There’s so much more to these four words than is attributed to them, so I thought I’d better write that down instead of leaving it in my head. Here’s hoping you learn something new.

Kary Mullis said a thing, and unsurprisingly, it was hijacked for clicks

“I don’t think you can misuse PCR, actually. Now, the results—the interpretation of it… see, if you can find this virus in you at all, and with PCR, if you do it well, you can find almost anything in anybody. It starts making you believe in the sort of Buddhist notion that everything is contained in everything else, right? I mean, because if you can amplify one single molecule up to something that you can really measure, which PCR can do, then there’s just very few molecules that you don’t have at least one single one of them in your body, okay?

No, not okay. Not even close to okay. This is a massive oversimplification of what PCR can do. For those of you who have used and worked with PCR for some time, there’s nothing more for you to learn here. You’re excused! For everyone else, please read on.

Quick things first

Giving Mullis credit – if you run enough PCRs each with different targets, or as multiplexed PCRs each detecting several targets, you can find a lot of different specific molecules in a human body. But only if there are enough target molecules present for your test to detect. PCR is not magic. It can’t detect too few of a thing, nor can it amplify a thing that isn’t in the tube. If that molecule isn’t in your body or in your collected sample, it will also never make it into the tube. So we’ll explore those ideas here.

Technically, a PCR can detect a single copy if that copy is in the reaction tube – usually containing a total reaction volume of between 20 μL and 50 μL, including the sample containing that one molecule. Realistically, PCR struggles to do that. If you look at many commercial PCR kits, you’ll see that they often state a real-life detection limit equal to about 3–80 copies in the reaction tube.

What are we talking about when we say “molecule”? We’re talking about a stretch of RNA or DNA that has been chosen as the target for PCR primers (and for real-time PCR probes, which broadened PCR’s reach after Mullis’s contributions). A DNA molecule is a long chain of nucleotides – themselves molecules made up of carbon, hydrogen, oxygen and phosphorus atoms.

So, a PCR product of about 150 base pairs, if laid out straight, would be 15 nm long (15 billionths of a meter; 100 pm per nucleotide base pair). For comparison, a red blood cell is 400 to 533 times bigger (RBC width).

A molecule of almost everything in us? No.

The key thing to focus on here is the statement that “just very few molecules that you don’t have at least a single one of them in your body“. To be fair, Mullis said “very few” – so he’s implying there are things that won’t be in your body to start with. But given how the anti-science cult has run with this quote, I think it’s relevant to dig in equally deeply. They believe it means a PCR will always be positive, you just have to run more cycles (that’s a bit they’ve taken on, not what Mullis said here) to detect the ever-present one molecule. But I think it’s a much smaller window of chance that you’ll have one molecule of all but very few. I don’t think I have any Andes hantavirus or Ebola Bundibugyo virus molecules in me. I don’t think I have any yak, mole, or squirrel genes, or Tyrannosaurus rex, Red-lipped batfish, panda, Lobster claw, Borderea chouardii, or Handkerchief Tree DNA in me right now. I don’t think I have the distinct sequences of your genomes floating around in my system either. As you can see…..that’s already a lot of molecules that I very likely don’t have.

This part of the statement is a ridiculous oversimplification.

Amplify one single molecule? A theoretical utopia

Yes, in a perfect world, in a perfect reaction in a perfectly prepared PCR tube containing one stable and purified single target molecule for two probes and one or two probes to bind to, a PCR could amplify that up enough for a PCR machine to manage to detect the millions of copies somewhere before the end of the PCR run at 40-45 cycles.

The reality is that these many conditions are seldom met. Why? Two broad reasons. One technical, the other about sampling. Technically, keeping that one molecule intact before PCR starts is asking a lot, even for PCR. And if that single molecule should survive degradation, it could still be destroyed during the initial cycles, or misprimed in a way that excludes the probe or other priming site – that target is then toast.

