Before the machines: how DNA sequencing began

Long before anyone read a genome, scientists learned to read a protein. The chemistry that made the Human Genome Project possible is almost fifty years old — and for twenty of those years, it was done almost entirely by hand.

Reading a protein, one piece at a time

The idea of "sequencing" something didn't start with DNA. It started with a protein, and with a scientist who spent the better part of a decade on a single molecule. In the late 1940s, Frederick Sanger, working in Cambridge, set out to show that proteins had a fixed, specific order of building blocks rather than being loosely organised blobs, as many biochemists then assumed.

He chose insulin partly because it was available in a pure, medically useful form. Breaking the molecule into fragments, separating them on paper, and piecing the order back together took Sanger and his colleagues most of a decade. By 1955 he had the complete sequence of both insulin chains, the first time anyone had read a full protein sequence^cite. It earned him the first of his two Nobel Prizes in Chemistry, in 1958. Nobody had sequenced DNA yet. Nobody was entirely sure how you would.

1977: two methods, one answer

It took another twenty years for DNA to catch up with protein. In 1977, two groups published, within months of each other, methods for reading the order of bases in a strand of DNA. At Harvard, Allan Maxam and Walter Gilbert developed a chemical method that cut DNA at specific bases using hazardous reagents. In Cambridge, Sanger's group took a different route: a chain-termination technique, now usually just called Sanger sequencing, that used modified building blocks to stop a growing DNA strand at a known point.

The two methods solved the same problem, but Sanger's proved easier to scale and safer to run, and it became the standard the field settled on for the next three decades — including, decades later, the entire Human Genome Project.

The same year, Sanger's group used the new method to read the genome of bacteriophage phiX174, a small virus with about 5,400 bases. It was the first time any DNA genome, of any organism, had been read in full.

What sequencing actually looked like

It's worth being concrete about what this meant day to day, because "sequencing" in 1980 bears almost no resemblance to the word today. Reactions were run with radioactively labelled DNA, separated on a slab gel, and the result exposed onto X-ray film. A researcher then read the sequence off the film by eye, ladder by ladder, writing down one letter at a time.

A few hundred bases was a good result for a serious piece of lab work. In 1982, Sanger's group sequenced the genome of bacteriophage lambda — about 48,500 bases — and it stood as a landmark precisely because it was so much bigger than anything done before. For comparison, a single human chromosome can run past 200 million bases.

From bench to machine

The chemistry that made the Human Genome Project feasible was Sanger's, essentially unchanged, but the project needed something the 1977 version didn't have: automation. Through the 1980s, fluorescent labels replaced radioactive ones, and machines took over the job of separating and reading fragments that a person used to do by hand, letter by letter, off a sheet of film.

Even with that automation, the first human genome was slow and expensive by any later standard: thirteen years and roughly $2.7 billion^cite, all using the method Sanger had published in 1977.

What's left of 1977 today

The remarkable thing isn't that the chemistry was good enough for one genome. It's how long it stayed the standard. Sanger's method, developed to read a few hundred bases off a gel by eye, was still the method used to sequence the first human genome a quarter of a century later. A full human genome today, at research-grade depth, takes a matter of minutes^cite and a few hundred pounds, on technology — long-read sequencing among it — that owes its existence to the chain-termination trick a Cambridge lab worked out in 1977.

Sources: MRC Laboratory of Molecular Biology archive (insulin, 1955); Wellcome Sanger Institute blog, 25th-anniversary piece (HGP cost and modern sequencing speed, 2025). Dates for the 1977 publications and the phiX174 and lambda genomes are well-documented historical facts; see the linked Wikipedia articles for further reading and primary citations.