Chemical Nobel Prize for Alphafold AI developers to predict protein structures
Chemical Nobel Prize for Alphafold AI developers to predict protein structures
for the first time - and certainly not for the last time - a scientific breakthrough that was made possible by artificial intelligence was awarded a Nobel Prize. The Chemistry Nobel Prize 2024 was awarded to John Jumper and Demis Hassabis by Google Deepmind in London for the development of a groundbreaking Ki-Tools for predicting protein structures called Alphafold , as well as David Baker from the University of Washington in Seattle for his work at the work computer-aided protein design that in recent years was revolutionized by Ki .
The effects of alphafold, the was unveiled a few years ago , are nothing less than transformative. The tool made the prediction of protein structures - often, but not always, highly accurately accessible to researchers and enabled experiments that were unthinkable a decade ago. Biologists now speak of an era "in front of Alphafold" and "to Alphafold".
"It was a dream for a long time to be able to predict the three -dimensional structure of proteins, based on their amino acid sequences. This was considered impossible for decades," said Heiner Linke, the chairman of the Nobel Committee and NanoSh science at the University of Lund in Sweden, during the pronouncement. This year's award winners "cracked the code", he added. The three winners share a price of 11 million Swedish crowns ($ 1 million).
awarded Ki
Deepmind presented Alphafold 2018 when a competition of protein structures, the Critical Assessment of Protein Structure Prediction (CASP), won every two years. But it was the second version of the deep neuronal network, which was presented at the end of 2020 that triggered an earthquake in the life sciences.
Many of the predictions of Alphafold2 at CASP were so precise that they could not be distinguished from experimentally determined protein structures. This was caused by John Moult, co-founder of Casp and computer biologist at the University of Maryland in College Park, In 2020 to explain that "the problem is in some way solved".
Hassabis, co-founder and CEO von Deepmind, and Jumper, head of the Alphafold team, led the development of Alphafold2. In order to predict protein structures, the neural network integrates similar structures from databases with hundreds of thousands of experimentally determined structures and millions of sequences of related proteins - contain information about their forms.
In 2021, Deepmind The underlying code of Alphafold2 free of charge , together with the data required for training the model. A Alphafold database , which was created in cooperation with the European Laboratory for Molecular Biology and the European Bioinformatics Institute in HinxTon, UK, Now contains the structures of most proteins from all organisms represented in genetic databases : A total of around 214 million predict. This year the company presented a Third version of Alphafold , which can also model other molecules that interact with proteins, such as medication.
The revolution that Jumper, Hassabis and her colleagues unleashed is still in the beginning, and the full effects of Alphafold on science may only be known in years. But the tool is already helping scientists to gain new knowledge.
A pioneering team used the tool, together with experimental data, to the nuclear pore complex to map, one of the largest machines in our cells, the molecules in the cell nucleus. Last year, two teams analyzed the entire Alphafold database to discover the deepest corners of the protein universe, identified new protein families and folds as well as surprising connections in the machine of life.
Many researchers hope that Alphafold and other AI tools that inspired it will transform medicine. However, it is still unclear , how, or whether Alphafold will rationalize the costly and multi -stage process of developing safe new medication.
create new proteins
More than a decade before Deepmind began to work with Alphafold, the computer -aided biophysicist David Baker from the University of Washington in Seattle and his colleagues developed software tools for modeling protein structures based on physical principles and are called Rosetta. The tool had early successes When designing new proteins .
Over the years, Baker's Team Rosetta turned to predict protein structures-it was one of the best participants in numerous casps before Alphafold's recent dominance-as well as for the design of new proteins such as enzymes and self-putting protein nanoparticles.
When Alphafold2 was announced-but not yet published-Baker and his team, including the computer-aided chemist Minkyung Baek, who is now working at Seoul National University in South Korea, went to understand the software and to apply some of its tricks to an earlier AI-based version of Rosetta. The first version of the resulting rosett fold network performed almost as well as Alphafold2. Since 2021, both networks have been continuously improved by their developers and other scientists in order to overcome new challenges, for example the prediction of the structure of complexes from several different interacting proteins.
In recent years, Baker's team has been particularly productive in using machine learning to the Raison d’être of his laboratory: to create new proteins that have never been seen in nature . A recently developed tool from Baker's team that combines rose stock with image-generating diffusion neuronal networks has led to a quantum leap in the ability of the researchers to design protein design.
Although computer -aided tools such as Alphafold are not a substitute for experimental studies, they are an accelerator, say scientists. "This will enable a new generation of molecular biologists to ask more advanced questions," said Casp-Richter Andrei Lupas, an evolutionary biologist at the Max Planck Institute for Development Biology in Tübingen, Germany, 2020 to Nature .