JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

PCR: Copying DNA in a Tube

How a single faint stretch of DNA becomes a billion clean copies in two hours — no cell required. Heat, cool, copy; repeat; and watch one target double, and double, and double again.

The cell's copying trick, borrowed for a tube

You already know, from far down this ladder, how a cell copies its DNA: a polymerase reads one strand and builds the complementary partner, laying down A across from T and G across from C. The previous rung borrowed the cell's scissors and glue so you could *cut and paste* DNA. This guide borrows its *copier*. The polymerase chain reaction — PCR — takes that same templated-copying chemistry out of the living cell and runs it in a plastic tube, on purpose, over and over, until one chosen stretch of DNA has been copied a billionfold. It is, quite simply, the most-used technique in the molecular lab, and it is the reason a swab, a hair, or a drop of blood can be made to speak.

Here is the cleverness in one sentence: a copy made in one round becomes a template in the next, so the copies make copies. That is what the word *chain* means — a self-feeding chain reaction. The cell copies its whole genome once per division and stops; PCR instead aims at one small target and copies *that* relentlessly, doubling it every cycle. To pull the cell's copier out of the cell, though, you have to solve two problems the cell solves with proteins it cannot share: how to pry the two strands apart without an enzyme, and how to tell the polymerase *which* stretch to copy. PCR's two key ingredients — heat and a pair of primers — are the answers to exactly those two questions.

Three steps, repeated: denature, anneal, extend

One PCR cycle is just three temperatures, held for a few seconds each, in a machine called a thermal cycler that heats and cools the tube on a schedule. First, denaturation: the tube is heated to about 95 degrees Celsius. The two strands of the double helix are held together only by hydrogen bonds — weak, and many — so heat shakes them loose and the helix unzips into two single strands. This is the same melting you met as a property of DNA; here it replaces the helicase enzyme the cell would use. Heat is the molecular crowbar.

Second, annealing: the tube is cooled to roughly 50 to 65 degrees. Now the two short, lab-made DNA pieces called primers — added in vast excess — find their matching sequences on the single strands and stick by base pairing (A-T, G-C). Cooling lets pairing happen; but because the tiny primers are everywhere and the long original strands are rare, the strands grab a primer far faster than they find each other again. Third, extension: the tube is warmed to about 72 degrees, the polymerase's favourite working temperature, and the enzyme settles onto each primed spot and copies along the template, adding nucleotides one at a time in the 5'-to-3' direction. One round of denature, anneal, extend, and every original target strand has become a fresh double strand.

  1. Denature (~95 C). Heat snaps the hydrogen bonds and unzips every double helix into two single-stranded templates — no helicase enzyme needed.
  2. Anneal (~50-65 C). Cool so the two primers base-pair to their matching ends of the target, one on each strand — this is the step that aims the reaction.
  3. Extend (~72 C). The heat-stable polymerase grabs each primer's 3' end and copies along the template, building a new complementary strand.
  4. Repeat 25-40 times. Each cycle's products become the next cycle's templates, so the target roughly doubles every round.

Doubling, doubling: exponential growth

Now run that cycle again. Every strand made in cycle one is a template in cycle two; both products of cycle two are templates in cycle three. The number of target copies roughly doubles each cycle, so it grows like 1, 2, 4, 8, 16, 32... — geometric, not steady. After n cycles you have on the order of 2-to-the-n copies. The numbers are startling: 10 cycles gives about a thousand-fold, 20 cycles about a million-fold, 30 cycles about a billion-fold. That is why a single target molecule, far too faint to detect, becomes a visible, workable mountain of identical DNA after an afternoon in the cycler. Exponential growth is the whole engine of thermal cycling.

one PCR cycle = three temperatures:

  DENATURE  ~95C   ===||===   ->   ===   +   ===   (strands part)
  ANNEAL    ~55C   ===>            <===            (primers stick to ends)
  EXTEND    ~72C   ===>------>     <------<===     (polymerase copies)

repeat -> copies become templates -> exponential doubling:

  cycle:   0     1     2     3   ...   10      20        30
  copies:  1     2     4     8   ...  ~1e3    ~1e6     ~1e9
           (after n cycles: about 2^n copies of the target)
Each cycle's new strands are next cycle's templates, so the target grows like 2-to-the-n: ~1,000-fold by cycle 10, ~a billion-fold by cycle 30.

