More than five years after the COVID-19 pandemic began, researchers are still trying to predict the next global disease outbreak—one that could claim millions of lives and cause economic chaos. Polina Brangel, head of research, development and data innovation at the Oslo-based Coalition for Epidemic Preparedness Innovations (CEPI), is working to help prevent it.

Brangel began studying biotechnology engineering at Ben-Gurion University of the Negev, Israel, in 2007. She graduated from the combined programme in 2012 with both a bachelor’s and a master’s degree. Later that year, she completed a PhD in bioengineering and biomedical engineering at Imperial College London.

Nature Spotlight: Infectious Diseases

Although her role initially focused on developing biosensors for tumour diagnosis, Brangel later became involved in creating rapid diagnostic, or lateral-flow, tests for cancer biomarkers. When the Ebola outbreak occurred in West Africa in 2014, the World Health Organization (WHO) called for a quick, easy and accurate test for the disease. Brangel realized that the tools she had developed could be repurposed for pandemic response.

That decision led to a career in pandemic preparedness. Brangel joined CEPI in 2023. She has developed tests to study immune biomarkers in people who recovered from Ebola, investigated the immune response triggered by COVID-19 and is now focused on protecting people from “Disease X”.

Brangel spoke with Nature about pandemic preparedness, lessons from COVID-19 and the role artificial intelligence could play in preparing for future outbreaks.

What is Disease X?

The WHO has created a list of priority pathogens for which there are no available countermeasures, such as vaccines, diagnostic tests or treatments, and which are considered likely to cause a pandemic. The list includes the viruses that cause Ebola haemorrhagic fever, Lassa fever and Rift Valley fever.

What has COVID-19 taught us about pandemic preparedness?

As a serology expert at the WHO, I worked to understand the immune response caused by COVID-19. In infectious-disease research, scientists conduct seroepidemiological studies by taking blood samples from a group of people and testing their antibody responses. Researchers can then estimate how many people have been infected and extrapolate the results to the wider population.

Because many different tests were used to detect antibodies, estimates of infection rates varied between countries—from almost zero to about 90%. The problem was not necessarily that the tests were inaccurate. Rather, the different testing systems could not be compared directly.

Working with the former government agency Public Health England, we introduced international standards for benchmark assays.¹ Previously, each laboratory selected its own assay, making it difficult to compare results.

This experience showed us how important it is to harmonize data when using it as evidence for decision-making. Sound, strong science is not enough if the results cannot be compared.

How can researchers prepare for unknown pathogens?

CEPI has a programme focused on Disease X. Humans can be infected by viruses from 25 known virus families, most of which originate in animals. These include filoviruses, the family that includes Ebola; coronaviruses, such as SARS-CoV-2, which causes COVID-19; and orthoflaviviruses, including the viruses that cause dengue, yellow fever and Zika.

It is not possible to develop vaccines for all 25 virus families. Instead, researchers prioritize the pathogens they believe are most likely to infect humans and generate as much knowledge as possible about their genome sequences.

Learn more about the virus families that could pose future threats.

How can artificial intelligence support pandemic preparedness?

Mutations can help pathogens move from animals to humans. Researchers therefore use computer simulations to predict which pathogens might mutate in ways that enable them to infect people. By inputting the most likely candidates into an AI tool, scientists can begin developing and testing vaccine sequences and designing clinical trials.

AI models are being used to track zoonotic diseases. Will they be able to prevent the next pandemic?

AI brings three important capabilities: speed, which is particularly valuable during a pandemic response; the ability to integrate data from different sources; and federated access to information held by governments, commercial companies or other data owners.

What is CEPI’s Pandemic Preparedness Engine?

CEPI’s Pandemic Preparedness Engine for Disease X will be similar to a chatbot. Imagine that researchers in Malaysia identify a new Nipah virus and want to develop a vaccine. By entering the virus’s genome sequence and location, an AI agent could suggest ways to create or modify specific molecules, known as antigens, to trigger the appropriate immune response.