ON A solid understanding of the structure of a protein can provide crucial insights into the mechanism of certain biological processes or provide a starting point for the development of a new drug. AlphaFold, an artificial intelligence program run by UK company DeepMind, has made significant strides in reducing the time it takes to predict the structure of a protein with unprecedented accuracy from months to minutes. Well, a paper published July 22nd in nature reports that a collaboration between AlphaFold and the European Molecular Biology Laboratory (EMBL) has created a public database of more than 350,000 protein structures.
“This understanding means that we are better equipped to decipher the molecular mechanisms of life and accelerate our efforts to protect and treat human health and the health of our planet. Innovation for scientists around the world,” says Edith Heard, Director General of EMBL The guard.
The human proteome – all proteins that human DNA is known for – consists of around 20,000 proteins. Laboratory analyzes have confirmed the structures of only about 17 percent of these molecules. Before the emergence of neural networks and modern computer processors, computational predictions of structures took a long time and were often imprecise. DeepMind reports that the new database contains structures for 98.5 percent of the human proteome with confidence or a high degree of confidence in accuracy. Proteins from 20 model organisms, including Caenorhabditis elegans and Drosophila melanogaster, are also included in the database, bringing the total to 350,000 structures.
See “DeepMind AI speeds up the time it takes to determine protein structures”
AlphaFold won the Biennale last December Critical evaluation of the prediction of protein structure (CASP) competition and is the first program with an accuracy of over 90 percent. It has already been a boon to some scientists who have used AlphaFold in their research.
“It’s just the speed – the fact that it took us six months per structure and now a few minutes. We couldn’t really predict it would happen that quickly, ”says structural biologist John McGeehan of the University of Portsmouth BBC. “When we first sent our seven sequences to the DeepMind team, we already had the experimental structures for two of them. So we were able to test these out when they came back. It was one of those moments – to be honest – when the hair on my neck stood up because of the structure [AlphaFold] produced were identical. “
DeepMind claims it will be able to expand the database from 350,000 structures to 130 million by the end of this year.
Beyond researching existing proteins nature According to reports, access to this fund could also facilitate the development of synthetic proteins, as it could be more reliably predicted how they would interact with other proteins.
AlphaFold isn’t the only protein folding program on the market. For example, RoseTTAFold, which was inspired by AlphaFold, builds on this technology to calculate the information in different ways. It was only opened to the public last week, and its creators believe it will benefit from the new database.
“It’s fantastic that they made this available,” says David Baker, one of the architects at RoseTTAFold science. “That will really increase the pace of research.”