DeepMind, a London-based subsidiary of Google’s owner Alphabet Inc, has been praised by the global scientific community after solving a 50-year-old challenge in biology.
Its artificial intelligence system AlphaFold has figured out what shapes proteins fold into, the so-called ‘folding problem’.
It is a major scientific breakthrough because it allows to better understand what a protein does and how it works, since its shape is closely correlated with its function.
Proteins are the ‘building blocks of life’ because they underpin the biological processes in every living thing.
There are currently around 200mln known proteins and another 30mln is found every year.
Each of them has their own shape and it is often expensive and time-consuming to find their 3D composition, so we know only a fraction of the millions known to science.
Proteins are made of amino acids, which make the protein to fold when they interact, meaning there are nearly infinite possibilities for shapes.
The 3-D shape into which proteins fold themselves determines just about everything in biology. To predict that shape from the 1-D sequence of amino acids is a truly stunning achievement. If a computer program were eligible for a Nobel Prize . . .https://t.co/n8QxINE5M0
— Richard Dawkins (@RichardDawkins) December 1, 2020
AlphaFold was trained on the sequences and structures of 100,000+ proteins mapped out by scientists around the world and can now predict a protein’s shape based on the sequence of amino acids.
As a result, scientists worldwide will have extra help in finding solutions, such as developing treatments for diseases or finding enzymes that break down industrial waste, because of the key role of proteins.
The system was officially recognised as a solution to the issue by the biennial Critical Assessment of protein Structure Prediction, a community created in 1994 by scientists that were looking to solve the protein folding problem.