45
$\begingroup$

A number of countries are using test kits for detecting new cases of nCoV (2019-Coronavirus) and apparently China is running low.

What exactly is in a nCoV "Test Kit" — How does it work?

(Surely they also differ, so in which way do they differ?)

| improve this question | | | | |
$\endgroup$
54
$\begingroup$

The CDC has made available online its nCoV test kit. Briefly,the kit contains primers and probes for real-time reverse-transcriptase PCR, as well as instructions for appropriate use and (critically) controls and guidelines to avoid false positives and negatives. Kits from different countries may use slightly different primers and probes, though since they are all working from the same sequences and the same principles they should be broadly quite similar.

Explaining how quantitative PCR works and the details of the primers and probes is out of the scope of this SE. A layman's introduction was written by John Timmer at Ars Technica.

| improve this answer | | | | |
$\endgroup$
  • 6
    $\begingroup$ Can it also be used on urine/fecal specimen? On the one hand, if you use the kit outside its guidelines, the accuracy, sensitivity, specificity, etc will not be defined. That said, in general conditions for qPCR on most clinical specimens are well understood, and specialists will probably be able to predict conditions that are compatible with the primers. So the answer is probably, It's not recommended for routine lab testing, but specialists with a solid technical understanding will undoubtedly be able to adapt it. $\endgroup$ – iayork Feb 4 at 18:16
  • 3
  • 4
    $\begingroup$ I don't think you'd expect to find much trace of a respiratory virus in stool or urine. $\endgroup$ – swbarnes2 Feb 4 at 22:47
  • 10
    $\begingroup$ There are reports that it’s been found in feces, raising the possibility of fecal/oral spread. Unlikely, perhaps, but dismissing it based on weak theoretical grounds seems unwise. $\endgroup$ – iayork Feb 5 at 2:15
  • 2
    $\begingroup$ Yes, it an live a long time in stool and urine, please see the links in my nCov Environement question. $\endgroup$ – not2qubit Feb 5 at 15:11
24
$\begingroup$

There are actually 3+ types of test kits widely used to diagnose diseases caused by viruses. We can check for the nucleic acid of the virus, the antigen of the virus that would cause an immune response or the antibody produced during the immune response of the patient. Since we've already got the whole sequence of the nCoV, the test kits normally used this time (at least in China) fall into the nucleic acid class.

To diagnose a patient with this type of test kit, the patient's nasopharyngeal swab, sputum or alveolar lavage fluid would firstly be collected to get a sample of the virus. Some reagents would then be used to extract the nucleic acid in it. After a reverse-transcription which would produce a DNA sequence based on the original single-stranded RNA gene of the virus, a PCR would be performed to amplify it to a large enough amount to eliminate the interference of other genetic materials in the sample (e.g. human DNA, genes of some normally existing bacteria and/or viruses). Here is where the nucleic acid detection kit would take place: the kit contains several characteristic genes serving as a guideline of the PCR, and it would have two primers that flank the target regions for each gene. The kit would also check if the amplification process goes well: it contains fluorescence molecules whose intensity would increase exponentially with the number of amplification (if target fragment exists in the original sample). The whole set of the testing process is called qPCR (quantitative polymerase chain reaction, or real-time polymerase chain reaction) and it requires highly precise thermal control and real-time fluorescence measurement. The labs where those qPCR instruments are in also need to be high-standard cleanrooms with negative air pressure in order to ensure that samples would not be contaminated by previously amplified DNA products. Unfortunately, most hospital labs in China are not qualified enough to perform tests of this level and that's why it takes so long to truly diagnose a patient with nCoV, even if tons of test kits are already sent to the infected area.

Recently a new method called POCT rapid detection is being tested and it is said that it takes only 15 min to get the result. This method takes a colloidal gold-based immunochromatographic assay to detect the nCoV antibody produced by the patient. Going through the strip, the material tested would pass through a layer of the colloidal gold marker that binds to the antibody, a detection line that contains the corresponding antigen, and a control line that would change colour once it detects any liquid that contains the colloidal gold. This method feels kind of like a pregnancy test, where two lines are bad (i.e. antibodies is detected in the sample as they bind with antigens) and one line is good. It's much faster than the nucleic acid test and it's easier to apply, but the problem is that you can get false positives or false negatives due to the different amounts of antibodies produced by patients. It can be used as a supplementary method for nucleic acid detection, but for clinical use we may want to be more careful with it.

