These awards recognize what has become the most significant advance in Artificial Intelligence. The methods implemented by Demis Hassabis and John M. Jumper for protein structure prediction from their sequence - AlphaFold based on DeepNN - have become an indispensable resource in biotechnology and biomedicine. A fundamental difference with respect to other AI developments, for example, the popular ChatGTP, is that these structure predictions are accompanied by a confidence index in the quality of the result. No less important is the contribution of David Baker, who for years has been the leader in the field of applying AI to the design of new proteins, with impressive results in the design of proteins with new properties with applications in biotechnology. His research has broken new ground in the exploration of the protein space with practical and scientific implications.

It is noteworthy that David Baker has led both the movement for the open release of the software and for the responsible use of these new technologies. Unfortunately, the work of Demis Hassabis and John M. Jumper represents a less bright page in this regard since, although their first developments (AlphaFold 1 and 2) were open source, the subsequent ones (AlphaFold 3) have not been, creating a great deal of controversy in the scientific community. Modestly, I am very happy to see that my developments of the first methods for obtaining constrictions from sequence alignments (correlated mutations) in the 90's are recognized by the community as the key piece on which these new methods are built.

The 2024 Nobel Prize in Chemistry recognizes the spectacular progress achieved in recent years in protein engineering through theoretical methods. The academy acknowledges the work of the field's pioneer, D. Baker, who has contributed to continued and systematic advances in our ability to predict protein folding, as well as the more recent work of Jumper and Hassabis, the lead authors of AlphaFold, a disruptive structure prediction algorithm based on artificial intelligence, which in just a few years has revolutionized and democratized the field of protein structural biology.

The advances awarded by the academy have enabled not only the assignment of structure to amino acid sequences, but also the design of new proteins with sequences and functions unexplored by nature, opening up unforeseen possibilities in molecular engineering.

It seems to me one of the most deserved Nobel Prizes. Baker has been the leader of the greatest revolution in bioscience in recent years, being able to begin to master the language of proteins. We are now able to predict the spatial structure and design of synthetic proteins, both of which were impossible just a short time ago.

The two lines of research recognized with this year's Nobel Prize in Chemistry are clearly disruptive. A protein’s function depends on its structure, and this, in turn, depends on the sequence. While sequences can be easily read, determining structures requires a great deal of effort. The transition from an amino acid sequence to a three-dimensional structure in the cell fundamentally depends on the laws of physics, which are known but involve millions of interactions and possibilities.

For decades, we believed we could eventually reconstruct how an amino acid chain folds by following these laws, but all attempts fell short. Hassabis and Jumper took a shortcut, using artificial intelligence models trained on databases of structures already determined by crystallography. Although this method is essentially a black box that doesn’t explain the folding process, it is capable of accurately predicting structures solely from sequences.

AlphaFold is now an indispensable tool in biological research, where it has already opened up new horizons. On the other hand, Baker has demonstrated the power of designing folds that don’t exist in nature using the same building blocks that nature uses. Ultimately, he has opened the door to a new chemistry inspired by nature, with the ability to generate synthetic proteins with interesting properties.

It’s wonderful to see this area of research getting well-deserved attention. We are seeing the first glimpses of a new era in biology. Massive publicly available datasets, combined with deep learning, is enabling discoveries which are transforming healthcare, drug discovery, material science, and so much more. This is just the start.

It's great to see computational work recognised for its impact in biology and biochemistry. The winners have for the most part been using data that is publicly available; the impact and novelty of their work really stems from the modelling approaches. Their work is a great example of how basic science can have an impact that ripples across diverse applications. And it also a reminder of the opportunity for discovery in these fields with the technological advances we are seeing in the past decades and the quantity and quality of data being generated by the community, this really is the century of biology, and in particular quantitative and predictive biology.

The problem with predicting the three-dimensional structure of a protein from its amino acid sequence, known as the 'protein folding problem', has been a central challenge not only for biology but also for chemistry and physics. Its importance lies in the fact that understanding how proteins fold is crucial to understanding their function in organisms and, by extension, in life itself. Furthermore, this understanding has significant practical applications, such as the design of optimized enzymes for industrial processes and the development of antibodies to combat various diseases.

The reason it is so important to know the three-dimensional structure of a protein is that its function primarily depends on its shape and not just on the specific amino acid sequence. Very different sequences may lead to similar shapes with practically identical functions, just as small changes in a protein sequence can denature it and destroy its function. For decades, physicists have tried to predict these structures by modeling the interactions between amino acids. However, the challenge is twofold: first, it is necessary to accurately model these interactions, which requires very well-calibrated force fields; second, even with good modeling, finding the minimum energy structure (i.e., the equilibrium state) is extremely slow from a computational perspective. This is because protein folding is a highly complex optimization problem, with many interactions that can be oppositional in nature. To date, molecular dynamics simulations have only been able to effectively reproduce the structures of very small proteins.

In the last decade, the approach to the protein folding problem has radically changed, primarily due to the massive accumulation of protein sequences in databases, made possible by the drastic reduction in the costs of genomic sequencing. The new idea was simple but innovative: although we do not fully understand how to model the interactions between amino acids, we now have access to a vast amount of data on protein sequences and their viable mutational variations, meaning those that have survived evolutionary pressure.

