Re:Israel’s Covid-19 boosters are preventing infections, new...
by datameta on Wednesday September 01, 2021 @12:34PM
(#28382069)
I wonder how accurately we can run mutation simulations updated based on best current data to correct the best guess optimization function(s). It would be a breakthrough if we can predict the likeliest mutation. I'm aware of the enormous number of variables making this tremendously difficult...
Re:Israel’s Covid-19 boosters are preventing infections, new...
by nradov on Wednesday September 01, 2021 @01:03PM
(#28382541)
We don't have the technology to stimulate that accurately in software. It would have to be done using transgenic animal models and/or human tissue cultures. Essentially it would be a form of gain-of-function research, and there are concerns that a lab accident could result in an even more dangerous variant escaping into the wild.
Re:Israel’s Covid-19 boosters are preventing infections, new...
by datameta on Wednesday September 01, 2021 @01:36PM
(#28383041)
With ML-assisted drug discovery in, say, kinase inhibitors we search the molecule variation space for test candidates. What stops us from applying this to the protein scale by folding binding sites to discover variant candidates that would be more virulent? I'd like to understand this better.
If we vaccinate against the modified spike, it's not clear how many more mutations exist like Delta. Maybe we can still get ahead of it, but I don't think anyone knows yet.
I do wonder if we'll see reformulated boosters against Delta's spikes, though.