A PCR test can call you positive weeks after you stopped being infectious. That is the test working correctly

A PCR test does not look for a virus. It looks for one specific stretch of genetic sequence, and then it copies that stretch until there is enough of it to see. Everything people found confusing about these tests during the pandemic — why someone stayed positive long after they felt fine, why one lab's number meant something different from another's, why a test taken too early missed an infection that was obviously present — comes from that single sentence, and none of it is a flaw in the method.

Three temperatures, repeated thirty times

The reaction is startlingly simple for something that reshaped biology. In a tube you put the sample DNA, a heat-stable copying enzyme, a supply of the four DNA letters, and two short pieces of designed DNA called primers. Then you cycle the temperature.

At around 95 °C the double helix comes apart — denaturation — leaving single strands. Cooled to somewhere between 50 and 65 °C, the primers anneal, each latching onto one strand at the exact position its sequence matches. Warmed to about 72 °C, the enzyme extends from each primer, building a complementary strand.

One cycle turns each target molecule into two. Thirty cycles, if every one were perfect, turns one into over a billion. That exponential is the whole point: it takes a quantity of DNA far too small to detect by any direct method and, in about two hours, produces enough to measure easily.

The step that made this practical was borrowed from a hot spring. Ordinary copying enzymes are destroyed at 95 °C, so early PCR required fresh enzyme by hand after every cycle. Taq polymerase, taken from a bacterium that lives in near-boiling water at Yellowstone, survives the denaturation step — which is what turned a tedious manual procedure into a machine that runs unattended.

The primers are the entire specificity

Only the region bracketed by both primers gets amplified. The enzyme has no interest in what organism the DNA came from; it copies whatever the primers land on. So a PCR test's target is a design decision, written into two short sequences typically twenty or so letters long.

That has two consequences worth understanding. First, a well-designed pair is chosen to sit in a region that is conserved within the target organism and absent from everything else likely to be in the sample — which is why a test can distinguish two closely related bacteria while ignoring the human DNA vastly outnumbering them.

Second, if the target region mutates, the primers may bind poorly and that target drops out. This is not hypothetical: during the pandemic, several variants carried a deletion that caused one of the three targets in a common multi-target assay to fail while the others still worked. Laboratories turned that failure into a free signal, using the pattern as an early proxy for which variant was circulating. It is also the reason serious assays use more than one target — so a single mutation cannot turn a positive into a negative.

What the Ct number actually is

Modern testing does not wait until the end and inspect the product. Real-time PCR measures fluorescence after every cycle, using a probe that lights up only as the specific target accumulates. The machine reports the cycle number at which that glow crosses a set threshold: the Ct, or cycle threshold.

Because each cycle roughly doubles, Ct runs backwards from intuition. A low Ct means the signal appeared early, which means there was a lot of target to begin with. A high Ct means it took many cycles of doubling before anything was detectable, which means there was very little.

It is tempting to read Ct as a viral load, and it is a rough guide to it — but it is not a measurement you can carry between contexts. Ct depends on the assay's primers and probe, the instrument, the extraction chemistry, and above all on how the sample was taken. A shallow, badly-collected swab produces a high Ct from a person carrying plenty of virus. This is why laboratory bodies advise against treating Ct as a clinical quantitative result, and why comparing a Ct from one lab with a Ct from another is close to meaningless.

PCR answers exactly one question: is this specific sequence present in this tube, and roughly how much. It does not report whether that sequence came from something alive.

Why a positive is not the same as infectious

This is the point that caused the most public confusion, and it is pure chemistry.

The reaction amplifies nucleic acid. It cannot tell whether that nucleic acid came from an intact, replicating pathogen or from debris left over after the immune system destroyed one. Fragments persist. After a respiratory infection has resolved, detectable RNA is routinely still being shed for weeks, while attempts to grow live virus from the same person typically stop succeeding much earlier — often within the first week or so in mild illness.

