Deepmind’s Alphafold: a solution to the 50-year-old challenge of protein folding

DeepMind is a community of scientists, engineers, machine learning experts and more, working together to advance the state of the art in artificial intelligence. The ultimate vision behind DeepMind has always been to build AI and then use it to help further our knowledge about the world around us by accelerating the pace of scientific discovery.

An AI-driven solution to the protein folding problem

The first proof point in this thesis came just last week when our AI system, AlphaFold, was recognised as a solution to the 50-year-old grand challenge of protein structure prediction, often referred to as the ‘protein folding problem’ – according to a rigorous independent assessment.

Proteins are essential to life, supporting practically all its functions. They are large complex molecules, made up of chains of amino acids, and what a protein does largely depends on its unique 3D structure. Figuring out what shapes proteins fold into is known as the “protein folding problem”, and has stood as a grand challenge in biology for the past 50 years. This DeepMind video explainer gives more details about protein folding.

The work started back in 2016, shortly after our system AlphaGo became the first computer program to defeat a world champion at the ancient game of Go. 

In 2018, a team of DeepMind researchers participated in CASP13 (Critical Assessment of protein Structure Prediction), a biennial blind assessment to catalyse research, monitor progress, and establish the state of the art in protein structure prediction. At the time, our initial version of AlphaFold achieved the highest accuracy among participants.

This year’s CASP took place in more difficult circumstances, with participants working from home against the backdrop of a global pandemic. Nonetheless, it produced very exciting results, with CASP recognising AlphaFold as a solution to the protein-folding challenge.

This breakthrough demonstrates the impact AI can have and the potential it has to dramatically accelerate progress in some of the most fundamental fields that explain and shape our world. We’re indebted to CASP’s organisers and the whole community, whose work enabled this rigorous assessment.

Next steps

While it’s still early days, we are hopeful that AlphaFold will help drive advances in areas like drug design and environmental sustainability, in alignment with the OECD principle that AI should benefit people and the planet. Alongside working on a peer-reviewed paper, we’re exploring how best to provide broader access to the system in a scalable way. We are also working with specialist groups to look into how protein structure predictions could contribute to our understanding of specific diseases. We’re exploring additional next steps, and we are excited to collaborate with others to learn more about AlphaFold’s potential in the years ahead. 

We encourage everyone to learn more on DeepMind’s blog and Twitter, and we look forward to keeping this community updated on our progress.