Science & Technology

Deep learning dreams of new protein structures. – Science Inquirer

Troy, NY — Just as you can use artificial intelligence to create compelling images of cats, you can now use similar tools to create new proteins.In the report Nature, A team of researchers University of Washington, Rensselaer Polytechnic University, When Harvard University We will discuss the development of a neural network that “translucents” proteins with a new stable structure.

Sci_Tee limited to Scientific Inquirer! Science. make. life. Better. (Dancing).

“The potential to illusion brand new proteins that bind to specific biomolecules or form the desired enzyme active site is very exciting,” said Gaetano Monterione, a professor of chemistry and chemical biology at Lenselia. I am saying. The network has been analyzed.

Proteins are filamentous molecules found in all cells that spontaneously fold into complex three-dimensional shapes. These folded shapes are key to almost every process in biology, including cell development, DNA repair, and metabolism. However, the complex shape of proteins makes them difficult to study. Biochemists often use computers to predict how a protein string or sequence will fold. In recent years, artificial intelligence technologies such as neural networks and deep learning have revolutionized the accuracy of this task.

[送信]By clicking, you agree to share your email address with the site owner and Mailchimp to receive marketing, updates, and other emails from the site owner. Use these email unsubscribe links to opt out at any time.


success! You are on the list.

husband! An error occurred and the subscription could not be processed. Please reload the page and try again.

“This project created completely random protein sequences and introduced mutations into them until we predicted that the neural network would fold into a stable structure,” said co-lead author Postdoctoral fellow at Baker Institute. Said Ivan Anishchenko. Protein Design Institute At the University of Washington School of Medicine. “The software never led to specific results. These new proteins are exactly what computers dream of.”

In the future, the team believes they should be able to manipulate artificial intelligence to produce new proteins with useful functions. “We want to use deep learning to design functional proteins, such as protein-based drugs and enzymes,” said Samperok, a postdoctoral fellow at Baker Institute.

The research team generated 2,000 new protein sequences that were predicted to fold. Over 100 of these were manufactured and studied in the laboratory. Detailed analysis of these three proteins confirmed that the computer-predicted shape was actually achieved in the lab.

“Our solution NMR studies show the amazing accuracy of protein designs created by the illusion approach, along with the X-ray crystallography determined by the Washington University team,” said a senior research scientist at the Monte Rione Institute. Teresalamelot, who is, says.Within Renseria Biotechnology and Interdisciplinary Research Center..

Monterione said: Related protein sequences. ”

“This approach greatly simplifies protein design,” said senior author David Baker, who won the 2021 Breakthrough Prize in Life Sciences. “Previously, in order to create new proteins of a particular shape, people first carefully studied the relevant structures in nature, came up with a set of rules, and applied them to the design process. For each new type of fold. Needed a new set of rules. Here, by using a deep learning network that already captures the general principles of protein structure, we eliminate the need for fold-specific rules and functional parts of proteins. Expands the possibility of focusing directly on just that. ”

“Finding the best way to use this strategy for a particular application is an active area of ​​research today, and we look forward to the next breakthrough,” Baker said.

Image Credit: Rensselaer Polytechnic Institute

Deep learning dreams of new protein structures. – Science Inquirer Deep learning dreams of new protein structures. – Science Inquirer

Back to top button