ARTIFICIAL INTELLIGENCE CALCULATES STRUCTURES OF ALMOST ALL KNOWN PROTEINS

Every living thing from bacteria to weeds to humans is made of proteins. Last year, the artificial intelligence (AI) AlphaFold, created by researchers at Google parent Alphabet, fleshed out the catalog of human proteins by completing chemical structures that scientists had only been able to calculate in part.

The way a protein is structured or “folded” determines how it functions and which other proteins it can interact with and how.

Now AlphaFold has topped itself: it has detailed the structure of virtually all of the 200 million proteins known to exist, leaving less than 190,000 unsolved.

With the new database, “you can look up the 3D structure of a protein almost as easily as you can do a keyword Google search,” one researcher told Science magazine.

More than 500,000 researchers have already tapped the database, Science said.

Some are using it to figure out the mechanisms cell membranes use to decide which substances to allow into the cell and which to exclude. Others have used the newly revealed protein structures to find candidate proteins for a malaria vaccine and to discover that a large group of plant proteins are related, which implies common vulnerabilities and, perhaps, common treatments. 

TRENDPOST: Being able to look up a protein structure instead of having to experiment to try different possibilities could speed the development of medical treatments by an order of magnitude, especially for complex conditions.

More broadly, the new database will enable scientists to make discoveries, think of questions, and devise experiments they weren’t able to envision before.