather than moved onto an indefinite operating budget, while the additional €8 million remains a competitive call. The programme provides a longer and more coordinated period of support without deciding how such infrastructure should be funded over decades.

That unresolved issue will become harder to ignore as Europe increases investment in biotechnology and AI. Research policy has traditionally rewarded new discoveries, facilities and projects more visibly than the maintenance of resources created by earlier programmes. Yet scientific systems become more dependent on those resources precisely because they have survived long enough to accumulate trusted information.

The Commission has linked the intervention to its wider life sciences and research infrastructure strategies, which put more emphasis on the systems required to translate research into medicines, biotechnology products and commercial services. Data now sits alongside laboratories, instruments and computing capacity within that infrastructure.

There is also a resilience argument. When a research community depends on a small number of internationally recognised resources, instability in their funding can affect experiments, software and commercial development far beyond the organisations maintaining them. Recreating years of curated scientific relationships would be considerably harder than moving files between hosting providers.

Europe’s latest biotechnology investment is therefore unusually focused on preserving what already works. Of the €65 million announced, most will maintain three resources with established international use rather than create new platforms. The decision recognises that AI and biotechnology need dependable scientific knowledge underneath them, and that maintaining that knowledge has become infrastructure work in its own right.