Main menu


AI predicts the shape of nearly every protein known to science

featured image

In 2020, an artificial intelligence lab called DeepMind unveiled a technique that can predict the shape of proteins. It is the microscopic mechanism that drives the behavior of the human body and all other living things.

A year later, the lab shared a tool called AlphaFold with scientists to publish predicted shapes for over 350,000 proteins, including all proteins expressed by the human genome. It quickly changed the course of biological research. If scientists can identify the shape of proteins, they can accelerate their ability to uncover the mysteries of life on Earth, such as understanding disease and developing new drugs.

DeepMind has now published predictions for almost every protein known to science. On Thursday, a London-based laboratory owned by the same parent company as Google said he had added more than 200 million predictions to an online database freely available to scientists around the world.

With this new release, the scientists behind DeepMind hope to accelerate the study of more obscure organisms and open up a new field called metaproteomics.

Demis Hassabis, CEO of DeepMind, said in a telephone interview.

Proteins begin as a set of chemical compounds that twist and fold into three-dimensional shapes that define how these molecules bind to other molecules. If scientists can identify the shape of a particular protein, they can decipher how it works.

This knowledge is often an important part of fighting disease and disease. For example, bacteria resist antibiotics by expressing specific proteins. If scientists can understand how these proteins work, they can start fighting antibiotic resistance.

Previously, extensive experimentation with X-rays, microscopy, and other tools was required on the lab bench to pinpoint the shape of a protein. Now, given a set of compounds that make up a protein, AlphaFold can predict its shape.

Technology isn’t perfect. However, independent benchmark tests show that it can predict protein shape with accuracy comparable to physical experiments about 63% of the time. With predictions, scientists can verify their accuracy relatively quickly.

Clement Velva, a researcher at the University of California, San Francisco, who uses the technology to understand coronaviruses and prepare for similar pandemics, said the technology “enhances” the study, often resulting in months of pain relief. He said that it saved the experimental time of Others use this tool to combat gastroenteritis, malaria, and Parkinson’s disease.

The technology is also accelerating research beyond the human body, such as efforts to improve the health of bees. DeepMind’s expanded database will help the larger scientific community enjoy similar benefits.

Like Hassabis, Verba believes this database will provide new ways to understand how proteins behave across species. He also sees it as a way to educate a new generation of scientists. Not all researchers are familiar with this kind of structural biology. A database of all known proteins lowers the barrier to entry. “It can bring structural biology to the masses,” he said Verba.

This article was originally published in The New York Times.