Remember when I told you about the MIT prof that was seeking new antibiotics using AI (artificial intelligence). You knew he couldn’t be the only one.
He’s not. A prof at UC Riverside has been developing his own machine learning drug pipeline. But, his focus is on COVID-19. I’m not sure how similar the algorithms are to MIT’s- just how similar the results will be. Dr. A. Ray and his grad student, J. Kowalewski, published their initial results, Predicting novel drugs for SARS-CoV-2 using machine learning from a >10 million chemical space, in Heliyon.
They studied known ligands (a ligand is a compound that attaches itself to larger compounds, in this case to 65 human proteins) that demonstrated their activity with the coronavirus (often via the ACE-2 receptor). This roster was then expanded to FDA-approved drugs, looking to repurpose them. Since they already have FDA approval for a (or multiple) use(s), obtaining approval for a new use should be relatively rapid. The research goal was to find small-molecule inhibitors and activators (the ligands),by examining their 3-D structures.
The model went further- it computed toxicity, which is critical to eliminate any drugs that would hurt the patient more than help them. It also sought out drugs that could inhibit more than one protein target. The problem is that there is not sufficient data to muster those compounds, since we are only certain of the ACE-2 interaction by SARS-CoV-2.
One of the unique aspects of their AI system was to include vapor pressure as a parameter. (Sorry- I have to get a little technical here.)
Why vapor pressure? It’s indicates the tendency of a compound to convert to vapor or gas state. In other words, it’s a measure of how volatile a compound may be. And, since SARS-CoV-2 involves proteins that are overexpressed in the respiratory tract, this parameter will outline chemicals (drugs) to serve as inhaled therapeutics to directly attack the virus where it has attached to the cells of the respiratory tract.
Let’s hope this approach can truly identify some great drugs to treat COVID-19.
That would be wonderful if one or both of their studies helped get drugs to combat Covid-19.
Certainly would, Martha. That’s always the goal.
Really enjoyed this information. I had no idea what some of the steps are that go into developing a treatment or how many factors are involved. An eye opener for me thanks for sharing.
ChefWilliam recently posted..Dessert time Banana Cream Frozen Yogurt
Thanks for reading it, William! And, thanks for your comment.
As always, I hope for rapid success. Husband recently had another cousin who contracted COVID-19 (she had a few rough days but never needed hospitalization). New York is in a good place right now but I think that’s about to end. You can’t keep human nature down forever and we need successful treatments in the worst way.
Alana recently posted..A Final Blueberry Moment
Hopefully, we will get this pandemic under control, without waiting for a vaccine.