UT Austin’s EvoRank, a groundbreaking AI model inspired by natural evolution, is set to transform biomedical research with its innovative approach to protein-based drug and vaccine development.
Researchers from The University of Texas at Austin have launched an innovative AI model that promises to revolutionize biomedical research. Known as EvoRank, this artificial intelligence mechanism taps into the principles of evolution to design protein-based therapies and vaccines that could lead to more effective and less toxic treatments in medicine.
Unveiled at the International Conference on Machine Learning, and detailed in Nature Communications, EvoRank employs a novel approach by emulating nature’s evolutionary processes.
“Nature has been evolving proteins for 3 billion years, mutating or swapping out amino acids and keeping those that benefit living things,” Daniel Diaz, a research scientist in computer science and co-lead of the Deep Proteins group at UT, said in a news release. “EvoRank learns how to rank the evolution that we observe around us, to essentially distill the principles that determine protein evolution and to use those principles so they can guide the development of new protein-based applications, including for drug development and vaccines, as well as a wide range of biomanufacturing purposes.”
A grant from the Advanced Research Projects Agency for Health will provide nearly $2.5 million to this UT team. Working in collaboration with vaccine-maker Jason McLellan, a professor of molecular biosciences at UT, and the La Jolla Institute for Immunology, the funding will focus on utilizing AI in protein engineering to develop vaccines targeting herpesviruses.
“Engineering proteins with capabilities that natural proteins do not have is a recurring grand challenge in the life sciences,” Adam Klivans, co-lead of Deep Proteins and a professor of computer science, said in the news release.
Unlike Google’s AlphaFold, which predicts the 3D structures of proteins, EvoRank, alongside the Stability Oracle framework, harnesses evolutionary variations within existing proteins. It allows the model to discern the biochemical viability and stability necessary for protein-based applications.
“The models have come up with substitutions we never would have thought of,” added McLellan. “They work, but they aren’t things we would have predicted, so they’re actually finding some new space for stabilizing.”
McLellan’s team is already synthesizing and testing these AI-generated viral protein designs.
The protein therapeutics market, already valued at approximately $400 billion, is expected to grow by over 50% in the next decade. However, the journey from drug design to FDA approval is painstakingly slow and costly. EvoRank’s advancements hold the potential to streamline this process dramatically, offering new ways to approach genetic engineering decisions traditionally made through laborious trial-and-error methods.
Diaz and his team aim to continue refining EvoRank and its broader framework by developing a “multicolumn” version for evaluating multiple mutations simultaneously. Incorporating cutting-edge machine learning algorithms, this work could orient protein engineering towards unprecedented breakthroughs in drug development.