April 18, 2024

Medical Trend

Medical News and Medical Resources

David Baker’s team achieved a milestone breakthrough in AI design protein

David Baker’s team achieved a milestone breakthrough in AI design protein


David Baker’s team achieved a milestone breakthrough in AI design protein, creating new and functional protein nanoparticles through reinforcement learning.

If life is compared to a precise mechanical clock, then proteins are internal gears, large and small, and as the main bearer of life activities, they play a key role like the cornerstone of life. The structure of a protein is the basis of its function

. If we can freely modify proteins or even design new proteins from scratch, this will change or even subvert the development of the entire molecular biology.


The proteins that make up organisms evolve gradually after a long period of evolution. It is very difficult to design a brand new protein out of nothing, because the limited computing power of the human brain cannot predict the spatial structure of proteins and the functions determined by these structures.

With the development of artificial intelligence (AI) technology, de novo protein design is becoming a reality step by step.


On April 20, 2023, the well-known protein design expert, Professor David Baker of Washington University School of Medicine and postdoctoral fellow Wang Shunzhi published in the journal Science entitled: Top-down design of protein architectures with reinforcement learning (top-down design based on reinforcement learning) Research paper on protein structure design under .


The study developed a reinforcement learning-based protein design software and demonstrated its ability to create functional proteins. The breakthrough could usher in a new era of protein design, with positive implications for cancer treatments, regenerative medicine, potent vaccines and biodegradable everyday products.


Professor David Baker , the corresponding author of the paper , said: This study showsReinforcement learning can do more than master board games. While trained to solve long-standing puzzles in protein science, it also excels at creating useful protein molecules. If this approach is applied to the right research questions, it can accelerate progress in a variety of scientific fields.


David Baker's team achieved a milestone breakthrough in AI design protein


At present, artificial intelligence (AI) has shined in many fields and penetrated into our daily life. From AlphaGo in the field of Go to AlphaFold for predicting protein structures , from AI painting to ChatGPT , which is popular all over the Internet , artificial intelligence, as an emerging disruptive technology, is gradually releasing the huge energy accumulated in the technological revolution and industrial transformation, and will It has profoundly changed the way of life and thinking of human beings.


AlphaGo’s ability to defeat top human professional Go players relies on a machine learning system called reinforcement learning , in which computer programs learn how to make the most correct decisions by repeatedly trying and receiving feedback.


Going back to protein design, if a protein is compared to a Go game, then the protein domain is a Go pattern . From this point of view, artificial intelligence software based on reinforcement learning can also be applied to the de novo design of proteins-through a lot of training, a powerful new protein design software is finally obtained.


David Baker's team achieved a milestone breakthrough in AI design protein

Bottom-up (left) and top-down (right) protein assembly design strategies


In order to create such an AI software that can be used for protein design, the research team input the sequence and structure information of millions of simple proteins into the computer.

Then, this AI software made tens of thousands of attempts, and each time feedback improvement , in order to achieve the predetermined goal – to design a new protein from scratch. In this process, the computer lengthens or bends the proteins in specific ways until it learns how to fold them into the desired shape.


David Baker's team achieved a milestone breakthrough in AI design protein

Top-down protein design strategies and computational pipelines


The research team designed hundreds of proteins through this reinforcement learning software, and performed gene cloning, protein expression and structure determination in the laboratory.

To measure the accuracy of the software, they determined the actual structure of these AI-designed proteins through equipment such as an electron microscope, and found that it was very consistent with the protein structure predicted by the software.


The research team focused on designing novel nanoscale structures composed of many protein molecules, which required the proteins they designed to have chemical interfaces that allow the nanostructures to self-assemble.

So the research team looked at the nanostructure of the AI-designed protein and found that each atom was in its intended position. In other words, this reinforcement learning software is capable of designing with atomic precision, with deviations between expected and actually realized nanostructures being on average smaller than the width of a single atom.


David Baker's team achieved a milestone breakthrough in AI design protein

The near-atomic-resolution cryo-EM structure of the designed protein capsid matches the design model


In addition, the research team also showed through the primary cell model of vascular cells that this reinforcement learning software can also optimize the protein scaffold structure. For example, it could be more effective at promoting blood vessel stability by allowing denser aggregation of cell receptors on more compact scaffolds.


Application of viral protein capsids designed by reinforcement learning software


All in all, the research, published in Science , shows that artificial intelligence software based on reinforcement learning can be applied to protein design, creating completely new proteins “top-down” and accurately predicting their shape and structure.

This study is a milestone in the use of artificial intelligence for protein science research, and its potential applications are enormous. Through this technology, scientists can efficiently create therapeutic proteins, vaccines and other molecules, and will profoundly change the research field of disease treatment, cell development and aging and biodegradability.


Professor David Baker is known as the ” Father of Protein Folding ” in the research field , and won the 2021 Scientific Breakthrough Award·Life Science Award .


Professor David Baker founded a company called Monod Bio based on laboratory research results, which is also the first company to design proteins from scratch for biosensors and medical diagnostics. Enzyme platform and other core technologies.

The company completed a $25 million seed round in August 2022.






Paper link :


Checkmate, Proteins! Reinforcement Learning Transforms Molecular Biology

David Baker’s team achieved a milestone breakthrough in AI design protein

(source:internet, reference only)

Disclaimer of medicaltrend.org

Important Note: The information provided is for informational purposes only and should not be considered as medical advice.