Hacker News Reader

Comment by rurban | Hacker News Reader

This is not rapid, only 10 tests/day, whilst the current overly slow sequencers can do 200/day. The accuracy of 5% is also worse than other faster sequencers of 3%.

And compare that to modern ThermalFischer parallel PCR sequencers, which operate in realtime, 25 tests in parallel, for 2.500€. That's a factor of 1000. With 10 modern sequencers per city you could test 10.000 per day, eg with drive-in.

https://www.thermofisher.com/de/de/home/clinical/clinical-tr...

View on HN

Agreed. Also the Cepheid one is faster (45 minute for point-of-care results) and has FDA approval ... not sure other advantages/disadvantages of each system though (instrument and reagent cost, # deployed units, etc).

https://www.cepheid.com/coronavirus

That said, it’s great to see all the companies that are putting real effort into this. This is their time to serve.

My own employer has assembled a task force for repurposing some kits we have for covid research. Everyone working on it is happy to have someway they can try and help out.

cool good luck with your task force!

I don't work in an area where my skills can contribute but I'm in the process of finding out if I can sort insurance out to go out and do medication deliveries on motorcycle for local pharmacises, with the lockdown and the number of vulnerable people (including my mother and best friend) who can't leave their houses there is >1.5m people who can't leave at all for health reasons most of whom require prescription medication monthly.

At the moment I feel so incredibly fucking useless lurking at home playing video games.

Every day you don't fill a hospital bed can save countless lives. From the doctors who would have to diagnose and treat you to the people that you would infect when you don't realize you have the virus.

Sometimes being a hero is boring.

If you’re healthy you could also reach out to a local shelter, neighborhood not for profit, charity or church. They may have needs.

Also, don’t forget to socialize in whatever way you can. Skype your relatives, call your friends, send a letter to your neighbor. Go on a walk and ask people (from 6 Ft away or more of course) how they’re doing. Take time to do the human things. Anyway to make someone else feel slightly normal is this time is a big help!

In addition to all information above this test is able to detect antibodies only not virus. So it is not recommended tests by WHO. This test is only useful in the later part of infection by the time which patient might have in contact with others thus spreading the disease. We need tests which can detect single of very few virus particle in the early stage of infection and that will help to quarantine the patient thus reducing virus spread.

...able to detect antibodies only not virus...This test is only useful in the later part of infection

Actually, getting a sense of the penetration of infection into the general population, among asymptomatic people and mild cases would be very valuable information! Antibody tests could be used to sample test the public at large.

Plus having a point of care test for triaging seems extremely useful as well.

I don't think this is meant to replace large labs running rtPCR tests.

Even if you have unlimited virus tests, antibody tests are extremely useful. The one combined with the other demonstrates recovery as opposed to pre-infection. Supposedly antibody tests can also be made field-deployable much easier than virus tests.

I imagine antibody tests are very very useful for spotting people have already contracted the disease, recovered and may therefore have a degree of immunity.

I'm not sure this is true. Bosh's site says "Which detection method is used in Bosch’s COVID-19 test? The COVID-19 test is based on a combination of sample preparation (including process controls): Multiplex PCR (Polymerase-Chain-Reaction), μArray-detection to allow the identification of SARS-CoV-2."

I believe that line in the Q&A is referring to the "other COVID-19 tests, where the test result is available in a few minutes". It's ambiguous and confusingly phrased, but it seems like Bosch directly detecting the pathogen, in contrast to other rapid tests (which don't, presumably, use PCR, as that takes longer than "a few minutes").

The accuracy of 5% is also worse than other faster sequencers of 3%

Is that false positives or false negatives or both so to speak? I.e. what is the confusion matrix of these techniques?

This is low throughput but high speed; there's a use-case for that.

Isn't that the same thing?

He means low throughput and low latency, both of which are related to a concept of speed, but not the same.

> Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway.

— Andrew S. Tanenbaum

No, this is faster than e.g. the Roche tests mentioned here last week, it will give you the first test result after two and a half hours, while the Roche one will take three and a half hours to give you the first result. So it is faster.

But it will give you the second test result after five hours, while the Roche machine will give you the second result after three and a half hours, too. As well as the third, the fourth, and the fifth. Oh, and the onehundredandninetysecond one, too. In the 24 hours the Bosch machine produces ten results, the Roche one will produce more than fourthousand results.

I heard that this is about testing in the periphery? MDs could put it in their offices and test their employees daily, for example. That’s one possible use case.

I'd assume most general practitioners aren't taking patients in person anymore except in exceptional circumstances. It could definitely be useful for testing hospital personnel before they go on shift, tho.

Tell us more. Another comment has a link to a news article that says this Bosch machine costs 15.000€ with cartridges costing 50-100€. The Bosch machine has a faster response (2.5hr), but lower overall throughput.

How much does the ThermoFischer sequencer cost? Is it affordable enough that every hospital just buys one and becomes a regional testing lab? How much specialized training does it need? (in every major city there are tons of biology and biology-adjacent majors who've done lab work but discovered there's no living to be made in bio so have switched to other fields).

Also it might be more expensive, but if you look at total cost of ownership - supporting infrastructure lab, skilled lab technicians to run it, supplies required, etc - how does it compare?

You're forgetting the difference between bandwidth and latency.

I don't get this.

Example of low latency connection: Morse code via flashlight.

Example of high bandwidth connection: van full of books.

These are orthogonal properties and must be considered separately.

The massively parallel machines will give you 200 samples per day, but in the very best case you have to wait until the end of the day to get the first one. This technology gives you a result a couple of hours after taking the sample.

The machine can only do one test at a time, so throughput = 1 / latency.

But this isn't true if you have multiple machines. Throughput goes up, but latency stays the same. In that case low latency counts more.

Not to mention the expensive cost of consumables. This is more of a marketing play than a practical solution to the global pandemic.

I think you mix up latency and throughput.

Honest question, are you speaking from experience? We run PCRs at our company for our vaccine research and having to run them back-to-back day-in-day out, without any manual errors, is going to be super stressful for any lab. Not to mention the costs of training healthcare personnel on how to run a PCR.

Thats what seems like the difference to me. If anybody can put samples in a cartridge in a machine and get results that seems like a whole different game.

Reply on news.ycombinator.com