‘Once in a generation advance’ as Google AI researchers crack 50-year-old biological challenge

A three-dimensional digital rendering of a protein. The 50-year-old "protein folding problem" may have been cracked by artificial intelligence created in the UK by Google-owned AI lab DeepMind, paving the way for faster development of treatments and drug discoveries.

Credit: PA/DeepMind

A "50-year-old biological challenge" has been cracked by AI researchers at Google, which could “significantly accelerate” drug development for cancer and other diseases.

The so-called "protein folding problem" has long been one of biology’s biggest hurdles, as researchers have sought to predict the shape of proteins to understand how they perform specific tasks in the body.

Proteins start off in a cell as a long chain and then fold upon themselves into a 3D shape in order to perform their biological function.

But sometimes this folding process fails, which can impact the health of the cell or cause other misfolded proteins to clump together.

This failure can cause several known diseases, scientists say.

There are 200 million known proteins at present but only a fraction have been “unfolded” to fully understand what they do and how they work.

Google-owned AI lab DeepMind, based in London, says its AlphaFold program has solved the issue and is capable of predicting the shape of many proteins.

They say the system is able to determine “unparalleled levels of accuracy” in the structures in a matter of days.

DeepMind has worked on the AI project with the 14th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP14), a group of scientists who have been looking into the matter since 1994.

"Proteins are extremely complicated molecules, and their precise three-dimensional structure is key to the many roles they perform, for example the insulin that regulates sugar levels in our blood and the antibodies that help us fight infections," Dr John Moult, chair of CASP14, said.

"Even tiny rearrangements of these vital molecules can have catastrophic effects on our health, so one of the most efficient ways to understand disease and find new treatments is to study the proteins involved.”

During the latest test, DeepMind said AlphaFold determined the shape of around two-thirds of the proteins with accuracy comparable to laboratory experiments.

The discovery could change the way we approach biology, the researchers said, by speeding up drug development for certain diseases and "opening up new avenues of exploration".

Arthur D. Levinson, founder and CEO Calico, said: “AlphaFold is a once in a generation advance, predicting protein structures with incredible speed and precision.

"This leap forward demonstrates how computational methods are poised to transform research in biology and hold much promise for accelerating the drug discovery process.”

But researchers behind the project say there is still more work to be done, including figuring out how multiple proteins form complexes and how they interact with DNA.

DeepMind is planning to submit a paper detailing its system to a peer-reviewed journal to be scrutinised by the wider scientific community.

You may also like...