So "PCR positive at day 20" and "not infectious at day 20" are both true, and there is no contradiction between them. It is why isolation guidance moved towards fixed time periods and symptom resolution rather than a negative PCR, and why culture, not PCR, remains the reference for viability. The same logic applies well beyond viruses: a PCR-positive water sample tells you the organism's DNA is there, not that anything in it is alive.

The two ways it goes wrong

False negatives are usually not the chemistry's fault. The commonest causes are sampling — a swab that did not reach the right tissue — and timing, because a test taken before the pathogen has multiplied to a detectable level will correctly report an absence that will not survive contact with tomorrow. Substances in the sample can also inhibit the enzyme, which is why good assays run an internal control that must amplify; if it does not, the result is invalid rather than negative.

False positives have a much more interesting cause. A completed reaction contains billions of copies of exactly the sequence the next test will look for. Open that tube carelessly and you have aerosolised the perfect contaminant. This is the reason molecular laboratories are laid out the way they are: physically separated rooms for reagent preparation, sample handling and post-amplification work, with a one-way workflow that people and equipment never reverse. There is even an elegant enzymatic defence — building reactions with a modified DNA letter and adding an enzyme that destroys any carried-over product from previous runs before the new cycle starts.

The variants of the same idea

RT-PCR adds one step at the front. Viruses such as influenza and SARS-CoV-2 carry RNA, which the polymerase cannot copy, so a reverse transcriptase first converts it to DNA. Everything after that is ordinary PCR — the "RT" is the only reason a covid test was called RT-PCR rather than PCR.

Digital PCR splits the reaction into thousands of tiny droplets, most containing either one target molecule or none, and counts how many droplets light up. That converts an exponential process into a direct count, giving absolute quantification without a standard curve — valuable for rare mutation detection and for anything where "how much" must be precise.

Isothermal methods such as LAMP abandon the temperature cycling entirely and run at a single temperature, which removes the need for a thermal cycler and makes field and point-of-care testing possible at the cost of some flexibility in design.

And PCR remains a component of nearly everything else in the molecular laboratory: preparing libraries for sequencing, cloning, screening genetically modified crops, detecting pathogens in food and water, and generating the short repeat profiles that underpin forensic identification.

Why it matters for students and researchers

PCR is the closest thing molecular biology has to a universal tool, and it is worth learning as a lesson in what a measurement actually claims. The test does not detect disease, or infectiousness, or an organism. It detects a sequence, in a tube, above a threshold — and every sound interpretation of a result begins by respecting that boundary and asking what else could put that sequence there.

That habit generalises. Whether the instrument is a thermal cycler, a mass spectrometer or a trained model, the same discipline applies: know precisely what the signal is, know what it cannot distinguish, and be suspicious of any conclusion that quietly claims more than the method can support. For students in biotechnology and diagnostics, PCR is where that lesson is cheapest to learn and most often ignored.

Frequently asked questions

How does a PCR test work?

It repeatedly heats and cools a sample so that a chosen stretch of DNA is copied over and over. Primers define which stretch is copied, a heat-stable enzyme does the copying, and after about thirty cycles the target has been amplified enough — potentially a billionfold — to be detected.

What does the Ct value mean?

Ct is the cycle number at which the fluorescent signal becomes detectable. A low Ct means a lot of target was present at the start and a high Ct means very little, but the value depends on the assay, instrument and sample quality, so Ct values are not comparable between laboratories.

Why do people test PCR-positive after recovering?

Because PCR detects genetic material, not living organisms. Fragments of viral RNA continue to be shed for weeks after an infection resolves, long after live virus can no longer be grown from the same sample.

What causes a false negative PCR result?

Most often poor sampling or testing too early, before the pathogen has multiplied to detectable levels. Inhibitory substances in the sample and mutations in the primer binding region can also cause it, which is why good assays include internal controls and more than one target.

What is the difference between PCR and RT-PCR?

RT-PCR adds a reverse transcription step that converts RNA into DNA before amplification, which is necessary for RNA viruses such as influenza and SARS-CoV-2. The amplification itself is identical to standard PCR.