A possible new COVID-19 treatment

[robertmiles]

Engineered decoys trap virus in test tube study; healthcare workers at high risk even with protections

https://www.reuters.com/article/us-heal ... SKCN24Z2EF

[robertmiles]

Watchdog Demands to Know If Drug Maker Sitting on Possible Covid-19 Treatment Due to Patent, Profit Concerns

https://www.commondreams.org/news/2020/ ... due-patent

[bruce]

Follow the money.
[unsurprised]

[astrorob]

did these guys just beat the moonshot project the old fashioned way?

https://www.ucsf.edu/news/2020/08/41824 ... t-covid-19

sounds like they are already trying to get it into production which seems a bit premature... shouldn't there be a bunch of FDA testing on this?

[bruce]

Maybe,

Russia made a similar claim a day or two ago.

I have zero doubt that there are going to be other claims; initially all with minimal testing. Some will turn out to be 51% effective; others may be 88% or 96% effective. Some will be total hoaxes. Several will work but have undesirable side-effects. Without widespread testing, there's no way to know which is which.

[astrorob]

well UCSF isn't exactly russia, as far as credibility goes. i guess what bothers me about this (as a computer guy) is that they apparently accomplished this the old-fashioned way. i would have hoped that this massive computing power we brought to bear on these problems would have yielded results before the petri dish... but then again proteins do their thing in real-time so the speedup involved in trial and error is pretty significant as long as you can check the results reliably.

[Neil-B]

Old fashioned ways may not have the speed/scale of technological approaches - and are probably actually needed at the tail end of technological efforts anyway - but with years of experience and some inspirational insights there is no reason why they might not get there first ... As I see it (in my naive simplistic world) a brute force attack on someones combination lock might take longer than guessing the number is the same as their birthday (or some such). I guess there is actually a massive blending of knowledge going on across both technological and old fashioned approaches - but honestly I can say it really doesn't bother me who gets "solutions" (be they cures/vaccines/protections) out first or by what method - competing approaches can actually support and drive each other forward - If we end up with half a dozen good "solutions" and a vast amount of improve scientific knowledge I'll be happy

[JimF]

Except that this isn't an "old-fashioned" approach at all. It is as new-fashioned as you can get.
They have the advantage of a naturally-occurring starting point in the nanobodies, which saves them a few orders of magnitude computing time right there.

And they used computer power when needed:

To find effective candidates, the scientists parsed a recently developed library in Manglik’s lab of over 2 billion synthetic nanobodies. After successive rounds of testing, during which they imposed increasingly stringent criteria to eliminate weak or ineffective candidates, the scientists ended up with 21 nanobodies that prevented a modified form of spike from interacting with ACE2.

They didn't do that by hand.

Smart is good, as a general rule.

[astrorob]

well they don't say what they did to eliminate the synthetic nanobodies... but if they were using someone's supercomputer i'd think they would mention it.

anyway as i understood it the moonshot project wasn't making moon shots in the dark as it were, but they invited people with domain knowledge to try to design drugs which might interrupt the spike proteins or other proteins in the virus.

if this is real i'm certainly not pooh-pooing it but i had hoped that a historic amount of compute power would yield historic results ahead of more traditional methods, computer-based or not. i'm all for this thing if it turns out to be a real therapy/preventative.

[JimF]

if this is real i'm certainly not pooh-pooing it but i had hoped that a historic amount of compute power would yield historic results ahead of more traditional methods, computer-based or not. i'm all for this thing if it turns out to be a real therapy/preventative.

The monoclonal antibodies are in phase 3 testing now, and should be available from several sources by the end of the year.
And that is in addition to the repurposed anti-virals that are also in testing, and which have already been FDA approved (for some purposes).

I have been pointing out for some time that our computer studies (including Rosetta, WCG/OPN, and anything else you find) won't be the first line of defense.
It is more for the follow-on products, hopefully including the next-generation viruses.

[JohnChodera]

> did these guys just beat the moonshot project the old fashioned way?

If by "the old fashioned way" you mean "doing something amazing through the awesome power of yeast genetics", then yes! This is a really awesome piece of work, and I think could be an important tool in the defeat of COVID-19!

We're going to need a variety of tools to treat and eliminate COVID-19, and to eradicate the threat of future coronavirus pandemics, so fear not---the work the COVID Moonshot is doing is going to still be incredibly useful.
But it's fantastic to see this kind of science (from folks at my graduate institution, UCSF!) joining the fight!

~ John Chodera // MSKCC

[psaam0001]

We can at least provide models (with the calculations our computers are performing), as tools to help give those who are using the test tubes and petri dishes guidance towards possible outcomes.

Paul

[rbpeake]

This is amazing!

Perhaps it's a technique that could be used on other viral diseases (hopefully!)

[robertmiles]

‘AeroNabs’ Promise Powerful, Inhalable Protection Against COVID-19

https://www.ucsf.edu/news/2020/08/41824 ... t-covid-19

[Neil-B]

... link already provided in forth post in thread?