Yes, PCR tests can detect “the COVID virus”

I was asked to write some comments for a fact check article about some of the myths going around about PCR-based testing and whether PCR tests can detect “the COVID virus”. In particular, an agglomeration of them in the form of a Facebook video by a guy in his car. I wrote a few hundred too many words – so here are most of them.

Don’t listen to a car dude

First up – the comments which generated the need for this article were from someone who appeared to have zero understanding about PCR or PCR test design or PCR use. Some of the comments were wrongly attributed to the creator of the PCR test, Kary Mullis, who, as we know, died before this pandemic began.

The speaker – who I will not specifically oxygenate here – also seemed to have no scientific experience, or any expertise in science (he was unfamiliar with routine words used by anyone in the biological sciences) and I’d guess that he hasn’t ever worked in a laboratory, designed a pair of PCR primers, constructed, optimised or validated a PCR test, regularly consumed or written scientific literature or generated or interpreted any scientific data. In sort – this guy is not an expert so please take what he said with a grain of NaCl (you know what that is, right?). These are important factors when someone is lecturing at you and using their platform to pull down or threaten experts on the topic. This sets the stage. It provides context.

The purpose of these sorts of comments seems to be disruption and the erosion of trust in expertise and science. There is possibly something here about the speaker feeling that they know something special or something different from the real experts. I’m sure there’s a much longer treatise on who this person just wanting his five minutes of fame, that his mum told him he was special and that he felt unheard or misunderstood. But during a major public health event where lives are on the line, these selfish reasons aren’t enough.

The claim that PCR tests can be contaminated

Contamination of PCR is a real issue and I’ve covered this here just recently. Labs that routinely use PCR and RT-PCR are extremely careful to consider and account for and prevent that from being an issue. These expert labs also include controls in every run which tell them if a contamination problem has occurred.

LEARN STUFF
RT = reverse transcription. This is an enzymatic step that precedes the PCR (polymerase chain reaction) cycling, still in the same tube, producing a DNA copy of an RNA target, such as a small portion of the RNA which makes up the genes and genome of SARS-CoV-2 and most other respiratory viruses that infect humans. The PCR is an enzymatic process that produces a new chain of nucleotides that are a mirror image of the original target region’s nucleotide sequence (genetic code) through a process of heating and cooling controlled by a programmable thermal cycler. It’s a cyclical process which copies (amplifies) that small piece of DNA millions of times until it reaches a level we can record using fluorescence detectors. Without the amplification, there is too little viral genetic material to detect.

The claim that PCR tests only detect a small part of the virus

Each RT-PCR test is designed to be specific to the intended virus (there are examples where that may not be the intention). So labs have a number of these tests; reference labs may have dozens!

The types of RT-PCR used to test for virus only seek out a small part of the complete genome sequence. This is to make the tests work as quickly (an hour or less) and efficiently as possible.

woman touching wall
This is not in any way related to a fluorescence producing probe.
Photo by Victoria Borodinova on Pexels.com

In particular, the “real-time” PCR method has been used widely since the 2009 fluA/H1N1 pandemic. It works best with shorter target regions. Real-time RT-PCR (RT-rPCR) includes a fluorescence-producing (the fluorogenic bit) DNA probe. The probe adds another layer of target-specificity to the test. It also generates a signal when the viral RNA is present (but not, when it isn’t). The probe doesn’t take part in the PCR amplification – it just binds to the DNA as it builds up…in real-time! When enough DNA is present, the probe will produce a detectable fluorescent signal. The signal grows a bit more during each cycle until the PCR finishes.

That small target sequence is usually – but not always – chosen to be highly specific to the target; both primers and probe bind only to the intended viral target.

PCR gleans its extreme specificity from the primers. At each and every position of a new DNA primer, we have one of four nucleotides to choose from, dATP, dCTP, dGTP and dTTP.

Deoxynucleotide triphosphates. 2’deoxycytidine-5′-triphosphate (dCTP; C9H16N3O13P3),2′-deoxyguanosine-5′-triphosphate (dGTP; C10H16N5O13P3),2′-deoxyadenosine-5′-triphosphate (dATP; C10H16N5O12P3) and 2′-deoxythymidine-5′-triphosphate (dTTP; C10H17N2O14P3). From https://virologydownunder.com/pcr-primers-a-primer/

If we designed a sequence-specific primer of 20nt nucleotides in length (a ’20mer’), the chance that the same exact sequence will occur randomly in nature would be (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4) x (1/4). That means there’s a 1 in 1012 chance of a 100% identical sequence occurring randomly in nature. Pretty small chance. That last bit is relevant to how incredibly specific RT-PCR tests can be.