That’s why PCR has a “limit of detection” and, as I said above, that limit is usually higher than a single molecule. The ‘one molecule’ claim implies that the one molecule gets from my body to the PCR tube with its primer and probe target regions intact – or at least that’s how I interpret it from the engagements I’ve seen and had. For that one molecule to remain intact and present during a standard pathology laboratory nucleic acid extraction and purification workflow is bordering on impossible. Many of those steps will remove some nucleic acids, damage others, and leave them sitting at non-ideal temperatures – all before the magic molecule is even added to the tube with the PCR ingredients. Once it is, there’s also some lag before the PCR even starts – more opportunity for degradation to occur. When you start with hundreds or thousands of target molecules, these hurdles are much less in play. One molecule? Most of the time, this just does not happen.

Next, we’ll follow the initial target molecule’s journey, and the ridiculousness should become very clear very fast.

Collecting a sample with one molecule.

Let’s use blood as our sample. We collect an adult blood serum-separating tube with 8mL of blood in it.

Failure 1: Separating the cells from the serum

You mix the fresh sample, let it sit, then centrifuge it to separate out the plasma.

Shit. My one copy got stuck somewhere in the gel.

No target. Negative PCR. This step requires that you acknowledge PCR can’t amplify if there’s no target. Kary Mullis knew this – he published a few peer-reviewed research papers, including one on the use of his early crude form of PCR for HIV DNA detection, showing that PCR was specific to HIV and not to other human retroviruses, HTLV-I or -II. He knew the target had to be present to be detected. If it wasn’t there, it couldn’t be amplified.

Failure 2: separating the nucleic acids from the serum and purifying them

Okay. Let’s say Failure 1 didn’t happen.

How much blood is in the average female body? Let’s say 5,000 mL. We took 8mL of that 4,500 mL into our collection tube (a 1:625 dilution). Next, we took 0.2mL (200 μL) of the – let’s say – 5 mL of serum (1:25 dilution) we left after separating the cells by centrifugation into our lab tube to perform nucleic acid extraction and purification.

Over multiple steps, we diluted, lysed, bound (to silica in some form tha may include a magnetic collection step), washed, rinsed, and eluted 0.0001 mL (100 μL) of purified nucleic acids and buffer/water. Shit, in one of these many handling and washing steps, my one copy degraded or was lost in the waste we discarded.

No target. Negative PCR.

Failure 3: Dilution effects betrayed us

Okay. Let’s say the target wasn’t lost. I mean, it’s ludicrous to skip over these things. Anyhoo.

We took 0.16% of the patient’s blood volume (sample collection; 1:562), then 4% of that (extraction; 1:25), 50% of that (purification; 1:2) and finally tested 5% of the eluate (5 μL) in the PCR (1:20).

Now, I hate dilution factor maths, so apologies if I messed this up, but we tested one 625,000th of the patient’s blood (0.00016%).

The odds of our one copy still being present, even without all those technical issues I just listed above, are getting pretty slim.

Failure 4: THE BIG ONE: we never collected the one copy, it was in our toe

Let’s do all those steps, and don’t miss out on collecting our one molecule of template.

We took 8 mL of blood from the patient’s arm. But her “one molecule” of target we most likely all have all the time was currently circulating at centimetres per second through a dorsal digital vein in her foot. We never collected it.

Oh, it just got degraded/ingested/pee’d out.

No target. Negative PCR.

Being a bit too pedantic?

Sure, this is all a bit much. But it’s NOTHING like reading people make claims based on a world with physical, biological, and technical laws that don’t even exist. And with an absolute megaton of evidence for what’s real, all of which is meaningless to these folk who have NEVER performed this work and become expert in it. We’re all so sick of this kind of claim being made on the back of corrupting off-the-cuff comments or poorly worded scientific paper titles, then building whole cults and conspiracy campaigns around it.

This is a lot of words to say “Nope”. This one-molecule thing is rejected outright.

Update

1. 13MAY2026. Added paragraph two to the “Quick Things” section to discuss that while PCR can technically detect one copy of the target, it almost always requires 3 or more copies for a reliable signal and under the “utopia” section – my thanks to Prof Cane for raising her concerns in the comments.