AI-aided vaccine platform aims to protect against virus families

Cambridge University researchers say an AI-aided vaccine platform could help protect against whole families of related viruses rather than one strain at a time. If the approach holds up in larger studies, its developers say it could shorten the lag between the emergence of a dangerous virus and an effective vaccine response.

Professor Jonathan Heeney, who leads the project at Cambridge’s Lab of Viral Zoonotics in the Department of Veterinary Medicine, told AFP that conventional vaccines often trail the pathogen they are meant to stop. He said vaccines are built from earlier strains, which means the virus circulating months later may differ from the version used to design protection.

Heeney compared the new method to having a master key for an apartment building. Rather than focusing on a single viral variant, he said, the platform uses AI to identify features shared across related viruses that the immune system can recognize.

How the platform is meant to work

According to Heeney, the Cambridge team combined available information about different viruses and used early AI tools to compare the parts most relevant to immune responses. The goal was to separate what changes between variants from what remains similar enough to serve as a broader vaccine target.

Heeney told AFP the strategy is designed to reduce the variability that makes strain-specific vaccines hard to keep current. He described the aim as a vaccine that the immune system can recognize across a range of related threats.

The technology has been used to make a universal Sarbeco coronavirus vaccine, developed by Cambridge scientists and the biotechnology company DIOSynVax. A trial involving 39 volunteers, sponsored by University Hospital Southampton and published in the Journal of Infection, reported no significant safety concerns, according to AFP.

The vaccine is now set to move into larger tests, AFP reported. The early study does not establish that the vaccine prevents infection or disease across a virus family; it addresses initial safety findings in a small human trial.

Ebola outbreak shaped the work

Heeney began work on the idea after the 2013 to 2016 Ebola outbreak in West Africa, where he was based at the time, AFP reported. Ebola had previously been associated with the Democratic Republic of Congo in Central Africa, and Heeney said the West African outbreak was initially mistaken for illnesses including Lassa fever, gastroenteritis and cholera.

The World Health Organization has said the outbreak killed about 11,300 people. Heeney told AFP that several months were lost identifying the cause before vaccine work could begin, during which the disease spread from Guinea into Sierra Leone and Liberia and killed many health workers.

Heeney said that experience convinced his team that vaccine development needed to change. He also pointed to population growth, cross-border movement and human intrusion into animal habitats as factors that can bring viruses from animal hosts into people who lack existing defenses.

Focus on future outbreaks

Heeney told AFP that influenza is one of his most pressing concerns because of its difficulty as a vaccine target. He also cited historic pandemics, including the Black Death and the 1918 to 1920 influenza pandemic, which he said killed an estimated 25 million to 50 million people worldwide.

Heeney said newer AI systems are now being used by his team to build a faster platform that can handle more data. He told AFP he hopes the work can help prove that the technology is safe, more effective and useful for a new era of vaccine manufacturing.

This story draws on original reporting from Medical Xpress.