Nobel Prize in Chemistry 2024 to David Baker, Demis Hassabis and John Jumper for revealing the secrets and techniques of proteins with AI and computing | Science | EUROtoday

The Royal Swedish Academy of Sciences awarded this Wednesday the 2024 Nobel Prize in Chemistry, with half to David Baker for “the design of proteins with computation” and the opposite half collectively to Demis Hassabis and John Jumper for “the prediction of the structure of proteins through the use of artificial intelligence.” The committee that awarded the award has highlighted the potential functions of its scientific achievements in quite a few processes through which proteins are concerned, from the sooner growth of vaccines to the invention of recent nanomaterials, by the design of focused medicine to deal with the most cancers or the evolution in direction of a greener chemical trade. To give an thought of ​​the brand new universe that the advances of Baker, Hassabis and Jumper speak in confidence to science, the president of the committee, Heiner Linke, identified that “if we wanted to understand how proteins work, we first had to know what they look like” and pressured within the nice advantages for humanity that this information will deliver.

The British Demis Hassabis (London, 1976) and the American John Jumper (Little Rock, 1985) have efficiently used synthetic intelligence to speed up a scientific profession that had been open for half a century, from the corporate Google Deepmind of which Hassabis is the CEO . With their synthetic intelligence mannequin AlphaFold2, offered in 2020, they’ve managed to foretell the three-dimensional construction of just about all of the proteins recognized to this point – about 200 million –, based mostly solely on the amino acid sequence that types their chain. These very lengthy chains fold into tremendously advanced shapes, which decide the perform and organic exercise of proteins, and predicting these buildings was a pending problem for science for the reason that Seventies.

The American David Baker (Seattle, 1962) has realized to grasp the so-called constructing blocks of life – the 20 amino acids – and to create fully new proteins that didn’t exist in nature. He achieved it for the primary time in 2003 and, since then, the imaginative cocina of his analysis group within the biochemistry division on the University of Washington has designed proteins on demand, for use as medicines, as new vaccines – already with success in animal fashions – or as tiny sensors – that permit, for instance, to detect the presence of the drug fentanyl within the surroundings. After being congratulated by the Nobel committee that awarded him the prize, Baker declared that he felt very “excited and honored” and thanked the researchers who preceded and accompanied him in his subject, citing the traditional phrase that Isaac Newton wrote in 1675: “If I have seen further, it has been by standing on the shoulders of giants.”

For his discovery, Baker developed a computing instrument referred to as Rosetta, and in the present day he thanked the contributions of yesterday’s Nobel Prize winners in Physics – for his or her developments in machine studying with neural networks – as elementary to creating the prediction a actuality. of the construction of proteins with AI, for which Hassabis and Jumper have received the opposite half of the Nobel Prize in Chemistry. The three new Nobel Prize winners in Chemistry additionally shared the Frontiers of Knowledge award in 2023, awarded by the BBVA Foundation, within the Biology and Biomedicine class. Demis Hassabis and Geoffrey Hinton – new Nobel Prize in Physics – had been additionally two of the 4 winners in 2022 of the Princess of Asturias Prize for Scientific Research, “for their developments of fundamental neural networks in all fields of science and technology.”

The first AI that revolutionizes science: advantages and dangers

In a press convention provided by Google DeepMind, each Demis Hassabis and John Jumper expressed their “incredible excitement and honor” that the Swedish Academy has acknowledged the analysis to which they’ve devoted their scientific careers. “My life’s task has been to search for applications of artificial intelligence that improve people’s lives and, in particular, that accelerate scientific discoveries,” stated Hassabis, recalling how his childhood curiosity in chess and computer systems led him to develop AlphaGo, an AI that defeated the world champion of Go in 2016, a board recreation whose complexity of doable strikes exceeds that of chess.

AlphaGo was the primary massive success of the corporate DeepMind, co-founded by Hassabis in 2010 and purchased by Google in 2014, and laid the foundations for AlphaFold, the primary synthetic intelligence instrument that has revolutionized scientific analysis. Hassabis made a basic reflection on AI, hours after studying that he had obtained the Nobel Prize in Chemistry for the appliance of that expertise to the research of proteins: “Artificial intelligence is perhaps the most powerful technology that humanity has ever had. It will affect all areas of our lives and, regarding its use, I define myself as a cautious optimist: it has enormous potential for beneficial applications, but it can also be used to cause harm. “This can create unprecedented problems for us and we will need to face the risks of AI,” stated the pc scientist, who took benefit of the event to ask for worldwide cooperation to be ready for the problem that “the next generation of AI tools, which will arrive in 10 years or so: they will have more risks and we will have to be more prepared. “We are going in the right direction with the current precautions, but we have to go faster.”

