Google accelerating new drug discovery

Google DeepMind has launched AlphaFold 3, an improved version of AlphaFold 2, that can not only accurately predict the structure of proteins but also the structure and behavior “all of life’s molecules” which will transform our ability to understand, at a molecular level, how our bodies complete the endless variety of tasks, which will aid new treatment and drug discovery.

What can AlphaFold 3 do and how will it be used?

AlphaFold 3–which is 50% more accurate than its predecessors–will help researchers in medicine, agriculture, materials science, and drug development by taking over expensive, lab-based methods, and accelerating research and discoveries across these industries.

Google has partnered with Isomorphic Labs (a drug discovery company) which is currently using AlphaFold 3 internally to improve its understanding of new disease targets, build drug discovery programs, and overcome real-world drug design challenges.

How does AlphaFold 3 work? 

Google has been working with domain experts and biosecurity, research, and industry specialists to develop and test AlphaFold 3 and make sure its deployed responsibly. Researchers have been adding lists of the molecules they want to combine, to an existing library of molecular structures. AlphaFold 3 then uses a diffusion technique (which is what AI image generators like Stable Diffusion use to generate images) to create a 3D model of a new structure.

‍