David Baker, Demis Hassabis and John Jumper revolutionized the study and design of proteins with Artificial Intelligence and promoted the development of new treatments against multiple diseases through this technology. Baker, a professor of biochemistry at the University of Washington and an investigator at the Howard Hughes Medical Institute, is the creator of the program. RoseTTAFoldwhile Hassabis and Jumper, CEO and senior researcher, respectively, of Artificial Intelligence company DeepMind, are the authors of AlphaFold2, two computational systems that are revolutionizing the knowledge of proteins. “Scientists are now using these new methods to predict interactions between proteins, design completely new proteins and find new drug targets”, emphasizes the jury of the BBVA Foundation Frontiers of Knowledge Award in Biomedicinean award that these three scientists have just obtained.
“Until now -says Óscar Marín, from the Medical Research Council at King’s College London (United Kingdom) and secretary of the jury-, years of very hard work in the laboratory were needed to predict the structure of a single protein, but thanks to advances achieved by the three awardees now we only need a few minutes on the computer to achieve this goal. Contributions from Baker, Hassabis and Jumper, he adds, “They will allow us to advance much faster in the development of therapies for multiple diseases».
In addition to predicting how naturally occurring proteins fold, the Baker-powered RoseTTAFold computer program has also proven very useful for designing completely new proteins from a simple description of the functions they are intended to perform. Thus, the program makes it possible to obtain proteins to block proteins from the flu virus or Covid-19 to cancer cells, and the results have been successfully verified in the laboratory.
“Novel proteins can be improved medicines, so there are many exciting new medical applications. For example, it will be possible to create new vaccines or new medicines to treat cancer”, explains Baker.
Thanks to these tools, the structure of almost all documented proteins is now known, not only human but also animal, plant and even bacterial. This knowledge has an immediate application in the creation of new drugs and vaccines..
“We have already seen that AlphaFold is applied to a wide variety of problems,” says Hassabis. “Some of the things we’re most excited about are using it for drug discovery, for example, to fight antibiotic resistance, or to try to find cures for diseases like malaria.”
New vaccines or new drugs may be created to treat cancer
David Baker
Howard Hughes Medical Institut
For Gonzalo Jiménez Osés, from CIC bioGUNE, one of the facets with the greatest potential for this contribution is in the biomedical area. “An immediate bypass is going to be the design of new drugs,” he points out, while, in the field of classic drug development, “we are going to discover new therapeutic targets.”
In his opinion, “the most important thing is that it will be possible to know more precisely the interaction networks between proteins that take place in a disease, such as cancer and processes of the immune system, much more quickly, and that It will lead to new treatments because computer simulations of these complex processes are going to be much more reliable.”
«AlphaFold has already impacted biological research with great impact in a very short time», highlights Hassabis for his part. “We know that more than a million researchers have used the structures predicted by AlphaFold in their research, and virtually every pharmaceutical company in the world has used AlphaFold in their drug discovery programs.”
For the moment, the greatest impact for the creation of new vaccines and drugs is the design of artificial proteins with bespoke properties. The latest update to RoseTTAFold even allows you to create proteins from simple descriptions.
Virtually every pharmaceutical company in the world has used AlphaFold in their drug discovery programs.
Thus, Jumper has collaborated with a research group from the Oxford University who was working on development for a malaria vaccine.
“I think AlphaFold really represents the first powerful example of how deep learning is able to capture the complexity of biological systems and really develop mathematical understanding of extraordinarily complex things,” Jumper explains.
“We believe that almost all of medicine will be transformed by the protein design revolution,” Baker says. Most of today’s medicines are made by introducing small modifications to proteins that already exist in nature. Now that we can design entirely new proteins, we can develop much more sophisticated and improved drugs that, for example, can treat cancer without the side effects, be made very quickly in the event of a new pandemic outbreak, and will generally be more precise and robust”.
AlphaFold truly represents the first powerful example of how deep learning is able to capture the complexity of biological systems.
Jiménez Osés believes that the real potential of these technologies “is totally unpredictable”, in a positive sense.
Although these Artificial Intelligence programs have not completely replaced experimental techniques, for now they have burst into force as a complement to them, revolutionizing the field of biology.
Thus, hundreds of research groups throughout the world are applying the predictive power of AlphaFold to their lines of research, acknowledges Jiménez Osés, since both RoseTTAFold and AlphaFold2 are free access tools for the scientific community, and the improvements that are have recently implemented have almost equalized the computing times that each one needs.
For example, the Spanish researcher points out, “at CIC bioGUNE we are already working on a project focused on a protein related to a hereditary disease that damages the nervous system called Ataxia de Friedreich».