Applying AI in the Quest for New Vaccines
Credit: Nolan Zunk
AJIT RAMAMOHAN, Biochemistry, ’24
Ajit Ramamohan, a senior biochemistry major, is hoping to pursue an M.D./Ph.D. degree.
Interviewed by Sowmya Sridhar.
Tell me about your research on the applications of AI to vaccine development.
In the lab of vaccine researcher Jason McLellan, we are collaborating with some of the folks who work on artificial intelligence here at UT. The researchers have developed models to find mutations predicted to stabilize a given protein. This matters a lot for vaccine development. Certain vaccines include the viral proteins most responsible for infection rather than the whole virus. However, these proteins are unstable when “in the wild” or not incorporated in the virus. Vaccine developers need to know how to stabilize this part of the virus before they can make an effective vaccine, but that can take a lot of time-consuming testing in labs. AI can help speed up the process. We’re trying to apply AI-generated data to proteins like the coronavirus spike protein to see if we can quickly and effectively design stabilized versions of those proteins for new vaccines.
How would you describe your research journey at UT?
Research has been an essential part of my experience at UT. I found a great mentor and learning environment in the Biobricks stream of the Freshman Research Initiative (FRI). I then joined two faculty labs – first, a plant developmental science lab and, later, a neuro-oncology lab at Dell Medical School. I was excited by the work being done in both labs, but after spending a year in those labs, I thought, “I don’t know if I want to do research anymore.” But I had been interested in the McLellan lab since hearing about them when I got to UT in 2020 and decided to give it another shot. It’s been two years since, and I’ve gotten fantastic mentorship and been allowed to thrive as a scientist. It was this experience that has solidified my desire to pursue a Ph.D.
“We’re trying to apply AI-generated data to proteins like the coronavirus spike protein.”
What’s one thing that you wish more people knew about AI?
I think it’s fascinating how there’s so much hype around AI now. In many ways, it will change everything, but I think it’s differentiated from actual solutions to problems in that it will not always be right. Someday you may get a “perfect AI model,” but we’re quite far from that. People should keep in mind that AI does have limitations while appreciating the great tool that it is.
What interested you in the research you’re doing now?
Overall, it’s essential basic biochemistry, and we get to see the impacts of our research. I’ve gotten to characterize some antibodies for the coronavirus spike protein and other viral proteins. This has included determining structurally where and how those antibodies bind and investigating the mechanisms they use to stop viral infection in your body. That information can be leveraged to create antivirals or more effective antibody therapies.
I’ve also worked on some projects where we’re trying to stabilize viral proteins of other viruses that could hopefully serve as a vaccine for those viruses. There’s a virus called human cytomegalovirus. It infects most of the adult human population. For most cases, it's not deleterious, but it is the leading infectious cause of birth defects. So, if you get it from your mother congenitally, it could be very harmful.