[ad_1]
An artificial intelligence system with the name AlphaFold, developed by the British company Deep mind, its subsidiary Google, managed for the first time to solve one of his great mysteries of biology for decades: to quickly and accurately predict how to fold protein, what three-dimensional shape they will take and therefore what will be their function.
The achievement, which was marked as a milestone by the scientific world, will help to better understand various diseases and accelerate the development of new drugs, including Covid-19.
In a related competition, won the “smart” AlphaFold system, as it was able to rapidly predict at least two-thirds of the structure of proteins, based solely on the original sequence (chain) of their amino acids, something that humans need a costly and time-consuming process to achieve procedures in the laboratory. In a few days, the algorithm did work that scientists need years in their lab.
“I was very surprised when I saw it.” I never expected to see it as long as I lived, “said Dr. John Mult, a structural biologist at the University of Maryland, one of the organizers of the international scientific competition CASP (Critical Assessment of Structure Prediction protein) biennial.
Vital role
The tens of thousands of different proteins play a vital role in the health and diseases of the body, as cells are made up of proteins, the complex shape of which plays an important role. For example, the coronavirus that causes Covid-19 disease enters human cells because a protruding spiky protein has a structural match, like a padlock key, with a receptor protein in our cells.
The shape of each protein is determined by sequence 20 different amino acids, which form chains to make proteins (for the relevant discovery, Christian ‘Anfinsen was awarded the Nobel Prize in 1972). It is easy for biologists to “read” this sequence, as it is determined by the DNA that encodes it, but it is almost impossible, despite the efforts of about half a century, to predict the shape of a protein based on its tens or hundreds. of amino acids.
To date, scientists have turned to experimental techniquessuch as X-ray crystallography, cryoelectron microscopy and MMR spectroscopy, but they are very difficult as they take months or years and do not always work properly. But the new AlphaFold deep learning system, which began testing in 2018 and has since improved significantly, can now associate each amino acid sequence with a different protein structure. However, it still has room for improvement, for example, it does not do as well with proteins whose structure is affected by interactions with other proteins.
According to the evolutionary biologist Andrei Lupas of the German Max Planck Institute for Developmental Biology, researchers will continue to “check” the accuracy AlphaFold forecasts, but then it will depend solely on artificial intelligence calculations. As he said, the real revolution will occur, when scientists can now use only computers to predict how proteins interact with other molecules. “This will completely change the face of medicine,” he said.
For example, the AlphaFold algorithm already correctly predicted the shapes of several SARS-CoV-2 coronavirus proteins. In the future you will probably have the ability to predict which of the thousands of existing drugs binds properly to these proteins, so that they can have a therapeutic effect, without the need for scientists to carry out experiments as expensive and complex as now.
Currently, according to “New Scientist”, the BBC and “Science”, the creators of AlphaFold have not revealed much Details for their system, but stated that they will soon make a relevant scientific publication. DeepMind has not yet clarified how scientists around the world will be able to take advantage of the new system, but has promised to ensure widespread access.
