Artificial Intelligence: DeepMind’s AlphaFold AI Software Releases Database Of 200M Protein Structures
The massive AlphaFold database will be useful for scientific applications across the globe.
In a defining moment for AI computing, a 200x expansion in the number of proteins included in Alphabet’s Deepmind’s database was announced Thursday, which the company says will significantly speed up drug development and other innovation. This week, Deepmind added over 200 million protein structures to its database, up from 1 million, and including predicted structures for plants, bacteria, animals, and other organisms. (GIZMODO)
Deepmind’s AlphaFold AI technology foresaw the formation of certain protein structures in 2020 with unprecedented accuracy. Last year, Deepmind published some of these structures in an open database, starting with the structures of twenty species and 98 percent of all human proteins. Scientists could use the structures Deepmind released this week to assist with sustainability issues and food scarcity. All the data is available for bulk download via Google’s Cloud Public Datasets.
Prior to AlphaFold, protein prediction was a laborious and time-consuming process of experimenting that utilized tools including X-rays, microscopes, and other devices. According to Scripps Research Translational Institute Founder and Director Eric Topol, AlphaFold has slashed the time required to accurately predict protein structure from months or years down to some seconds. In a statement, he said. AlphaFold is a neural network system that trained on previously known models of proteins, but thereafter learned itself to predict protein structures.
The size of the circles in the picture reflects the amount of data that was added this week. While the predicted protein structure for all of the organisms listed increased dramatically since last year, the largest portion of data applies to animals. Plants are then followed by bacteria.
“AlphaFold has already accelerated and enabled massive discoveries, including cracking the structure of the nuclear pore complex,” Topol said. “And with this new addition of structures illuminating nearly the entire protein universe, we can expect more biological mysteries to be solved each day.”
According to European Bioinformatics Institute Director Emeritus Dame Janet Thornton, the AlphaFold database could prove invaluable for applications in medicine, agriculture, biotech … almost everything.
“It’s like a shop you can go in and just get your favorite protein and look at it, instantly.”
Isomorphic Labs, a new company created by Alphabet (NASDAQ: GOOGL) late last year, will utilise AlphaFold data to develop new drugs.
Meanwhile, globally the science and research community is already using AlphaFold models for advances in their respective fields.
Related Story: Alphabet To Commercialize AlphaFold2 AI For Drug Discovery
Image Credit: Flickr
Latest Alternative Investment News
Ripple Labs, the blockchain payments company that created the XRP cryptocurrency, is interested in purchasing assets of Celsius Network, the crypto lender that was forced into bankruptcy amidst the meltdown…
Nuri, a digital banking platform and app, offered users crypto transactions as well as normal banking activities including a debit card. The firm announced it had filed for insolvency on…
Telexistence Inc., the Japanese robotics company, announced the launch of its TX SCARA restocking robots in 300 FamilyMart convenience stores in Japan later this month as part of its AI-based…
US ETF issuer GraniteShares, which first launched leveraged ETFs on single stocks in the UK in 2019, and now offers 106 such products across Europe, has listed a suite of…