DeepMind reveals the structure of 200 million proteins in a scientific leap forward | deep mind

Artificial intelligence has solved the structure of nearly every protein known to science, paving the way for the development of new drugs or technologies to address global challenges such as starvation or pollution.

Proteins are the building blocks of life. It consists of chains of amino acids, folded into complex shapes, the three-dimensional structure of which largely determines its function. Once you know how a protein folds, you can begin to understand how it works and how to change its behaviour. Although DNA provides the instructions for making a chain of amino acids, predicting how they will interact to form a three-dimensional shape has been more difficult, and until recently, scientists had only deciphered a fraction of the 200 m or so of proteins known to science.

In November 2020, the Artificial Intelligence Group deep mind It announced that it had developed a program called AlphaFold that could quickly predict this information using an algorithm. Since then, he’s been crushing the genetic codes of every organism whose genome has been sequenced, predicting the structures of the hundreds of millions of proteins they contain collectively.

Last year, DeepMind published the protein structures of twenty species – including Approximately 20,000 proteins are expressed by humans – Open Database. Now he has finished the task, releasing the predicted structures of more than 200 million proteins.

“Essentially, you can think of it as covering the entire protein world,” said Demis Hassabis, DeepMind and DeepMind founder and CEO.

Scientists are already using some of his previous predictions to help develop new drugs. In May, researchers led by Professor Matthew Higgins of Oxford University announce They used AlphaFold models to help determine the structure of a key malaria parasite protein, and figure out where antibodies that could prevent transmission of the parasite are likely to bind.

“Previously, we used a technique called protein crystallography to see what this molecule looked like, but because it’s so dynamic and moving, we couldn’t handle it,” Higgins said. “When we took the AlphaFold models and combined them with this experimental evidence, suddenly it all made sense. This insight will now be used to design improved vaccines that induce antibodies that are more effective in preventing transmission.”

AlphaFold models are also being used by scientists at the University of Portsmouth’s Center for Enzyme Innovation, to identify enzymes from the natural world that can be modified to digest and recycle plastic. Professor John McGeehan, who is leading the work, said. “There is a complete paradigm shift. We can really speed up where we are going from here — and that helps us direct these precious resources to the things that matter.”

Professor Dame Janet Thornton, Group Leader and Chief Scientist at European Molecular biology European Institute of Laboratory Bioinformatics said: “AlphaFold protein structure predictions are already being used in countless ways. I expect this latest update to lead to a flood of exciting new discoveries in the coming months and years, all thanks to the fact that the data is available for everyone to use.”