The Nobel Prize for Chemistry was accorded on Wednesday to Demis Hassabis and John M. Jumper of Google Deepmind and University of Washington professor David Baker, who made breakthroughs in their work with proteins through “supercharged,” artificial intelligence-powered capabilities.
According to the Encyclopedia Britannica, proteins are “directly involved in the chemical processes essential for life,” underscoring the significance of the scientists’ work: Hassabis and Jumper were able to predict the structure of proteins in the millions, allowing insight into how substances would bind with them, while Baker found a way to develop novel, useful proteins.
Johan Aqvist, member of the Nobel Committee for Chemistry, described the works’ impact as “truly huge.” He reasoned that comprehending protein functions requires one to know its shape, which the laureates have been able to do.
With the help of AI, such feats that take from months to decades have been fast-tracked to mere hours or even minutes. Deepmind’s AI program Alphafold has been tapped in the acceleration of pharmaceutical studies, having been cited over 20,000 times.
“A.I. is changing the way we do science. It is supercharging our ability to explore previously intractable problems,” California Institute of Technology chemical engineering professor and 2018 Chemistry Nobel Prize laureate Frances Arnold said.
Deepmind, which was acquired by Google for over half a billion dollars in 2014, aimed to construct AI as functional as a human brain, branch out onto innovations which ultimately advance into that goal or solve specific scientific problems—such as Alphafold.
Alphafold is based on a neural network which enables it to study a myriad of data and pick up on how to perform a multitude of tasks. This is the same technology behind facial and voice recognition, as well as chatbots online. Alphafold in particular has been used to make predictions on the shape of proteins found in the human body. Through this, it may be deduced how the molecules of other substances will attach themselves to the protein, which is one way to develop drugs.
Baker’s work with “generative AI” in building novel protein structures has led to the development of medical remedies such as an antiviral COVID-19 nasal spray and a COVID vaccination; the latter was approved for use in South Korea in 2022.
In 2003, Baker was able to develop the first new protein called Top7, an unprecedented accomplishment as the only proteins known of had been filtered through “millions or billions” or years of evolution, he told The New York Times.
He designed a protein structure and made use of a computer model, Rosetta, to go through databases of recognized proteins in search of an amino acid sequence which might yield the designed structure. He shared that it had been an “amazing moment” to find that the protein he had put together from the given sequence had turned out nearly identical to his design.
Another Nobel Prize, this time in Physics, was awarded to John J. Hopfield and Geoffrey E. Hinton for advancements in AI.
