PCR Versus qPCR: What Real-Time Detection Changes

People use the terms interchangeably, and laboratories rarely correct them, but PCR and qPCR are not two different chemistries. The enzymes are the same. The primers are the same. The temperature cycling is very nearly the same. What changes is the moment you choose to look at the reaction, and that choice determines whether you end up with a yes-or-no answer or a number.

This distinction matters far beyond terminology. It explains why a conventional PCR result cannot tell you how much target was present, why quantitative results are reported as an odd unit called a Ct value rather than a concentration, and why two laboratories running the same assay can report different numbers for the same sample.

The Three Temperature Steps of a PCR Cycle

Every polymerase chain reaction is a repeated three-step temperature programme, and each step exists to force one specific molecular event.

Denaturation comes first, typically near 95 degrees Celsius. At this temperature the hydrogen bonds holding the two DNA strands together break, and the double helix separates into single strands. This is purely physical. Nothing is being synthesised; the reaction is simply making the template accessible.

Annealing follows, usually somewhere between 50 and 65 degrees. The temperature drops enough for short synthetic DNA sequences, the primers, to find and bind their complementary sequences on the template. This step is where specificity is won or lost. Too low a temperature and primers bind loosely to sequences they only partly match, producing amplification of the wrong thing. Too high and they fail to bind at all.

Extension completes the cycle, generally around 72 degrees, which is close to the optimum for the thermostable polymerases used in these reactions. The polymerase binds where a primer has annealed and synthesises a complementary strand, extending outward from the primer along the template.

Then the cycle repeats. Each round doubles the number of copies of the target region, at least in principle. Twenty cycles produce roughly a million-fold amplification, thirty cycles roughly a billion-fold. This exponential character is the entire point of the technique and also the source of every complication that follows.

Why Endpoint Detection Loses Information

Conventional PCR examines the reaction only once, after all cycling has finished. The product is typically run on an agarose gel, where a band of the expected size confirms that amplification occurred.

The problem is that PCR reactions do not amplify exponentially forever. Every reaction eventually plateaus. Reagents deplete, the polymerase loses activity through prolonged heat exposure, and the sheer concentration of product means strands reanneal to each other faster than primers can bind. By the final cycles, the reaction has essentially stopped producing new material.

That plateau is a levelling force. A reaction that started with a million template copies and one that started with a hundred both reach the plateau, and both reach roughly the same final concentration. They arrive at different times, but if you only look at the end, you cannot tell them apart. Two samples differing by four orders of magnitude in starting material produce indistinguishable bands.

This is why conventional PCR is a presence-or-absence test. It answers whether the target sequence exists in the sample above the detection threshold. It cannot answer how much, and attempts to read quantity from band brightness are unreliable enough that they are rightly treated with suspicion.

There is a second, subtler loss. Endpoint detection cannot distinguish a reaction that amplified efficiently from one that struggled and eventually caught up. Inhibitors present in the sample, which are common in blood, soil and stool extracts, slow early cycles. Given enough cycles the reaction may still reach plateau, hiding the inhibition entirely.

Fluorescent Chemistries That Watch in Real Time

Real-time PCR solves this by measuring product accumulation during every cycle. The instrument reads fluorescence from each well after each extension step, building a curve rather than a single data point.

Two chemistries dominate. The simpler uses an intercalating dye that fluoresces strongly when bound to double-stranded DNA and weakly when free in solution. As product accumulates, more dye binds, and fluorescence rises. This is inexpensive and works with any primer pair, which makes it the default for method development.

Its weakness is that it cannot tell what it is binding to. Primer dimers, where two primers anneal to each other and get extended, are double-stranded DNA and fluoresce exactly as brightly per base pair as genuine product. Non-specific amplification products do the same. The dye reports total double-stranded DNA, not target.

Probe-based chemistries solve specificity by adding a third oligonucleotide that binds between the two primers. In the most common design, the probe carries a fluorescent reporter at one end and a quencher at the other; while intact, the quencher absorbs the reporter’s emission. When the polymerase extends through the region, its exonuclease activity degrades the probe, separating reporter from quencher and releasing a fluorescent signal. Because signal only appears when the polymerase has copied the specific region between the primers, non-specific products contribute nothing.

Probes cost more and require additional design work, but they enable something dyes cannot: multiplexing. Different probes carrying spectrally distinct fluorophores let several targets be quantified in one well, which is how a single respiratory panel can report on multiple pathogens simultaneously.

What a Ct Value Actually Represents

The output of a real-time reaction is an amplification curve, and the number extracted from it is the cycle threshold, written Ct or sometimes Cq.

The instrument sets a fluorescence threshold in the region where amplification is still exponential and the signal has clearly risen above background noise. The Ct is the fractional cycle number at which each well’s curve crosses that threshold.

