Artificial intelligence has solved one of the most difficult problems in biology
The long-standing and extremely complex scientific problem of the structure and behavior of proteins has been solved by the new AlphaFold artificial intelligence system. DeepMind scientists have shown that the artificial intelligence they create can predict which three-dimensional shape proteins will form from amino acid sequences.
Predicting what three-dimensional shape a protein will take has been a mystery to scientists for half a century. The ability to accurately predict the structure of proteins based on their amino acid sequence would be a great boon to life sciences and medicine. It would greatly accelerate efforts to understand the building blocks of cells and enable new drugs to be developed more quickly.
The team from DeepMind developed artificial intelligence has dealt with the problem. It is a Google affiliate that has had many successes in developing advanced algorithms. A few years ago you did AlphaGo program the master of go played several times. Another of their artificial intelligence, known as AlphaStar, turned out to be better than 99,8% of the players in the real-time strategy game StarCraft II. However, achieving their new artificial intelligence - AlphaFold - surpasses the good results in games.
AlphaFold has made a giant leap towards solving one of the greatest challenges in biology - determining the three-dimensional shape of a protein based on its amino acid sequence. The artificial intelligence developed by the DeepMind research team came out on top in a biennial CASP competition (Critical Assessment of Structure Prediction) against around 100 other teams to develop systems that can predict three-dimensional protein shapes. In the competition, research teams receive amino acid sequences for around 100 proteins and should develop their structure on this basis.
The results achieved by the teams are compared with the results of the laboratory work.
The experts are impressed with the work done by AlphaFold. They point out that this will have far-reaching implications, including a radical acceleration of new drugs. - What the DeepMind team has achieved is fantastic and will change the future of structural biology and protein research, "says Janet Thornton, retired director of the European Bioinformatics Institute. - This was a problem 50 years ago. I thought it was going to in my lifetime, "added John Moult of the University of Maryland at Shady Grove and co-founder of CASP.
AlphaFold first took part in the CASP in 2018. She took first place at the final table of the competition. But London’s DeepMind’s algorithms baffled everyone this year. Even the competition showed that the AlphaFold works so amazingly well that it could herald a revolution in biology.
- It's a breakthrough in the game. It will transform medicine, research and bioengineering. It will change everything, "emphasizes Andrei Lupas, biologist at the Max Planck Institute for Developmental Biology in Tübingen, who has evaluated the performance of various teams at the CASP.