Oct. 17, 2024
Comments from RIKEN researchers regarding the 2024 Nobel Prize in Chemistry
We would like to extend our heartfelt congratulations to Dr. David Baker, Dr. Demis Hassabis and Dr. John M. Jumper upon winning the Nobel Prize in Chemistry. The following are messages from RIKEN researchers.
- Yuji Sugita, Team Leader, Computational Biophysics Research Team, RIKEN Center for Computational Science (Deputy Director, RIKEN Center for Computational Science)
The prediction of the three-dimensional structure of proteins using amino acid sequence information and the molecular design of proteins that do not exist in nature were major challenges for the life sciences. I would like to extend my heartfelt congratulations to Dr. David Baker, Dr. Demis Hassabis, and Dr. John Jumper upon their Nobel Prize in Chemistry for their elucidation of these issues. I hope that research in the life sciences using AI and machine learning will continue to develop solutions to unresolved life science problems, drug discovery, the advancement of medical care, and more. We would like to actively engage in research on the development of AI for scientific research and the scientific use of AI (known as AI for Science) at RIKEN. We are also working on research and development for the next-generation supercomputer “Fugaku NEXT" (tentative name) that will form the basis for AI for Science in Japan.
At the same time, I think we must not forget the contributions of experimentalists who have carried out experiments to generate the big data used in AI and machine learning, and researchers who have worked to elucidate the mechanisms of three-dimensional structure and molecular design.
- Yasushi Okuno, Director, HPC- and AI-driven Drug Development Platform Division, RIKEN Center for Computational Science
Following the Nobel Prize in Physics, I was both surprised and delighted to see that the Nobel Prize in Chemistry was also awarded for research related to machine learning and AI. Until now, machine learning and AI have been considered to be convenient auxiliary tools in the life sciences and other areas of natural science, but I believe that this award is a historic moment demonstrating that machine learning and AI are powerful technologies that can transform the natural sciences and that they are the equivalent of wonderful inventions and discoveries.
AlphaFold, the AI which was one of the subjects of the award, has predicted the three-dimensional structure of all the proteins currently known (over 200 million), and these findings have been published in a database. It would have taken decades to determine the structures of over 200 million proteins through experimentation, but by using AI to automate the process, the results were obtained in just one or two years, in a true testament to the power of AI.
AlphaFold can also predict the three-dimensional structure of proteins that do not exist in nature, or artificial proteins, and it is expected to be a very useful tool for designing pharmaceuticals. In fact, we have implemented the public version of AlphaFold, OpenFold, on Fugaku, and are working on applications for drug discovery and life science research.
On the other hand, AlphaFold also has its limitations, and in reality there are many proteins for which the prediction results are uncertain. RIKEN is still working on this difficult problem, and we have been encouraged by the award to continue pushing forward with further research and development.
- Makoto Taiji, Program Director, Advanced General Intelligence for Science Program (AGIS), Transformative Research Innovation Platform of RIKEN platforms (TRIP) Headquarters (Deputy Director, RIKEN Center for Biosystems Dynamics Research (BDR) )
The success of AlphaFold and RoseTTAFold, AIs developed for predicting protein 3D structures, was the first example of a grand challenge problem in science being solved by AI, and it was an amazing achievement that opened up the possibility of using AI for science. In addition to the accumulation of high-quality experimental data, the rapid development of machine learning beginning with the development of deep learning, and improvements in computational power using GPUs, etc., I think that the ability to grasp the structural characteristics of proteins was also made possible by the accumulation of knowledge by experts in this field. This event made us keenly aware of the great potential of intensive collaboration between top researchers in AI and other fields of science, and it was one of the reasons that led us to launch an AI development program for science at RIKEN. At RIKEN, we also believe in the potential for collaboration between researchers from different fields, and we will continue to work towards the realization of accelerated research through AI for science.