Our test is designed to detect a particular virus. We can look at the genetic sequences of all the other known viruses and gauge, with quite a bit of confidence, that this test will not interact with their genetic material. This helps us avoid one form of false result (rarely are things absolute in biology).

A nucleotide sequence alignment comparing the sequences of a few influenza A virus genes. Whenever a sequence in the alignment differs from the highlighted (yellow) sequence, we see a black line using this computer program. Where the sequences are identical, we see a grey line. If I wanted to design a test to detect all these virus variants, I’d probably fail to detect the first one because of all the differences. But there are some grey regions in which I could design my primers to sit, and they seem to have the same sequence – a ‘conserved region’ – across all the other virus variants in this alignment.

So rather than next to impossible, it’s actually highly probable that expert primer and RT-PCR test designers do detect what we design our primers and probe (see below) to detect.

The claim that PCR tests don’t detect “the COVID virus”

RT-PCR tests in use today are extremely effective at very sensitively and specifically detecting SARS-CoV-2, the virus that causes COVID-19. Some are a bit better than others, but they very much can detect SARS-CoV-2. Some early RT-rPCR tests were designed intentionally to allow for sequence variation just in case the virus did a lot of changing early on (it didn’t).

Additionally, SARS-CoV-2 tests do not detect other coronaviruses (see that caveat above and it really only holds for other closely related SARS-like and bat CoVs), or rhinoviruses or adenoviruses, parainfluenza viruses, influenza viruses, respiratory syncytial viruses and so on, that may be found in the same patient sample.

But we test developers also check this in the real world. We test other related and unrelated viruses (like those👆) and different types of human samples (snot, saliva, sputum, swabs, liquids etc) during a process called validation. This way we know what the primers – and thus the RT-PCR test – can do and can’t do. This is all part of the expertise developed over decades of our use of PCR methods.

Of course, a badly designed RT-PCR can have problems, as can anything in science, or life. False positives and negatives can happen because of test design issues (not just because of sampling and timing issues), but professional laboratories watch for this, minimise the risk of this and can identify and rectify this. And they spend time on test design, optimisations and that validation thing.

Expertly designed and used SARS-CoV-2 tests have been well-proven to detect this virus. The leading tests have also been tested on SARS-CoV-2 RNA purified from virus isolated and grown in cells in the lab – so we know it is this virus that we’re detecting. We can also design and order viral sequences in the lab and have them made commercially, so again, we know the tests detect the viral sequences we expect them to and not other viruses.

The claim that PCR tests don’t detect the whole genome

Is correctamundo.

But that is by design, not an error or an oversight. We have only been able to quickly and relatively easily (compared to even a decade ago) detect entire viral genomes – a key part of the science of genomics – in very recent times.

It is still costly, less sensitive and slower to determine the entire sequence of a virus compared to simply detecting its presence using an RT-rPCR test though.

For cluster, outbreak, epidemic and pandemic control and for patient management – detection the presence of viral RNA (which real-time RT-PCRs do) is all that is needed. For more extensive tracking of virus change over time and to better understand where an outbreak may have come from – genomics is great, but only a small fraction of viruses usually get this treatment. That will change with time.

A positive result doesn’t mean virus is present – but we can be pretty sure it was

A positive RT-PCR result doesn’t prove that replicating (also called infectious or viable) virus is present. It proves that its genetic material – the RNA – is present.

But we have learned much about this virus in 7 months and SARS-CoV-2, like other viruses we detect with RT-rPCR, is present before and during the peak symptomatic period. These are also the times we get sampled and tested because we feel sick. We know from experience that when detectable viral RNA is present in a patient’s sample, there is often an infectious virus present as well. RNA is not so stable that it hangs around without a constant source producing more of it.

Even though it’s really sensitive, RT-rPCR has a limit of detection – as we have been seeing in the story of “false negatives” among patients being tested for SARS-CoV-2 well after they became clinically well.

So we use the RT-rPCR result as a surrogate indicator of the presence of the infectious virus.

Our other option would be to try and “grow” virus in the lab, but this has problems. It is slower, more expensive (lots of labour costs), affected by poor sample handling/storage/cold chain, needs a high level of expertise (actually a dying art) and is much less sensitive than real-time RT-PCR.

For some non-experts, including medical Doctors in some cases, this caused a big problem in the early years of PCR as it moved from a research tool to an aid for patient diagnosis.

They couldn’t understand what a PCR positive result meant because they were used to virus culture results….which were less often positive because the method was less sensitive.

