In any living organism, there are thousands of different proteins, each with its own unique shape. For decades, the exact formation of these shapes has been difficult for scientists to understand. How exactly does a protein, which starts out as a chain of amino acids, fold into the funky 3D shapes you might recognize from diagrams? AlphaFold, an AI from DeepMind, may have an answer. It can predict, with unprecedented precision, the shape a protein will take.
AlphaFold has been put to the test in a global competition called Critical Appraisal of Protein Structure Prediction, or CASP, which DeepMind CEO Demis Hassabis calls the “Protein Folding Olympics.” in a video. During competition, systems like AlphaFold receive amino acid chains for proteins whose shapes have already been identified in previous experiments but which have not yet been published. Judges compare the forms of protein produced by the systems with what they know the forms should be.
At the end of the competition, AlphaFold had the most accurate predictions of all CASP participants in its 25 year history by a wide margin. Even predictions that weren’t precise enough to be considered “competitive” with experimental results were only a few atomic widths away. The full data is yet to be peer reviewed and published, but the DeepMind team is excited about the results so far, saying in a blog post that they are “optimistic about the impact AlphaFold can have. on biological research and the world in general ”.
It can take years in the lab for scientists to identify individual protein forms. Neural networks like AlphaFold could help speed up biological research and drug development in the future. The AI method isn’t perfect yet, and it won’t take over any time soon for flesh-and-blood researchers, but it could be a major milestone in the eternal marathon of scientific progress. To learn more about AlphaFold and to see a group of scientists and engineers become stunned by the results of the competition, watch the video below.