One honest wrinkle: the doubling is never perfectly efficient, and it cannot go on forever. Each cycle copies somewhat less than 100 percent of the available templates, and as the reaction runs out of primers and nucleotides — and as so many product strands crowd the tube that they re-pair with each other faster than primers can bind — the curve flattens into a plateau. So PCR does not give you literally 2-to-the-n; it gives you exponential growth at first that levels off near the end. That detail looks like a footnote, but it matters enormously for *measuring* DNA, which is the door into the next idea.

The primers are the aim — and the two heroes

Why does PCR copy only one chosen stretch out of a whole genome, and not the entire DNA? The answer is the primer pair, and it deserves to be understood, not memorised. A primer is a short single strand of DNA, about 18 to 25 letters long, that you design and order to be complementary to one end of your target. A reaction uses two: a *forward* primer matching the start of the target on one strand, and a *reverse* primer matching the other end on the opposite strand — they face inward, like two bookmarks bracketing exactly the passage you want. The polymerase can only extend *forward* from each primer, into the region between them, so only the DNA lying between the two primer sites is ever copied. Change the two primers, and you point the same machine at a different gene. The primers *are* the address.

The second hero is the enzyme. An ordinary DNA polymerase would be cooked to death by the first 95-degree denaturation step, and early PCR really did require adding fresh enzyme by hand every single cycle. The fix came from an unlikely place: a bacterium, *Thermus aquaticus*, that lives in near-boiling hot springs. Its polymerase, [[taq-polymerase|Taq]], stays folded and active at those temperatures, so it survives every denaturation and keeps copying all the way through 30-plus cycles — add it once, walk away. Like every polymerase you have met, Taq needs a primer's 3' end to start from and builds only 5'-to-3'. One caveat worth stating plainly: standard Taq has no proofreading, so it makes occasional copying mistakes; when an exact sequence matters, labs swap in a high-fidelity proofreading enzyme instead.

Its power is its peril: PCR copies any matching DNA

Here is the most important honest caveat in this whole guide, and it follows directly from how PCR works. The reaction has no idea where its template came from. It will faithfully copy *any* DNA in the tube that the primers can bind — including DNA you never meant to be there. A single stray molecule of contaminating DNA, a flake of skin, a trace of someone else's sample, or, most insidiously, a few product molecules drifting over from a previous PCR, can each become the seed of a billionfold amplification. The very exponential power that lets PCR detect one target molecule means it can also turn one contaminant molecule into a strong, confident, completely false signal.

Why PCR is the workhorse — and what comes next

PCR earned its place as the workhorse of the molecular lab by being fast, cheap, and astonishingly versatile. The same plain three-step cycle can detect a virus from a swab, type DNA at a crime scene, diagnose a single-gene disease, prepare a gene for cloning, or check whether an edit worked — you simply choose new primers each time. The basic idea also branches: feed it RNA first turned into DNA by reverse transcriptase, and RT-PCR lets you measure gene *expression* or detect an RNA virus. Watch the copies pile up live, cycle by cycle, instead of only at the end, and quantitative PCR (qPCR) turns the reaction into a *measuring* instrument — the reason that plateau detail from earlier matters, since the early exponential phase is where honest quantities live.

Step back and see what PCR did *not* solve. It can make a billion copies of a stretch of DNA — but it cannot tell you the order of letters inside that stretch. PCR *amplifies*; it does not *read*. Together those are the two halves of the invention that turned biology into a data science, exactly as this rung promised: copying a chosen piece of DNA a billionfold, and reading the letters. You have the copier; the next guides hand you the reader. And the two team up beautifully — most modern sequencing starts by using PCR-like copying to make enough material to read in the first place.