back to article Bio-boffins devise potentially fast COVID-19 virus test kit out of a silicon wafer and machine-learning code

Boffins have demonstrated that machine-learning algorithms may be able to help scientists identify viruses, and could even be used to develop more efficient tests for the presence of the COVID-19 coronavirus in the near future. Researchers at Nagoya University, Osaka University, and the National Institute of Advanced …

  1. Duncan Macdonald

    Messy samples

    Have they tested this system with the typical mix of thousands of different compounds in a typical biological sample? There is a big difference between identifying a virus when it is the only thing in a sample and identifying the same virus in a sample containing thousands of other substances.

    1. Sgt_Oddball

      Re: Messy samples

      My thoughts exactly, there's going to be alot of noise as well as multiple viruses (not all of them are harmful to us, so should be filtered out).

      Will the samples be processed first to remove as much non-virus material floating around in the samples?

    2. jmch Silver badge

      Re: Messy samples

      That's what I thought too... Not only other biological crud but fine particles of inorganic matter too

    3. Anonymous Coward
      Anonymous Coward

      Re: Messy samples

      There's a lot of this technology coming on stream, the most advanced application

      currently is DNA sequencing. The buzz-word is "nanopores" as most of it consists

      in looking at electrical profiles as particles are squeezed though tiny holes in a membrane of some sort. You need a lot of data processing to go from that to an identification, so yes, you do need more clean-up for messy samples.

  2. Flak
    Go

    Lateral thinking

    Brilliant to apply combined mechanics, physics and AI to a biological challenge!

  3. IGotOut Silver badge

    Nice idea...

    but cost?

    Bearing in mind these I would think are disposable, they need to be cheap for mass use.

  4. Pascal Monett Silver badge

    "Tiny holes 300nm across are drilled into the metal"

    To me, that means that the system can only detect viruses that are a maximum of 300nm 299nm in size.

    Can a virus get bigger than that ?

    What happens if the virus is 302nm in size ?

    1. JDX Gold badge

      Re: "Tiny holes 300nm across are drilled into the metal"

      Stop scaremongering about the risks of mega-virus!

    2. Cuddles

      Re: "Tiny holes 300nm across are drilled into the metal"

      "Can a virus get bigger than that ?"

      Yes, but it's fairly unusual to be bigger than that in more than one dimension. The largest ones are aroun 400nm, and there are only a few know that are over 300nm. They're unusual enough that the first really big one to be found wasn't even recognised as a virus at first. None of the big ones appear to be relevant to humans; they're pretty much all water-borne and infect amoeba.

      On the other hand, there are viruses that are long and thin, that can get to over 1um long. They'd fit through the holes lengthwise, but I don't know how this method would cope with long strings flapping about the place instead of neat little blobs that block holes up. That includes fun things like ebola and related viruses, so potentially quite relevant to humans.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like