Artificial intelligence has solved a biology ‘grand challenge’ that has defied scientists for half a century.
The challenge involves working out how proteins fold into their unique three-dimensional structures.
This is a far more complex question than it might sound as there are an astronomical number of possible configurations for each protein.
The challenge stems from work by American biochemist Christian Anfinsen, who won the Nobel Prize in Chemistry in 1972 after showing that it should theoretically be possible to determine the shape of proteins from the sequence of their amino acid building blocks.
In 1994, an experiment known as CASP (Critical Assessment of protein Structure Prediction) was launched.
Every two years, research groups from the scientific community across more than 20 countries use computers to predict the shape of a set of proteins based on their amino acid sequences.
Other scientists work out the actual 3D structures using techniques such as X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy.
The CASP team then compares the predictions and the experiment-derived result, awarding an accuracy score of between 0 and 100.
DeepMind AI accurately predicted the shape of a series of proteins
In this year’s challenge, a new algorithm from London-based AI lab DeepMind appears to have finally solved the problem.
DeepMind has previously made headlines with its game-playing algorithms, but AlphaFold used new deep learning architecture that represents protein structures in the form of complex spatial graphs.
It was able to achieve a score of 90 on the accuracy test metric, known as the global distance test.
This is comparable to the accuracy of the experimental techniques, meaning that the AI was able to predict the shape as well as scientists were able to determine it in the laboratory.
CASP co-founder Dr John Moult said: “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.”
The achievement is not just about solving the problem for its own sake.
Knowing exactly how proteins fold would have huge potential applications in fields including drug design and discovery and medical understanding of numerous diseases such as cancer and dementia and viruses such as COVID-19.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
