Member-only story
|AI|BIOLOGY|PROTEIN DESIGN|
DeepMind’s AlphaProteo: Revolutionizing Protein Design with Machine Learning
Harnessing AI to Create High-Affinity Protein Binders in a Single Step
Biology is the most powerful technology ever created. DNA is software, protein are hardware, cells are factories. — Arvind Gupta
Over the past four years, deep learning has accelerated research in biology, especially in the field of structural proteomics. AlphaFold2 was a breakthrough moment, changing the entire scientific community's attitude toward these models. For the first time, a deep learning model could produce the structure of a protein in seconds with unprecedented accuracy. This changed everything because obtaining the structure of a protein is an economically expensive and extraordinarily laborious process. AlphaFold2 made this process easy and affordable.
AlphaFold2, however, is not perfect. It is not even the end point, because for many applications knowing the structure of the protein is only the beginning. For example, the structure gives us different information but does not give us the function, with whom it interacts, and how to activate or block it. Because of this, research has not stood still, and several papers have come out in the past three years trying to solve the problems left open by AlphaFold2