Now you are more expert than you were

So now you may know a bit more about this subject than you did before. Hopefully, I’ve explained that PCR tests can detect “the COVID virus”. I feel confident in saying that you will know a lot more than a certain guy in a certain car in a certain video. Always be willing to learn a new thing. Thanks for reading.

Hits: 3537

55 thoughts on “Yes, PCR tests can detect “the COVID virus””

  1. Hi there! Great post, really informative! I just have a small suggestion. I’m sure you understand exactly what you’re talking about, but the description of the statistic 10^-12 seems to indicate that the odds are 10^-12 that the sequence would be matched anywhere in nature, where the actual calculation is the odds that a particular 20-length sequence will be a match. This appears to be related to the ‘prosecutor’s fallacy’. Again, really great work- really informative!

  2. I am wondering whether pcr tests performed at vivid stations where people are sitting in cars may be contaminated at times? If the personnel are using same gloves for a number of hours but handling swabs, they may not only be contaminating the swabs but may be also spreading the virus as they are bringing their hands to the next persons mouth each time. Just a thought.

      1. When this whole business started, they said that PPE was supposed to be changed out with every patient. I recently was watching an instructional video for health care professionals on how to do a swab. They specifically instructed to use NON-STERILE gloves. I found this very disturbing and disgusting. If you’re going to test for this particular disease they should darn well be using sterile gloves rather than risk a false positive that would require a person to be quarantined and re-tested after 2 weeks all for nothing due to someone accidentally contaminating something. Outrageous.

        1. I’m pretty sure they will either be using a new set of gloves or sanitising the outer surface of the gloves between swabs. Could you link to this video?

  3. Total falcification of the reality; PCR tests are non-specific, they can detect traces of a corinavirus any coronavirus but cannot specify if this is ordinary flu, avian flu, Sars, Mers or Sars Cov 2. In addition it cannot tell if it is an active infection or traces from a past infection and as a result they keep classify them as asymptomatic cases. Scientific articles that could be found on the internet stipulating the limitations of the PCR test have been “sanitized” and the limitations removed.

      1. This is an open evidence review:
        https://www.cebm.net/covid-19/infectious-positive-pcr-test-result-covid-19/

        “PCR detection of viruses is helpful so long as its accuracy can be understood: it offers the capacity to detect RNA in minute quantities, but whether that RNA represents infectious virus may not be clear.”

        “The relation with infectiousness is unclear and more data are needed on this.
        If this is not understood, PCR results may lead to restrictions for large groups of people who do not present an infection risk.”

  4. Thank you for the article. Can you explain how the nucleotide sequence was initially identified as being specific to COVID19?

  5. False negative rates in sick people (up to 30%) and false positive rates among asymptomatic up to 50% and even 80% according to Kary Mullis. As an integrative medicine practioner for the final 6 years of a 49 year pediatric career (retired March 16) and 11 years in academic medicine. I understand how to study medical literature.

    Mullis did warn that PCR should never be used for diagnosis due to arbitrary boundaries determining cutoff between neg and positive, in an interview before his death. The are many publicized cases where physicians have described doing serial tests on sick CoVID 19 patients, and seeing them bounce back and forth between positive and negative.

    The almost universal practice of tracking “cases” based on PCR tests (which often counts all tests done on one patient, with the false positive problem in asymptomatic renders this data as virtually useless. And yet public health restrictions are being made largely on this flawed and misleading data.

    Bose Ravenel, MD, FAAP, Retired

  6. It is my understanding that currently there is no ‘gold standard test for the presence of SARS-CoV-2. Do you now consider the RT-PCR test to be the gold standard?

    1. Culture is the gold standard at detecting infectious virus but is slow, insensitive, expensive and requires specialised expertise. PCR-base tests have proven themselves time and again over decades at detecting the presence of an acute infectious pathogen.

  7. Hello!

    I saw other claims that current PCR tests are limited because we currently lack a gold standard to compare to (i.e. it would be the isolated and purified virus).

    For disclosure reasons, my background is legal so I have no technical knowledge on this issue. But I do think that a minimum understanding of these basic facts are necessary since everybody is using the precautionary principle to justify the current public health policies.

    Cheers,

    Mauro

  8. When was the SARS CoV 2 first sequenced? Which/whose sequence data set is used for the primers and probes? The test is only as good as the specificity of the primers and probes. Being that SARS CoV 2 shares much of the same genome with other coronaviruses, which portion(s) of the genome are distinctly SARS CoV 2 and when were these sequences determined?

    Thanks

    1. 11th Jan The first sequence went online. That was the basis of many tests but these tests were also checked against subsequent genome sequence (tens of thousands now) and they still match well. Most of the genome is unique.

