In the new work, the authors have created a neural network that determines what form it will takeThis or that protein
The protein shape is now calculated usinga particle accelerator, which gives three-dimensional photographs of protein molecules, or with the help of a supercomputer, which calculates their structure in accordance with the laws of chemistry and quantum physics.
The authors created an algorithm called evoformer, which attempts to determine the structure of individual segments of protein molecules by representing them as a three-dimensional tree of graphs, a mathematical abstraction that consists of from a set of objects that are connected to each other in pairs.with each other, relying on already known examples, and gradually changesthe structure of connections and the location of nodes, approaching the optimum.
Next, they combined such algorithms and created the AlphaFold2 neural network .
Last year, we unveiled the first version of our system, AlphaFold, which was able to predict the structure of proteins with near-atomic precision as part of the CASP13 competition.Now we have created a new version of it, which is noticeably superior to all competitors in terms of speed and accuracy.At the same time, its source code is completely open.
Demis Hassabis, CEO of Deepmind
As a result, the neural network reconstructs a three-dimensional shape with atomic precision with an error of 0.096 nanometers for each atom inside the protein in about 10 minutes.
Read more:
Explained how the universe is reflected near black holes
Mass poisoning and new versions of the death of civilization: how our knowledge about the Maya changed
Changes in the Earth's orbit contributed to the emergence of complex life on the planet