Google DeepMind Reveals A New AI Model AlphaFold 3 Capable Of Accurately Predicting Protein Structure And Its Interaction With Molecules

Google has unveiled the third version of the “Alpha model” an artificial intelligence model capable of predicting the protein structure and interactions of all life’s molecules with unprecedented accuracy. This model comes with the increased accuracy for the interactions of proteins with other molecule types and for some of the important categories of interaction it has doubled prediction accuracy.

The AlphaFold 3 given the list of molecules can generate the joint 3D structure of it. It can also model chemical modifications to these molecules which control the healthy functioning of cells, that when disrupted can lead to disease. To perform such a complex task AlphaFold 3 after processing the input assembles its predictions using a diffusion network, similar to those found in AI image generators. It’s a crucial step to understand the structure of a particular drug and how it interacts with different molecules and accuracy is something that can’t be compromised in it.

Google does mention in its blog post not once but multiple times that this model achieves unprecedented accuracy in predicting drug-like interactions, including the binding of proteins with ligands and antibodies with their target proteins. Its sister company Isomorphic Labs is working on drug design for internal projects as well as with pharmaceutical partners. Scientists can access the majority of its capabilities, for free, through the newly launched AlphaFold Server.