I am IC, assistant professor of medicine. My background is in ML,AI, and data science, and in my research group I look at data intensive problems that range from screening for cmpelcx diseases to patterns emerging in evolving pathogens to modeling complex social interactions.
Modeling how complex selection pressures constrain and shape evolutionary trajectories, and we have developed tools that can model that from data, large databases of unique strains, purely from that, learn what these constraints are, and 1) be able to predict what the future changes are. Briefly this makes it possible to probabilistically predict future mutations, and amongst other things 2) predict which strains are at the edge of emergence into humans, and 3) design escape-resistant vaccines.
In the context of research that has been accelerated by the cvid pandemic, a lot of work has been going on in the space of pandemic prevention. The ideas I am talking about falls under the umbrella of “Infection Prevention and Control Projects’. Note however while many interesting ideas are being explored, including opne platforms for pandemic modeling, what is perhaps missing is this capability to model this evolutionary processes directly, -- how do we rank order the many strains that are collected? We need to somehow scale up this risk analysis, and cannot use the very slow process that for example CDC uses for their IRAT analysis .. takes months for each sequence, we can speed it up exponentially.