Hope this answers your questions!

| improve this answer | | | | |
$\endgroup$
  • 3
    $\begingroup$ Excellent addition to accepted answer! Welcome to Biology SE! $\endgroup$ – not2qubit Feb 6 at 10:36
  • 2
    $\begingroup$ Can you provide a link describing the POCT method, because it's an ungooglable general term for Point of Care Test, which I don't think is what you intended. $\endgroup$ – not2qubit Feb 6 at 10:47
  • 1
    $\begingroup$ @not2qubit - not quite excellent, as it is not sourced in any way. $\endgroup$ – AliceD Feb 22 at 19:35
  • 1
    $\begingroup$ @AliceD I know what you mean, but I think we need to give people a little bit of slack for having to source something that have just very recently become available and possibly not even official or marketed yet. (Also some test kits have already been shown to be very poor with only 60% accuracy etc.) $\endgroup$ – not2qubit Feb 22 at 20:11
  • 1
    $\begingroup$ @AliceD My bad, it's my very first answer on bio SE and I didn't quite know how things work here. Thanks for pointing that out! $\endgroup$ – SyouTono242 Mar 1 at 2:55
11
$\begingroup$

There are several proposed tests for 2019-2020 novel coronavirus responsible for the Wuhan outbreak. The main techniques these are based on are:

  • PCR or qPCR targeted amplification

  • CRISPR-Cas9 genome editing

PCR-based tests are the current standard for diagnosis of viral infections (as these are faster than protein staining and can be made widely available as a kit). Most tests in use are based on PCR targeting sequences unique to novel coronavirus from the publicly release viral genome sequence. These protocols are recommended by the CDC and kits are commercially available or in development from Novacyt SA, or Liferiver Bio-Tech, or Altona Diagnostics. Before kits were available, various companies such as GeneSig offered primer design or primers specific to 2019-nCov which can be performed with standard PCR or qPCR kits. Bio-Rad has been cleared by the FDA to provide a test using Droplet Digital PCR (ddPCR).

PCR-based tests have the advantage that anyone with experience with standard molecular biology techniques or clinical testing can perform them, they do not need specific expertise in microscopy or pathology. Controls are important but there are still issues with low specificity with these kind of tests. For example, PCR tests for influenza type A (a similar virus) have high relatively false-positive rate for a clinical test. Considering the prevalence of the disease, these should only be used to triage people showing symptoms for further diagnosis. Results should be interpreted carefully. It is difficult to design specific PCR-based tests for viruses as they can mutate quickly (as 2019-nCov has been shown to do), especially in emerging epidemic. Note that as an RNA virus "reverse transcription" is required before PCR can be performed. These tests take several hours and it can take days to get results if a lab has many samples to process.

A proposed alternative is tests based on CRISPR. Several startups are working on this including Mammoth Biosciences in San Francisco and Sherlock Biosciences in Cambridge, Massachusetts. However, further testing is required to demonstrate the performance of these tests and none have been FDA approved yet. The advantages of this approach is that the test could be performed faster during an outbreak or epidemic. This is an emerging technology that has not been thoroughly tested.

Updates

Recently, another testing technique has been announced (in English and Japanese) by Drs Hayashizaki and Usui at the RIKEN institute. This proposed testing technique uses the Smart-Amp2 technology which has been commercially released. This technique is based on cycles of amplification similar to PCR but has the advantages of being more cost effective and able to perform the test as fast as 30 minutes.

Disclaimer: while I am currently affiliated with RIKEN, I was not involved in the development of this technology.

We should also bear in mind that these tests are all subject to false positive and false negative results. Many current media reports do not acknowledge this limitation. Any reports based on a single case deserve skepticism due to the inaccurate nature of some of these tests. For example, the State of New York is not using the CDC recommended test due to problems that they’ve identified with it.

| improve this answer | | | | |
$\endgroup$
  • 1
    $\begingroup$ synthego.com/blog/sherlock-crispr-diagnostics#Unlocked suggests that crispr-cas13 (not cas9) is just replacing the fluorescent probe. ncbi.nlm.nih.gov/pmc/articles/PMC5961727 says the theoretical advantage is that a cas9 based probe needs much smaller concentrations $\endgroup$ – reuns Feb 8 at 9:34
  • $\begingroup$ Less PCR cycles required? Running many cycles can take hours. Of course sample preparation and the RT step will take time as well. I’m not sure but this is the benefit claimed by the companies proposing CRISPR-based tests. I doubt it’s cheaper. $\endgroup$ – Tom Kelly Feb 8 at 9:38
  • $\begingroup$ For nCov the sequence amplified by WHO' primers is 100 nucleotides long, they recommand 45 cycles of 45s. Now maybe for more common viruses they need longer sequences ? $\endgroup$ – reuns Feb 8 at 9:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.