Scientific discoveries rarely happen in isolation. A finding from one laboratory can inspire an experiment in another, while an observation buried in a decades-old paper might provide the missing piece of an entirely different research question. Sharing scientific findings and building on what others have learned is fundamental to advancing modern biology.

One influential example of this approach came during the Human Genome Project. In 1996, participating scientists established the Bermuda Principles, which called for newly generated human genome sequence data to be made publicly available as quickly as possible. This commitment to rapid data sharing helped establish a model for scientific collaboration that continues to shape genomics research today.

But making scientific information available is only the first step. Researchers also need ways to find it, understand it, and connect it to what is already known.

That’s where biological knowledgebases come in. These specialized databases bring together decades of scientific research, organizing findings about genes, proteins, and biological processes into reliable, searchable resources. They help scientists find connections between discoveries that might otherwise remain scattered across thousands of publications.

In a new Perspectives piece published in GENETICS, researchers warn that these essential resources are facing a growing funding crisis. Just as scientific research and artificial intelligence (AI) become increasingly dependent on their information, the databases themselves are struggling to survive.

Connecting the dots of scientific discovery

Valerie Wood, a researcher at the University of Cambridge and lead author of the commentary, helps maintain one of these resources. The knowledgebase known as PomBase organizes research on the fission yeast Schizosaccharomyces pombe, a tiny organism that has helped scientists understand fundamental biological processes, including how cells grow and divide.

Although yeast might seem far removed from humans, approximately 70% of its proteins have evolutionarily related counterparts in humans. This makes fission yeast an important model for studying cellular processes that are shared across species and relevant to human biology and disease.

Wood recalls a time when a graduate student might spend four years studying a single gene. Today, researchers increasingly investigate complex biological questions involving dozens or even hundreds of genes and proteins. Resources like PomBase support that work by connecting findings from individual laboratories into a larger picture of how cells function.

For example, when a cell divides, a ring of proteins assembles around the middle and tightens to help separate it into two cells. Understanding this process requires studying numerous proteins, their interactions, and what happens when particular genes are altered. A researcher investigating this process can use PomBase to identify genes associated with abnormal assembly of this ring, examine the effects of different mutations, and explore connections to other proteins involved in cell division. Rather than searching through countless publications to piece together these relationships, researchers can explore the accumulated evidence in one place. These connections help scientists identify gaps in knowledge, develop new hypotheses, and determine which experiments to pursue next.

But PomBase is more than an archive of past discoveries. Wood describes it as “a living system that grows and changes day by day.” As researchers uncover new information, its expert curators continually update the database, connecting new findings to existing knowledge and correcting earlier assumptions when the evidence changes.

The work requires considerable scientific expertise. Curators read published studies, interpret experimental results, standardize terminology, and resolve inconsistencies between findings. This process, called biocuration, transforms individual observations into information that researchers and computer programs can reliably use. And its impact extends well beyond fission yeast.

Information curated by PomBase and FlyBase, a similar resource for fruit fly research, contributes to Gene Ontology, an international system for describing gene functions. Through comparisons between evolutionarily related proteins, annotations produced by each resource support more than a million additional annotations in UniProtKB, a major international protein knowledgebase. In other words, discoveries involving yeast and fruit flies help scientists interpret biological information across many other species, including humans. This interconnected system is becoming even more important with the rapid expansion of AI in scientific research.

Why AI needs human-curated knowledge

AI tools can analyze enormous amounts of biological information, identify patterns, and help researchers generate new hypotheses, but the reliability of the information they generate depends on the quality of the information they use. Without accurate, regularly updated descriptions of genes and proteins, consistent terminology, and connections to experimental evidence, computational systems risk reproducing or amplifying incorrect information.

Wood reports that PomBase is now accessed approximately 50 times more than two years ago, an increase she believes is largely driven by automated AI activity. 

While AI is sometimes presented as a technology that could replace human work, many biological AI applications depend on information that human experts have carefully organized, interpreted, and verified. Yet the resources providing that information face serious financial challenges.

What happens when scientific infrastructure loses funding

PomBase is maintained by a core team of just three people: two curators and one developer. Its current funding is scheduled to end in January 2027, and Wood is working to secure additional support. Other major resources have suffered funding cuts and disruptive funding gaps. In 2025, FlyBase lost funding for five curator positions following the termination of National Institutes of Health overseas subawards. The HUGO Gene Nomenclature Committee (HGNC), which maintains standardized names for human genes, also experienced a prolonged funding gap before securing a new NIH award.

These databases serve different purposes, but their importance reaches across the scientific ecosystem. HGNC, for example, ensures that researchers, databases, and clinical systems use consistent names for human genes, such as BRCA1. Without standardized names, connecting genetic information across studies and institutions becomes much more difficult. And when funding disappears, the consequences extend beyond losing access to a website.

Wood explains that PomBase’s existing information could be preserved even if active funding ends. However, without expert curators, new discoveries would no longer be incorporated at the same pace, outdated information could remain uncorrected, and valuable connections between studies might go unnoticed. Researchers would then have to spend more time finding and interpreting information themselves, potentially duplicating work already performed elsewhere. The specialized expertise developed by database curators over many years would also be difficult to replace.

Why are such valuable resources struggling to survive? Part of the problem is that scientific funding systems often prioritize new research over maintaining existing infrastructure. As Wood puts it, funders “want to fund the new building that’s fancy, but they don’t want to repair the roads.”

While their financial support typically comes from individual countries or institutions, biological knowledgebases also serve researchers worldwide. Their benefits are shared globally, but the responsibility for maintaining them often falls to a relatively small number of organizations.

Wood and her colleagues argue that biological databases should be treated as essential scientific infrastructure rather than short-term research projects. They recommend sustained funding, greater recognition of the scientists who maintain these resources, and funding decisions that consider how losing one database can affect many others. The stakes are particularly high as governments and research institutions invest heavily in genomics, precision medicine, and AI-driven drug discovery. All of these ambitions depend on access to reliable, interconnected biological knowledge.

The Human Genome Project demonstrated the power of sharing scientific information so others could build upon it. Biological knowledgebases help extend that principle by transforming individual findings into an evolving body of knowledge that researchers around the world can use.

As scientists investigate increasingly complex biological questions, these resources are becoming more important than ever. Biological knowledgebases hold decades of hard-won data. Protecting them keeps that data working for discoveries no one has yet imagined. To realize that potential, funders, research institutions, and scientific societies must recognize these resources as essential infrastructure and invest in the experts who keep them current.

, , ,

Jenny Montooth is a science communicator and co-founder of The Science Underground. Jenny is passionate about using creative and innovative strategies to make complex topics engaging and accessible to wider audiences. You can follow Jenny on LinkedIn.

View all posts by Jenny Montooth »

Read more in