Advances in Structural Biology enabled by AI

Professor Matt Higgins, Professorial Fellow at Merton and EPA Abraham Professor of Structural Biology at the University of Oxford uses AI with his team on a daily basis at the Higgins Lab.  

The Higgins lab studies parasites such as those that cause malaria or other parasitic diseases. They study these at a very detailed molecular level and the lab team are particularly interested in host parasite interaction: so how molecules on the surface of the parasites interact with molecules from the infected human.  

These interactions are essential for the life cycle of the parasite. For example, malaria parasites need to get inside our red blood cells before they can make us ill and members of the lab focus on how this process of invasion works. 

Other members of the lab team then use this information to design therapies for parasitic diseases, which in some instances take the form of vaccines. 

Professor Higgins commented: 

‘The way in which proteins adopt their structure is that they’re each made up of a linear chain of building blocks. There are 20 different building blocks, and those building blocks can be organised in any order in a long chain. Each one of these building blocks has a slightly different chemistry and the order of the chemistry along these chains then determines how it all folds up into its three-dimensional structure which in turn determines its function.  

All of the known protein structures form a huge data set and being able to predict the structure of a protein from its DNA sequence alone has been one of structural biology’s greatest challenges. This is where AI now comes into play.  

We can now use an AI programme called AlphaFold. This a protein prediction package involving an algorithm which can learn from the huge dataset of known protein structures to then predict structures of unknown proteins. AlphaFold can help us predict structures of parasite proteins and work out where antibodies that could block transmission of the parasite are likely to bind. It takes away some of the laborious experimental work, speeding up the overall vaccine design process. Scientists will still need to confirm a protein’s structure, but they’ll get an impressive head start through the use of AI.  

If we understand the molecules on the surface of the malaria parasites that are essential for its life cycle, we can work out which of the important parts are targeted by antibodies. We can then design specific vaccines which cause the body to produce good quality protective antibodies. Protein design can also be aided by AI technology as it allows us to predict the structures of novel designed proteins, generating totally new molecules, such as vaccine immunogens. The real-world benefit of AI here is focused and specifically designed vaccines which induce the most potent transmission-blocking antibodies.’ 

The Higgins team have been using AlphaFold as well as protein design package, Rosetta, for nearly three years now. The AI technology has brought about a huge change in structural biology. 

Learn more about the work of Matt and his team at the Higgins Lab.