AI and the Future of Vaccines

Photo by Hakan Nural on Unsplash

The novel coronavirus has devastated communities across the globe, creating an unprecedented global health crisis. Yet, there is light at the end of the tunnel: a vaccine.

Traditional vaccines work by introducing a weakened portion of the virus into the bloodstream. These components, known as immunogenic sites, elicit an immune response by the body. Such a site is the “spike protein” that exists on the exterior of the coronavirus. Spike proteins give coronaviruses their definite characteristic and allow the virus to infiltrate host cells. But to do this, researchers would have to synthetically grow the virus in a lab, which takes time and effort.

But the vaccine has been so successful due to breakthrough research in mRNA technology. Using this technology, researchers were able to isolate the spike protein gene and incorporate it into the Covid vaccine. Instead of researchers having to grow the virus, the body creates and replicates the virus within. The coronavirus vaccine is the first implementation of this decades-long research.

The mRNA technology associated with the vaccine is revolutionary, but another tool helped researchers along the way: AI. For the first time, AI and machine learning played an immense role in vaccine development.

The spike protein was an immediate starting point for researchers trying to develop an mRNA vaccine, but they used DeepMind’s Alpha Fold algorithm to aid them in their research. Alpha Fold can predict the intricate protein structure created by chains of amino acids (the building blocks of proteins). AlphaFold allowed researchers to visualize the protein in its entirety, and from this, machine learning algorithms were able to detect the most effective immunogenic sites on the spike protein.

But the spike protein is highly susceptible to mutations: changes in RNA that create new strains of the virus. Research shows that the Covid strain discovered in Europe has a unique spike protein composition. Its spike protein is more effective at infiltrating host cells making this strain even more transmissible. And so, a vaccine developed using the spike protein of the virus found in China may not be effective against other strains (i.e Brazil and South Africa).

It’s a similar dilemma with the flu. The public is encouraged to take a flu vaccine every year due to the increasing number of flu strains discovered annually. To combat this, researchers working on influenza have tried looking at inner proteins that are less susceptible to mutations, rather than spike and other surface proteins, to create a universal flu vaccine.

Fortunately, with the AI models employed to develop the Covid vaccine, researchers can identify proteins, and subsequently segments of code, that are less likely to mutate. As more strains of the virus are discovered, AI can find similarities between all known strains and identify a common protein that can be used as a basis for a universal vaccine. This information, in combination with protein modeling structures created by AlphaFold, can provide reliable target immunogenic sites. The mRNA technology used for the Covid vaccine has made this possible. Given the success of the coronavirus vaccine to date, researchers will definitely emphasize using mRNA technology to combat viruses in the future.

Countries around the world poured billions of dollars to aid pharmaceutical companies in developing the Covid vaccine. Yet, the research that allowed for the creation of mRNA vaccines was conceived by federal-funded research. It’s this work that big pharma has built upon today. But federal R&D spending has decreased from 5% of total GDP 60 years ago to 0.7% in 2018; to make up for this deficit, the government is relying on investment from the private sector. But the interests of corporations like Google and Pfizer aren’t always aligned with the public’s. Pharma is expected to make billions off of vaccines, and universal vaccines could undermine these profits. And often, corporate R&D spending is directed towards creating a new product or process, not advancing technological growth and scientific discovery.

Instead, the government needs to be spearheading research so that breakthroughs like the one made on mRNA technology can take place. If the responsibility of research is left to big corporations, the government cannot protect the interests of the people. Specifically, it should encourage research into AI algorithms such as AlphaFold, which can be used in biotechnology and healthcare. The coronavirus has fueled research in vaccine technology, but it’s also given us the chance to change the research landscape in the US and around the world.

Interested in the intersection between tech and ethics.