Using artificial intelligence, scientists said they can now predict with “unprecedented accuracy” how all of life's molecules interact with each other.
The program, called AlphaFold 3, could help accelerate the search for new drugs and treatments for some of humanity's most devastating diseases, including cancer, Parkinson's disease, malaria and tuberculosis, its creators say. says one Google DeepMind.
AlphaFold 3 explores how the complex shapes and networks of molecules present in every cell in the human body are connected, and how even the slightest changes in these can affect biological functions that can cause disease. You can imagine the impact.
It also helps scientists predict how these molecules will interact with potential treatments such as antibodies and drugs.
Sir Demis Hassabis, founder and chief executive of London-based DeepMind, said the program will allow researchers to “drastically speed up the drug discovery process” and “improve our understanding of the biological world.” It said it would provide a “toolset” that would enable people to “transform.”
All living cells run on what researchers describe as “molecular machines,” made up of proteins, DNA, small molecules known as ligands, and more.
“By observing how they interact across millions of combinations, we can truly understand life's processes,” the DeepMind team wrote in a blog post.
When provided with a list of molecules, AlphaFold 3 can generate a combined 3D structure of them and predict how they will all fit together.
The team says AlphaFold 3, which powers the company's free tool known as AlphaFold Server, is 50% more accurate than the best traditional method available, reducing predictions that typically take months or years to within seconds. It is said that it can be generated.
Dhavanthi Harharan, product manager at DeepMind, said AlphaFold Server is a “one-stop solution to generate large quantities of biomolecules with just a few clicks of a button.”
“This is currently the world's most accurate tool for predicting how proteins will interact with other molecules.”
Lord Demys said the tool builds on the “huge milestone in structural biology” of AlphaFold 2, an early model that predicted the structure of almost every protein made in the human body.
In December 2020, AlphaFold was recognized by the organizers of the Critical Assessment of Protein Structure Prediction (Casp) as a solution to the grand challenge of protein structure prediction, spanning 50 years.
Details about AlphaFold 3, which also included a team of experts from Isomorphic Labs, an independent company from DeepMind, were published in the journal Nature.
The researchers said further work would be needed to improve the model's accuracy and would “incur additional computational costs.”
Matthew Higgins, EP Abraham Professor of Structural Biology at the University of Oxford, who is using AlphaFold to research malaria vaccine candidates, says the tool will “make a huge difference to the ability of scientists across biomedical research to investigate You will understand how the machine works.” Cells work. ”
Commenting on AlphaFold 3, Dr Nicole Wheeler, Birmingham Fellow at the Institute of Microbial and Infectious Diseases at the University of Birmingham, said: “It extends what can be done with these AI tools for biological understanding and engineering, including biological design. “It brings a lot of expectations to the top.” These are the parts that control gene expression and the parts that design small molecules to treat diseases. ”