In the identical look, fellow award-winner John Jumper highlighted that the key of AlphaFold having solved the issue of protein construction is that “artificial intelligence is incredibly good at unraveling patterns that people cannot see. But that is not achieved by ‘just clicking the AI ​​button’. The development of AlphaFold has been a very complex iterative process, developed over years by a great team and the award recognizes the potential of computational biology.” For Jumper, the secret’s that these instruments will assist us perceive illnesses higher and, above all, achieve this a lot sooner. His colleague Hassabis acknowledged that, on this new scientific period that’s opening, “we will have to be cautious: continue using appropriate scientific procedures and, in addition, make sure that we understand these systems and that we can control them.”

Double-edged limitations and transparency

José Antonio Márquez, researcher on the European Molecular Biology Laboratory (EMBL), has expressed his satisfaction for “a well-deserved Nobel Prize that was awarded very shortly after the advances now recognized were achieved.” But for Márquez, who works elucidating protein buildings on the EMBL synchrotron in Grenoble (France), this velocity has not been a shock. “This award was something we expected, because the scientific community has absorbed these advances in a very short time. All laboratories now apply this methodology by default: thanks to AlphaFold, knowing the most probable structure of a protein takes seconds or minutes, while before it took months or years of experimental work,” explains this scientist, who clarifies that for this to be As possible, open access to protein models and the software code used by both David Baker and Demis Hassabis and John Jumper has been essential.

This transparency with the scientific community was the hallmark of the first two versions of AlphaFold, which in just two years managed to predict the structure of 200 million proteins; when in the previous 60 years only the 3D shape of about 200,000 proteins had been determined. However, AlphaFold3 has generated some controversy because its free access is limited to researchers and its code is not available. The great novelty of the third version of this AI is that it now also allows us to predict the interactions between proteins and the rest of the essential molecules of life. This will accelerate the development of new drugs, but it is no longer something that is freely available to the scientific community.

For Alfonso Valencia, director of Life Sciences at the Barcelona National Supercomputing Center (BSC), the 2024 Nobel Prize in Chemistry “recognizes what has become the most significant advance in artificial intelligence. The methods implemented by Demis Hassabis and John Jumper for predicting the structure of proteins from their sequence have become an essential resource in biotechnology and biomedicine.” Valencia has defined, in statements to SMC Spain, {that a} elementary distinction between AlphaFold and different widespread AI instruments, comparable to ChatGTP and its not very dependable responses, “is that these construction predictions [de las proteínas] “They are accompanied by an index of confidence within the high quality of the consequence.”

Márquez recalls that, “despite the high degree of reliability of AlphaFold in predicting the structure of 80% of proteins, there are some types of them with which it is much less reliable.” Furthermore, this researcher recalls another fundamental limitation of these AI tools that use neural networks: “They are very skilled at finding patterns, but they do not explain to us how they do it. So for us it is a brutal advance to have reliable predictions of the structures, and so quickly, but we still do not understand why proteins fold that way and not another way.”

The award is worth 11 million Swedish crowns, about 950,000 euros. After the Medicine and Physics awards, announced on Monday and Tuesday, the round of announcements continues with the Chemistry award, which will culminate on Thursday with the Literature award and on Friday with the Peace award. In 2023, the winners were the Frenchman Moungi Bawendi, the American Louis E. Brus and the Russian Alexei Ekimov, for the discovery of quantum dots.

Since 1901, the Nobel Prize in Chemistry has been awarded 116 instances and has been vacated eight instances—the final in 1942. As this award will be shared every year between as much as three laureates, a complete of 195 individuals have obtained it. Only eight of them are girls: the primary was Marie Curie, in 1911; and the final, Carolyn Bertozzi, in 2022. And solely scientists Frederick Sanger (in 1958 and 1980) and Barry Sharpless (in 2001 and 2022) have received the Nobel Prize in Chemistry twice.