The relationship is inverse and logarithmic, which trips up almost everyone at first. More starting template means fewer cycles are needed to reach the threshold, so a low Ct means a high concentration. In a perfectly efficient reaction, where every cycle exactly doubles the product, a tenfold difference in starting material shifts the Ct by about 3.32 cycles, because that is how many doublings a factor of ten requires.

Several things follow from this. A Ct value is not a concentration and carries no units. It is meaningful only relative to other wells run under the same conditions with the same threshold setting. Comparing raw Ct values between laboratories, between instruments, or even between runs with different threshold placements is not valid.

Amplification efficiency deserves attention here. Real reactions rarely double perfectly. Efficiency is calculated from the slope of a standard curve, and an assay running well typically falls between 90 and 110 percent. Outside that band, the arithmetic converting Ct differences into fold-changes breaks down, and results calculated as though efficiency were perfect become progressively wrong as differences grow.

Standard Curves and Absolute Quantification

Turning a Ct into an actual quantity requires calibration, and the standard curve is how that is done.

A series of samples with known target concentrations, usually a tenfold dilution series spanning five or six orders of magnitude, is run alongside the unknowns. Plotting Ct against the logarithm of starting quantity produces a straight line. Unknown samples are then read off that line.

The quality of this calibration determines the quality of every result derived from it. The slope should be close to negative 3.32, corresponding to full efficiency. The correlation coefficient should be very close to one across the whole range. If the lowest standards fall off the line, that marks the practical limit of quantification, below which results should be reported as detected but not quantified.

Standards themselves are a common weak point. Plasmid DNA behaves differently from genomic DNA in solution, tends to adsorb to plastic at low concentrations, and can carry over between runs as a contaminant. Synthetic oligonucleotide standards avoid some of this but do not replicate the structural context of genomic targets.

Where absolute numbers are not required, relative quantification is simpler and often more robust. The target is compared against one or more reference genes assumed to be stable across the conditions being tested, and results are expressed as fold change. This cancels out variation in input amount and extraction efficiency, but it stands or falls on the reference genes genuinely being stable, which should be verified rather than assumed.

Melt Curves and Non-Specific Product

When intercalating dyes are used, a melt curve analysis run immediately after cycling provides a specificity check that costs almost nothing.

The instrument slowly raises the temperature while continuously monitoring fluorescence. As the temperature climbs past the melting point of a particular double-stranded product, its strands separate, dye is released, and fluorescence drops sharply. Plotting the negative first derivative of fluorescence against temperature converts each drop into a peak.

A clean, specific reaction produces a single sharp peak at the expected melting temperature. Primer dimers, being short, melt at noticeably lower temperatures and appear as a separate peak. Non-specific products of different length or base composition appear as additional peaks or shoulders.

This turns a potentially misleading quantitative result into an interpretable one. If a sample shows a low Ct but the melt curve reveals the signal came largely from primer dimers, the quantification is meaningless and the assay needs redesign rather than reinterpretation. Melt analysis is not available with probe-based chemistries in the same form, since probe fluorescence does not depend on double-stranded structure in the same way.

Where Digital PCR Goes Further

Digital PCR takes a different route to quantification, and in doing so removes the need for calibration altogether.

The reaction mixture is divided into a very large number of tiny partitions, either as droplets in an emulsion or as wells in a fabricated chip. Partitions are small enough that, at appropriate dilution, most contain either one target molecule or none. Amplification then runs to completion in every partition, and each is scored simply as positive or negative at the end.

Counting positives and applying Poisson statistics, which corrects for partitions that happened to receive more than one molecule, yields the absolute number of target molecules in the original sample. No standard curve is involved. The measurement is a count.

This confers real advantages. Sensitivity to small differences improves, because the method is counting rather than inferring from a curve. Tolerance of inhibitors improves, since partition-level amplification either works or does not, and partial inhibition affects results less severely. Detection of rare targets against a large background of similar sequence, such as a small population of variant molecules among many normal ones, is substantially better.

The costs are real too. Instruments and consumables are more expensive, throughput is lower, and the dynamic range of a single reaction is narrower, meaning samples of unknown concentration may need dilution and repetition. For most routine work, real-time qPCR remains the sensible default. Digital PCR earns its place where absolute accuracy matters, where rare targets must be found, or where inhibitors defeat conventional approaches.

The practical upshot is that the choice between these methods is really a choice about what question is being asked. If the question is whether a sequence is present, conventional PCR answers it at the lowest cost. If the question is how much, real-time detection is the minimum requirement, and the Ct value it produces is only as meaningful as the calibration and controls surrounding it. If the question demands an absolute count without reference to standards, digital partitioning is the method built for it.