DeepMind Claims Major Breakthrough in Understanding Proteins

The breakthrough, called by Alphabet’s DeepMind AlphaFold 2, uses computers to predict how proteins fold. Scientists have been stuck on the protein-folding problem for almost 50 years, as proteins can take, in theory, as many as 10x300 different shapes. The breakthrough could help researchers design new drugs and understand diseases. In the longer term, predicting protein structure will also help design synthetic proteins, such as enzymes that digest waste or produce biofuels. Researchers are also exploring ways to introduce synthetic proteins that will increase crop yields and make plants more nutritious. Pharma companies have been stepping up their own investments in AI to solve the problems of the industry’s narrowly focused research. For example, UK-based GSK recently set up an AI lab in London’s King’s Cross. It, along with AstraZeneca and others, is partnering with Nvidia, which is in the process of building the UK’s most powerful supercomputer to aid drugs research. The commercial impact of DeepMind’s triumph is not yet clear. It says it is still thinking through how to apply its findings. Speeding up drug discovery promises to improve one of the industry’s most costly and failure-prone activities.