Can say that Life is to search for order in the disorder. In the universe, seeking entropy, life creates small molecules that combine and order against the global cosmos trend. For this, accepts the most basic works and connects them together, forming long chains. In the case of the polysaccharidesson sugar eg glucose and fructose, in Adv These basic works are called nucleotides, and in squirrels They have by name amino acids.
After joining, long chains allow you to store energy, information either becomes a structural or functional part of the cells. A Protein, thanks to their versatility, are some of the most complex cellular structures Because they can form chains of thousands of links in which 23 different amino acids are combined.
But finding the order of these amino acids is only the first step in finding a protein function. After formation to be functional, It needs to be folded very specifically In the molecular origami, called the tertiary structure.
This composition, which sometimes occurs automatically and another time, helping with special proteins called Shaperon, is one of the most difficult puzzles. To find out where each amino acid is in three dimensions of space The physical and chemical properties of each amino acid must be taken into account. Which is a protein and achieve what is considered a minimum energy condition.
The solution of these puzzles with the help of gross force requires a huge computing power, but certain shortcuts can be found with the help of artificial intelligence tools.
The most complicated puzzle
During the last decade they have developed Puzzle games To avoid using so much computing power. In these games, anyone who can, through a personal computer, use their brain to collaborate with other players and find this minimum energy condition in proteins.
Huge success platforms, It had thousands of active players.
But in 2020 everything changed, because only in the second year participated in one of these competitions Artificial intelligence swept the competition. Alphafold, developed by Deepmind Google, managed to find a structure of 200,000 proteins with a precision similar to the experiment.
Also in the 2021To make any laboratory in the world to use its conclusions. Out -with this huge jump in their capabilities, its developers David Baker, Demis Hasabis and John Jamper win the Nobel Chemistry Prize in 2024.
But not everything stops, The third version This II not only allows us to find the structure of proteins, but can also reveal how two molecules can interact with each other. This new ability is used to search Possible undesirable interactions between drugsand promises revolution in accurate medicine
Where Alphafold does not reach
But all the discoveries that seem endless, sooner or later, getting into the barrier. In this case, the Achilles heel from the essential deposit is the proteins whose structure is not fixed.
Counted that at least 30% of human proteins are not perfectly stable, But they constantly change the shape of the external stimuli. Thus, they can respond to the environment and find out whether everything works properly or emit signals if it is not. Therefore, they are necessary works for life. These proteins called “Actually dissolved proteins” create an even more problem for both the most powerful II and the people who study them.
But the team, consisting of Harvard members and the north -western, has developed a method that can develop these essentially disordered proteins from scratch, taking into account the energy of different states. To do this, it uses the type of machine learning called “Automatic Differentiation”, which is traditionally used to train neural networks.
Thus, the proteins they develop through this method are “differentiated” and They are not based on the decision that found the II And it may not be perfect but we also take into account the people Physical and chemical properties of atoms and the behavior of protein in nature. The team hopes to use their findings to find new drugs and finally find out this A huge set of tiny pieces that allow life as we know: ordered to a certain extent.
(Tagstotranslate) National (T) Geographical (T) Science (T) Character (T) History (T) Journey
