lswcap
DeepMind, an artificial intelligence company affiliated with Google, announced that it has open-sourced AlphaFold, a protein three-dimensional structure prediction analysis algorithm that it has been developing for a long time, and has released the source code to GitHub for free. DeepMind has made AlphaFold open-sourced, revealing that users can unravel the three-dimensional structure of proteins with unprecedented precision.
It is a high molecular compound that combines 20 kinds of amino acids decoded from the nucleotide information genome that composes protein genes that perform various functions in the human body. However, only the primary sequence information of which amino acids are linked can be known from the base information. To analyze how a protein works in the body, it is necessary to know the three-dimensional structure of the protein. However, the three-dimensional structure cannot be known only from the amino acid sequence information. The problem of how to elucidate the conformation of this protein is working on protein folding.
The three-dimensional structure elucidation itself is already in progress, and in the past, the three-dimensional structure of a protein was identified by nuclear magnetic resonance, low-temperature electron microscopy, X-ray diffraction, etc., but no one had a difficult problem unless it was a research team and a medical institution with a high budget and high cost. In addition, a lot of computational power is required to elucidate the actual three-dimensional structure from data, and it is necessary to analyze distributed computing operations such as supercomputers and Folding@home, so there is not only money but also time costs.
Alpha Fold, developed by DeepMind in 2018, is a technology that predicts the three-dimensional structure of a protein from gene sequence information using machine learning. Therefore, it is possible to elucidate the three-dimensional structure of a protein without incurring the same cost as before. In fact, it was reported that the evaluation of Alpha Fold at the protein structure prediction contest surpassed other computer programs.
Also, on July 16, 2021, DeepMind co-founder and CEO Demis Hassabis announced that Alphafold version 2.0 would be open-sourced and the source code released on GitHub. At the same time, a paper on Alphafold was published in the journal Nature. Regarding AlphaFold open-source, Google CEO Sundar Pichai, the parent company of DeepMind, said he was delighted that DeepMind AlphaFold technology was released along with open-source code. said. Related information can be found here.
Add comment