Kary Mullis inventor of PCR test - Covid19 false positives

[Epaphroditus]

https://www.idsociety.org/covid-19-real-time-learning-network/diagnostics/RT-pcr-testing/

We are talking about PCR not shutdown. Please don't be obtuse.

[therealnickb]

https://www.idsociety.org/covid-19-real-time-learning-network/diagnostics/RT-pcr-testing/

We are talking about PCR not shutdown. Please don't be obtuse.

Tell that to purple boy newsome.

[Epaphroditus]

I've met Kary Mullis. I've been to his cabin in Andersen Valley. He is a nutjob. Doesn't mean he's not excruciatingly intelligent and can go on and on about PCR and be exactly right - but, technical details matter.

There are myriad ways even a properly configured PCR reaction can "go wrong" and the more cycles the more likely things go wrong.

[Epaphroditus]

Despite Dr Mullis statements about any molecule the fact is PCR amplifies DNA- not every type of molecule. So, no, you can't detect anything in anyone no matter how configured your reaction.

There are countless millions of molecules that will denature the polymerase enzymes and kill the PCR - never to be detected.

[duenor]

PCR can isolate a small number of infected cells and multiply them exponentially through a process of raising/lowering temperatures. Small amounts of infected cells doesn't mean we are sick. Our immune system fights off infection all the time.

Your OP reads, "it can find any molecule in anyone". As someone who designs PCR machines for a living, as you said, what do you mean by that statement then? Because it certainly seems to imply that you can take a person who hasn't been infected, run the test a certain way (and your later post implies 40 cycles might be that certain way), and you'll get a positive result.

If a person has a small number of infected cells, then that person is indeed infected. Maybe he will fight them off, maybe not. Or, maybe he was sick before and is now recovered. These are nuances that you have chosen to to omit from your previous statements.

[five.five-six]

I wouldn’t be surprised to find that at this point, every one of us has had at least 1 viable COVID19 virus infected cell since this all started.

[UCT]

Your OP reads, "it can find any molecule in anyone". As someone who designs PCR machines for a living, as you said, what do you mean by that statement then? Because it certainly seems to imply that you can take a person who hasn't been infected, run the test a certain way (and your later post implies 40 cycles might be that certain way), and you'll get a positive result.

If a person has a small number of infected cells, then that person is indeed infected. Maybe he will fight them off, maybe not. Or, maybe he was sick before and is now recovered. These are nuances that you have chosen to to omit from your previous statements.

It looks like you don't understand the nuances yourself. PCR testing does not look for infected cells. It looks for discrete pieces of genetic material. It doesn't distinguish between whole virus particles capable of causing infection (virion) or pieces incapable of causing infection. It works on the principle that the less number of amplifications before a positive test the higher the probability of a positive test includes virions. If I recall correctly, after amplification of 30 times or so the percentage of positive tests with virion as opposed to viral debris is only about 8 percent. So at 40 amplifications (which is the standard cut off) a positive test is almost certainly not positive for active infection.

[freespool]

It looks like you don't understand the nuances yourself. PCR testing does not look for infected cells. It looks for discrete pieces of genetic material. It doesn't distinguish between whole virus particles capable of causing infection (virion) or pieces incapable of causing infection. It works on the principle that the less number of amplifications before a positive test the higher the probability of a positive test includes virions. If I recall correctly, after amplification of 30 times or so the percentage of positive tests with virion as opposed to viral debris is only about 8 percent. So at 40 amplifications (which is the standard cut off) a positive test is almost certainly not positive for active infection.

What's with the fascination with high cycle count positives? Just what is the great insight that flows from them? Are they a significant proportion of positive test results?

The test requires adequately intact viral RNA to span the distance between the primers, and adequate quantity to amplify to positive given the cycle limit. That's all it tests for. It produces a binary call : infected or not. It means you've had replicating SARS CoV2. Trying to read the cycle count as some kind of clinical correlate is bogus. There are other reasons related to sampling and sample handling that increase the cycle count needed to show positive. You could even be too early rather than too late in the infection.

[UCT]

What's with the fascination with high cycle count positives? Just what is the great insight that flows from them? Are they a significant proportion of positive test results?

The test requires adequately intact viral RNA to span the distance between the primers, and adequate quantity to amplify to positive given the cycle limit. That's all it tests for. It produces a binary call : infected or not. It means you've had replicating SARS CoV2. Trying to read the cycle count as some kind of clinical correlate is bogus. There are other reasons related to sampling and sample handling that increase the cycle count needed to show positive. You could even be too early rather than too late in the infection.

Cycle times aren't reported. That is the point. If the CT is 10 you definitely have virions If it is 40 you almost certainly do not.

