Photo by Nicole Berro on Pexels.com
Fans of our old Facebook blog (what’s Facebook?-ed) will recall a significant posting about how Artificial Intelligence was used to scour the databases of old papers to come up with a new class of antibiotic, called Halicin. We hoped that this team of workers would receive a Nobel Prize . God knows they deserve it.*
Could AI help to dig us out of this terrible hole we are in due to SARS-Cov-2? Especially, will the virus mutate, and if so, how? And if it does, will all these wonderful new vaccines stop working all of a sudden? Uh-oh.
There may be hope. We know that DNA contains “hotspots” where the chances of a mutation are higher. A Google search will give you lots on this, so I’ll include just one ref to give you a jumping off point.* Up to know, this work has mainly been the preserve of workers in fields like cancer and autism, but that’s fair enough, as we didn’t have a virus pandemic until this year.
But will it work for the RNA in a coronavirus? According to a post in Engineering.com by Vincent Charbonneau an AI start-up called Graphen are doing exactly that. Using data on whole genome sequencing from over 30 countries, they are using AI to plot how the virus mutates, spreads and propagates. This will lead both to better drugs, and predictions of the most harmful variants of the virus.
A thought occurs, albeit in the realms of Science Fiction. Can someone use AI to look at hotspots and map the future mutation potential of other viruses? What about other organisms, such as bacteria?
We await with anticipation.
*J Stokes R Barzailly J Collins et al A deep learning approach to antibiotic discovery CELL vol 180 pp 688-702 Feb 20 2020
*https://www.sciencedaily.com/releases/2018/07/180730132924.htm
#graphen #engineering.com #mutationhotspots #ArtificialIntelligence
Learning, Science and Society 