A place to hang out and discuss Coronavirus related matters. Hit the public chat button to join!
Not a member? You should
Already have an account?
To make the experience fit your profile, pick a username and tell us what interests you.
based on your interests.
Choose more interests.
Hackaday's "How You Can Help" list:
Create an account to leave a comment.
Already have an account?
If you're looking for activities to pass the time in lockdown here's one candidate.
The intrepid founder of Bellingcat geolocates a feelgood photo of swans in Venice demonstrating "nature recovery". Turns out the photo is from Burano, where they are regularly seen, part of metropolitan Venice so technically correct but not quite what people have in mind when they think Venice. The water in Venice is clearer these days though, with no boats to stir up sediment.
So much that can be done with the abundance of photos online. When I was retroblogging I used to geolocate photos I took when this meant film, and Brin and Page were still kids
Are you sure? yes | no
Hi all, I dropped a comment in the public chat about possibly doing a "Brainstorming COVID-19" Hack Chat on 4/8. Sounds like everyone here might find that interesting. If so, drop into the chat and respond so I can gauge interest. Thanks!
And BTW, I used to work in the pharma industry, in high-throughput screening on candidate compounds against disease targets to generate leads for the medicinal chemists. We used to routinely add multiple compounds to each assay and deconvolute the results informatically, which is pretty much the scheme that @Donnie Agema is suggesting.
Good to know that the idea has already been institutionalized. I guess I was a bit naive in thinking it could have been overlooked.
Have a bit of upside: https://www.theguardian.com/world/2020/mar/20/the-only-way-is-up-this-newsletter-will-not-mention-the-c-word then go make somebody's day.
The current infection rate should be known by the lab. Assuming 10%, mix 5 samples into one and, on average, you eliminate tests at a rate of 10 to 7 without any iterations and any of thereto resulting complications.
The idea here is not to name all of the reasons why it will not work, but to show the way we need to go to make it work. Think positive. Our entire culture is at stake!
I have to think about it.
As long as the infected percentage of the samples is 5% or less, one should expect to exclude 10 samples at the cost of 3 tests after the second iteration, another 5 excluded at the cost of 2 additional tests in the third iteration. Stopping there would cost an additional 5 tests for the remaining positive tested group of 5. That results in performing 10 tests instead of 20, or a 50% reduction in testing. Obviously, there is a curve of diminishing returns as the percentage of infections increases, so the process would be most effective at the beginning of the infection explosion, so the sooner, the better. I cannot even guess how much additional effort would be needed to do the combining process, for I am not lab expert.
What about your implicit assumption of non-interference between samples, to name just one aspect? Keep going, I had a list of about 5 or 6.
I have to leave the details of extraction/combination to the lab experts. Each infection swab must contain at least hundreds of virus entities, if not millions. If modern science cannot replicate samples from that, it would be very disappointing.
That question was about the possibility of type I and type II errors caused by mixing. The markers they are seeking are amplified by the process.
Next point: One positive result will mean "innocent" swabs will be retested, requiring more 4 hour waits. Is this acceptable? Will they have to take swabs again? Can they divide up swabs for more samples? In fact is 4 hours significant in the scheme of things? Are there other bottlenecks such as lack of kits which are more pressing? What about the news that they have got a test that takes a few minutes, does this make this shortcut moot?
I don't have any answers, do you? It's your idea after all.
My standpoint is, that sometimes experts are blinded by their work and thus unable to see outside of the established routines. An idea might be unpractical, but it might just as well be life saving. I just wanted to get the idea out, hopefully getting somehow into the expert domain for consideration. I cannot do more than than that.
How can we get COVID-19 test labs to play "20 questions" with potential infection samples? What I mean is as follows: take 20 test samples, "combine" them into 1 sample. When that 1 sample is tested negative, the lab saves 19 unnecessary tests. If that sample tests positive, the original 20 samples are divided into 2 groups of 10 each, and the process repeated on each of the 2 new combined samples, and so on ... This must be easily possible and would surely save tests (and lives!).
I was going to post a list of conditions for your shortcut to be valid as well as a mathematical analysis but instead I would like to see a follow-up from you elaborating those. An ability to scrutinise one's own proposals is invaluable in engineering.
Sorry this approach will not work for the swab test for corona virus as you cannot mix the samples without loosing the whole original sample. So is not available for the secondary and teriary tests which would be required. Even the swab used to collect the origibal sample is is short supply ( at least here in Australia.) I am told the swabs we use are made in Milan.
Are you certain there is no way of diluting the swab sample and using droplets to perform the test iterations? I cannot believe we are in such a low-tech culture.
Become a member to follow this project and never miss any updates
© 2020 Hackaday
Yes, delete it
You are about to report the project "Coronavirus chat", please tell us the reason.
Your application has been submitted.
Are you sure you want to remove yourself as
a member for this project?
Project owner will be notified upon removal.