United Kingdom • Covid-19 • Mar 22, 2020
DeepMind's AlphaFold is going after Coronavirus
DeepMind's AlphaFold is a deep learning system designed to predict protein structure with its 3D model predictions are more accurate than any previous models.
Just as the late Dr Li Wenliang began warning Chinese officials about an epidemic in Wuhan in December last year, an Artificial Intelligence server at BlueDot in San Francisco too started issuing alerts about a potential epidemic outbreak. The server monitors patterns of infectious diseases outbreaks globally. But both the man's and the machine's warnings fell to silent ears globally, and today, three months later, the world is battling the pandemic of COVID-19.
DeepMind's AlphaFold is a deep learning system designed to predict protein structure. Now it has been deployed to predict 3D models of protein structures of the COVID-19 virus without having any prior information on similar proteins. This is essential because how a protein looks is crucial to create drugs to defeat the virus. It is like a drug has to truly "fit" into the protein of a virus to destroy it.
Currently, due to the urgency, several giant players have made their findings open-access. The Chinese have put out the genomic blueprint of the virus, and others have released their findings on protein structures. This is helping AlphaFold look at a handful of understudied proteins, using machine learning. AlphaFold stands out as a union of decades of deep learning progress but guided by expertise from protein structure databases in the public domain.
In a nutshell, AlphaFold uses genome sequences (available for COVID-19 and relatively easy to get) to predict the properties of resulting proteins that actually do the work, by looking at the "distance" of each component that makes up a specific protein. It then offers a quick insight into how the virus protein may look. While DeepMind itself stresses tha "these structure predictions have not been experimentally verified," the automation of drug discovery may become an intrinsic part in the next epidemic.
Chinese company Baidu is also using an algorithm to predict the structure of mRNA - another important biomolecule that links the genome to proteins of a virus. By eradicating mRNA, the link is broken, destroying the virus. Several other AI innovations are playing their part from helping in diagnosing the virus through scans, help medical front liners, there is still not enough data to train the AIs yet.
Source: Singularity Hub
Image by NIAID-RML via Flickr
AI can prevent disruptions in fusion devices
Intel introduces Pohoiki Springs, its most powerful neuromorphic research system till date