Baseimmune makes future-proof vaccines with AI

September 29, 2022 - 6 minutes
Image/Anastasiia Slynko

Baseimmune is using artificial intelligence (AI) to design vaccines that can protect against future variants of infectious diseases including COVID-19 and malaria.

Big data has taken off in the last decade, with dramatic improvements in bioinformatics, data science and AI. These advances allow researchers to store and handle more and more data generated by research in genomics, biochemistry and immunology.

The COVID-19 pandemic proved a major catalyst for the development of big data, especially when it comes to designing vaccines.

According to Joshua Blight, CEO and co-founder of the U.K. startup Baseimmune, most COVID-19 vaccines are aimed at the spike protein, a surface protein that the coronavirus uses to enter cells. New viral variants such as Omicron have been adept at evolving their spike protein to make it harder for vaccines and the immune system to recognize it. This means vaccine developers must chase the latest variants with newer, more specific vaccines.

“One of the mantras that I always say is: ‘Don’t forget that every disease is evolving to compete with the human immune system,’” said Blight. “If our vaccines aren’t able to reflect that, they’re not going to work long-term.” 

Baseimmune aims to develop vaccines that are one step ahead of infectious diseases such as COVID-19. The company crunches data from genomics, proteomics, epidemiology, evolution and clinical research. With the help of AI, Baseimmune turns this knowledge into future-proof vaccines that target what a pathogen might evolve into, not just how it looks now.

Content continues below

Related Content

Blight started off his career doing a PhD at the University of Oxford’s Jenner Institute, which is famous for its central role in developing the COVID-19 vaccine Vaxzevria. He met Baseimmune’s future co-founder, Ariane Gomes, at this center. But prior to setting up the company, Blight and Gomes continued their academic career in post-doc roles.

“During that time, I worked with the third co-founder, Phillip Kemlo, whom I’ve known since I was 14 in secondary school. He’s a computer whizz,” recalled Blight. “I said: ‘Let’s make something amazing.’ So we decided to build a completely new vaccine platform from scratch and formed Baseimmune.”

For Blight, forming a startup means that his team can be versatile, and develop many more vaccines than is normally possible in an academic setting.

“Academia is fantastic. The ability that it gives to research and explore things — I absolutely love it,” he said. “But I wanted to have a huge impact and I didn’t want to be restricted by one kind of research.” 

A large array of different vaccine delivery technologies came to the fore in COVID-19 vaccine development, including mRNA, DNA and adenoviral vectors. However, the design of antigen targets for vaccines — which is Baseimmune’s focus — hasn’t kept up.

“I like to think of a vaccine as having two parts. One is the delivery system, which is effectively a car. The antigen, which represents the disease, is the driver,” said Blight. “You need to have a great car and a great driver. We make the driver.”

Content continues below

Related Content

Rather than using a single protein as the vaccine antigen, as do many current vaccines, Baseimmune uses its AI know-how to create synthetic antigens that contain a mixture of the most important signals of the pathogen. This means it’s harder for the pathogen to mutate a single protein and avoid detection. The antigen can then be slotted into whichever delivery platform is most appropriate for the disease in question, and can cut the preclinical stage work from years to months.

Baseimmune is currently running preclinical programs for three vaccine candidates against African swine fever — a lethal viral disease in pigs — COVID-19 and malaria. 

In the last months, COVID-19 has been circulating at more controlled levels than at the height of the pandemic, leading president Joe Biden to recently conclude that the pandemic is ending in the U.S. Nevertheless, Blight cautioned that coronavirus vaccines are still needed.

“Coronaviruses circulate every year; some cause common colds. Whether they jump [to humans] from other species, we don’t know, but there’s evidence to suggest that it happens a lot,” he explained. “If we look at the big picture, the likelihood that we’re going to have another coronavirus causing a pandemic is significant.”

Baseimmune’s AI-driven strategy showed promise at the start of the pandemic in 2020. At this early stage, Baseimmune’s research predicted 80% of the mutations that were eventually present in key variants of the coronavirus, including in alpha, delta and omicron.

“One of the frustrations for us was that we were at a stage where we were still building the company so we didn’t have the ability to run forward with that quickly enough,” said Blight.  However, it did demonstrate that Baseimmune’s concept is feasible: you can vaccinate people against mutations that aren’t circulating yet. 

AI-driven drug and vaccine discovery is heating up fast. One of the most advanced AI drug discovery firms, Exscientia, raised $510 million in an initial public offering (IPO) on the Nasdaq and accompanying private placement in late 2021, and sealed a $5.2 billion collaboration deal with Sanofi in January 2022.

In the arena of AI-assisted vaccines, the U.S. startup Apriori Bio was launched with $50 million in July 2022. The aim of the company is to use AI to predict which antibodies protect the best against future viral variants, allowing the development of future-proof vaccines and antibody drugs.

Along with the increasing investor interest in AI-assisted drug and vaccine development, there are also a number of challenges. For example, AI-designed vaccines may still need to undergo years of clinical testing to ensure they are safe and effective, even if the discovery stage is faster. Additionally, caution is always required to ensure that the datasets used to train the AI are unbiased and of a high quality.

Add to this the challenges of getting vaccines approved fast enough to tackle newer variants. The first omicron-targeted COVID-19 vaccine booster was approved 9 months after the variant was first detected — time enough for omicron to form multiple sublineages. According to Blight, there is room to speed up the approval of better vaccines.  

“Where there’s a will, there’s a way,” he said. “COVID-19 is showing us that, if approval is needed, it can happen quickly.”

Baseimmune raised $4.8 million in a seed round in November 2021 to fund the development of its vaccine technology. The company is currently focused on advancing its pipeline, and is keeping tight-lipped for now about its upcoming milestones.

“I’m really excited about our three vaccine candidates,” said Blight. “All I will say is watch this space.”

You might also be interested in the following: