PCR copied a stretch of DNA by just heating and cooling it

Almost every use of a DNA fragment — diagnosing a disease, sequencing a gene, or fingerprinting a person — needs far more copies of it than any sample provides. The workhorse used to be cloning into bacteria, which is slow and fiddly, so a fast way to amplify a chosen stretch in a test tube was a genuine prize.

Mullis conceived PCR on a drive in April 1983, when he combined his knowledge that DNA separates and re-anneals with heat with the idea of an iterative loop. Two primers flank the target, and a heat-stable polymerase copies them; heating separates the strands, cooling lets the primers bind, and the polymerase extends across the target. Because the new copies are also copied, the target doubles each cycle.

Mullis's first successful experiment came in December 1983 after a group at Cetus was formed to perfect it, and the first patent application was filed in March 1985. A few hours of thermal cycling yield billions of identical copies, which is exactly the amplification a diagnostic or sequencing read needs.

Let alternating heat double a target DNA segment

Doubling the target each cycle gives exponential amplification, turning one molecule into billions in hours.

Two primers flank the desired segment so only that stretch, not the whole genome, is copied.

It needs no living cells, so it is fast, cheap and works on a tiny degraded sample.

When you need a lot of something you have only one of, do not try to harvest it; instead, build a loop that doubles the amount each pass, and exponential growth does nearly all the work.

PCR became the foundation of molecular biology, earned Mullis the Nobel Prize in 1993, and generated roughly $2 billion in royalties; Cetus patented and licensed it broad, and it now underpins DNA diagnostics, forensics, genomics and pandemic testing.