  9. Would you be so kind to give a link to the validation of this test. The virus seems to have been isolated last month.

  10. BS! You must have a “golden standard test” to compare it too. You must isolate the virus. Simple as that! If you dont have isolated it, then it does not excist!

    This confusion is not a matter of “we need to improve the tests” or “we need to understand the immune system, virus, mutations, etc., better.” This circumstance is a failure of the very concept of surrogate testing, which is what’s being used with Covid-19. You simply can’t use a surrogate test EVER for diagnosis unless you have a definitive (called “gold standard”) test to compare it to.

    Thats a fact!

  11. Assuming that the nucleotides are equally distributed in the RNA is not correct.

    Counting them in different published virus genomes showed that t is about 2 times more frequent than c.
    Also sequences cccccc, gggggg do not exist although in my trial they should be present theoretically 43 times.

    So the specificity of an rt-PCR is not that easy to explain with the low theoretical probability, especially as there are many different genomes and fraction of them present in a throat swab which which could be used for the multiplication.

    Also the existance of fraction of SARS-COV2 may trigger rt-PCR positive without that the presence of virulent virus can be proven by cell culture.

    Still a long way to learn …

  12. Thanks for taking the time to write this article I appreciate it. The only thing we need to know for sure is, do we (or even you) know for sure which type of PCR tests are being used at the so called ‘testing stations’ and where they were obtained and manufactured? Everything else is irrelevant, especially the statistics regarding infections and fatalities propagated in the screaming media.

  13. What a load of propaganda only idiots would believe this bovine feces . They CG’d the virus in the lab from a random patient that did not actually have covid, as the PCR tests for covid on 26th dec had not been developed as they had not even seen the genome ( no agents/primers unknown).
    Its just a research tool not a diagnostic tool.

  14. Thank you Ian for explaining the general procedures of primer & RT-PCR test design and the significant amount of self-checking processes that go into creating a reliable test. I hope that you spending your precious time writing this easy to understand article will help to spread more confidence in the RT-PCR test as a diagnostic tool, since we live in an age where negative (mis)information about rare malpractice spreads like wildfire and good news on a job well done is seldomly remembered.

    I guess the harshest of critics can only be silenced when an in depth scientific article is published explaining how one of the actual RT-PCR tests that specifically detects key markers of SARS-CoV-2 was designed, showing them in full detail what defining part(s) of the viral RNA are targeted and what corresponding primers are used, but with patents an financial interests involved this will ofcourse never happen. So unless you have the freedom to uncover the full details there still is a certain amount of trust involved.

    What do you personally think about the way the Australian government is currently translating the observed trend in test results into active measures? Does the severity of these measures match up to the actual danger present and how do these measures compare to the actual amount of hospitalizatons and fatalities?

  15. Thank you Ian for explaining the general procedures of primer & RT-PCR test design and the significant amount of self-checking processes that go into creating a reliable test. I hope that you spending your precious time writing this easy to understand article will help to spread more confidence in the RT-PCR test as a diagnostic tool, since we live in an age where negative (mis)information about rare malpractice spreads like wildfire and good news on a job well done is seldomly remembered.

    I guess the harshest of critics can only be silenced when an in depth scientific article is published explaining how one of the actual RT-PCR tests that specifically detects key markers of SARS-CoV-2 was designed, showing them in full detail what part(s) of the viral RNA are targeted and what corresponding primers are used, but with patents an financial interests involved this ofcourse will never happen. So unless you have the freedom to uncover the full details there still is a certain amount of trust involved.

    What do you personally think about the way the Australian government is currently translating the observed trend in test results into active measures? Do the measures match up to the actual danger present and how do they relate to the actual amount of hospitalizatons and fatalities?

  16. Dear Ian,

    In your article you write about the reference of PCR test that “The leading tests have also been tested on SARS-CoV-2 RNA purified from virus isolated and grown in cells in the lab – so we know it is this virus that we’re detecting.”

    Do you mean nu this that there is a golden standard that is used as a reference?
    And if so, van you link of post the scientific paper to that?

    Thanks in advance.
    Kind regards,
    Cees

  17. Hello,

    To quote you:

    “The leading tests have also been tested on SARS-CoV-2 RNA purified from virus isolated and grown in cells in the lab”

    Can you help me out with a link to a paper/study describing how this was done.

    Can you also comment on the relevance of Rivers postulates to SARS COV 2 and if there is a paper/study describing it’s application to establish the causal link between the virus and the disease.

    Thanks!

