DeepMind says it will release the structure of all the proteins that science knows
In recent months, Baker’s team has been working with biologists who were previously stuck trying to figure out the shape of the proteins they were studying. “Pretty pretty biological research has accelerated a lot,” he says. A well-prepared public database of hundreds of thousands of shapes should have an even bigger accelerator.
“It’s impressive,” says Tom Ellis, a synthetic biologist at Imperial College London, who studies the yeast genome and is happy to test the database. But he cautioned that most of the announced forms have not yet been verified in the lab.
In the new version of AlphaFold, the predictions come with a confidence score used by the tool to indicate how far each predicted shape is from the real thing. Using this measure, DeepMind found that AlphaFold predicted shapes for 36% of human proteins, with accuracy up to the level of individual atoms. That’s enough to develop the drugs, Hassabis says.
Previously, after decades of work, only 17% of the proteins in the human body have been identified by structures in the laboratory. If AlphaFold’s predictions are as accurate as DeepMind says, the tool has more than doubled that number in just a few weeks.
Predictions that are not completely accurate at the atomic level are still useful. For more than half of the protein in the human body, AlphaFold has predicted that researchers should be fit enough to know what the function of the protein is. The rest of today’s AlphaFold predictions are incorrect or belong to a third of the human body’s proteins that have no structure until they bind to others. “They’re discos,” Hassabis says.
“It’s a tremendous thing that can be applied at this level of quality,” says Mohammed AlQuraish, a systems biologist at Columbia University who developed his software to predict the structure of proteins. He also noted that having the structures of most proteins in an organism will allow us to study how these proteins function as a system, not just in isolation. “That’s what I find most exciting,” he says.
DeepMind releases its tools and predictions for free and will not say if it intends to make money in the future. He doesn’t rule out the possibility, however. To set up and launch the database, DeepMind is collaborating with the European Molecular Biology Laboratory, an international research organization that already has a large database of protein information.
For now, AlQuraishi can’t wait to see what researchers do with the new data. “It’s pretty impressive,” he says, “I think none of us thought we’d be here fast. It’s disturbing.”