The Nobel Prize in Chemistry was awarded Wednesday to three scientists who have helped unravel some of the enduring mysteries of proteins, the building blocks of life.
While Demis Hassabis and John Jumper of Google’s DeepMind lab used artificial intelligence techniques to predict the structure of proteins, biochemist David Baker managed to design entirely new proteins never seen in nature.
These breakthroughs are expected to lead to many advances, from the discovery of new medicines to enzymes that break down pollutants.
Here’s an explanation of the science behind the Nobel win.
What are proteins?
Proteins are molecules that serve as “factories for everything that happens in our bodies,” David Calabiro, a protein researcher at the University of Birmingham, UK, told AFP.
DNA provides the blueprint for every cell. The proteins then use this information to change that cell into something specific – such as a brain cell or a muscle cell.
Proteins are made up of 20 different types of amino acids. The order in which these acids are introduced determines which 3D structure they will fold and fold into.
Mary Carroll, president of the American Chemical Society, compared it to an old-fashioned telephone cord.
“So you can stretch that telephone cord, and then you just have a one-dimensional structure,” he told AFP.
“Then it will come back into 3D shape,” he said.
So if chemists wanted to master proteins, they needed to understand how 2D sequences turned into these 3D structures.
“Nature already provides thousands of different proteins, but sometimes we want them to do something they don’t yet know how to do,” said French biochemist Sophie Sacquin-Mora.
What did AI do?
The work of previous Nobel laureates had demonstrated that chemists should be able to look at amino acid sequences and predict their structure.
But it was not that easy. Chemists struggled for 50 years—there was even a biennial competition called the “Protein Olympics,” where many failed the prediction test.
Enter Hassabis and Jumper. They trained their artificial intelligence model AlphaFold on all known amino acid sequences and related structures.
When given an unknown sequence, AlphaFold compares it to previous sequences, gradually reconstructing the puzzle in three dimensions.
After the new generation AlphaFold2 crushed the 2020 Protein Olympics, organizers thought the problem was solved.
Models have now predicted the structure of almost all of the 200 million proteins known on Earth.
What about new proteins?
The American biochemist Baker started from the opposite end of the process.
First, they designed an entirely new protein structure never seen in nature.
Then, using a computer program called Rosetta, which he had developed, he was able to work out the amino acid sequence as it began.
To achieve this, Rosetta explored all known protein structures, searching for small protein fragments similar to the structure it wanted to create.
Rosetta then made changes to them and proposed a sequence that could end up as the structure.
What is all this for?
Mastering small machines as fundamental and important as proteins could have a huge number of potential uses in the future.
“This allows us to better understand how life functions, including why certain diseases develop, how antibiotic resistance occurs or why certain microbes can decompose plastics,” the Nobel website said.
Creating all-new proteins could lead to new nanomaterials, targeted drugs and vaccines, or more climate-friendly chemicals, it says.
Asked to choose a favorite protein, Baker pointed to a protein he “designed during the pandemic that protects against coronavirus.”
“I am very excited by the idea of a nasal spray of a tiny designed protein that would protect against all possible pandemic viruses,” he said at the Nobel ceremony via videolink.
Calabiro emphasized how “transformational” this research would be.
“I think this is the beginning of a whole new era.”
(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)