  18. Hi. Very useful info. “But we have learned much about this virus in 7 months and SARS-CoV-2, like other viruses we detect with RT-rPCR, is present before and during the peak symptomatic period. These are also the times we get sampled and tested because we feel sick. We know from experience that when detectable viral RNA is present in a patient’s sample, there is often an infectious virus present as well. RNA is not so stable that it hangs around without a constant source producing more of it.” This is a crucial area surrounding the current ‘casedemic’ in Europe. Virtually all positive PCR tests are of asymptomatic people and the data shows they are mostly young people. For how long does the viral RNA remain detectable after the infection has passed? I have read studies where infectious and non-infectious RNA has been identified.

  19. How are you sure that the “small part of the genome” that the test is looking for belongs exclusively to this virus and does not come from something else ?
    As far as i know humans and monkeys share 96% of their genome, so chances are that a small part of this virus genome might be shared with something else.

    1. By specifically checking that it does. There is a public sequence database (https://blast.ncbi.nlm.nih.gov/) that is accessed using its online tool called BLAST. We can interrogate all the sequences on it, from all the animals and viruses and sequenced so far. Plus we test the primers as discussed in the blog.

  20. I’m worried by this article. “We’re pretty sure” this means hypothesis. PCR is very specific yes, but detects a snipped of the hyped genome. 40 PCR cycles are the gate opener for nonspecific amplification. The PCR results, (a) do not provide evidence of the virus, (b) do not provide evidence for a active infection, (c) do not explain the “wonder” of asymptomatic people. I’ve seen some comments here not answered, please provide reference of isolation and culture of the virus, then please provide reference of an active infection from one object to another and then finally provide reference for a correlation of disease progression and virus quantification. Thanks

    1. Hi Craig, there is no one single Ct. The PCR is cycled 30-50 times (varies) and the result (which could be anything from 10-40something cycles) is then carefully interpreted in suitable context, and the diagnosis made by a medical doctor using all this info, then passed onto the patient. The media seem to be running with this idea that a lab has the only say in this process. It’s generally a bit more complex and well through than that

  21. Interesting article and comments until…your refeence of the isolation of the virus refers to a twitter feed. I think what we are looking for is a direct reference to a peer reviewed piece of research document that shows the isolation of this virus

  22. Two problems

    a.) we talk always about known Viruses, but the most of them are unknown. So the primer is only specific against known Viruses.

    but that’s not the main problem.

    b.) infected or ill?
    The PCR Test only detect the possibilty presence of a Virus. But there are always and everday a lot of Viruses present in our body. Our immune defence is always fighting against them. And that’s the point here. Since the most people are not ill and many show no sympthoms, it seems that the virus try to infected the first cells, but never had a chance to spawn.

    There are a couple of studies which see a possible big T-Cell answer ( e.g. https://www.researchsquare.com/article/rs-35331/v1%20 ) and in this case the PCR Test is also positive, but it’s doubful if there is a real infection (it depends wich definition you count.) neither a transmission, because the virus is not augmentable anymore.

    Also this T-Cell response may responsible for the big difference between the dead cases in the regions. For example compare, Peru and Singapore – the later had a big Sars outbreak in the past and a possible cross immunity
    https://biospectrumasia.com/news/26/16310/singapore-scientists-uncover-sars-cov-2-specific-t-cell-immunity.html

    So, the question still remains, did the PCR Test really test an infection?

    1. a.) The primers are indeed specific to what they are intended to detect. That’s how they’re designed. In this case, SARS-CoV-2.
      b.)The PCR test detected an infection. Yes. The virus may be at such low levels once finally tested that the [person is not a transmission risk *now* – but they were at some point. So then you can track their close contacts and ask them to quarantine for up to 2 weeks to make sure they don’t become ill and carry on the spread. If testing takes too long, this won’t be useful of course.

  23. The problem with the test isn’t if it can detect the virus, it is whether the virus they’re detecting is at an infectious level. The virus is almost always NOT infectious above 30 cycles of the test, yet 37-40 cycles is being used to detect positive cases. That leads us to believe at least 63% of positive tests shouldn’t be.

    1. To be clear, the virus detected below 30 cycles isn’t always able to be cultured using cell culture nor is that detected above 30 cycles never able to be cultured.
      Reasons may include the specimen being from the wrong anatomical area, specimen sat around too long, it wasn’t kept at the right temperature, it wasn’t protected by the addition of buffered with culture medium, the people trying to grow virus are not suitable expert, it was frozen and thawed a few times first, and so on. Culture failing to “grow” virus doesn’t mean there is not infectious virus there – its a bit more complex than that.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.