Coronavirus: 7 of 27 PCR tests contain mismatches

Some of the templates used for coronavirus tests no longer match sequences of the virus itself...
15 June 2020

Interview with 

Kashif Aziz Khan, York University


Since the start of the pandemic, whenever someone has been tested for coronavirus, it’s almost always been using something called a PCR. This machine looks for the genetic code of the virus itself, and compares it against a ‘gene sequence’ it uses as a template so that it can say, “yes, this is genuine SARS-CoV-2 right here”. But a recent analysis found that a number of those templates that are being widely used actually no longer match versions of the coronavirus. That is - its genes have mutated in tiny ways, and so some of the tests might be missing it - what are called ‘false negatives’. Phil Sansom heard from study author Kashif Aziz Khan…

Kashif - 7 tests out of 27 we studied have mismatches with the virus genome. Not necessarily that they are wrong because not every mismatch would lead to false negative, but they is a possibility. And especially certain tests were developed earlier in the outbreak when only few sequences are available. While in our study, we use 17,000 sequences. So there is room for improvement of the test.

Phil - That seems like a pretty big problem, no? How many cases of coronavirus do you think might have been missed thanks to this?

Kashif - It's difficult to like really give a number to these mismatches compared to other factors, but there are estimates of up to 30% false negatives. We cannot put all the false negatives due to mismatches. One of the major factors is timing off sampling, but I believe part of that might be due to these mismatches.

Phil - And these are tests that are looking for the viruses, what genetic code?

Kashif - Yeah. So these are called PCR. In PCR, DNA is amplified. In the case of coronavirus it's RNA, so we convert RNA to DNA, and then DNA is amplified using PCR.

Phil - So you're saying that those PCR tests, they're looking for little bits of the viruses genetic code, its RNA. And actually the virus now looks completely different.

Kashif - I would not say that all these tests are not detecting the virus at all or virus is completely different now, but there are small changes. These small mutations, we should not really worry because not all these mutations are bad. So these are small changes. It has not developed into a separate strain.

Phil - Why are there 27 different tests though?

Kashif - Earlier in the outbreak, different national organisations around the world developed certain tests. In addition to these 27 that I found in the literature, there are many other being developed by different commercial manufacturers.

Phil - These 7 out of the 27 that you found the mismatches in, where are they getting used right now?

Kashif - Actually, most of these are used around the world. Like one of the tests, which is widely used, it has one mismatch. So we need to improve it. That's what we have suggested. And in addition, many different countries are using commercial tests. And for those, I could not include those in my study because the information for those tests is not available.

Phil - And does that apply to tests that aren't these PCR tests as well? Cause I know people are definitely interested in looking for something that's quicker or something that's cheaper or whatever.

Kashif - So that would apply to any molecular test, even if it's not a PCR. If the test is targeting the exact region, we can go check how much that region is variable. That was the objective of this study to put this information there. And we have a pipeline which is easy to use, anybody can use on a regular computer because we use all open source free software.

Phil - And that that's the message that you're trying to get out. You're saying, come to me, check that you're looking for the right bit of the virus.

Kashif - Yeah.


Add a comment