For several years now, John M. G. Kighan, a biologist in Portsmouth, England, and director of the Center for Enzyme Innovation, has been searching for a molecule that could break 150 million tons of soda bottles and other plastic waste worldwide.

Working with researchers on both sides of the Atlantic, he has found a few good options. But its function is that of the most demanding locks: to direct chemical compounds that turn into microscopic shapes on their own and that they fit perfectly into the molecules of a plastic bottle and can separate them, like a door opening key.

Determining the exact chemical content of a given enzyme is a fairly simple challenge nowadays. But identifying its three-dimensional shape may involve years of biochemical experiments. So last fall, after reading through an artificial intelligence laboratory in London, a system was created that automatically predicts the shape of enzymes and other proteins, Dr. McGeehan asked the lab if it could help with his project.

Towards the end of a workweek, he sent Deepmind a list of seven enzymes. The following Monday, the lab returned shapes for all seven “This moved us a year further to where we were, if not two,” Dr. McGeehan said.

Now, any biochemist can accelerate their work in the same way. On Thursday, DeepMind released predictive shapes of more than 350,000,000 proteins – microscopic mechanisms that drive the behavior of bacteria, viruses, the human body and all other living things. The new database includes three-dimensional compositions for all proteins expressed by the human genome, as well as proteins found in 20 other organisms, including the mouse, the fruit fly, and the E. coli bacterium.

This vast and detailed biological map – which provides approximately 250,000 shapes that were previously unknown – has the ability to understand diseases, develop new drugs and recreate existing drugs. It can lead to new types of biological equipment, such as enzymes that effectively break down plastic bottles and turn them into materials that can be easily reused and reused.

This will take you forward in time – affect the way you think about problems and help solve them faster, said Girabha, an assistant professor in the Department of Cell Biology at New York University. “Whether you study neuroscience or immunology – whatever your field of biology – this can be useful.”

This new knowledge has its own key: If scientists can determine the shape of a protein, they can determine how other molecules will bind it. This reveals how bacteria resist antibiotics – and how to cope with that resistance. Bacteria resist antibiotics by expressing specific proteins; If scientists were able to identify the shapes of these proteins, they could develop new antibiotics or new drugs that suppress them.

In the past, indicate the shape of proteins on lab benches for months, years or even decades associated with X-rays, microscopes and other tools. But Deepmind can significantly shrink the timeline with its AI technology, known as alphafold.

When Dr. McGeehan sent Deepmind a list of his seven enzymes, telling the lab that he had already identified the shapes for two of them, but he said nothing about those two. This was a way to test how the system works; Passed the alphafold test, accurate prediction of both shapes.

It was more significant, said Dr. McGihan, the forecasts have come in a few days. He later learned that Alphafole had in fact completed the task in just a few hours.

Alphafold predicts protein structures known as neural networks, a mathematical system that can learn functions by analyzing large amounts of data – in this case, thousands of known proteins and their physical shapes – and extraplating into the unknown.

This is the same technology that recognizes barking commands in your smartphone, recognizes faces in photos you post on Facebook, and translates one language into another on Google Translate and other services. But many experts believe that Alphafold is the most powerful application of the technology.

“It shows that AI can do useful things amidst the complexities of the real world,” said Jack Clarke, one of the authors of the AI ​​Index, an attempt to track the progress of artificial intelligence technology around the world.

According to Dr. When McGeeh was discovered, it could be significantly more accurate. Alphafold can predict the shape of a protein with approximately 63 percent of the time compared to physical experiments, according to independent benchmark tests that compare its predictions to known protein structures. Most experts assume that this powerful technology is still years away.

“I thought it would take another 10 years,” said Randy Reed, a professor at Cambridge University. “This was a complete change.”

But the accuracy of the system varies, so some of the predictions in DeepMind’s database will be less useful than others. Each prediction in the database comes with a “confidence score” that shows how accurate it is. Deepmind researchers estimate that the system provides “good” predictions about 95 percent of the time.

As a result, the system will not be able to completely replace physical experiments. When experiments fail, it is used in conjunction with working on a lab bench to help scientists decide which experiments should be performed and fill in the gaps. Using Alphafold, researchers at the University of Colorado Boulder recently helped identify a protein structure they had struggled to identify for more than a decade.

With the hope of advances in biological science, Deepmind’s developers have chosen to freely share its database of protein structures rather than sell rather than sell. “We’re interested in maximum performance,” said Demis Hasabis, chief executive and co-founder of Deepmind, which owns a parent company similar to Google, but operates more like a research laboratory than a commercial business.

Some scientists have compared DeepMind’s new database to the Human Genome Project. Completed in 2003, the Human Genome Project provided a map of all human genes. Now, DeepMind has provided a map of the approximately 20,000 proteins expressed by the human genome – another step in understanding how our bodies function and how we can react when things go wrong.

The hope is that technology will continue to evolve. A laboratory at Washington University in Washington has created a similar system called RosTTFfield, and like Deepmind, it has openly shared the computer code that runs its system. Anyone can use technology, and anyone can work to improve it.

Even before Deepmind began sharing its technology and data openly, Alphafold was feeding a wide range of projects. Researchers at the University of Colorado are using technology to understand how bacteria such as E. coli and Salmonella develop resistance to antibiotics and to develop ways to combat this resistance. At the University of California, San Francisco, researchers have used this tool to improve their understanding of coronavirus.

Coronavirus destroys the body by 26 different proteins. With the help of Alphafold, researchers have improved their understanding of a key protein and hope the technology can help increase the understanding of another 25 people.

If this is late in affecting the current epidemic, it can help prepare for the next. “A better understanding of this protein will help target not only this virus but other viruses as well,” said San Francisco-based researcher Clement Verba.

The possibilities are numerous. After Deepmind, Dr. McGeehan shaped seven enzymes that could potentially rid the world of plastic waste, he sent a list of 93 more to the lab. “They’re working on this now,” he said.