Sign In

Communications of the ACM

ACM TechNews

ML Model Finds SARS-COV-2 Growing More Infectious


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
The coronavirus.

A novel machine-learning model developed at Michigan State University suggests SAR-CoV-2 has become more infectious due to genome mutations.

Credit: Centers for Disease Control

Researchers at Michigan State University have developed a novel machine-learning model that suggests SAR-CoV-2 has become more infectious due to genome mutations.

The model analyzed SARS-CoV-2 genotyping from more than 20,000 viral genome samples and determined that five of the six known virus subtypes are now more infectious.

Researchers focused their analysis on mutations to the spike protein, which causes infection when it interacts with a human host cell receptor.

The advanced neural network analyzed more than 8,000 protein interaction records to gauge the impact of the current known mutations on the binding affinity—the binding interaction between the spike protein and host receptor during the initial stage of infection—of the spike protein.

The researchers found that viral infectivity increases if the binding affinity strengthens. Further, the model predicts that multiple residues on the receptor-binding motif have high chances to mutate into more infectious Covid-19 strains.

From MSU Today
View Full Article

 

Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


 

No entries found