Instead of trying to model the interactions at a physical level, researchers began to statistically study families of 'homologous proteins', that is, sequences with similar functions in different but evolutionarily related organisms. From this data, they were able to infer two key things: first, which amino acids could not mutate in isolation without denaturing the protein; and second, which pairs of amino acids needed to be in contact in the three-dimensional structure, as a mutation in one would destabilize those critical contacts and, consequently, the structure.

This bioinformatics approach, completely 'data-driven', combined with improved models that allowed for the identification of correlations beyond pairs of amino acids, enabled effective learning of 'important mutational couplings', that is, the constraints on how amino acids could change without altering the function of the protein. Subsequently, this strategy was combined with supervised 'machine learning' methods, where models learned to predict the three-dimensional structure of proteins whose structures were already known, using their sequences as a training base. 

This approach culminated in a historic milestone in 2020 during the CASP (Critical Assessment of Structure Prediction) competition when AlphaFold2 was able to predict the structures of proteins that had never before been experimentally resolved with great accuracy. Surprisingly, this included proteins with very different sequences from those studied previously, where traditional methods failed spectacularly. Thus, the protein folding problem was practically solved, not through detailed physical modeling of its components, but by imitating patterns from stored evolutionary data.

This achievement has truly revolutionized computational biology, where the combination of large volumes of data with the power of artificial intelligence has surpassed decades of attempts based solely on physical models.

The Nobel Prize in Chemistry awarded to David Baker, Demis Hassabis, and John M. Jumper represents a crucial moment in the field of structural biology and protein design. The combination of computational protein design and accurate prediction of their structures marks the beginning of a new era in synthetic biology and precision medicine. The ability to create and predict protein structures not only constitutes a significant technical advancement but also has the potential to revolutionize key fields such as biomedicine, biotechnology, and the development of personalized therapies.

I am very pleased that the field of computational protein design is being recognized with the Nobel Prize in Chemistry along with the development of the AlphaFold2 tool. AlphaFold2 demonstrated at the protein-structure prediction challenge (CASP) conference in 2020 its ability to predict the 3D structure of proteins with high accuracy, and since then the number ofdeep learning-based methods for its application in the field of protein design has increased exponentially. “It will change everything”, ‘it's a game-changer’, these are some of the headlines that were published at the time.

The reality is that the field of protein design is advancing at a great speed, every day scientific articles focused on artificial intelligence are published, and in the last 4 years we have advanced a lot, protein designs that previously needed years of research now thanks to AlphaFold2 and the methods developed by Baker's group (among others) can be solved in a matter of minutes.

Computational protein design is based on prediction algorithms for both the structures of proteins with known sequences and the amino acids that stabilize these structures. The implications of these methodologies in the fields of structural biology, biomedicine, and new materials are extraordinary. On one hand, it is exponentially accelerating the elucidation of the three-dimensional shape of receptors, molecular signaling molecules, transcription factors, enzymes, antibodies, and more; on the other hand, it has irreversibly opened up new areas of research dedicated to creating protein structures that do not exist in nature with 'custom-made' properties, such as biological drugs, vaccines, catalysts, and more.

The 2024 Nobel Prize in Chemistry awarded to Baker, Hassabis, and Jumper recognizes the enormous relevance and utility of computational tools for advancing knowledge in structural biology and their application in the generation of new therapies, materials, and bioreactors.
 

Proteins are made up of small building blocks, called amino acids, which are organized to create a variety of shapes that enable them to perform their functions. Until the advances made by the 2024 Nobel Prize in Chemistry recipients, there was a dogma that the structure of a protein, and therefore its function, could not be predicted by simply 'reading' its amino acid sequence, in the same way as our genetic code. Biophysical studies were then necessary, such as X-ray crystallography, nuclear magnetic resonance or electron cryo-microscopy, the latter also awarded the Nobel Prize in Chemistry in 2017 for its advances, which did not always come to fruition.

The Nobel Prize in Chemistry 2024 discoveries and advances are capable of translating that amino acid readout, predicting its structure and, in turn, helping to understand its function, all computationally, and supported by other experimental techniques for validation. The impact of these advances spans all fields of fundamental biology (human, animal or plant), medicine or drug development. These advances, together with recent advances in computation and artificial intelligence, could position society at a moment of progress as important as that experienced during the industrial revolution.

The Nobel Prize awarded to Baker, Hassabis and Jumper is certainly a well-deserved recognition and highlights among other things the great potential of computational capabilities to solve scientific questions. Their recent contributions have radically transformed the field of structural biology by accelerating, among other things, research in drug design. It is not possible to live this moment without being amazed to see that what used to involve years of research can now be obtained with a simple click, although behind that click there are also many years of research and work. Today is undoubtedly the demonstration with the greatest impact of the use of artificial intelligence that solves one of the most complex issues in biology. It is also a recognition of the area of computational biology and bioinformatics, which in my opinion will continue to transform scientific research.

This work provides an answer to one of the great challenges in biology, how to predict the structure of proteins from the sequence encoded in DNA with a high degree of reliability. Structure predictions for hundreds of millions of different proteins are now available worldwide in the EMBL EBI databanks and the prediction programmes and algorithms developed by these teams have become an indispensable tool in most biology laboratories. They allow us, for example, to understand the relationship between the structure and function of a protein, why certain genetic mutations can lead to disease, or how proteins interact with each other to carry out more complex functions. It was an expected award and demonstrates how new artificial intelligence technologies can contribute to solving very complex scientific problems.