[freespool]

Has that been demonstrated so categorically? But OK let's throw out all the >35 positive tests. What changes in understanding does that bring?

[duenor]

It looks like you don't understand the nuances yourself. PCR testing does not look for infected cells. It looks for discrete pieces of genetic material. It doesn't distinguish between whole virus particles capable of causing infection (virion) or pieces incapable of causing infection. It works on the principle that the less number of amplifications before a positive test the higher the probability of a positive test includes virions. If I recall correctly, after amplification of 30 times or so the percentage of positive tests with virion as opposed to viral debris is only about 8 percent. So at 40 amplifications (which is the standard cut off) a positive test is almost certainly not positive for active infection.

No, I am not mistaken. The "infected cells" bit came from what the OP wrote. I'm just quoting his own misinformed stance.

Originally Posted by surfdesigner View Post
PCR can isolate a small number of infected cells and multiply them exponentially through a process of raising/lowering temperatures. Small amounts of infected cells doesn't mean we are sick. Our immune system fights off infection all the time.

[UCT]

Has that been demonstrated so categorically? But OK let's throw out all the >35 positive tests. What changes in understanding does that bring?

When you test with cycle times up to 40 like we are doing you get a lot of false positives. That leads to a lot of bad data everywhere. For example, a high percentage of false positives makes it appear there are lots of cases of asymptomatic patients. It also makes it appear that the number of cases of asymptomatic transmission is lower than it is. It also makes it appear that the death rate per infected case is lower than it is.

At 25 ct, about 70 percent of the swabs have live virus. At 30 ct it is only 20 percent. The maximum ct should be in between. https://academic.oup.com/cid/advance...aa1491/5912603

[UCT]

No, I am not mistaken. The "infected cells" bit came from what the OP wrote. I'm just quoting his own misinformed stance.

okay, if I misread your post, sorry.

[freespool]

When you test with cycle times up to 40 like we are doing you get a lot of false positives. That leads to a lot of bad data everywhere. For example, a high percentage of false positives makes it appear there are lots of cases of asymptomatic patients. It also makes it appear that the number of cases of asymptomatic transmission is lower than it is. It also makes it appear that the death rate per infected case is lower than it is.

At 25 ct, about 70 percent of the swabs have live virus. At 30 ct it is only 20 percent. The maximum ct should be in between. https://academic.oup.com/cid/advance...aa1491/5912603

Your point about interpreting the results is a good one, it's just not correct to call culture-negative but PCR-positive samples as false positives. The term has a clearly defined meaning, and that people are trying to over-interpret the results, or that a binary call from a PCR test is a poor clinical correlate doesn't mean a problem with the test false positivity spec, or with the use of 40 cycle counts. The test is not a useful forward-looking transmission risk indicator, whether you use 20 counts or 40. I think that's likely the reason that after several months of supposed contact-tracing practices around the country, nobody is crowing about the great results in transmission reduction. The likelihood is that the timeliness of the testing and tracing is far too slow to be actionable. They are constantly chasing it, but never catching up.

Even in that paper, only 3% of their PCR-positive sample set was Ct>35. If that's currently representative, it's not much of a distorting factor, and for those population statistics you refer to, it's probably not close to the biggest problem in estimating those. There's no systematic collection of symptomatic data, and the PCR testing is not being conducted in a manner which provides good population survey numbers.

It was interesting to see that even in samples that were PCR-positive at much lower Ct, they failed both to isolate virus and even to culture virus from those isolated, for significant percentages of samples. That probably says a whole lot more about their isolation and culturing protocols than about the specificity of the PCR method.

[surfdesigner]

Virus infected cells are lysed (exploded) from which the RNA is extracted (then converted into DNA) for the RT-PCR assay. In Dx lab machines with blood they use magnetic beads which "stick" to the proteins on the outside of the infected cells and then use magnets to isolate them, then lyse them. Then they add "mastermix" reagents, before putting them through the PCR thermocycling to exponentially multiply and create the amplification curve (viral load/concentration/# of cycles). As an example from 0-10 cycles not much is happening (the baseline), then exponential amplification starts to occur say from 15-25 cycles. Depending on the type/concentration of reagent used the curve will start to plateau.

I learned a few things from one of our scientists today...The key is what they call the CT or crossing threshold and where that line is marked. This is the line used to determine a positive infection. CT should be low enough to maximize limit of detection but not too high as to call negative curves positive. A big variable.

Depending on reagent concentration used, thermocycling ramp rates and assay design, the curve can be greatly influenced by number of cycles, crossing the CT.

He also mentioned contaminated test kits/primers/reagents, contaminated labs could also lead to false positives.