Proteins, which sustain life and support all functions, are large molecules made up of chains of amino acids. The shape of a protein is related to its function and the ability to predict this structure can provide greater insights into how it works. Its functionality relies on unique 3D structures, and understanding what shapes proteins can fold into is known as protein folding – a grand challenge in biology for half a century now.

In a major scientific advance, the latest version of Alphabet’s AI system AlphaFold has been recognised as a solution to the grand challenge of the biennial Critical Assessment of protein Structure Prediction (CASP).

This breakthrough demonstrates the impact AI can have on scientific discovery and its potential to dramatically accelerate progress in some of the most fundamental fields that explain and shape our world, stated Alphabet in a release.

Findings can be accessed in a research report - High Accuracy Protein Structure Prediction Using Deep Learning by John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Kathryn Tunyasuvunakool, Olaf Ronneberger, Russ Bates, Augustin Žídek, Alex Bridgland, Clemens Meyer, Simon A A Kohl, Anna Potapenko, Andrew J Ballard, Andrew Cowie, Bernardino Romera-Paredes, Stanislav Nikolov, Rishub Jain, Jonas Adler, Trevor Back, Stig Petersen, David Reiman, Martin Steinegger, Michalina Pacholska, David Silver, Oriol Vinyals, Andrew W Senior, Koray Kavukcuoglu, Pushmeet Kohli, Demis Hassabis.

“We have been stuck on this one problem – how do proteins fold up – for nearly 50 years. To see DeepMind produce a solution for this, having worked personally on this problem for so long and after so many stops and starts, wondering if we’d ever get there, is a very special moment,” Professor John Moult, Cofounder and Chair of CASP, University of Maryland.

This has been a focus of intensive scientific research for many years, using a variety of experimental techniques to examine and determine protein structures, such as nuclear magnetic resonance and X-ray crystallography. These techniques, as well as newer methods like cryo-electron microscopy, depend on extensive trial and error, which can take years of painstaking and laborious work per structure, and require the use of multi-million dollar specialised equipment, according to the statement.

In the results from the 14th CASP assessment, released yesterday, the AlphaFold system achieved a median score of 92.4 GDT overall across all targets. This means that our predictions have an average error (RMSD) of approximately 1.6 Angstroms, which is comparable to the width of an atom (or 0.1 of a nanometer). Even for the very hardest protein targets, those in the most challenging free-modelling category, AlphaFold achieves a median score of 87.0 GDT

These exciting results open up the potential for biologists to use computational structure prediction as a core tool in scientific research. These methods may prove especially helpful for important classes of proteins, such as membrane proteins, that are very difficult to crystallise and therefore challenging to experimentally determine.

Also Read: DeepMind's AlphaFold marks major milestone in predicting protein structure